Functional characterization and target discovery of glycoside hydrolases from the digestome of the lower termite Coptotermes gestroi

Biochemical characterization

Here we describe the first global study of an enzymatic repertoire involved in plant polysaccharide degradation by C. gestroi. The crude extract was highly active on hydrolyzing glucose-containing polysaccharides, in both the linear and branched forms (Figure 1A), such as β-glucan from barley; lichenan, a β-1-3:1-4 glucose linked polysaccharide; and laminarin, a β-1-3:1-6 glucose linkage polysaccharide. On these substrates, enzyme activity was over 70% active on a wide range of pH, from pH 4 to pH 8 (Figure 1A). The crude enzyme extract showed activity on carboxymethyl cellulose (CMC), a linear β-(1,4) polymer of D-glucose (Figure 1A). The relative enzymatic activity on CMC decreased to 73% at pH 4 and about 50% when incubated on pH 7 and pH 8. The enzymatic activity was also detected on starch, which is an α-linked glucose polysaccharide, and retained more than 77% of its enzymatic performance at pHs ranging from pH 4 to pH 8 (Figure 1A).

The termite crude extract was also competent in hydrolyzing hemicellulosic substrates (Figure 1B), such as mannan and galactomannan, as well as xylan from birch wood, a decorated β-1-4 xylose-containing polysaccharide. The enzymatic activity was also observed on xyloglucan and arabinoxylan, two β-1-4 glucose- and β-1-4 xylose-containing polymers, with lateral sugar branches of xylose and arabinose respectively. The enzymatic hydrolysis of pectin, which is composed of 1,4-linked α-D-galactosyluronic acid residues, was evidenced as well. Relatively low activities were observed over arabinogalactan and debranched arabinan. The optimum pH of operation in most of these substrates was between pH 5 and pH 6; nevertheless, more than 80% of activity at pH 4 and pH 8 was retained (Figure 1B).

Hydrolysis of p-nitrophenyl-carbohydrate (pNP) substrates had its most significant result against pNP-glucopyranoside (pNP-G), followed by pNP-cellobioside (pNP-C) and pNP-mannoside (pNP-M) (Figure 1C). For all pNP derivatives evaluated in this study, the optimal pH of operation occurred at pH 6 (Figure 1C). However, a drastic drop in the enzyme activity was observed at pH 7, of about 80%. pNP-α-L-arabinofuranoside (pNP-A) was barely hydrolyzed by the crude enzyme extract and no action was detected on pNP-xylopyranoside (pNP-X) in this study.

Metaproteomic analysis

A metaproteomic analysis from the whole body of the lower termite C. gestroi was carried out. Our strategy allowed the identification of a hundred positive fractions that degrade at least one of the carbohydrates listed above. Following the functional screening, the positive fractions were subjected to trypsin digestion and ultra performance liquid chromatography (UPLC) followed by tandem mass spectrometry analysis. Since there is very little genomic sequence data available for C. gestroi, the LC-MS/MS spectra were analyzed by Mascot Ions-Search Software (Matrix Science, London, GB, UK)for protein identification using a database containing all non-redundant proteins derived from the CAZy website [20], a specialized database (up to 67,000 protein sequences) that describes the families of structurally-related catalytic and carbohydrate-binding modules (or functional domains) of enzymes that degrade, modify or create glycosidic bonds. After spectra acquisition, using a stringent cut-off (cut-off > 27; significant at P < 0.05) allowed the explicit assignment of 121 unique peptides as GHs and CBMs (Additional File 1).

Our proteomic approach successfully identified a large number of polypeptides in the C. gestroi digestome. A total of 55 different protein matches were identified and classified in 29 CAZy families (Figure 2). The sequence and distribution of the peptides identified through proteomic analysis, as well as their categorization in different GH and CBM families, are described in detail in Additional File 1. Using a confidence criterion of two or more unique peptides of different peptide sequences [21], three protein matches for GH members were identified. Potential cellulolytic enzymes were identified using this criteria, such as the match to probable β-glucosidase (GH3) from Dictyoglomus turgidum, two GH7 protein matches, one derived from Pleurotus sp. and another from the uncultivated symbiotic protist Hodotermopsis sjoestedti, as well matches to GH9, from Reticulitermes flavipes and C. formosanus, which are endogenous cellulases from termites [19, 22–25]. Two protein matches with high confidence were assigned to CBM families 6 and 27, derived from Salinispora arenicola and Thermotoga sp., respectively (Additional File 1). CBM6 is often coupled to a GH16 catalytic domain [26], which has been described as acting against substrates that posses β-1,3 linkages such as lichenan, laminarin and curdlan. CBM27 is a protein domain that binds to mannose polymers, and is usually tied to GH26, β-mannosidase or β-xylosidase, forming a multidomain protein [20, 27].

Based on the total number of peptides identified, the majority of components found in the C. gestroi digestome were cellulose degrading enzymes, including β-glucosidases (GH1 and GH3), which accounted for up to 10% of the protein matches, followed by GH5, GH7 and GH9 families, which made up 22% of the total peptides. These are enzymes that preferentially hydrolyze internal β-1-4 glycosidic bonds. Protein assignments for GH16 (3%), probable endo-β-1-3-glucanase, were also present in the termite extract (Figure 2). Xylanolytic enzymes, members of GH10 and GH11 families, also counted as an important component, comprising up to 6% of the total peptides identified. Three peptides with matches for family GH43, with known activities including β-xylosidase and α-L-arabinofuranosidase, were found in our analysis. Potential mannan-hydrolytic enzymes, members from families GH2 and GH38 (three peptides identified), as well as polygalacturonase(GH28), beta-galactosidase (GH42), fucosidases (GH29) and fructofuranosidases (GH32) were identified (one peptide match of each, respectively). A remarkable number of peptides with matches for starch-degrading and branching enzymes were found; matches for families GH4, GH13, GH31, GH57, GH63, GH65, GH85 and GH94 counted up to 22% of the total peptides matches. Protein matches for putative lysozymes were also found and GH25 and GH28 members, one peptide for each family, were identified. About 3% of the peptides matches found (two distinct protein matches) were identified as chitinases, based on GH18 classification.

The enzyme hydrolytic activity verified on chromatographic fractions selected for mass spectrometry analysis corroborated with their protein identification by LC-MS/MS, thereby validating the identification of proteins that have a unique peptide sequence found in our proteomic analysis (Additional File 1). Cellulolytic protein matches to GH1, GH3, GH5, GH7 and GH9 were mostly derived from the chromatographic fraction active on hydrolyzing CMC and lichenin. Among the peptides found with low confidence criteria, which were confirmed by the expected enzymatic hydrolysis, were members of families GH10, GH11 and GH43 (Additional File 1). Those were identified by their hydrolytic activity on xylan. The unique peptide match for GH16 was derived from a positive fraction with activity against lichenin, and another unique match for GH2 was derived from mannan degrading fractions.

With the aim of further validating our proteomic results, we analyzed a couple of ion-exchange (IEX) fractions through SDS-PAGE and ran a carbohydrate degradation profile by capillary zone electrophoresis. As shown in Figure 3A, the analysis by capillary electrophoresis of cellopentaose degradation by the IEX4-37 fraction, with just one peptide match identified as GH7, showed the derived hydrolysis products as being cellobiose and cellotriose. The enzyme source specifically cleaved internal glycosidic bonds of 8-aminopyreno-1,3,6-trisulfonic acid (APTS)-reducing-end-labeled-cellopentaose (Figure 3A), which was the expected mode of action for a GH7 type endoglucanase. The SDS-PAGE analyses for fraction IEX4-37 reveled one unique band around 53 kDa (Figure 3C), the predicted molecular weight for this protein match as suggested by the Mascot ion search. The purified fraction IEX1-26, which was identified as a single peptide match for the GH2 family, was able to hydrolyze mannohexaose, producing mannotriose and mannobiose, as expected for typical endo-β-1,4-mannanase hydrolysis (Figure 3B). The SDS-PAGE for IEX1-26 reveled protein bands of about 80 kDa (Figure 3C), compatible for mannanases derived from fungi and bacteria [27, 28].

Termites

Workers of C. gestroi were collected from field colonies with traps of corrugated cardboard and from incipient colonies reared in the Termite Laboratory of the Biology Department, UNESP, Rio Claro, São Paulo, Brazil (22823'S, 47831'W). Termites were kept at room temperature with nest materials and worker termites manually selected and utilized for all experiments.

Crude enzyme extract preparation

Total protein extractions for biochemical characterization were prepared from 50 whole termite bodies. They were homogenized using Potter-Elvehjem Tissue Grinder (Wheaton, Millville, NJ, USA) with 1 mL of 50 mM McIlvine [52] buffer extraction at pH 5.0. After extraction, the mixture was centrifuged at 20,100 g for 20 minutes at 4°C. The supernatant was collected and is here referred as crude enzyme extract. Protease inhibitor cocktail (Anresco - Solon, OH, USA) was added to the crude extract produced (1 μL/mL). The concentration of the protein in crude extract was determined by the Bradford method [53] All experimental procedures were performed on ice.

Enzymatic assays

The assays for biochemical characterization were conducted with 10 μL of crude enzyme extract (3.2 μg/μL) incubated with 40 μL of 50 mM McIlvine buffer (pH 4 to pH 8) and 50 μL of 0.5% specific substrate (in water), in 96-well PCR plates, in triplicate, at 37°C, for 20 minutes to 60 minutes, depending on the substrate. Substrates were purchased from Megazyme (Wicklow, Ireland) and Sigma Aldrich (St, Louis, MO, USA). Enzymatic assays were stopped by the addition of 100 μL of dinitrosalicylic acid [54] and heated at 99°C for 5 minutes. The measurement of color change was performed at 550 nm using a microplate reader (InfiniteM200 Tecan - Männedorf, Switzerland). The enzymatic assays with pNP were performed as follows: 10 μL of crude extracts were incubated with 10 μL of 5 mM pNP substrates, 30 μL of 50 mM McIlvine buffer (pH 4 to pH 8) and 10 μL MilliQ-Water (Millipore - Billerica, MA, USA). Assays were stopped by the addition of 100 μL of 1 M sodium carbonate. The measurement of color change was performed at 400 nm using a microplate reader. The enzymatic assays were done in triplicate. Glucose and pNP were used for the standard curve construction. The enzymatic activity assays results were expressed in terms of mM of glucose equivalents produced

Chromatographic separation steps through ion-exchange and gel filtration

The metaproteomic analysis on the whole body of C. gestroi was carried out to reveal polysaccharidases contained in the termite digestome. To reduce the complexity of the sample, the termite crude extract was submitted to chromatographic separation followed by enzymatic activity screening of purified fractions. Four types of polysaccharides were used as substrates for enzymatic activity assays, namely CMC, lichenan, mannan and xylan.

Prior to FPLC-LC-MS/MS analysis, about 500 termites were used for crude enzyme extract preparation, as described above, using 100 mM phosphate-potassium buffer (PPB), pH 6.2. The ion-exchange (IEX) purification chromatography was carried out using a Resource Q anionic exchange column on an AKTA Purifier System (both GE Healthcare Life Sciences - Piscataway, NJ, USA). Initially, 1.5 mL of crude extract was cleared on a Millipore filter of 0.45 μm and then 1 mL was applied onto the Resource-Q column, equilibrated with 100 mM PPB, pH 6.2. The termite crude extract was fractionated through a linear gradient (100 mL) of 0 to 1 M sodium chloride in 100 mM PPB, pH 6.2, at a flow rate of 1.0 mL/min and collecting 2-mL fractions. IEX purifications were repeated five times, generating 250 eluted fractions. These fractions were submitted to enzymatic assays against the polysaccharide substrates at room temperature for 24 hours. Fractions with positive enzymatic activity were selected for LC-MS/MS analysis.

Fractions derived from IEX positively selected for lichenanases, xylanases, CMCase and mannanases activity screenings were applied onto Hi-Load 16/60 Superdex 75 - Gel Filtration Column (GE Healthcare), equilibrated in 100 mM PPB, and eluted in 100 mM PPB, pH 6.2 containing 0.5 M sodium chloride, at a flow rate of 0.1 mL/min with 2-mL fractions. Fractions of gel filtration chromatography with positive activity against the substrates described above were selected for LC-MS/MS analysis.

Mass spectrometer analysis (LC-MS/MS)

About 500 μL of eluted fractions from chromatography steps were precipitated with ice cold acetone (Merck - Darmstadt, Germany) (1:2, v:v), then incubated at -20°C for 2 hours, followed by centrifugation at 4°C. The supernatant was recovered and the pellet was dried with a Speed-Vac (Eppendorf - Hauppauge, NY, USA) for 30 minutes and then resuspended into 100 μL of 50 mM ammonium bicarbonate. Proteins were reduced, alkylated and the reduction step repeated. Proteins were then digested with 1 μg trypsin (Porcine Prep Grade Sequence, Promega - Madison, WI, USA), in 1 mM calcium chloride) at 37°C overnight. The reaction was stopped by adding 0.1% formic acid (Merck) and dried in the Speed-Vac. Each sample was resuspended in 12 μL formic acid, 0.1%, and an aliquot (4.5 μL) of the resulting peptide mixture was separated by C18 (100 μm × 100 mm) RP-nanoUPLC (nanoAcquity, Waters) coupled with a Q-Tof Ultima mass spectrometer (Waters) with a nano-electrospray source at a flow rate of 0.6 μL/min. The gradient was 2% to 90% acetonitrile in 0.1% formic acid over 60 minutes. The instrument was operated in the 'top three' mode, in which one MS spectrum is acquired followed by MS/MS of the top three most intense peaks detected. The spectra were acquired using the software MassLynx v.4.1 (Waters - Milford, MA, USA) and the raw data files were converted to a peak list format (mgf), without summing the scans, by the software Mascot Distiller v.2.3.2.0, 2009 (Matrix Science Ltd.). These were then searched against the CAZy Database (67,651 sequences; 33,940,339 residues [20]) using Mascot v.2.3.01 engine (Matrix Science Ltd.), with carbamidomethylation as the fixed modification, oxidation of methionine as a variable modification, one trypsin missed cleavage and a tolerance of 0.1 Da for both precursors and fragment ions. Only peptides with a minimum of five amino acid residues, with a significance threshold of P < 0.05 (Mascot-based score) were considered as a product of peptide cleavage. The peptide was considered as unique when it differed in at least one amino acid residue and covalently modified peptide, including N- or C-terminal elongations (that is, missed cleavages). Different charging states of the same peptide and modifications were not counted as unique.

Capillary electrophoresis

Cellopentaose and mannohexaose (Megazyme) were derivatized with APTS by reductive amination, as described by dos Santos et al. [55]. Enzymatic reactions were performed as described for standard enzymatic activity, except that 1 pM of APTS-labeled specific oligosaccharides were used as substrates. Capillary zone electrophoresis was performed on a P/ACE MQD (Beckman Coulter - Indianapolis, IN, USA) with laser-induced fluorescence detection. A fused silica capillary (TSP050375, Polymicro Technologies - Phoenix, AZ, USA) of internal diameter 50 mM and length 31 cm was used as separation column for the oligosaccharides. Samples were injected by the application of 0.5 psi for 0.5 seconds. Electrophoresis conditions were 10 kV/70-100 mA with the cathode at the inlet, 0.1 M sodium phosphate, pH 2.5, as a running buffer, and a controlled temperature of 20°C. The capillary was rinsed with 1 M sodium hydroxide followed by running buffer to prevent carryovers. APTS-labeled oligomers were excited at 488 nm and emission was collected through a 520-nm band pass filter. Because of the small volumes of capillary electrophoresis combined with small variations in buffer strength, retention times varied slightly when comparing separate electrophoresis runs. The combined information, obtained from the electrophoretic behavior and co-electrophoresis with oligosaccharides, was used for the identification of formed products.