Protein binding site similarity analysis method based on topological water network
The method analyzes protein binding site similarity using a topological water molecule network to address the challenge of determining ligand activity differences, facilitating drug repurposing and polypharmacology studies by simplifying free energy calculation and identifying enzyme selectivity.
Patent Information
- Application Number
- PCT/KR2024/001698
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-10
- Filing Date
- 2024-02-06
- Publication Date
- 2024-11-14
AI Technical Summary
Current protein binding site similarity analysis methods are challenging to determine the most appropriate approach for specific studies, as they fail to accurately account for differences in ligand activity despite similar protein structures, hindering drug repurposing and polypharmacology studies.
A method analyzing protein binding site similarity based on a topological water molecule network (TWN) is developed, involving preprocessing of protein structures, molecular dynamics simulations, extraction of water molecules, and calculation of interaction potential energy to generate matrix data and classify binding site similarities.
This approach simplifies free energy calculation, enables the design of phosphorylase inhibitors, and identifies selectivity of phosphorylase enzymes, demonstrating high applicability to drug repurposing and polypharmacology studies by providing accurate binding site similarity analysis.
Smart Images

Figure KR2024001698_14112024_PF_FP_ABST
Abstract
Description
A protein binding site similarity analysis method based on a topological water molecule network
[0001] The present invention relates to a protein binding site similarity analysis method based on a topological water molecule network, and more particularly, to a protein binding site similarity analysis method that analyzes the binding site similarity calculated based on a topological water network (TWN).
[0002] The three-dimensional structure of a protein is most important for understanding its biological function, and analysis of protein structure, especially the study of binding sites, plays a key role in new drug design.
[0003] Binding sites are closely related to protein function, and understanding their characteristics is a crucial step in structure-based drug design. Many proteins possess similar binding sites, and calculating binding site similarity provides valuable information for repurposing known drugs, identifying potential off-target side effects, and conducting polypharmacology studies.
[0004] To date, various computational methods have been developed to compare ligand binding sites in protein structures. Methods for calculating protein binding site similarity can be categorized into residue-based, shape-based, and interaction-based.
[0005] Amino acid residue-based methods utilize structural descriptors, conformation-based methods utilize graphs, and interaction-based methods utilize interaction fields. While binding site comparison methods continue to be developed, determining which method is most appropriate for a particular research area is becoming increasingly difficult.
[0006] For example, there was a problem that while the protein structure and sequence were similar, there were differences in the activity values of the ligand (or drug).
[0007] The technical task to be achieved by the present invention is to provide a protein binding site similarity analysis method that analyzes the protein binding site similarity calculated based on a topological water network (TWN).
[0008] A protein binding site similarity analysis method according to an embodiment of the present invention for achieving such a technical task includes the steps of collecting protein information files (Protein Data Bank, PDB) from the RCSB website, preprocessing the collected PDB files to prepare a protein structure, extracting a trajectory file after a molecular dynamics simulation experiment from the prepared protein structure, extracting water molecules in the binding site using the extracted trajectory file, performing TWN analysis using the extracted water molecules to generate matrix data of distance values between amino acid residue atoms in the binding site and TWN centers of gravity, and then rearranging time trajectories of two proteins to be compared, and calculating the similarity for the binding site using the distance values between the TWN centers of gravity.
[0009] The step of rearranging the above time trajectories can calculate the interaction potential energy between water molecules using the following mathematical formula in each time trajectory file.
[0010]
[0011] Here, v(a, b) represents the interaction energy between water molecules a and b, represents the distance between oxygen atoms, represents the partial charge at position i, represents the partial charge at position j, Is and represents the distance between i and j, A represents the repulsive force between i and j, and C represents the attractive force between i and j.
[0012] The step of rearranging the above time trajectory can extract TWN using the potential energy, and obtain the TWN center of gravity coordinates as the average value obtained by dividing the x, y, and z axis values of the coordinates including oxygen and hydrogen of the water molecules constituting the extracted TWN by the number of each atom.
[0013] The step of rearranging the above time trajectories can be performed by searching the inside of the protein binding site by specifying a radius of 5 to 6 Å from the coordinates of the co-crystallized ligand, calculating the three-dimensional Euclidean distance between the backbone atoms of each amino acid residue located inside the binding site in the TWN center of gravity coordinates, and generating matrix data based on the distance values calculated from all time trajectory files.
[0014] The above matrix data can represent the TWN time trajectory as rows and the amino acid residue numbers within the binding site as columns.
[0015] The step of rearranging the above time trajectories is performed by calculating the difference in the distance values between the TWN center of gravity coordinates and amino acid residue backbone atoms in all time trajectories of the second protein based on the time trajectories in the matrix data of the first protein from the matrix data generated based on the two proteins to be compared, and then selecting the time trajectory of the second protein that shows the minimum value, thereby rearranging the time trajectory of the first protein and the time trajectory of the second protein that has the minimum matrix data value.
[0016] The step of calculating the similarity for the above binding site can be performed by classifying into multiple classes using the distance values between TWN and amino acid residue backbone atoms, and calculating the binding site similarity value using the classified classes.
[0017] In this way, according to the present invention, free energy calculation is simple, a kinase inhibitor can be designed, and the selectivity of kinases can be determined, so there is an advantage that it can be applied to drug repurposing.
[0018] Figure 1 is a flowchart illustrating a protein binding site similarity analysis method based on a topological water molecule network according to an embodiment of the present invention.
[0019] Figure 2 is an exemplary diagram illustrating the protein binding site similarity analysis method illustrated in Figure 1.
[0020] FIG. 3 is a diagram for explaining the rearrangement process of molecular dynamics time trajectories in step S130 illustrated in FIG. 1.
[0021] Figure 4 is a heat map showing trends according to differences in activity values through a protein binding analysis method according to an embodiment of the present invention.
[0022] Figure 5 is a table showing the results shown in Figure 4.
[0023] Figure 6 shows the correlation of the data shown in Figures 4 and 5.
[0024] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. In this process, the thickness of lines and the sizes of components depicted in the drawings may be exaggerated for clarity and convenience of explanation.
[0025] Furthermore, the terms described below are defined based on their functions within the present invention. These terms may vary depending on the intended use or customs of the subject of measurement or operator. Therefore, the definitions of these terms should be based on the overall content of this specification.
[0026] Hereinafter, a network-based protein binding analysis method according to an embodiment of the present invention will be described in more detail using FIG. 1.
[0027] Figure 1 is a flowchart illustrating a protein binding site similarity analysis method based on a topological water molecule network according to an embodiment of the present invention.
[0028] As illustrated in Fig. 1, the protein binding site similarity analysis device collects protein information files (Protein Data Bank, PDB) from the RCSB website and preprocesses the collected PDB files to prepare a protein structure (S110).
[0029] The PDB files are for X-ray crystal structures of sunitinib bound to KIT (PDB: 3G0E), PDGFRA (PDB: 6JOK), VEGFR2 (PDB: 4AGD), PHKG2 (PDB: 2Y7J), ITK (PDB: 3MIY), HPK1 (PDB: 6NFY), MST3 (PDB: 4QMZ), PAK6 (PDB: 4KS8), and CDK2 (PDB: 3TI1).
[0030] Next, the protein binding site similarity analysis device performs a preprocessing step to remove ligands from the collected PDB files, thereby preparing the protein structure. At this time, the protein binding site similarity analysis device adds amino acid self-contained molecules and hydrogen atoms, assigns atomic bond orientations and formal charges, and performs ionization calculations based on pH 7.4.
[0031] When step S110 is completed, the protein binding site similarity analysis device extracts a trajectory file from the prepared protein structure after a molecular dynamics simulation experiment, and extracts water molecules within the binding site using the extracted trajectory file (S120).
[0032] The protein binding site similarity analysis device performs a molecular dynamics simulation experiment using the prepared protein structure as input data while running GROMACS to extract a trajectory file.
[0033] To elaborate, the protein binding site similarity analysis device is assigned a CHARMM27 force field and performs solvation using the TIP3P model.
[0034] And, the protein binding site similarity analysis device performs energy minimization and equilibration. Energy minimization is performed using the steepest descent algorithm. And, the protein binding site similarity analysis device equilibrates for 100 picoseconds (ps) using the v-rescale thermostat under NVT (constant particle number, volume, and temperature) conditions, and equilibrates for 200 picoseconds (ps) using the Parrinello-Rahman barostat algorithm under NPT (constant particle number, pressure, and temperature) conditions.
[0035] However, the protein-binding site similarity analysis device performs the final molecular dynamics simulation at 298 K for 10 nanoseconds (ns). Protein motion is restricted, long-range electrostatic interactions are performed using PME, and short-range electrostatic interactions are set to 12 Å. In addition, all bond lengths are set using the LINC algorithm, and molecular dynamics trajectories are extracted every picosecond (ps).
[0036] Next, the protein binding site similarity analysis device performs TWN analysis using the extracted water molecules to generate matrix data of distance values between amino acid residue atoms at the binding site and the TWN center of gravity, and then rearranges the time trajectories of the two proteins to be compared (S130).
[0037] FIG. 2 is an exemplary diagram for explaining the protein binding site similarity analysis method illustrated in FIG. 1, and FIG. 3 is a diagram for explaining the rearrangement process of molecular dynamics time trajectories in step S130 illustrated in FIG. 1.
[0038] First, TWN represents a ring-shaped network of water molecules. The protein binding site similarity analysis device extracts a tricyclic (3-ring) TWN. The potential function uses the TIP3P water molecule model and calculates the interaction energy between water molecules using the Leonard-Jones force and Coulomb force. The interaction potential energy is calculated using the following mathematical equation (1).
[0039] (mathematical formula)
[0040]
[0041] Here, v(a, b) represents the interaction energy between water molecules a and b, represents the distance between oxygen atoms, represents the partial charge at position i, represents the partial charge at position j. Also, Is and represents the distance between i and j, A represents the repulsive force between i and j, and C represents the attractive force between i and j.
[0042] The protein binding site similarity analysis device obtains the TWN center of gravity coordinates by dividing the x, y, and z axis values of the oxygen and hydrogen of the water molecules that make up the extracted TWN by the number of each atom.
[0043] Next, the protein binding site similarity analysis device searches the inside of the protein binding site by specifying a radius of 5 to 6 Å from the coordinates of the co-crystallized ligand, and calculates the three-dimensional Euclidean distance between the backbone atoms (N, CA, C, O) of each amino acid residue located inside the binding site in the TWN center of gravity coordinates.
[0044] In addition, the protein binding site similarity analysis device generates matrix data based on the three-dimensional Euclidean distance values calculated above from all time trajectory files.
[0045] Here, the matrix data represents the TWN time trajectories in rows and the amino acid residue numbers within the binding site in columns.
[0046] Next, as illustrated in Figure 3, the protein binding site similarity analysis device rearranges the TWNs at specific positions within the protein binding site according to different trajectories of the proteins, and the second protein rearranges according to the molecular dynamics trajectory of the first protein. At this time, the rearrangement process is performed according to the minimum value of the distance of each MD trajectory.
[0047] To elaborate, the protein binding site similarity analysis device calculates the difference in distance values between the TWN center of gravity coordinates and amino acid residue backbone atoms in all time trajectories of the second protein based on the time trajectories in the matrix data of the first protein in the matrix data generated based on the two proteins to be compared.
[0048] Additionally, the protein binding site similarity analysis device selects the time trajectory of the second protein showing the minimum value and rearranges the time trajectory of the first protein and the time trajectory of the second protein with the minimum matrix data value.
[0049] When step S130 is completed, the protein-binding site similarity analysis device calculates the similarity for the binding site using the distance value between the centers of gravity of the TWN (S140). The protein-binding site similarity analysis device classifies the matrix data of the corresponding protein into multiple classes using the distance values between the TWN and the amino acid residue backbone atoms. For example, if the distance value between the TWN and the amino acid residue backbone atoms is less than 5 Å, the protein-binding site similarity analysis device classifies the matrix data of the corresponding protein into class A. If the distance value between the TWN and the amino acid residue backbone atoms is greater than 5 Å and less than 7.5 Å, the protein-binding site similarity analysis device classifies the matrix data of the corresponding protein into class B. In addition, if the distance value between the TWN and the amino acid residue backbone atoms is greater than 7.5 Å and less than 10 Å, the protein-binding site similarity analysis device classifies the matrix data of the corresponding protein into class C. In addition, if the distance value between the TWN and the amino acid residue backbone atoms is greater than 10 Å, the protein-binding site similarity analysis device classifies the matrix data of the corresponding protein into class D.
[0050] Next, the protein binding site similarity analysis device calculates the binding site similarity value using the classified classes.
[0051] Here, the similarity can be defined according to the class type of the matching number to the class type of the total number.
[0052] Hereinafter, experimental results for a protein binding analysis method according to an embodiment of the present invention will be described in more detail using FIGS. 4 to 6.
[0053] Fig. 4 is a heat map showing trends according to differences in activity values through a protein binding analysis method according to an embodiment of the present invention, and Fig. 5 is a table showing the results shown in Fig. 4.
[0054] As shown in Figure 4, the horizontal axis of the heat map represents the difference in sunitinib activity values of the two proteins as a log value, and the vertical axis represents the similarity method within each protein binding site.
[0055] TWN-RENCOD described on the vertical axis represents a protein binding site similarity analysis method based on a topological water molecule network according to an embodiment of the present invention, IsoMIF represents an interaction-based method, ProBis represents a shape-based method, PoecketMatch P-max, P-min represents an amino acid residue-based method, Sequence similarity represents amino acid sequence similarity, and Sequence Identity represents amino acid sequence identity.
[0056] Each cell marked with color represents the similarity, and it can be seen that the method based on TWN-RENCOD shows the greatest tendency according to the difference in activation value.
[0057] Figure 6 shows the correlation of the data shown in Figures 4 and 5.
[0058] As shown in Fig. 6, the difference in the sunitinib activity value of each kinase protein and the similarity within the binding site were calculated, and it can be seen that the method based on TWN-RENCOD according to the embodiment of the present invention shows the highest value.
[0059] Thus, the protein binding analysis method according to the present invention can be applied to drug discovery, as it can design kinase inhibitors and elucidate the selectivity of kinases without requiring free energy calculations. Furthermore, according to the present invention, multiple homologous proteins that bind the same ligand can be analyzed using a binding site similarity calculation method.
[0060] While the present invention has been described with reference to the embodiments illustrated in the drawings, these are merely exemplary, and those skilled in the art will understand that various modifications and equivalent embodiments are possible. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the following claims.
[0061] The protein binding site similarity analysis method based on a topological water molecule network of the present invention has industrial applicability.
Claims
1. A step of collecting protein information files (Protein Data Bank, PDB) from the RCSB website and preprocessing the collected PDB files to prepare the protein structure. A step of extracting a trajectory file from a prepared protein structure after a molecular dynamics simulation experiment and extracting water molecules within the binding site using the extracted trajectory file. A step of performing TWN analysis using the above-mentioned extracted water molecules to generate matrix data of distance values between amino acid residue atoms at the binding site and the TWN center of gravity, and then rearranging the time trajectories of the two proteins to be compared, and A protein binding site similarity analysis method comprising a step of calculating the similarity for a binding site using the distance value between the centers of gravity of the above TWNs.
2. In paragraph 1, The step of rearranging the above time trajectory is, A protein binding site similarity analysis method that calculates the interaction potential energy between water molecules using the following mathematical formula in each time trajectory file: Here, v(a, b) represents the interaction energy between water molecules a and b, represents the distance between oxygen atoms, represents the partial charge at position i, represents the partial charge at position j, Is and represents the distance between i and j, A represents the repulsive force between i and j, and C represents the attractive force between i and j.
3. In paragraph 2, The step of rearranging the above time trajectory is, A protein binding site similarity analysis method that extracts TWN using the above potential energy and obtains the TWN center of gravity coordinates by dividing the x, y, and z axis values of the coordinates including oxygen and hydrogen of the water molecules constituting the extracted TWN by the number of each atom.
4. In paragraph 3, The step of rearranging the above time trajectory is, The interior of the protein binding site is searched by specifying a radius of 5 to 6 Å from the coordinates of the co-crystallized ligand, and the three-dimensional Euclidean distance between the backbone atoms of each amino acid residue located inside the binding site is calculated in the TWN center of gravity coordinates. A protein binding site similarity analysis method that generates matrix data based on distance values calculated from all time trajectory files.
5. In paragraph 4, The above matrix data is, A protein binding site similarity analysis method in which TWN time trajectories are represented as rows and amino acid residue numbers within binding sites are represented as columns.
6. In paragraph 4, The step of rearranging the above time trajectory is, In the matrix data generated based on the two proteins to be compared, the difference in the distance values between the TWN center of gravity coordinates and the amino acid residue backbone atoms in all time trajectories of the second protein based on the time trajectories in the matrix data of the first protein is calculated, and then A protein binding site similarity analysis method that rearranges the time trajectory of the first protein and the time trajectory of the second protein with the minimum matrix data value by selecting the time trajectory of the second protein that shows the minimum value.
7. In paragraph 1, The step of calculating the similarity for the above binding site is: A protein binding site similarity analysis method that classifies into multiple classes using distance values between TWN and amino acid residue backbone atoms and calculates binding site similarity values using the classified classes.
Citation Information
Patent Citations
Method and apparatus for predicting structure ofunknown protein
KR1020050064644A
Protein-ligand docking method using 3-dimensional molecular alignment
KR1020120099922A
Protein Ligand Binding Site Prediction Method
KR1020160078537A
Manufacturing Method of Transfer Print Film for Efficient Production
KR1020210041842A
Method for analysis of protein binding site similarity based on topological water molecule network
KR102622760B1
