Characterization of a new bifunctional endo-1,4-β-xylanase/esterase found in the rumen metagenome

Metagenomic data mining of the Nellore cattle rumen microbiota identified a new bifunctional enzyme, endo-1,4-β-xylanase/esterase, which was subsequently overexpressed in E. coli BL21 (DE3). This enzyme was stable at pH intervals of 5 to 6.5 and temperatures between 30 and 45 °C, and under the test conditions, it had a Vmax of 30.959 ± 2.334 µmol/min/mg, Km of 3.6 ± 0.6 mM and kcat of 2.323 ± 175 s−1. Additionally, the results showed that the enzyme is tolerant to NaCl and organic solvents and therefore is suitable for industrial environments. Xylanases are widely applicable, and the synergistic activity of endo-1,4-β-xylanase/esterase in a single molecule will improve the degradation efficiency of heteroxylans via the creation of xylanase binding sites. Therefore, this new molecule has the potential for use in lignocellulosic biomass processing and as an animal feed food additive and could improve xylooligosaccharide production efficiency.

www.nature.com/scientificreports/ new carbohydrases, including xylanases. In this study, we performed metagenomic data mining of the rumen microbiota of Nelore cattle and identified the xylr gene, which is responsible for the production of a bifunctional xylanase esterase that is tolerant to salt and organic solvents. This gene was cloned into an expression vector and purified, and the enzyme was characterized.

Results
New xylanase gene prospected from the Nelore cattle rumen metagenome. With the aid of a sequence-driven approach to evaluating the metagenomic data of the Nelore cattle rumen (SRX818104), the xylr gene encoding an endo-1,4-β-xylanase was identified. The xylr gene contains 2196 bp and 723 amino acids and has a signal peptide of 21 amino acid residues at the N-terminus with a predicted molecular mass of 81.6 kDa 14 .
Using the BLASTp tool 15 , endo-1,4-β-xylanase was shown present 64% sequence similarity with the endo-1,4β-xylanase/feruloyl esterase enzyme from Prevotella ruminicola (Table S1) and 66% similarity to a 2015 United States patent for hemicellulose-degrading enzymes (US 9012186). Multiple sequence alignments and secondary structure predictions demonstrated that the enzyme had the following conserved regions with xylanase GH10: 159 WDVVNEA 163 and 280 TELD 283 (Fig. 1). Regarding the esterase domain, the presence of the 624 GLSMG 628 motif with a serine (S) was observed at the center of this conserved region (Fig. 2).
XylR was cloned, expressed in the soluble fraction and purified. The xylr gene was amplified directly from the rumen metagenomic DNA in the Laboratory of Biochemistry of Microorganisms and Plants (LBMP) database at São Paulo State University, Jaboticabal Campus. Thereafter, it was cloned and expressed in E. coli BL21 (DE3) ( Figure S1) using the pET28a( +) vector. A fusion protein with His 6 at the C-terminus was obtained. The cloning was confirmed by colony PCR and Sanger sequencing.
The optimum conditions for the expression of the enzyme in E. coli BL21 (DE3) were obtained with a final concentration of IPTG of 0.1 mM and grown at 30 °C for 22 h at 180 rpm. Under these conditions, the recombinant protein was obtained from the cell lysate supernatant soluble fraction (Fig. 3). The XylR enzyme was purified by immobilized metal ion affinity chromatography (Fig. 3) and later by gel filtration chromatography (Fig. 4). SDS-PAGE analysis showed that the purified enzyme had a molecular weight of 75 kDa. The expression and purification of the recombinant enzyme was confirmed by western blot analysis with anti-His6 antibodies, which confirmed that the purified protein was expressed heterologously ( Figure S2).

Figure 1.
Fragment of the multiple alignment between the amino acid sequence of XylR and xylanases belonging to GH10. The conserved regions are marked with ★, and the conserved catalytic residues of glutamate (E) are highlighted with ▲. The alignment included xylanase sequences from Prevotella ruminicola (D5EY13), Prevotella ruminicola (P48789), Zunongwangia profunda (ADF53358) and a noncultivable bacterium (AAL06078). Similar amino acids are highlighted in red. Multiple alignment was performed using the Clustal Omega tool 16 , and assembly was performed in EsPript 3.0 17 . Figure 2. Fragment of the multiple alignment between the XylR amino acid sequence and CE1 family esterases. The conserved regions are marked with ★, and the conserved catalytic residue of serine (S) is highlighted with ▲. The alignment includes the esterase sequences of Lactococcus lactis (AAM45148), Bacteroides intestinalis DSM 17,393 (5 VOL), Streptococcus pyogenes (4ROT) and a noncultivable bacterium (6RZO). Similar amino acids are highlighted in red. Multiple alignments were performed using, the Clustal Omega tool 16 and assembly was performed in EsPript 3.0 17 .  XylR activity at neutral pH and room temperature. Enzyme activity was observed with 0.1 M sodium acetate buffer at pH 5.5 to 6.5, and optimal hydrolysis occurred in this range (Fig. 7). The temperature effect showed that the optimal enzyme activity occurred at 37 °C and approximately 80% activity was maintained in intervals of 30 and 45 °C (Fig. 8). Additionally, the enzyme remained stable when subjected to heat treatments of 45 and 50 °C for up to one hour ( Fig. 9).
XylR is tolerant to high concentrations of NaCl. The activity of the XylR enzyme decreased by less than 15% in NaCl concentrations up to 2 M (Fig. 10). Moreover, enzyme activity was inhibited by 20% under  www.nature.com/scientificreports/ exposure to 2.5 M NaCl for up to 4 h, and at least 50% enzymatic activity was maintained with further exposure to this treatment for 8 h (Fig. 11).
XylR is tolerant to organic solvents. The effect of different metal ions and EDTA on XylR activity is shown in Table 2. Enzyme activity was reduced by less than 15% under exposure to Ca +2 , Mg +2 and Mn +2 ions and by 60% under exposure to Co +2 . Moreover, complete inhibition of XylR enzymatic activity was observed with ions Zn +2 , Hg +2 and Cu +2 . The presence of EDTA did not inhibit the enzyme, whereas the studied organic solvents (methanol, ethanol, propanol, DMSO, Triton X-114, Triton X-100, Tween 20 and Tween 80) caused inhibitory effects on the enzymatic activity only at concentrations above 20% (v/v) (Fig. 12). Additionally, XylR     www.nature.com/scientificreports/ remained stable in the presence of detergents at the studied concentrations except for Tween 80 at 2%, which resulted in 60% inhibition of enzymatic activity.

Discussion
In this study, the XylR enzyme was identified from the rumen metagenome and characterized as a bifunctional xylanase that is tolerant to salt and organic solvents. This enzyme should be a promising candidate for applications in xylan degradation; moreover, its constituent radicals (ferulic acid or p-cumaric acid) are united by ester bonds and can be applied in industrial biotechnological procedures. The production of recombinant protein from the discovery of new genes has made it possible to leverage the production of lignocellulose-degrading enzymes for the best use of this abundant raw material. Among xylanolitic enzymes, the endo-β-xylanases of the GH10 and GH11 families are more efficient and have greater temperature and pH tolerance 18 . The GH10 family can act on low molecular weight cellulose substrates, such as aryl-cellobiosides, and some cell-oligosaccharides 4 and thus is a better candidate for the degradation of heterogeneous compounds. Because of the ability of GH10 xylanases to act on short, soluble and branched xylo-oligosaccharides, genes that transcribe these enzymes have been identified in metagenomic databases on organisms that are efficient in hydrolyzing lignocellulose, such as ruminants 11,12,19,20 and termites 21 . Such work facilitates the search for these types of enzymes and the identification of their hydrolysis mechanisms and structures.
The gene transcribed by xylr that was found in the Nelore cattle rumen metagenome shared 64% similarity with an endo-1,4-β-xylanase/feruloylesterase from Prevotella ruminicola, which is a gram-negative bacterium and one of the most abundant genera in the bovine rumen 22 . The presence of the two glutamate residues (Fig. 1) in the conserved regions 159 WDVVNEA 163 and 280 TELD 283 reinforces the catalysis mechanism characteristic of the GH10 family based on the retention of the β configuration of the anomeric carbon 4 and signals the presence of this enzyme's active site as well as the conservation of these sequences in different species 23,24 . The presence of serine in the conserved GXSXG region (Fig. 2) signals the active site characteristic of carboxylesterases 25 .
XylR cloning was performed without the signal peptide predicted by SignalP 4.0 26 , which optimized the process of obtaining the enzyme from the soluble fraction of E. coli BL21 (DE3), which was purified, and  www.nature.com/scientificreports/  www.nature.com/scientificreports/ heterologous enzyme expression was confirmed by western blot analysis. XylR had endo-β-1,4-xylanase and esterase activities in short chain esters (Fig. 6), which confirms the theoretical predictions of the conserved domains and their bifunctional activity. The kinetic constant values for xylan from beechwood substrate (Table 1) indicate a high efficiency in xylan degradation when compared with those isolated and characterized from species such as Thermotoga thermarum (289 mg ml −1 s −1 ) 27 , Remersonia thermophila CBS 540.69 (417.4 mg ml −1 s −1 ) 28 and Bacillus sp. SN5 (142.6 mg ml −1 s −1 ) 29 . The affinity for p-nitrofenyl acetate indicates that XylR has sterasic activity and is capable of hydrolyzing small molecules containing ester bonds 30 , although its efficiency was similar to that of previously characterized esterases, such as from the fungus Rhizomucor miehei (0.37 mM −1 s −1 ) 31 and the bacterium Bacillus pseudofirmus OF4 (3.4 µM −1 s −1 ) 32 . The specificity for short-chain esterase substrates relative to that of endo-β-1,4-xylanase indicates that this enzyme has an esterase action in addition to that of xylanase, which helps in the digestibility of heteroxylans. Enzymes with xylanase/esterase bifunctional activity have been previously described 33,34 , and the importance of the synergism between esterase and xylanase activities in hemicellulose degradation has been characterized 35 . The xylan main chain has a high number of ferulic and cumaric acid ester radicals, and the addition of esterases increases their digestibility during bioconversion 13 . The synergistic action of xylanases and esterases is due to the creation of new xylanase binding sites after the cooperative removal of ester radicals between enzymes 36 . In this sense, the enzymatic hydrolysis of ester bonds becomes an important step towards the complete degradation of this type of hemicellulose, which helps to fully utilize lignocellulosic biomass.
The optimal hydrolysis pH of XylR endo-β-1,4-xylanase occurred in 0.1 M sodium acetate buffer at pH 6.5 (Fig. 7), and 50% of its activity was maintained between pH 5.5 and 6.5. This result was similar to that found for endo-β-1,4-xylanases with different origins and biochemical properties, including a thermostable enzyme of Marasmius sp. (0.1 M sodium acetate buffer, pH 6.0) 37 ; a GH10 family enzyme from Bacillus subtilis B10 (0.1 M citrate-phosphate buffer, pH 6.0) 38 and Kitasatospora sp. (0.05 M sodium acetate buffer, pH 6.0) 39 ; a salt-resistant strain from Bacillus subtilis cho40 (0.05 M citrate buffer, pH 6.0) 40 ; and an enzyme obtained from the chicken cecum metagenome (0.05 M sodium phosphate buffer, pH 6.5) 41 . This characteristic reinforces that endo-β-1,4xylanases activity occurs at acidic to neutral pH values, thus indicating the importance of this range of action in different environments.
The endo-β-1,4-xylanase from XylR exhibited a capacity for hydrolysis at 30 and 45 °C, with optimum hydrolysis occurring at 37 °C (Fig. 8) and stable hydrolysis occurring at 50 °C for one hour (Fig. 10). Hydrolysis activity at room temperature is an important characteristic for the solubilization of hemicelluloses, which would not need thermal control in the range of 30 to 45 °C. This result is similar to that for xylanases obtained from other microorganisms, such as Paenibacillus xylanilyticus KJ-03 (40 °C) 42 , Sorangium cellulosum So9733-1 (30 °C) 43 , and Bacillus sp. SN5 (40 °C) 29 .
The Ca +2 , Mg +2 and Mn +2 ions did not inhibit XylR activity, which indicates that this enzyme is tolerant to these metals. Ca +2 and Mg +2 ions are macronutrients normally found in lignocellulosic biomasses 44 , and stable activity of XylR in the presence of these ions reinforces its potential application in hydrolyzing biomasses of different origins. Additionally, the absence of an inhibitory effect of EDTA indicates that XylR is not a metaldependent enzyme, which could be an advantage in large-scale biotechnological processes because this enzyme would not require the addition of metal ions for enzymatic catalysis, thereby reducing the operational cost. However, the metallic ions Zn +2 , Cu +2 and Hg +2 caused total inhibition of XylR activity at low concentrations (Table 2); therefore, these ions should be avoided in biotechnological procedures using this enzyme. The presence of Cu +2 -inhibiting xylanases has been described in previous studies 39,45 .
The tolerance of XylR to 2 M NaCl in the reaction medium ( Fig. 10) and 2.5 M for up to 8 h (Fig. 11) indicates an important characteristic of this enzyme, thus demonstrating its feasibility for use in the biotechnological sector. Halotolerant xylanases have been previously reported 40,46,47 as feed additives for broilers 48 and other farm animals because the diets of these animals are routinely formulated with the inclusion of NaCl; moreover, they are included in pretreatments containing high concentrations of NaCl, such as for cellulose solubilization and depolymerization 49 , because they increase the efficiency and synergy between these pretreatments. Additionally, the tolerance of XylR to organic solvents (Fig. 12), such as methanol, ethanol and propanol, at concentrations up to 20% (v/v) increases the range of biotechnological processes to which it can be applied, such as the continuous production of lignocellulosic ethanol, in which the saccharification and fermentation steps occur in the same bioreactor 50 . Interactions between the hydrophobic ends of a protein and organic solvents can affect an enzyme's stability; however, XylR showed a low degree of hydrophobicity, which corroborates its resistance to the tested organic solvents.
Xylanases are used in various industrial processes, such as second-generation ethanol production processes 3 , xylooligosaccharide production 8 and xylitol acquisition 9 . To use lignocellulosic biomass as a raw material, different treatments are required for its solubilization. Currently, efficient treatments to facilitate the use of this material are generally based on chemical compounds that cause environmental contamination 2 ; thus, xylanases has been used for the depolymerization of hemicellulose in biotechnological processes.

Conclusions and future perspectives
In conclusion, data mining of the rumen metagenome identified XylR as a bifunctional enzyme xylanase/esterase that is tolerant to NaCl and organic solvents, thus indicating the diversity of biomolecules that can be exploited in these environments to overcome current barriers for the production of lignocellulosic biomass. Here, we demonstrate that XylR is able to more efficiently degrade the xylan skeleton and its radical esters in biomasses with a high hemicellulose content. The tolerance of this enzyme to high concentrations of salt makes indicates its versatility and suitability for pretreatments that usually require high concentrations of salts. The use of XylR www.nature.com/scientificreports/ could extend the composition of enzymatic cocktails for the production of lignocellulosic ethanol and food additives in animal feeds and lead to more efficient production of xylooligosaccharides. In future research, tests will be performed to assay the efficiency of this enzyme in the hydrolysis of lignocellulosic biomass using agricultural waste (e.g., straw and sugarcane bagasse) as raw material and in pretreatments similar to those used in industry to determine its practical applicability. To further understand its structure, X-ray crystallography analysis should be performed to determine its bifunctional activity and its active sites. Additionally, this analysis would help to provide new insights about the behavior of this enzyme under high concentrations of NaCl.

Material and methods
Bioinformatics analysis. The functions of the xylanase sequence were inferred based on its similarity to the reference sequences of endo-β-1,4-xylanase from the GH10 family, which was extracted from the American database of the National Center for Biotechnology Information (NCBI), and the sequence was then submitted to the database eggNOG orthologists 51 . Prospecting was performed using the metagenomic database of the Laboratory of Biochemistry and Plant Microorganisms (LBMP) and the Illumina HiScanSQ sequencing data set from Nelore cattle rumen (public data access: SRX818104). ORFs (open reading frames) were selected with an e-value of -30, aligned with CLUSTALW 52 and verified for enzymatic domains in the Pfam database 53 . The sequences were screened based on predictions in silico in terms of the conserved domains 54 , presence of signal peptides 26 and similarity to sequences others available at the NCBI 15 . The secondary structure prediction was performed based on multiple alignment of the sequences by the Clustal Omega tool 16 (https:// www. ebi. ac. uk/ Tools/ msa/ clust alo/), and assembly and prediction was performed using EsPript 3.0 17 .  58 . The samples were previously incubated at 100 °C for five minutes in sample buffer (62 mM Tris HCl (pH 6.8) containing 20% glycerol, 4% SDS, 5% β-mercaptoethanol and 0.02% bromophenol blue) and applied in a 10% polyacrylamide gel containing SDS. Protein separation was carried out by applying an electric field (100 V) for 2 h. The gels were stained using the Coomassie Blue method (0.2% Coomassie Brilliant blue, 40% methanol, and 10% acetic acid). Zymogram. Electrophoresis was performed on a polyacrylamide gel (6%) copolymerized with 1% xylan from beechwood in the absence of SDS under 100 V for 2 h. The sample was previously diluted in a 1:1 ratio with 62 mM Tris HCl buffer (pH 6.8) containing 20% glycerol and 0.02% bromophenol blue. The gel was incubated at 37 °C in 0.10 M sodium acetate buffer at pH 6.5 for 1 h and then stained with 0.1% Congo red.

Construction of the recombinant vector.
Determination of protein concentration. The protein concentration was determined using a commercial Bio-Rad kit based on the method proposed by Bradford 59 using serum albumin as a protein standard.
Determination of xylanase activity. Xylanase activity was determined by adding 20 µL of the enzyme at 0.1 mg/ mL to 0.10 M sodium acetate buffer (pH 6.5) containing 0.2% (w/v) xylan from beechwood (Meganzyme) for 15 min at 37 °C. Reducing sugars were quantified using the 1-3-dinitrosalicylic acid (DNS) method 60 , and reading was performed on a spectrophotometer at 540 nm. One unit of enzymatic activity was defined as 1 µmol of reducing sugar per minute per mg of protein under the previously established standard test conditions. The tests were carried out in triplicate, and each reaction included a control without the enzyme to measure the spontaneous hydrolysis of the substrate.
Effect of temperature and pH. The optimum pH for xylanase activity was determined using 0.1 M sodium acetate (pH 3.0-6.5), 0.1 M Tris HCl (pH 6.5-7.5) and 0.1 M AMPOL (pH 8.5-10). The optimum temperature was determined by placing the enzyme in 0.1 M sodium acetate buffer at different temperatures (4-95 °C) for 15 min. The parameters km and Vmax were determined using xylan from beechwood at concentrations of 0.025 to 5 mg/mL in 0.1 M sodium acetate buffer (pH 6.5).
Effect of metal ions. The influence of the ions Co +2 , Cu +2 , Hg +2 , Mg +2 , Mn +2 , and Zn +2 on the activity of xylanase was evaluated at concentrations of 1, 2, 3, 4 and 5 mM in the reaction medium. The relative activity was calculated from the control of each test, which did not include the addition of ions (100%).
Effect of sodium chloride. The effect of NaCl on xylanase activity was evaluated by adding NaCl at concentrations of 0.1 to 2 M to the reaction medium and exposing the enzyme to NaCl at concentrations of 0.1 to 2.5 M for 2, 4 and 8 h. Relative activity was calculated by adding to each assay a control that did not include the addition of NaCl (100%).
Effect of temperature. The effect of temperature on xylanase activity was determined by preincubation of XylR in the absence of the substrate at 45, 50, 55 and 60 °C for 15, 30, 45 and 60 min. After incubation, the activity was determined at 37 °C and the residual percentage was estimated.
Determination of esterase activity. The esterase activity was determined discontinuously in a spectrophotometer at 405 nm. The reaction was initiated by adding 20 µL of the enzyme at 0.1 mg/mL to 0.1 M sodium acetate buffer at pH 6.5 and 1 mM p-nitrophenyl acetate for 15 min at 37 °C. Substrate affinity was determined using the following substrates (1 mM, Sigma-Aldrich): p-nitrofenyl acetate, p-nitrofenyl butyrate, p-nitrofenyl valerate, p-nitrofenyl octanoate, p-nitrofenyl dodecanoate, p-nitrofenyl myristate, and p-nitrofenyl palmitate. One unit of enzymatic activity was defined as 1 µmol p-nitrophenol released per minute per mg of protein under previously established standard assay conditions. The tests were performed in triplicate, and each reaction included a control without the enzyme to measure the spontaneous hydrolysis of the substrate.
Determination of kinetic parameters. The catalytic parameters K m (Michaelis-Menten constant), V ma x (maximum reaction speed), kcat (catalytic constant) and kcat. K m −1 (catalytic efficiency) was determined for endoβ-1,4-xylanase activity with xylan from beechwood as the substrate, and the concentration varied from 0.025 to 2.5 mg/mL. For esterase activity, p-nitrofenyl acetate was used as the substrate, and the concentration varied from 0.75 to 5 mM. The data were tested using the F test (p < 0.05) to determine the best kinetic model: H 0 = Michaelis-Menten and H 1 = sigmoidal. The normality of the residues was verified by the Shapiro-Wilk W test. The hypothesis test and nonlinear regression of the data by the Michaelis-Menten equation were performed using GraphPad Prism Software, version 5.00 for Windows (GraphPad Software, San Diego, California, USA).