Team:SMS Shenzhen/Model1
We apply another engineering design cycle aside from the cycle described in wetlab to improve the HBT affinity of Laccase CotA by rational design. The engineering circle we apply includes six steps (Figure 1): Research, Imagine, Design, Build, Test, and Learn. In this cycle, we use our research results to construct proper mutants that could increase the Laccase CotA's substrate (HBT) affinity utilizing Discovery Studio, a protein-design tool. Then those candidate mutants are actualized in wetlab quantitively to test the increase in HBT affinity. Based on testing results from both software and wetlab, we provide further improvements that can be done to develop better protein-design procedures. The improvement in substrate affinity can make our gum cleaning products work more efficiently, thus reducing the time and labor costs.
Figure 1 | The Steps of the Engineering Cycle
Research
For the "Research" stage, we investigate existing macromolecular compounds degrading proteins and evaluate their feasibility for our chewing gum degradation process, which eventually lead us to select the Laccase CotA due to its good performance on degrading PE and high thermal stability. A complete view of the "Research" stage can be seen on Wet Lab-Design.
Imagine
For the “Imagine” section, we consider the reason why we choose to enhance HBT affinity using rational design and provide valuable improvements that we can make to Laccase CotA.
Motivation for improving Laccase CotA
Laccase CotA will be implemented in our gum cleaning tools to clear the chewing gums on the ground. We want this enzyme to be as efficient and robust as possible from a convenient and environmental standpoint. To achieve this, we need to ask the following questions:
- Is the degrading speed of our Laccase CotA fast enough?
- Can unit amount of Laccase CotA degrade large amount of EVA at a given time?
- Are the operating costs of our cleaning tool low enough to be appealing?
From a biochemical perspective, these questions can be answered by improving the affinity and optimal operating conditions of Laccase CotA.
How to improve substrate affinity?
In order to know which method can be used to improve the affinity of the substrate, we must first understand the factors that determine the affinity of the substrate. Substrate affinity is controlled by the chemical interactions between a protein's binding residue and the substrate. Consequently, the factors that may influence binding affinity include:
- The type of the molecules.
- The shape of the protein and the ligand.
- The conditions of the enzymatic reaction.
First of all, since our enzymes and substrates——Laccase CotA and HBT——have been identified, we cannot change the type of the molecules. Second, since the reaction conditions are limited to the conditions inside chewing gums, the enzymatic reaction condition cannot be altered as well. As a result, the only way left for us is to change the shape of the binding site, which requires the mutation of amino acids.
According to our research, rational protein design and directed evolution are the methods that we can implement for amino acid mutation. Rational design mainly uses the understanding of protein and ligand's chemical interactions to predict specific protein mutation sites. In contrast, directed evolution relies more on the random mutation during the error-prone PCR. Thus, considering the significant differences between the two methods, we decide to apply both in our project to increase enzyme activity as much as possible. The following text illustrates our work on enhancing binding affinity through rational design (directed evolution can be seen on Wet Lab-Engineering.
The following are an overview on the steps of rational design:
- 1. Laccase CotA's three-dimensional structure determined by homology modeling is used as a starting point to begin selecting critical residues involved in ABTS coordination and binding. This is done by analyzing residues adjacent to the copper ions that are close enough to aid in the direct coordination of ABTS.
- 2. These residues are selected, and adjustments are made using the mutagenesis and measurement tool to determine possible amino acid substitutions that can be made to enhance the binding site affinity for ABTS.
- 3. Those possible mutation sites are actualized in wetlab.
Design
For the “Design” stage, we use the knowledge developed during the “Imagine” phase to propose specific changes to our Laccase CotA and prepare for simulated amino acid mutation.
Investigating the Three Dimentional Structure of Laccase CotA Using Homology Modeling
Based on our literature review, there is no existing 3D structure obtained through X-ray or Electron Microscopy. Therefore, we need to apply homology modeling to get the 3D structure of Laccase CotA for further investigation.
Protein homology modeling is that when a protein with an unknown structure and a protein with a known structure are relatively similar in the primary sequence, the protein with a general structure can be used as a template to predict the unknown protein's 3D structure through computer simulation and calculation. This simulation is reliable mainly because of two reasons: 1. Its unique amino acid sequence determines the structure of a protein. If the primary sequence is known, the secondary structure and tertiary structure can theoretically be obtained. 2. The tertiary structure of a protein is more conserved in evolution than its primary sequence. The research has shown that if the amino acid sequences of two proteins are 50% identical, then about 90% of the a- carbon atoms will have a positional deviation of no more than 3 Å. [1]
We utilize SWISS-MODEL to run homology modeling automatically by inserting the protein sequence of our Laccase CotA. The corresponding two results are shown in figure 2.
Figure 2 | Homology Modeling Result A from SWISS-MODEL
To select the most favourite simulation, we have chosen two scoring systems——GMQE (Global Model Quality Estimate) and QMEANDisCo Global——in SWISS-MODEL to estimate our model. Both of the scoring systems give an overall model quality measurement between 0 and 1, with higher numbers indicating higher expected quality. GMQE is coverage dependent, i.e. a model covering only half of the target sequence is unlikely to get a score above 0.5. QMEANDisCo, on the other hand, evaluates the model 'as is' without explicit coverage dependency.[10] Then, looking at the scoring of our two models (at the upright corner of each graph), we can conclude that the first one, which has 0.98 for GMQE and 0.92 for QMEANDisCo Global, is a better 3D model for our Laccase cotA since it has scores closer to 1. The simulated 3D structure of Laccase cotA can be seen below.
Figure 3 | 3D structure for Laccase cotA(This 3D structure is performed on PyMOL.)
Figure 4 | The four copper atoms (in yellow) in Laccase cotA's binding site
Once we obtain the 3D structure, we can better understand the oxidation mechanisms of Laccase CotA. In our model, Laccase CotA contains four copper atoms in their active site, which mediate the redox process and are classified into three groups according to their magnetic and spectroscopic properties. Figure 4 shows the three types of copper coordination in Laccase cotA: type 1 (Cu 1), type 2 (Cu 4), and type 3 (Cu 2 and Cu 3). Among them, type 1 copper is responsible for the substrate oxidation and redox potential of Laccase CotA. Consequently, an applicable mutation should generate changes in the electron transportation around Cu 1 atom. [2]
Discovery Studio, AMDOCK, PyMOL —— Protein Design Tools
In order to analyze the proposed design changes from the rational design process, software tools are used to provide initial analysis and modeling of the structure. In our project, among various computational protein design platforms, Discovery Studio is chosen to redesign the structure and conduct energy analyses; AMDOCK is chosen to conduct molecular docking; PyMOL is chosen to visualize proteins and ligands in different steps rational design. Since those software suite offer algorithms and protocols for macromolecular structure prediction and analysis of protein structures, they can significantly help us select the best mutation among numerous candidates.
Investigating HBT Binding Using Molecular Docking and Multiple Sequence Alignments
Since there is no specific information about HBT's binding site on Laccase CotA, we decide to utilize molecular docking to estimate a possible binding site. Molecular docking is a method of protein design based on the characteristics of the receptor and the interaction between the receptor and some small molecules. It is a theoretical simulation method that mainly studies the interaction between molecules (such as ligands and receptors) and predicts binding mode and affinity. In molecular docking, the software can indicate the potential binding sites on the protein according to the protein's cavities. Then, by conducting a molecular docking program in the software, one can test the feasibility of the simulated binding site.
It can be seen that molecular docking is a valuable tool to identify potential binding sites for HBT on Laccase CotA. Therefore, in the following content, we will show the detailed docking process.
The following is a sketch of our docking ligand, HBT (respectively, Laccase CotA is the receptor) using PyMOL.
Figure 5 | Structure of HBT
First of all, according to the 3D structure of Laccase CotA and HBT, ten possible binding sites are derived from molecular docking in AMDOCK. Table 1 shows the Grid Box volume (in cubic angstrom) and energy per volume (EPV) (in kcal/mol) for each possible binding site. Figure 6 shows the relative position of the then binding sites in Laccase CotA.
Table 1 | Properties of each possible binding site
Figure 6 | HBT binding sites in Laccase CotA (the one in pink cube is binding site 1)
Since EPV represents the energy released when binding at that position, the lower the EPV, the stabler the ligand-receptor system will be. However, this calculation is based only on the structure of Laccase CotA and HBT, so we conduct an actual molecular docking for those ten binding sites to obtain a reliable result. The following Table 2 shows the result of our molecular docking.
Table 2 | Molecular docking results
In the table, Affinity represents the binding energy of HBT and Laccase CotA, in which a lower affinity value means better stability. Estimated Ki (inhibitor constant) represents the degree to which protein and ligand bind. The smaller the Ki value, the tighter the two molecules combines. Ligand Efficiency (LE) is a critical information parameter in selecting a pilot complex. It is calculated using the following formula, in which delta G is free energy, HA is the number of non-hydrogen atoms in ligands. (Results with LE < -0.3 are considered to be potential binding sites.)
Based on the above analysis, 7 would be the most optimal binding site, and the following figure is the 3D structure of it.
Figure 7 | The most optimal HBT binding site
Additionally, by using the multiple sequence alignments which uses an optimization algorithm to identify regions of homology/similarity between different amino acid sequences, we re-test the validity of the binding site that we discover. We conduct the MSA for Laccase CotA from Bacillus sp. HR03 with the two other Laccase CotA proteins:
- WP_003219376.1: outer spore coat copper-dependent Laccase CotA [Bacillus subtilis subsp. spizizenii ATCC 6633 = JCM 2499
- NP_388511.1: outer spore coat copper-dependent promiscuous Laccase [Bacillus subtilis subsp. subtilis str. 168]
As the result shows below, all of the amino acids in our simulated binding sites (marked in yellow) are the same in the three Laccase CotA protein. Therefore, the fourteen amino acids shown in Figure 7 is the potential binding site of HBT.
Figure 8 |Multiple sequence of alignment of several Laccase CotA proteins
(This MSA is performed on Geneious using the Clustal Omega algorithm. Residues are coloured based on hydrophobicity, where a scale of red to blue = most hydrophobic to least hydrophobic. The amino acids related to HBT binding are painted in yellow boxes.)
Determine Residues Before And After Mutation
Since the amino acid residues that play a role in the binding are shown to be adjacent to HBT, mutations should be carried out in residues within a certain distance of HBT. Besides, these mutated amino acids are best not involved in the binding of Cu2+. This is because we usually do not want to alter the primary mechanisms of Laccase CotA.
Figure 9 |Structure of the copper coordination network of the Laccase molecule. [2]
As we can see in figure 9, the amino acids that are involved in Cu2+ binding is completely different from that around the HBT binding site. Therefore, we decide to select the fourteen labeled amino acids in Figure 7 as mutation candidates: THR-87, ILE-88, HIS-89, HIS-90, SER-91, ASP-379, GLU-380, TYR-381, ARG-383, ARG-392, TRP-393, HIS-394, TYR-500, ASP-501. To make the ultimate mutation result convincing, we mutate those candidates into twenty common amino acids, including ALA, ARG, ASN, ASP, CYS, GLN, GLU, GLY, HIS, ILE, LEU, LYS, MET, PHE, PRO, SER, THR, TRP, TYR, VAL.
Build
For the “Build” stage, we use the protein modeling software Discovery Studio to create the mutated proteins chosen in the “Design” stage, and perform a check for HBT affinity.
Single Residue Mutation
The heat map below shows the simulated single mutation energy terms derived from Discovery Studio. Each row corresponds to the targeted mutation residues that original residues are mutated into, and each column corresponds to the original residues.
Figure 10 |Heat map representing single mutation energy
Using this heat map, it is possible to rationally select mutations that can combine to further reduce the mutation energy. In the heat map, the redder the color is, the lower the mutation energy will be. However, we cannot choose more than one mutation from one column because one amino acid cannot be mutated into two other amino acids simultaneously. Therefore, as it is labeled in black border in the heat map, THR87>GLU (mutation energy: -1.51 kcal/mol) is the best single mutation.
Double Residue Mutation
To determine a suitable combination of two mutations, we simulate double mutation from the five best single mutations (THR87>GLU, ILE88>CYS, HIS394>PHE, TRP393>ARG, ASP501>PHE). The corresponding results are shown in Table 3 below. Since there are only three different combinations, there is no need of designing a heat map.
Table 3 |Mutation energy of double mutations
In the ascending table, A:THR87>GLU & A:TRP393>ARG has the lowest mutation energy among the ten combinations and is considered the best double mutation.
Triple Residue Mutation
From the mutation energy results for double mutation, we find that double mutation resulted in the even better. Therefore, we wonder if the combination of the three single mutations will result in better triple mutations as well. Thus, the simulation of triple mutation based on the five best single mutation is conducted in Discovery Studio, in which TRP393>ARG & HIS394>PHE & ASP501>PHE is the most optimal triple mutation.
Table 4 |Mutation energy of triple mutations
Protein Mutation Procedures
Using the official guide from BIOVIA Discovery Studio, we generate the following amino acid mutation protocols.
- Obtain the original Protein Data Bank (PDB) file of Laccase CotA and HBT from homology modeling (the latter one is used in our project).
- Prepare the structure of Laccase CotA.
- Delete water molecules from the protein.
- Clean Laccase CotA. (This includes removing protein's multiple conformation, supplementing incomplete amino acid residues, and adding hydrogens.)
- Apply the CHARMm forcefield to Laccase CotA.
- Select amino acids candidates and run single/double/triple mutation under pH=4.5 .
- Analyze the mutation energy result.
Test
For the "Test" stage, we implement our mutated candidates from "Build" stage in wetlab to further demonstrate the effect on increasing ABTS affinity.
The wetlab experiments can be parted in three——the expression and purification of Laccase CoA with mutation of THR87>GLU, THR87>GLU & TRP393>ARG, TRP393>ARG & HIS394>PHE & ASP501>PHE. Although, for a deliberate consideration of time and costs, we do not complete the quantitative enzyme activity tests for our mutations, those mutated Laccase CotA are proved to be active.
Learn
For the "Learn" stage, we reflect on the successes and failures of the approach used in the previous steps and suggested further work that can be done to develop even better Laccase CotA mutants.
Although we haven't got any results from wetlab, the above experiences of protein design still provide us with space for reflection. First, time constraints limit the number of amino acid residue candidates we can mutate with Discovery Studio. Therefore, we would like to investigate as many residues as possible. To achieve this goal, we will further consider other potential HBT binding sites. With a larger number of residue candidates, we will simulate feasible single mutations and double mutations as many as possible. Besides, we can continue to explore the functions of Discovery Studio on protein engineering. In this project, the Discovery Studio simulates molecular docking, amino acid mutation, and scoring mutations. An additional function that we can further investigate is the homology protein modeling, which we implement in SWISS-MODEL this year. Implementing this protocol in Discovery Studio can make the whole engineering process more convenient and coherent. Moreover, it may be helpful to compare the scores from other tools such as Rosetta and Alphafold.
References
[1] cloud.tencent.com/developer/news/579459
[2] Christopher, L. P., Yao, B., & Ji, Y. (2014). Lignin Biodegradation with Laccase-Mediator Systems. Frontiers in Energy Research, 2. http://doi:10.3389/fenrg.2014.00012