Analysis of the proteins secreted by Trichoderma harzianum P49P11 under carbon-limited conditions.

The wild type strain Trichoderma harzianum was able to synthesize enzymes that can catalyse the hydrolysis of p-nitrophenyl-β-D-glucopyranoside (PNPGase) in glucose-limited chemostat cultures. Fructose/glucose and sucrose conditions provided low levels of PNPGase activity. To investigate whether under these conditions other enzymes were produced, a shotgun proteomics analysis of their supernatants was performed. The analysis has indicated that the different carbon sources used influenced the amounts of proteins secreted including 1,3-beta-glucanosyltransferase, alpha-1,2-mannosidase, alpha-galactosidase and glucan 1,3-beta-glucosidase. The analysis has also suggested the presence of beta-glucosidase, which could also be represented by PNPGase activity. Intracellular metabolites were quantified during PNPGase production for the condition using 20 g/L of glucose in the feed and differences were observed, indicating that intracellular glucose could be inhibiting PNPGase production. SIGNIFICANCE: This work shows that sugars such as glucose, fructose/glucose and sucrose can be used as substrates for the continuous synthesis of different enzymes under carbon-limited conditions by Trichoderma harzianum. As far as we know, this is the first work about the continuous synthesis of enzymes under carbon-limited conditions suggesting that different easily assimilated carbon sources can be used to generate different enzymatic cocktails. Each enzyme or uncharacterized protein suggested by shotgun proteomics has the potential to become a promising product for biotechnological applications.


Introduction
Proteins are a diverse group of molecules containing different properties such as catalytic activity, molecular weight and solubility [1]. Proteins are also a biotechnological product and their identification and characterization are important for understanding their functionalities and possible applications. In proteomics, a method called shotgun analysis can be used to identify proteins [2]. Proteolytic enzymes such as trypsin are used to digest the proteins, and due to the specificity of this enzyme, it generates peptides of a size that is more readily analysed in the mass spectrometer and facilitates the identification of the protein [1]. Liquid chromatography can be used to separate the peptides, and then peptide sequencing can be performed by tandem mass spectrometry (MS/MS) [1]. In this method, peptides are ionized and selected peptide ions are subjected to sequencing, which is determined by MS/ MS spectra using a database search approach [2].
Beta-glucosidase is a group of enzymes responsible for catalysing the final step of the complete hydrolysis of cellulose into glucose [3]. Beta-glucosidase can present a high p-nitrophenyl-β-D-glucopyranoside (PNPG) specificity [3], thus the estimation of enzymatic activity using PNPG can be applied to verify the presence of enzymes that can catalyse the hydrolysis of glycosidic bonds like beta-glucosidase.
It was found that T. harzianum was able to synthesize extracellular enzymes that can catalyse the hydrolysis of PNPG (PNPGase) during steady-state under carbon-limited chemostat cultivations [4], especially when using glucose as the carbon source. When sucrose or an equimolar mixture of fructose and glucose were used, low levels of PNPGase activity were observed. To investigate whether under these conditions other enzymes were produced, a shotgun proteomics analysis of their supernatants was performed and it is presented in this work.
https://doi.org/10.1016/j.jprot.2020.103922 Received 23 March 2020; Received in revised form 22 July 2020; Accepted 24 July 2020 Additionally, analysis of PNPGase production was evaluated according to the results of samples from continuous cultures using 20 g/L of glucose in the feed under carbon-limited conditions. Intracellular metabolites were analysed to verify the possible influence of their concentrations on PNPGase production.

Culture conditions
Trichoderma harzianum P49P11 was isolated from the Amazon forest [5]. Culture conditions are described in Gelain [4]. The microorganism was grown on plates (potato dextrose agar) at 29 °C and after 5-7 days, the spores were harvested with sterilized water, distributed into vials of 1 mL and stored at −80 °C. Each vial was used to inoculate one shake flask (250 mL of medium). Spores were used to inoculate 500 mL shake flasks containing 250 mL of the medium: 10 g/L of glucose (carbon source), 2 g/L of KH 2 PO 4 , 5 g/L of (NH 4  Different limiting carbon sources were applied in continuous culture: 10 g/L of glucose (G), 10 g/L of sucrose (S), 5.26 and 5.26 g/L of fructose and glucose (FG). The medium composition was the same as described for shake flasks, only peptone was not added to the feed medium. Additionally, 20 g/L of glucose was also tested in the feed with the following modifications to the medium composition: 3 g/L of KH 2 PO 4 ; and 6 g/L of (NH 4 ) 2 SO 4 . The medium composition used for the batch phase of the condition using 20 g/L of glucose was the same as used for the shake flask cultivation. The experiments G, S and FG were performed in sequence and 20 g/L of sucrose was used as the carbon source in the batch phase. The medium was sterilized by filtration using a filter 0.2 μm.
The experiments were performed using a 7 L bioreactor (Applikon, Delft, the Netherlands) and a constant broth mass of 4 kg. The temperature was controlled by a water bath at 29 °C, and pH 5 was controlled by the addition of 2 M KOH and 2 M H 2 SO 4 . Sterile air was supplied via a mass flow controller (Brooks 58,505, calibration at 0 °C and 1 bar). The volume fraction of oxygen and carbon dioxide were measured by the NGA 2000 off-gas analyser.
All chemostat experiments were carried out with a dilution rate of 0.05 h −1 ± 0.003 h −1 . For the batch phase, the stirring speed was kept between 200 and 400 rpm and for the continuous culture, it was changed to a constant stirring speed of 600 rpm. The airflow of 1 L/min was used, and only for the condition using 20 g/L of glucose, the airflow was 1.5 L/min. A constant antifoam addition (Basildon BC antifoam 86/ 013) of approximately 7 μL/min was used. The achievement of the steady-state was assumed when the CO 2 production and mycelium concentration were constant for at least 6 residence times.

Enzymatic activity
The method for the estimation of PNPGase activity was adapted from Zhang et al. [6] using p-nitrophenyl-β-D-glucopyranoside (PNPG) as the substrate. The reaction mixture was composed of 80 μL of 1 mM substrate diluted in 50 mM citrate buffer (pH 4.8), and 20 μL of the diluted enzyme extract. After 10 min at 50 °C, the reaction was stopped by adding 100 μL of 1 M sodium carbonate. The measurement was performed at 400 nm.

Quantification of intracellular metabolites
The estimation of the metabolites was performed in duplicate using different samples withdrawn at the same time (10 s difference). Intracellular metabolites were extracted and analysed according to Lameiras et al. [7]. Broth (approx. 1.3 mL) was rapidly withdrawn into 10 mL of pre-cooled 40% (v/v) aqueous methanol solution (−20 °C) and after, the samples were weighted for the estimation of cell mass and kept at −20 °C until extraction of the metabolites. Methanol was removed by filtration and the samples were washed thrice with cold methanol solution (−20 °C). Then, boiling ethanol extraction was performed to disrupt the cell and inactivate the enzymes. 25 mL of ethanol solution (75% v/v) was first pre-heated at 75 °C, and after the filtration and washing of the samples, they were placed into the ethanol solution along with 100 μL of U-13 C-labeled cell extract of S. cerevisiae as the internal standard and moved to a water bath at 95 °C for 3 min. After the extraction, the samples were first placed on ice and then stored at −80 °C.
Before the quantification of the metabolites, ethanol was evaporated until almost dryness in a Rapid-Vap under vacuum for 240 min. After evaporation, the residues were suspended in 500 μL of Milli-Q water, and centrifuged at 1000 ×g for 5 min in a tube coupled with a filter (0.22 μm) to remove cell debris. The supernatants were stored at −80 °C until analysis. The concentrations of the intracellular metabolites were measured by isotope dilution mass spectrometry (LC-IDMS/ MS and GC-IDMS) according to the protocols of Dam et al. [8], Jonge et al. [9] and Cipollina et al. [10].

SDS-PAGE
The samples were concentrated using Amicon 10 kDa cut-off Eppendorf tube concentrators. Then, they were mixed with NuPAGE LDS sample buffer (4×) (Thermo Fischer Scientific, NP0007) at a ratio of 1 to 4 and heated at 70 °C for 15 min. For the SDS-PAGE analysis, a mini-protean polyacrylamide gel was prepared (10% resolving gel) (BIO-RAD, 1610182), 1 μg protein was loaded from each sample. SDS-PAGE was run using a MES buffer (Thermo Fischer Scientific, NP0002) at constant voltage (200 V) for 25 min. Following a short fixation, the gel was stained using Coomassie Brilliant Blue for 15 min and destained in 10% acetic acid over-night. The gel image was taken using a XY camera and the image was processed using the Microsoft Office Picture Manager, 2010.

Shotgun proteomics and label-free quantification
Protein was precipitated from the supernatant using 4 volumes of ice-cold acetone at −20 °C for 20 min, centrifuged at 14.000 rpm for 10 min, and the pellet was washed twice using ice-cold acetone. The protein pellet was reconstituted in 200 mM ammonium bicarbonate containing 6 M urea. The protein solution was further reduced using DTT and then alkylated using iodoacetamide according to the protocol by Herbert et al. [11]. The protein solution was diluted using 200 mM ammonium bicarbonate buffer to approximately 1 M urea and further digested using trypsin in a ratio of protein to trypsin of approximately 50 to 1, at 37 °C over-night [11].
The proteolytic digest was purified using an Oasis HLB solid-phase extraction plate (Waters) and approximately 250 ng of the proteolytic digest was then analysed by a one-dimensional reverse-phase gradient (Acclaim PepMap RSLC 50 μm × 15 cm, 2 μm, 100 A, Thermo) using an EASY-nLC 1200 coupled to a QE plus Orbitrap mass spectrometer operating in top 10 DDA mode. Further details regarding one-dimensional shotgun proteomics approaches are described in Köcher et al. [12].
Tandem-MS data were analysed using PEAKS Studio 8.5 against the TrEMBL Trichoderma harzianum protein database (un-reviewed, taxon 5544 http://www.uniprot.org/), and results were filtered for < 1% FDR for both, peptide and protein identification. 3 replicate injections for each condition were finally analysed using the label-free quantification option provided by PEAKS Studio 8.5 [13].

Secretome analysis
SDS-PAGE analysis was performed in the supernatant of the samples from carbon-limited chemostat cultivations using glucose (G), sucrose (S), and fructose/glucose (FG) (Fig. 1). For the glucose and fructose/ glucose-limited chemostat cultures, more intense bands were observed than for the sucrose-limited culture, indicating higher protein concentrations. The molecular weights of the majority of the proteins were between 70.3 and 92 kDa for the glucose and fructose/glucose-limited cultures. For the sucrose-limited culture, the molecular weight of the  Fig. 2 highlights spectrum intensity, indicating the relative abundance for specific groups of enzymes. For the glucoselimited condition, higher intensities were observed for most of the proteins, indicating higher protein levels. According to Owen et al. [14], the relative intensity of each spectrum peak is proportional to the peptide concentration; this allows a relative quantification across different experimental conditions. Table 1 shows the result of some enzymes related to Fig. 2. Several enzymes were found including four 1,3beta-glucanosyltransferase, two alpha-1,2-mannosidase, two alpha-galactosidase, two beta-glucosidase, two glucan 1,3-beta-glucosidase, two glucan endo-1,3-beta-glucosidase, thirteen glycosyl hydrolase and three mutanase. The complete table is presented in the supplementary material. It presents the coverage, the number of peptides and unique peptides identified, and intensities of the samples for each protein.
Beta-glucosidase is highlighted in Table 1, the first one had coverage of 37%, 26 peptides, 24 unique peptides, 93 kDa and the second one, coverage of 8%, 8 peptides, 7 unique peptides and 95 kDa.
Vale et al. [15] reported a secretome analysis of T. harzianum using batch cultivation on cellulose performed in Erlenmeyer flasks with a liquid volume of 20 mL. They identified 56 proteins based on at least 2 unique peptides. In this project, the proteins with at least 1 unique peptide are presented to visualize the possible enzymatic cocktail secreted. Nevertheless, there are 114 proteins identified based on at least 2 peptides and 93 with only 1 unique peptide. According to Vale et al. [15], nowadays, high MS accuracy, low ppm mass errors, and rich MS/ MS data, provided by mass spectrometers, offer excellent identification even based on single peptides. Fig. 3 indicates through a heat map, the presence of proteins in the supernatant of the three samples analysed (G, S and FG). The B sample corresponds to the blank (no protein present, bright green colour). The samples came from continuous cultures with a dilution rate of 0.05 h −1 and the differences observed are mainly due to the carbon sources used since no other variables were changed. Clear differences in protein abundances were observed between the three experimental conditions. If they would have been similar, this would indicate the constitutive production of these enzymes/proteins; however, the difference observed indicates that their production was possibly influenced by different inducers. Qualitative analyses such as ethanol precipitation, followed by hydrolysis of the precipitate confirmed the presence of extracellular polysaccharides in all conditions [4]. The presence of extracellular polysaccharides suggests that they were the source of inducers.
Gómez-Mendoza et al. [16] evaluated the T. harzianum secretome using glucose, carboxymethyl cellulose, xylan and an agricultural byproduct (sugarcane bagasse) as carbon sources. The experiments were performed in shake flasks containing 1% (w/v) of the carbon source with a liquid volume of 30 mL. Glucose was used as the promoter of catabolite repression of enzymes and the condition resulted in the secretion of 107 groups of proteins of which 40 were exclusively identified in this carbon source. The secretion of cellobiohydrolase and betaglucosidase were detected for this condition and their presence was attributed to the consequence of low constitutive enzyme production. In this project, however, the different colour intensities in Fig. 3 suggest that glucose condition was capable of inducing the production of proteins.
The difference between G and FG conditions is the lack of fructose in the G condition. This change resulted in different protein production as indicated in Fig. 3. It seems that the presence of fructose inhibited/ induced the production of different proteins, but the samples also share common ones. The sucrose-limited condition resulted in the lowest protein levels in the culture supernatant. The use of monosaccharides (fructose and glucose -FG) had a positive influence on the diversity of the proteins/enzymes than the disaccharide used. The glucose-limited condition showed the highest diversity of proteins. It seems that each condition provided a distinct enzymatic cocktail, which could be related to the different carbon sources used. However, more research is needed to identify the uncharacterized proteins suggested by shotgun proteomics analysis.

Analysis of intracellular metabolites during PNPGase production under glucose-limited conditions
The intracellular metabolites from the tricarboxylic acid cycle, glycolysis and pentose phosphate pathway, as well as nucleotides, were quantified for samples from glucose-limited chemostat cultures named A and B using 20 g/L in the feed. Both samples (A and B) were withdrawn during the steady-state of cells and CO 2 . The difference between them is that sample A was not at steady-state of PNPGase production. The Tukey test with 95% confidence interval (OriginPro 8 software) was applied to analyse the average values for each metabolite. Significant differences among metabolite levels could indicate Table 1 Description of some enzymes from shotgun proteomics analysis, beta-glucosidase is highlighted in grey frames, G -glucose at 10 g/L, FG -fructose/glucose, and Ssucrose.  metabolites that might be directly or indirectly related to PNPGase production. Statistically significant differences in concentrations of the metabolites are highlighted in blue ( Table 2). The high experimental errors observed for some metabolites could be attributed to the sample processing and/or analytical errors. Another explanation considered is a possible heterogeneity of cells inside the bioreactor caused by the consumption of fragments of polysaccharides under carbon-limited conditions. Considering that the highest shear rates in a mixing vessel are close to the blades of the impeller [17], there are regions with different shear rates inside the bioreactor. Fragments of polysaccharides from the cell wall can be released due to this shear stress [18], and because of different shear rate regions inside the bioreactor, it could also be expected that there exist regions with different concentrations of fragments. Since the cells can consume these fragments, the cells closer to the regions with a higher concentration of fragments could present different concentrations of some intracellular metabolites than the cells farther from that region.
Although samples A and B were at steady-state of cell concentration and CO 2 production, it seems that there was another parameter, perhaps polysaccharide fragments, influencing the intracellular metabolite concentrations and the changes highlighted in Table 2 were not considered as resulting from sample processing and/or analytical errors due to the low standard experimental errors. Table 2 shows a significant difference regarding intracellular glucose concentration that decreases from sample A to B and its concentration is inversely proportional to PNPGase activity (A = 1.1 ± 0.1 U/mL, B = 6.2 ± 0.4 U/mL). The higher glucose concentration for sample A could explain the lower enzymatic activity. Perhaps at this stage (A), the process was still unstable and glucose was inhibiting the inducer uptake. Additionally, 1.6 μmol/g (B) seems to be an intracellular glucose concentration that could prevent carbon catabolite repression.
Based on PNPGase activity and intracellular glucose concentration, it can be considered that sample A provides information on an inhibition-state and sample B provides information of an inducer-state for PNPGase synthesis. For sample B, the uptake rate of inducer fragments of PNPGase synthesis was probably higher than for sample A, meanwhile, the uptake rate of hydrolysed fragments was higher for sample A, provoking inhibition. Therefore, significant differences observed between some metabolite levels indicate changes in the metabolic pathway when there is a higher consumption of inducers (sample B) or inhibitors (sample A) of PNPGase production. The higher uptake of inducers resulted in higher concentrations of metabolites from the tricarboxylic acid cycle, part of glycolysis (from glyceraldehyde-3-phosphate to phosphoenolpyruvate) and lower values for metabolites from the pentose phosphate pathway. Nevertheless, more experiments have to be conducted to verify whether the metabolic profiles observed always follow this pattern.

Conclusions
Shotgun proteomics and SDS-PAGE analysis were performed for the proteins present in the supernatant of carbon-limited chemostat cultures using glucose, fructose/glucose and sucrose as carbon sources. The shotgun proteomics analysis has indicated that the different carbon sources used greatly influenced the amounts of secreted proteins, of which many of them are enzymes. The possible differences regarding the enzymes secreted could be related to the presence of different inducers. This brings the possibility of creating a hypothesis that different carbon sources easily assimilated by the cells under carbon-limited conditions could lead to the synthesis of different inducers and thus different enzymes. Several uncharacterized proteins were also suggested by the analysis and methods to identify them and their functions could reveal new promising products. Analysis of the intracellular metabolites has suggested that high intracellular glucose concentrations can inhibit the production of PNPGase.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.