Computational Assessment of Botrytis cinerea Lipase for Biofuel Production

The demand for ecofriendly green catalysts for biofuel synthesis is greatly increasing with the effects of fossil fuel depletion. Fungal lipases are abundantly used as biocatalysts for the synthesis of biofuel. The use of Botrytis cinerea lipase is an excellent approach for the conversion of agroindustrial residues into biofuel. In this study, phylogenetic analyses were carried out and the physicochemical properties of B. cinerea lipase were assessed. Furthermore, the protein structure of B. cinerea lipase was predicted and refined. Putative energy-rich phytolipid compounds were explored as a substrate for the synthesis of biofuel, owing to B. cinerea lipase catalysis. Approximately 161 plant-based fatty acids were docked with B. cinerea lipase in order to evaluate their binding affinities and interactions. Among the docked fatty acids, the top ten triglycerides having the lowest number of binding affinities with B. cinerea lipase were selected, and their interactions were assessed. The top three triglycerides having the greatest number of hydrogen bonds and hydrophobic interactions were selected for simulations of 20 ns. The docking and simulations revealed that docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid had stable bonding with the B. cinerea lipase. Therefore, B. cinerea lipase has the potential to be used for the transesterification of fatty acids into biofuels, whereas docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid can be used as substrates of B. cinerea lipase for biofuel synthesis.


Introduction
Fossil fuels are unable to deal with the increasing demand for energy, as they are nonrenewable [1][2][3]. Biofuels are preferred over conventional fuels because of their nontoxic, biodegradable, and renewable natures. Moreover, biofuels offer lower impact in terms of global carbon footprints [3]. Biofuel synthesis is the best alternative for coping with fossil fuel consumption. According to a report, biofuel global production was increased by 10 billion liters in 2018, compared to 2017, to meet a record of 154 billion liters [4]. A twenty-five percent increase in biofuel production is forecasted by the year 2024, due to the better growth statuses of biofuel-producing countries, including Brazil, the United States, and, especially, China [5]. Among all the countries, Brazil has the prominent agricultural backbone to convert agroindustrial waste into renewable energy products [6]. Biobased residues, including fats and oils, are most frequently used for the synthesis of biofuels.

Multiple Sequence Alignment and Phylogenetic Analysis of B. cinerea Lipase
The physicochemical properties of B. cinerea lipase are represented in Table 1. Multiple sequence alignment revealed that lipases in Botrytis share higher identity within the genus. The active site, in the form of catalytic residues (G-X-S-X-G), is conserved in all fungal lipases. For the phylogenetic tree, multiple sequence alignment was used to generate a maximum likelihood tree on MEGAX software, as shown in Figure 1. The phylogenetic tree, based on the protein sequences of isolates, is represented in Figure 2. The tree was divided into subgroups. B. cinerea lipase showed a close relationship with Botrytis cinerea lipase, and forms a sister clade with the lipases of other Botrytis species, and with the lipases of many different fungal species. However, it showed a distant relationship with the lipases of Rutstroemia sp., Lepidopterella palustris, and Coleophoma crateriformis. The lipase from Vibrio coralliilyticus was an outgroup in the rooted phylogenetic tree. Table 1. Physicochemical properties of B. cinerea lipase. Physicochemical properties, including amino acid residues, molecular weight, theoretical pI, positively and negatively charged residues, molecular formula, instability index, aliphatic index (%), and GRAVY, are represented in the table.

Sr. No
Physiochemical Properties of B. fungal lipases. For the phylogenetic tree, multiple sequence alignment was used to generate a maximum likelihood tree on MEGAX software, as shown in Figure 1. The phylogenetic tree, based on the protein sequences of isolates, is represented in Figure 2. The tree was divided into subgroups. B. cinerea lipase showed a close relationship with Botrytis cinerea lipase, and forms a sister clade with the lipases of other Botrytis species, and with the lipases of many different fungal species. However, it showed a distant relationship with the lipases of Rutstroemia sp., Lepidopterella palustris, and Coleophoma crateriformis. The lipase from Vibrio coralliilyticus was an outgroup in the rooted phylogenetic tree.

Signal Peptide Removal
For B. cinerea lipase, the signal peptide cleavage site was found between the 27 and 28 aa residues, as depicted in Figure 3. The Sec/SPI score was 0.9965, which indicated the presence of a signal peptide in the protein sequence of B. cinerea lipase. Similarly, the Philius web server also predicted the same cleavage site.

Signal Peptide Removal
For B. cinerea lipase, the signal peptide cleavage site was found between the 27 and 28 aa residues, as depicted in Figure 3. The Sec/SPI score was 0.9965, which indicated the presence of a signal peptide in the protein sequence of B. cinerea lipase. Similarly, the Philius web server also predicted the same cleavage site.

Prediction of Secondary Structure of B. cinerea Lipase
The secondary structure of lipase by PSIPRED is shown in Figure 4. A total of 18 helices were found. Six beta sheets were also found.

Prediction of Tertiary Structure of B. cinerea Lipase
All the models were evaluated for the Q-mean, the ERRAT score, and Ramachandran plots of the predicted models. The scores for the top models of each method used for prediction are provided in Table 2. Robetta-TR model 1 had the best scores among all the

Prediction of Secondary Structure of B. cinerea Lipase
The secondary structure of lipase by PSIPRED is shown in Figure 4. A total of 18 helices were found. Six beta sheets were also found.

Prediction of Secondary Structure of B. cinerea Lipase
The secondary structure of lipase by PSIPRED is shown in Figure 4. A total of 18 helices were found. Six beta sheets were also found.

Prediction of Tertiary Structure of B. cinerea Lipase
All the models were evaluated for the Q-mean, the ERRAT score, and Ramachandran plots of the predicted models. The scores for the top models of each method used for prediction are provided in Table 2. Robetta-TR model 1 had the best scores among all the

Prediction of Tertiary Structure of B. cinerea Lipase
All the models were evaluated for the Q-mean, the ERRAT score, and Ramachandran plots of the predicted models. The scores for the top models of each method used for prediction are provided in Table 2. Robetta-TR model 1 had the best scores among all the tested models, as its ERRAT score graph and Ramachandran plots show in Figures 5 and 6. Therefore, Robetta-TR model 1 was further refined using the 3Drefine webserver. The predicted model is shown in Figure 7. Furthermore, the predicted model of the B. cinerea lipase was superimposed on the sterol esterase of Ohiostoma piceae, as shown in Figure 7. A similarity of 80% was selected as the default, and the RMSD values for the alpha carbons and backbone were 2.90 and 2.86, respectively. The binding sites, predicted by CASTp, are shown in Figure 8. Table 2. Thread-based and ab initio scores for protein modeling. Scores obtained from output models of I-TASSER, Phyre2, and Robetta ab initio TR are shown. Robetta-TR models with best scores were selected. tested models, as its ERRAT score graph and Ramachandran plots show in Figures 5 and  6. Therefore, Robetta-TR model 1 was further refined using the 3Drefine webserver. The predicted model is shown in Figure 7. Furthermore, the predicted model of the B. cinerea lipase was superimposed on the sterol esterase of Ohiostoma piceae, as shown in Figure 7. A similarity of 80% was selected as the default, and the RMSD values for the alpha carbons and backbone were 2.90 and 2.86, respectively. The binding sites, predicted by CASTp, are shown in Figure 8.

Prediction of Binding Pocket Site of Protein
The binding sites of the protein were identified by CASTp, as represented in Figure  9, and were embedded inside the proteins. The protein had large and deep binding sites, consisting of 100 amino acid residues. Figure 9. Boxplot of the top ten plant triglycerides. The top ten plant triglycerides showed the differential binding affinities pattern. Dicranin, docosahexaenoic acid, and hexadeca-7,10,13-trienoic acid showing the smallest binding energy values, which dictate a stronger binding interaction with the lipase.

Prediction of Binding Pocket Site of Protein
The binding sites of the protein were identified by CASTp, as represented in Figure 9, and were embedded inside the proteins. The protein had large and deep binding sites, consisting of 100 amino acid residues.

Prediction of Binding Pocket Site of Protein
The binding sites of the protein were identified by CASTp, as represented in Figure  9, and were embedded inside the proteins. The protein had large and deep binding sites, consisting of 100 amino acid residues. Figure 9. Boxplot of the top ten plant triglycerides. The top ten plant triglycerides showed the differential binding affinities pattern. Dicranin, docosahexaenoic acid, and hexadeca-7,10,13-trienoic acid showing the smallest binding energy values, which dictate a stronger binding interaction with the lipase. Figure 9. Boxplot of the top ten plant triglycerides. The top ten plant triglycerides showed the differential binding affinities pattern. Dicranin, docosahexaenoic acid, and hexadeca-7,10,13-trienoic acid showing the smallest binding energy values, which dictate a stronger binding interaction with the lipase.

Molecular Docking of Lipase with Plant Triglycerides
The top ten triglycerides having the least number of binding affinities associated with B. cinerea lipase were selected, and their binding affinities are shown as boxplots in Figure 9. The IUPAC names, molecular weights, PubChem IDs, and the numbers of hydrogen and hydrophobic interactions are shown in Table 3. Docosahexaenoic acid possessed the lowest binding energy (−7.6 kcal/mol), followed by dicranin (−6.7 kcal/mol), and hexadeca-7,10,13-trienoic acid (−6.3 kcal/mol). The protein-ligand interactions of the top three triglycerides, including docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid, are represented in Figure 10. Docosahexaenoic acid and hexadeca-7,10,13-trienoic acid both contained two hydrogen bonds, while dicranin had three hydrogen bonds. Docosahexaenoic acid and dicranin both had fourteen hydrophobic interactions, while hexadeca-7,10,13-trienoic acid had 9 hydrophobic interactions with B. cinerea lipase. The protein-ligand complexes were proceeded by molecular dynamics simulation in order to explore conformational energy landscapes.

Molecular Dynamics Simulations
The molecular dynamics simulations of the top three complexes revealed structure stability and fluctuations during the 20 ns simulations. Hexadeca-7,10,13-trienoic acid showed minor fluctuations at the start, but stabilized from 4.5 ns to 11.5 ns, according to the RMSD values shown in Figure 11. The structure fluctuated from 11.5 ns to 14 ns and again stabilized onwards, whereas the dicranin complex with the lipase was only stable from 9 ns to 17 ns. It fluctuated at the start and end of the simulation. Doxosahexaenoic acid was unstable initially, but then stabilized after~8500 ps, and remained stable to~18,500 ps.

Molecular Dynamics Simulations
The molecular dynamics simulations of the top three complexes revealed structure stability and fluctuations during the 20 ns simulations. Hexadeca-7,10,13-trienoic acid showed minor fluctuations at the start, but stabilized from 4.5 ns to 11.5 ns, according to the RMSD values shown in Figure 11. The structure fluctuated from 11.5 ns to 14 ns and again stabilized onwards, whereas the dicranin complex with the lipase was only stable from 9 ns to 17 ns. It fluctuated at the start and end of the simulation. Doxosahexaenoic acid was unstable initially, but then stabilized after ~8500 ps, and remained stable to ~18,500 ps.
The radius of gyration represents the compactness of the structure in Figure 12. The structures fluctuated between 2.40 nm and 2.50 nm. The Rg for hexadeca-7,10,13-trienoic acid was lower at the start, representing that the structure is compact, wheras the Rg values for docosahexaenoic acid and dicranin showed overall fluctuations. However, in the end, the compactness was reduced as the Rg value increased.
The lipase complex with docosahexaenoic acid became more compact towards the end of the simulation. Similarly, the compactness of the lipase complex with dicranin increased towards the end of the simulation.
The root mean square fluctuation of docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid in the complex showed stability, as shown in Figure 13. Only a few residues at the start fluctuated, but overall the structure was stable. Other minor fluctuations were less than 0.75 nm, indicating the stability.  The radius of gyration represents the compactness of the structure in Figure 12. The structures fluctuated between 2.40 nm and 2.50 nm. The Rg for hexadeca-7,10,13-trienoic acid was lower at the start, representing that the structure is compact, wheras the Rg values for docosahexaenoic acid and dicranin showed overall fluctuations. However, in the end, the compactness was reduced as the Rg value increased.
The lipase complex with docosahexaenoic acid became more compact towards the end of the simulation. Similarly, the compactness of the lipase complex with dicranin increased towards the end of the simulation.
The root mean square fluctuation of docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid in the complex showed stability, as shown in Figure 13. Only a few residues at the start fluctuated, but overall the structure was stable. Other minor fluctuations were less than 0.75 nm, indicating the stability. Figure 11. Root mean square deviation (RMSD) plot interactions with dicranin (red), docosahexaenoic acid (blue), and hexadeca-7,10,13-trienoic acid (green) complexes, with protein's backbone.

Discussion
Fungal lipases are advantageous over other microbial lipases because of their potential to utilize free fatty acids, monoglycerides, diglycerides, and triglycerides during the transesterification reaction at higher activity and maximum yields. These lipases can catalyze the reaction in nonaqueous media at lower temperatures, within shorter times, and ease mass biofuel production [25,26]. Approximately more than 80% plant triglycerides are used for diesel fuels because of their availability, higher heat content, biodegradability,

Discussion
Fungal lipases are advantageous over other microbial lipases because of their potential to utilize free fatty acids, monoglycerides, diglycerides, and triglycerides during the trans-esterification reaction at higher activity and maximum yields. These lipases can catalyze the reaction in nonaqueous media at lower temperatures, within shorter times, and ease mass biofuel production [25,26]. Approximately more than 80% plant triglycerides are used for diesel fuels because of their availability, higher heat content, biodegradability, lower sulfur content, renewability, and lower aromatic content [27,28]. Current research reveals that plant triglycerides, such as dicranin, docosahexaenoic acid, and hexadeca-7,10,13-trienoic acid, were potent feedstock for B. cinerea-lipase-mediated biofuel production.
B. cinerea lipase represent a 50 to 60% homology with lipases from filamentous fungi, and a 35 to 45% homology with lipases from C. rugosa and G. candidum [23]. In the current study, the lipases from two different strains of Botrytis represented a 100% homology. Hence, the lipases from any Botrytis strain can contribute in similar manners for biofuel production, as these are very similar.
Because of the great advancements in in silico approaches, this study was designed to decipher the interactions between B. cinerea lipase and the plant triglycerides that have been previously reported [29]. Recently, fungal lipases obtained from Mucor circinelloides, f. circinelloides, and Rhizopus oryzae interacting with triglycerides were also modeled using computational approaches [30]. The three-dimensional protein structures of various lipases from ascomycetes, such as Ophiostoma piceae, Melanocarpus albomyces, and Candida rugosa, have been resolved [31][32][33]. However, B. cinerea lipase has not been structurally characterized. In earlier studies, protein homology modeling was used for the prediction of the tertiary structure for those not already resolved [34]. For this purpose, SWISS MODEL was chosen for homology modeling, and the resultant homology was found to be 58% with Ophiostoma piceae sterol esterase. The obtained value is higher than other recently modeled lipases: 32% for Arabidopsis thaliana lipase, and 32.11% for Streptomyces rimosus lipase [35,36]. The protein modeling for B. cinerea lipase was also performed using thread-based and ab initio approaches after first removing 17 amino acids from the signal peptide. I-TASSER, Phyre2, and the Robetta ab initio and TR methods were chosen for the determination of the protein tertiary structure, but the final model was selected from the Robetta-TR method having the highest scores, as shown in Table 2.
An ERRAT-based statistical analysis was performed for the quality validation of the selected model. This program analyses the nonbonded interaction patterns of various types of atoms [37]. The overall quality factor assigned to the B. cinerea lipase model was 93.236%, as shown in Figure 6, which is a good value for a predicted protein structure. Similarly, the quality factor for Glaciozyma antarctica lipase was found to be 91.3% [38]. A Q-mean-based quality estimation was also performed, and the best fit model represented a 0.68 value, which is above the 0.52 reported for Nocardiopsis alba lipase modeling [39]. Therefore, the quality of the selected model was further assessed by a Ramachandran plot. The Ramachandran plot obtained showed that 83.0% of the residues were found in the most favored regions. A total of 14.4% of the residues were in the additional allowed regions, 1.1% residues were found in the generally allowed regions, and 1.5% of the residues were found in the disallowed regions, as shown in Figure 7. Similarly, the crystal structure of feruloyl esterase from A. niger represented 81.1% of the residues in the most favored regions, 15.8% in the additional allowed regions, and 0.9% in the disallowed regions [34,39].
The isoelectric point predicted for B. cinerea lipase was 4.93, which is below 7.0, an indication of its acidic nature, as was reported very recently for β-glucosidase and xylanase from Trichoderma asperellum [40,41]. The instability index value was 36.20%, validating the stable lipase, while the aliphatic index was 92.21%, proving the thermal stability of the lipase because the higher aliphatic index (>40%) presents thermostable tertiary protein [42]. The grand average of the hydropathicity (GRAVY) indicates the sum of the hydropathy values of the protein. Negative and positive GRAVY values indicate hydrophilicity and hydrophobicity, respectively [43]. The values for the negatively and positively charged residues were 38 and 30, respectively. These values correspond to the hydrophilic or hydrophobic nature of the protein. The B. cinerea lipase protein had a positive GRAVY value, which corresponds to its hydrophobic nature, while the GRAVY values reported for the recombinant lipases, EXANL1 and A. niger F044 lipase lip1, were 0.025 and 0.05, respectively, which also corresponds to a hydrophobic nature [40,44].
Finally, the top ten fatty acids were selected having the least number of binding affinities from 161 fatty acids docked with B. cinerea lipase. The resultant affinity energy values of the lipase with triglycerides were consistent, and even greater than recently reported in the literature. Lower affinity values exhibit higher interactions between the fatty acid and the lipases [45,46]. The docking affinity values of the lipase, with different fatty acids, such as palmitic acid, stearic acid, lauric acid, and myristic acid, were −5.9, −3.7, −3.5, and −3.2 kcal/mol, respectively. The current resultant affinity values for docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid are −7.6, −6.7, and −6.3 kcal/mol, respectively. Therefore, the resultant values confirm that B. cinerea lipase shows promising interactions with plant triglycerides for biofuel production.
The molecular docking results were further assessed for quality estimation through docking accuracy and screening enrichment procedures [17]. Lipases contain the consensus sequence, G-X-S-X-G, which presents the active residue-forming catalytic site. The residues include active glycine (G) and serine (S), in combination with any other (X) [47]. The active residues shown in Figure 1 present G-X-S-X-G as a conserved motif, whereas the active residues of the B. cinerea lipase were comprised of the binding pockets of 100 amino acid residues containing the G-S-X-S-X-G motif, from residue no. 225 to 232, as predicted by DoGSiteScorer [48] and CASTp analysis [49].
The catalytic activity of hydrolases is known to depend upon the transition state stability caused by the hydrogen bonds involved. To assess the transition state stability, protein interactions with the top three ligands, docosahexaenoic acid, dicranin, and hexadeca-7,10,13-trienoic acid, were predicted through molecular docking. These interactions represent the active sites of B. cinerea lipase in the form of amino acid residues, which were directly interacting with ligands, accompanied by a number of hydrogen bonds and hydrophobic interactions, as shown in Figure 10. Docosahexaenoic acid has two hydrogen bonds interacting with Val 263 and Ala 265, and fourteen hydrophobic interactions. Moreover, hexadeca-7,10,13-trienoic acid also has two hydrogen bonds interacting with Val 263 and Ala 265, and nine hydrophobic interactions, whereas dicranin has three hydrogen bonds interacting with Val 286 and Ser366, and fourteen hydrophobic interactions with B. cinerea lipase. These results were in accordance with the reported studies where lipase-based transesterification depended on three hydrogen bonds for the stable transition state [50,51]. Hence, protein ligand complexes provide valuable information about the overall interactions taking place at the molecular level.
Molecular dynamics simulation was then chosen for providing insights into the dynamic structure and protein-ligand interactions for recognizing the structure-based functionality. The stability of these structures was confirmed by 20 ns simulation, through the MD simulation protocol [52]. The structural stability of the protein was evaluated through the RMSD and RMSF values of the given protein-ligand complex [53]. The compactness of the protein structure was indicated by the Rg value during the simulation, which has an inverse relation with the stability of the structure. B. cinerea lipase complex with dicranin and docosahexaenoic acid had compact structures towards the end of the simulation, whereas hexadeca-7,10,13-trienoic acid displayed reduced compactness towards the end. The RMSD values of dicranin presented stability from 9 ns to 17 ns, docosahexaenoic acid showed overall stability from 5 ns to 15 ns, with minor fluctuations at the start, whereas hexadeca-7,10,13-trienoic acid presented stabilized values from 4.5 ns to 11.5 ns, which was the highest stability without any fluctuations. This indicates that, as the complex was stabilized, the lipase protein's compactness was reduced.
The RMSF values of the three ligands showed major fluctuations at the start, around 100, and minor fluctuations of around 250 and 400, which were less than 0.75. The MD simulation revealed that hydrogen bonds significantly affect the intermolecular recognition and active center for the lipase-substrate interaction. It can be concluded that dicranin, docosahexaenoic acid, and hexadeca-7,10,13-trienoic acid indicated overall stable structural confirmations with B. cinerea lipase. Hence, B. cinerea lipase can be utilized as a potent biocatalyst to process feedstock substrates for biofuel production.

Materials and Methods
Detailed flowchart of methodology is represented in Figure 14.

Multiple Sequence Alignment and Phylogenetic Analysis of Lipase in Botrytis Cinerea
The FASTA protein sequence of the lipase from Botrytis cinerea was retrieved from NCBI (Accession ID: AAU87359.1) to assess its function in biofuel production.  [54]. The pairwise and multiple alignment score for the gap opening penalty was 10.00 for both, whereas the gap extension penalty was 0.10 and 0.20, respectively. Using multiple sequence alignment, phylogenetic analysis was performed using MEGAX (version 10.1.7) through the maximum likelihood statistical method and the Jones-Taylor-Thornton (JJT) model as an amino acid-based substitution model. The rooted tree was generated with the bootstrap values of 10,000 replicates.

Multiple Sequence Alignment and Phylogenetic Analysis of Lipase in Botrytis cinerea
The FASTA protein sequence of the lipase from Botrytis cinerea was retrieved from NCBI (Accession ID: AAU87359.1) to assess its function in biofuel production. , and Vibrio coralliilyticus (NOI18297.1) were retrieved from NCBI (https://www.ncbi.nlm.nih.gov/, accessed on 5 June 2021) for multiple sequence alignment. The multiple sequence alignment of all these lipases was carried out using ClustalW plugin in MEGAX software (version 10.1.7) [54]. The pairwise and multiple alignment score for the gap opening penalty was 10.00 for both, whereas the gap extension penalty was 0.10 and 0.20, respectively. Using multiple sequence alignment, phylogenetic analysis was performed using MEGAX (version 10.1.7) through the maximum likelihood statistical method and the Jones-Taylor-Thornton (JJT) model as an amino acid-based substitution model. The rooted tree was generated with the bootstrap values of 10,000 replicates.

Tertiary Structure Prediction of B. cinerea Lipase
Previously, the crystal protein structure of lipase from B. cinerea was not resolved. A homology search was carried out using the BLASTp, and B. cinerea lipase showed 58% similarity with Ophiostoma piceae sterol esterase. A lower homology with the template crystal structure can result in a low-quality model. Therefore, the three-dimensional structure of B. cinerea lipase was determined using thread-based and ab initio approaches. The B. cinerea lipase sequence was used for the tertiary structure prediction after removal of the signal peptide. I-TASSER, Phyre2, and the Robetta web server were used to determine the tertiary structure of the lipase. I-TASSER (https://zhanglab.dcmb.med.umich.edu/ I-TASSER/, accessed on 7 June 2021) is a thread-based tool that uses a hierarchy-based approach to predict the protein structure and annotate its structure-based function [58]. Phyre2.0 uses the remote homology detection method to predict the three-dimensional structure of proteins. The Robetta web server (https://robetta.bakerlab.org/, accessed on 9 June, 2021) was used to determine the ab initio and thread-based structure prediction of B. cinerea lipase [59].

Binding Site Prediction
The refined protein model was analyzed for binding site prediction using the CASTp webserver (http://sts.bioe.uic.edu/castp/index.html?1bxw, accessed on 13 June 2021) [49]. DoGSiteScorer web server (https://bio.tools/dogsitescorer, accessed on 13 June 2021) was used for the validation of binding sites [48]. Binding site prediction gives information on the active residues involved in the catalytic site for ligand attachment.

Protein Preparation
The refined protein structure was prepared and converted into .pdbqt format from .pdb format using the MGL tool (http://mgltools.scripps.edu/downloads, accessed on 19 June 2021) [65]. Water molecules were removed, whereas polar hydrogens were added to the protein.

Molecular Docking
Molecular docking was performed to investigate the appropriate ligands that energetically fit into the protein binding site. AutoDock Vina [65] was used for molecular docking. The top ten compounds having the least number of binding affinities were selected, and a boxplot was generated using R-studio with ggplot2 [66]. The output .pdbqt were combined with the respective protein, and the protein-ligand interactions were determined using Protein-Ligand Interaction Profiler (https://plip-tool.biotec.tu-dresden.de/plipweb/plip/index, accessed on 30 June 2021).

Molecular Dynamics Simulation
A molecular dynamics simulation of 20 ns was performed to investigate the backbone stability of B. cinerea lipase with the top three ligand-protein complexes having the least number of binding affinities. GROMACS 5.1 was used for the molecular dynamics simulations [67]. CHARMM 36-feb2021 forcefield was used for the protein topology preparation by using pdb2gmx. CHARMM General Forcefield CGenFF online server ( https://cgenff.umaryland.edu/, accessed on 5 July 2021) was used for the initial ligand topology. Ligand and protein were combined and centralized in the dodecahedron. 2 Na + counterions were added to neutralize the system. The steepest descent minimization algorithm was used to neutralize the energy of the system. To carry out the equilibrium of ions and the solvents, a position-restraining simulation was performed. NVT (constant particles number: N; volume: V; and temperature: T) was executed for 100 ps in a 300 K temperature bath with a constant coupling of 0.1 ps. NPT (constant particles number: N; pressure: P; and temperature: T) was executed at 1 bar constant pressure and the coupling constant of 0.2 ps. Final molecular dynamics simulation was performed for 20 ns. The root mean square fluctuations (RMSF), the root mean square deviation (RMSD), and the radius of gyration were calculated between the protein and ligands by using GROMACS.

Conclusions
The conversion of plant triglycerides into biofuel using B. cinerea lipase as a catalyst could be a promising approach for the assimilation of plant-based fatty acids as precursor raw material for environmentally sustainable biofuel or biodiesel production. In the current study, molecular docking and molecular dynamics simulations were performed for identifying the potential of B. cinerea lipase as a biofuel-catalyzing agent. The MD simulations results indicate the stable RMSD between the ligand and protein. Hence, it is inferred that B. cinerea lipase is a promising catalyst for plant fatty acid residual bioconversion into biofuel.