INTRODUCTION

Bioethanol can replace part of the consumption of traditional fuels such as oil or coal. It is possible to mix gasoline with bioethanol to offer savings in CO2 emissions of at least 50% compared to fossil fuels (Hernández Rodríguez & Hernández Zárate, 2008). However, be-cause bioethanol is traditionally manufactured with lignocellulose from crops such as corn, the price of food tends to increase, which becomes a reason to focus efforts on the use of lignocellulose from lignocellulosic waste for its production.

The lignocellulosic waste is composed of cellulose, hemi-cellulose and lignin and around 200,000 million tons are produced annually (Medina-Morales, Lara-Fernández, Aguilar, & de la Garza-Toledo, 2011; Saini, Saini, & Tewari, 2015). Colombia has a high production of lignocellulose, which can be used to produce bioethanol. If waste from Colombian industries such as citrus juice and rice husks were used, 136,045,500 L and 12,200,000 L of bioethanol would be obtained annually (Sánchez Riaño et al., 201 0). Taking into account the price of the gallon of ethanol of U$ 3.07 Colombian pesos (COP) (August 2018), approximately U$ 120,240,897 would be obtained (Sánchez Riaño et al., 2010).

Lignocellulose is currently hydrolyzed by mechanical breakdown followed by acid, alkaline and / or enzymatic hydrolysis (^{Persson, Tjerneld, & Hahn-Hágerdal, 1991}; ^{Cuervo, Folch, & Quiroz, 2009}; Castillo et al., 2012). Of these three methods, the most promising by specificity and high recovery of glucose is enzymatic hydrolysis, with the disadvantage that the production of enzymes has high costs (^{Gao et al., 2008}, G. ^{Liu, Qin, Li, & Qu, 2013}), which can be reduced if genetically modified microorganisms are used, to produce lignocellulolytic enzymes more efficiently in large quantities and in shorter times (Gao et al., 2008; ^{Zelena, Eisele, & Berger, 2014}).

Degradation of lignocellulose is carried out for a cellulolytic system composed of endoglucanase, exoglucanase and celubiosidase (Munjal, Jawed, Wajid, & Yazdani, 201 5).

According to the CAZy database, enzymes that degrade cellulose and other polysaccharides from plant cell wall, are classified into the glycose hydrolase families. In these GH families enzymes with the same enzymatic activity are not necessarily gathered, this classification is based on the sequence and structure of the protein, therefore families will understand different enzymatic activities (^{Busk, Lange, Pilgaard, & Lange, 2014}).

The objective of this work was to evaluate the expression of endoglucanase fungal gene in E. coli BL21 (DE3) bacterium for the saccharification of lignocellulose residues, through the modification and cloning of eukaryotic genes of lignocellulolytic enzymes.

MATERIALS AND METHODS

RESULTS AND DISCUSSION

Ten families sequences belonging to CAZY database were downloaded from corresponding database (Figure 1), the identification catalytic regions from endoglucanase enzymes from literature search was done, as a result, an eukaryotic gene was selected from Aspergillus nidulans (Table 2).

Figure 1 3D modeling for endoglucanase. (Left) l-TASSER best model for modified endoglucanase. C-score = 0.52 TM-score = 0.78 ± 0.09 RMSD = 5.9 ± 3.7 Á. (Right) Ramachandran graphic generated for the model. 

Endoglucanase presented only one intron and the modification of the genes resulted in a sequence with different length from the original (Berg et al., 2002).

I-Tasser program predicted ten template structures, all of which are cellulase enzymes with a z-score greater than 1, which indicates that the alignments have a high quality (^{Roy et al., 2011}). The program predicted 3 models for this enzyme, suggesting that the models used as a mold are remarkably similar to each other (Roy et al., 2011; J ^{Yang & Zhang, 2016}). The Ramachandran plot showed that the best model (Figure 1) presents most of the amino acids (which are not glycine or proline) in allowed areas, which indicates a good conformation (Mathews et al., 2002).

The models structurally similar to the problem protein (modified endoglucanase) were ten, which overlapped to the problem structure coincide almost in their entirety in their conformation, with TM-scores greater than 0.5, indicating the reliability of the prediction (J ^{Yang & Zhang, 2016}). The prediction of the function using the COACH server (Jianyi Yang et al., 2013) yielded 2 possible ligands: (i) Cellobiose. Identifier in the ChEBI, Pub-Chem and PDB databases: CBI (https://www3.rcsb.org/ligand/CBI). (ii) Beta-D-Glucose. Identifier in the ChEBI, PubChem and PDB databases: BGC (https://www3.rcsb.org/ligand/BGC).

The first of these substrates (ligands) corresponds to the cellulose monomer, which is the preferred substrate for endoglucanase enzymes and the second corresponds to the molecule that makes up the same cellobiose, so it also has affinity for this enzyme (^{Bohórquez Saval , 2012}). This shows that the active site of the protein was correctly modeled (^{Roy et al., 2011}). The residues in which the two similar ligands are located since both ligands share the amino acids located at positions 194, 218, 221, 232 and 376, which, as shown in Figure 2, are located in the central region of the protein.

Figure 2 Ligands identified for the modified endoglucanase protein. (Left) Docking site of modified endoglucanase. In blue, amino acids that share the three best ligands. In green, location of the CBI ligand. (Right) Binding sites of the best ligands selected by the l-Tasser for the endoglucanase 3D model. 

The identifiers for the biological function or Enzyme Commission Numbers (E.C) found by the I-Tasser were two: (i) 3.2.1.4 which corresponds to the family of the cellulase enzymes or endoglucanases. (ii) 3.2.1.91, which is the identifier of cellulose 1,4-beta-celobiosidases (non-reducing end). Both identifiers correspond to enzymes that degrade cellulose. The consensus of the Gene Onthology prediction yielded as result for the molecular function the term with the number GO: 0016162 which corresponds to the cellulose activity 1,4-beta-celobiosidase and the GO term: 0008810, which identifies the cellulase activity. For the biological process, Gene Onthology gave the term GO: 0030245 that corresponds to the catabolic process of cellulose. For the cellular component, the term GO: 0005576 was obtained, which indicates that the protein functions in the extracellular region. The predictions of Gene Onthology yielded results that prove the biological process and molecular function of cellulose degradation. The above corroborates that the modified endoglucanase enzyme that was modeled by the I-Tasser can perform the degradation function of the cellulose polymer.

The molecular docking of the best model obtained from the I-Tasser against the specific ligands selected yielded mixed results, being able to compare the positions obtained in the SwissDock with the actual location of the ligands obtained by crystallography (G. ^{Wu, Robertson, Brooks III , & Vieth, 2003}) for endoglucanase enzyme, with a good result (Figure 3).

Figure 3 Best predicted docking for cellobiose ligand with the modified endoglucanase enzyme. Green colored molecule is the cellobiose ligand. Violet molecule is docking predicted by the SwissDock. In dark gray color the residues involved in the interaction with the ligand are appreciated. 

PCR of the genes modified for cloning in pET151 TOPO vector, in the agarose gel, yielded bands of DNA fragments. eglB gene a band of ≈ 1300 bp was observed as expected. (^{Belancic et al., 1995}, Lockington et al., 2002, ^{Chávez et al., 2002}, Ruhl et al., 2013).

From each of the colonies obtained in this transformation, the colony PCR was conducted, verifying that the colonies contained the respective amplicon with the expected sizes. The sequencing of PCR products resulted in the sequencing of the genes of interest for one eglB gene in pET151 in E. coli Top10 identified as E4T3.

The other samples sent for sequencing showed exceptionally low quality of the sequencing and were not analyzed. The only sample obtained for the eglB gene coincided with the sequence of the eglB gene from the organism Aspergillus nidulans by 99% (Table 2).

The analysis of sample E4T3, using the CLC Main Work-bench program, to check the orientation of the gene, showed that it was in the correct orientation within the expression vector pET151 TOPO.

In the three tests of enzymatic activity endoglucanase in E. coli BL21 (DE3), the eight replicas evaluated from the experimental control of Bacillus xiamenensis (positive control) presented halos of hydrolysis of the medium, between 2.45 and 4 mm, after development with Congo red (^{Meddeb-Mouelhi et al., 2014}; ^{Montoya et al., 2014}), which was expected since their endoglucanase activity has already been proven (^{Lai et al., 2014}). (Figure 4). The eight replicates of the E. coli BL21 (DE3) transformed with the eglB gene in pET151 TOPO presented hydrolysis halos of between 0.24 and 0.46 cm on the CMC agar in the presence of the IPTG expression inducer (^{Studier & Moffatt, 1986}), similar to the hydrolysis rings obtained for experimental control.

Figure 4 hydrolysis haloes obtained from the third round of enzymatic activity of the endoglucanase recombinant enzyme. (a) Replies 1,2 and 3. Experimental control Bacilus xiamenensis. (b) Replicas 1,2 and 3. Negative control. E. coli BL21 (DE3) untransformed. (c) Replicas 1, 2 and 3. E coli BL21 (DE3) transformed with eglB gene in pET51 TOPO. (d) Replicas 1, 2 and 3. Positive control. E coli BL21 (DE3) transformed with pET151 TOPO vector. 

On the other hand, the eight replicas of the negative control, of E. coli BL21 (DE3) without transformation, presented faint hydrolysis haloes that were almost imperceptible, showing little growth of colonies on the agar, which may indicate contamination with another bacterium that is able to grow in the agar taking advantage of the sugars of the CMC, since this is a semi-selective agar, whose only carbon source is the CMC (^{Meddeb-Mouelhi et al., 2014}), which cannot be metabolized by the E.coli strain BL21 (DE3) untransformed. As regards the eight replicates of the positive control of E. coli BL21 (DE3) transformed with the plasmid pET151 TOPO with-out insert, they did not present any hydrolysis halo. (Figure 16). These results suggest that the endoglucanase enzyme was expressed correctly in E. coli BL21 (DE3) cells after the first transformation.

The statistical analysis was performed for the semiquantitative measurement of the hydrolysis of the recombinant endoglucanase enzyme. The ANOVA analysis resulted in statistical significative differences (p value <0.0001). Duncan's analysis showed the four treatments, grouped into three different groups. The values obtained by the comparison for each treatment were the following: (i) Treatment 1: 0 cm. (Group C) (ii) Treatment 2: 0.29625 cm. (Group B) (iii) Treatment 3: 0 cm. (Group C) (iv) Treatment 4: 0.39375 cm. (Group A). The treatment iv with the highest hydrolysis halo was E. coli BL21 (DE3) transformed with the modified eglB gene in the plasmid pET151 TOPO (Group A), followed by the experimental control Bacillus xiamenensis (Treatment 2, Group B), which presented the second largest hydrolysis halos.

CONCLUSIONS

The 3D structures predicted by the I-Tasser proved to be very close to the existing structures of similar enzymes in the PDB database, being able to demonstrate that they were structures that were correctly folded that did not undergo major changes in their conformation due to the elimination of introns and modification of codons, which shows that this process did not affect in silico the functionality of the enzymes. Regarding the ligation sites, predicted with the SwissDock, were highly conserved sites modeled correctly in the modified enzymes. The results of 3D modeling and docking allowed to corroborate in a preliminary way that these did not suffer important changes that affected their functionality during the process of gene modification.

The sequencing also allowed us to conclude that eglB genes cloned within the E. coli Top10 and E. coli BL21 (DE3) cells underwent very little alteration and remained almost identical to the originally modified genes. T7 promoter and the restriction analysis carried out for each of the constructs, allow to elucidate whether the genes were in the correct orientation within the plasmid for the eglB gene, which was in the correct orientation.

The expression of the modified endoglucanase enzyme was achieved in the expression cells E. coli BL21 (DE3), in soluble and functional form, demonstrated by hydrolysis of the CMC substrate. Cells from E. coli BL21 (DE3) transformed with the modified eglB gene in the pET151 TOPO vector were shown to form larger hydrolysis rings than the experimental control Bacillus xiamenensis. The expression and activity of enzymes sequences have to overcome many troubles such as possible incorrect orientation within the expression plasmid; the possible formation of inclusion bodies that could have prevented the secretion to the outside of the cell due to its size and recombinant nature; the natural defense mechanisms of E. coli to avoid toxic levels of enzymatic cofactors within its cytoplasm, which may prevent the correct folding of enzymes; the possibility that there has been an incorrect folding enzyme due to its recombinant nature. All these problems can be partially avoided if bioinformatics is applied to change amino acids without alter functionality and affinity for substrate. This is a step forward to produce endoglucanase in vitro, with a high efficiency because bacteria low cost and high producing system expression.