A method for screening sEH inhibitors based on pharmacophore model and molecular docking

Abstract

The application discloses a method for screening sEH inhibitors based on a pharmacophore model and molecular docking, and belongs to the field of food science and technology. The steps of the method mainly comprise the following steps: preparation of a training set and a test set, construction of a pharmacophore model, construction of a to-be-screened set, screening of the to-be-screened set by using the optimal pharmacophore model and molecular docking, and screening based on binding energy. The application solves the problem that rapidness and accuracy cannot be considered simultaneously in the current polypeptide screening process by jointly using a pharmacophore model and flexible docking.

Description

TECHNICAL FIELD

[0001] The present application relates to a method for screening sEH inhibitors based on pharmacophore model and molecular docking, in particular, a method for screening a polypeptide from gadusiformes having soluble epoxide hydrolase activity, and belongs to the field of food science and technology. BACKGROUND

[0002] Soluble epoxide hydrolase (sEH; EC 3.3.2.10) belongs to a class of epoxide hydrolases (EHs; EC 3.3.2.3) and is a hydrolytic enzyme that uses epoxide fatty acids as substrates in the human body. Human sEH is encoded by the EPHX2 gene located on human chromosome 8 and consists of a total of 555 amino acid residues, with a molecular weight of about 60 kDa. The hydrolysis catalytic site center of sEH is composed of three amino acid residues Asp333-His523-Asp495 and opens and hydrolyzes the epoxide ring group in epoxide fatty acids to generate the corresponding diols DHETs. It is widely present in various organs of mammalian systems and rapidly hydrolyzes the metabolite EETs of arachidonic acid. Recent studies have also confirmed that sEH is mainly distributed in human kidney, adrenal gland, liver, pancreatic islets and vascular endothelial cells and other tissue structures, and in addition, the presence of sEH can also be detected in the human brain. sEH usually exists in the form of a dimer in the human body, which has two ends with different structures, one end being a C-terminal structure and the other end being an N-terminal structure. The C-terminal region has epoxide hydrolysis activity, while the N-terminal region has lipid phosphatase activity and may have a stimulating effect on cell proliferation, but the specific function and mechanism of action are not yet clear. The main substrate hydrolyzed by sEH in the human body is unsaturated epoxide fatty acid, which is a product obtained by the metabolism of polyunsaturated fatty acids (PUFA) through the cytochrome P450 (CYP450) pathway. The most commonly studied epoxide type is epoxyeicosatrienoic acid (EETs), which has 4 isomers according to the position of the epoxide ring group. It has a preventive effect on many chronic diseases, such as cardiovascular and cerebrovascular diseases, metabolic diseases and muscle pain, but it is hydrolyzed by sEH into relatively inactive diol fatty acids DHETs. Therefore, inhibiting the enzyme activity of sEH can also increase the level of EETs in the human body to achieve a preventive effect.

[0003] The substrate EETs of sEH in human body has many biological effects, such as vasodilation, lowering blood pressure, treating inflammation, etc., and the hydrolysis product DHETs loses the corresponding activity. Therefore, the currently used sEH inhibitors have shown great potential in the prevention and treatment of hypertension, inflammatory diseases, cardiovascular diseases and diabetes, etc. Current research shows that in various animal disease models, by using sEH inhibitors or sEH gene knockout, it has been proved that sEH plays an important physiological activity in myocardial hypertrophy, diabetes, inflammation, hypertension and kidney disease, etc. Therefore, by using sEH inhibitors to increase the level of EETs in the human body, it has been regarded as a new target for the prevention and treatment of various diseases to be studied.

[0004] Recent scientific research has also confirmed that sEH plays an important role in regulating blood pressure and treating inflammation, and inhibiting sEH can be an effective measure for preventing and treating diseases such as hypertension, atherosclerosis, stroke and chronic renal failure. Chiamvimonvat et al. study shows that due to the genetic deletion and pharmacological regulation of sEH, the level of plasma EETs can be increased, and the efficacy of EETs can be improved, so inhibiting sEH may have anti-hypertensive and anti-inflammatory effects. Another study confirmed that rats with spontaneous hypertension characteristics have higher sEH expression levels in the body during the same period from non-hypertensive to hypertensive growth. In the study of IBD-related diseases, Zhang et al. studied the effect of sEH expression gene-deficient mice on dextran sulfate sodium (DSS)-induced colon cancer, and the results showed that the gene-deficient mice had significantly lower levels of IL-6, TNF-α and other inflammation-related indicators associated with UC, as well as tumor incidence compared with normal sEH-expressing mice. Based on the function of sEH in mammals in the metabolic process of endogenous substrates and the role of EETs in clinical practice, inhibiting sEH also provides a new direction for the treatment of various diseases, opening up new prospects, and sEH inhibitors have potential significance in the field of clinical research drugs.

[0005] Scientists have long explored sEH inhibitors. From the discovery of early epoxy sEHI to the design and synthesis of the third generation urea sEHI, sEHI has undergone a 30-year development process. From the original human sEH inhibitor with only μM-level inhibitory activity in vitro and poor or no activity in vivo to the current human sEH inhibitor with nM-level activity in vitro and excellent pharmacokinetic properties in vivo. The current sEH inhibitors are mainly directed against the catalytic site of sEH, most of which are artificially synthesized inhibitors, but their process is complex and expensive, such as t-AUCB, AUDA, TPPU, etc. In addition, there are inhibitors isolated from natural products, but most of them are not as effective as artificially synthesized ones, such as capsaicin isolated from chili peppers. Current research has found that most sEH inhibitors have amide bond structures, and the molecular weight is mostly distributed in the range of 300-500 Da. In addition, Dan's research shows that tryptophan isolated from the plant grass stone insect also has certain inhibitory activity on sEH, so polypeptides have the potential to become sEH inhibitors.

[0006] Computer simulation is used more and more widely in current research, especially in the field of medicine. Computer simulation and screening have become essential means for rapid drug development. However, many studies often limit virtual screening to small molecules. However, for macromolecules or polypeptides with strong flexibility, there has been no ideal method. Although some researchers have tried to use flexible docking methods to screen polypeptides with inhibitory activity, this method is only suitable for a small number of polypeptides. For polypeptides with large molecular weight or large sample size, it often takes days or even tens of days to complete. Therefore, how to quickly screen polypeptides with potential inhibitory activity has become a problem to be solved. SUMMARY

[0007] [TECHNICAL PROBLEM]

[0008] The technical problem to be solved by the present application is the problem of the existing virtual screening method in screening polypeptides: (1) The existing virtual screening method for polypeptides is basically molecular docking, of which the flexible docking algorithm is mainly used. However, this method sacrifices time to improve accuracy, especially when screening a large number of polypeptides (more than one hundred or one thousand), which often consumes a lot of time; (2) The existing screening method often has good accuracy for small molecules and molecules with poor flexibility, and cannot be applied to virtual screening of macromolecules and molecules with strong flexibility; (3) Conventional rapid screening often cannot consider the flexibility and spatial structure of polypeptides, so it is often less accurate when used for screening of peptide substances, and sometimes it has no reference value.

[0009] [TECHNICAL SCHEME]

[0010] The application provides a method for screening sEH inhibiting peptides based on a pharmacophore model and molecular docking, comprising the following steps:

[0011] (1) Preparation of training set and test set

[0012] 29 compounds with sEH inhibiting activity and clear chemical structure were collected from RCSB Protein Data Bank (PDB) and literature, and the inhibiting activity values IC 50 The value range across five orders of magnitude (0.0018-2200 μM) ensures the accuracy of the model; 18 molecules are randomly selected from the 29 compounds as a training set to construct the pharmacophore model, and the remaining 11 compounds are used as a test set to verify the matching of the pharmacophore model constructed based on the training set.

[0013] (2) Extraction of pharmacophore model characteristic elements

[0014] The training set obtained in step (1) is subjected to the Edit and Cluster Pharmacophore Features module, Feature Mapping program in Discovery Studio 2017R2 software (DS 2017R2, Biovia, San Diego, CA, USA) to extract the pharmacophore characteristic elements and their spatial arrangement forms required for generating the Hypogen pharmacophore;

[0015] (3) Construction of pharmacophore model

[0016] The pharmacophore characteristic elements obtained in step (2) are subjected to the 3D QSAR Pharmacophore Generation program in the Pharmacophore module of the DS software, and the hydrogen bond acceptor (Acceptor), hydrogen bond donor (Donor), hydrophobic center (Hydrophobe), ionizable positive center (Ionizable Positive) and aromatic ring center (Ring Aromatic) are selected as the characteristic elements, and the Fast algorithm of the Conformation Generation module of the DS software is used to generate the Hypogen pharmacophore model.

[0017] (4) Screening and verification of the best pharmacophore model

[0018] The statistical index cross-validation correlation coefficient (q2) and root mean square error (RMS Error) of the ten pharmacophore models obtained in step (3) are screened to obtain the best pharmacophore model; at the same time, the test set molecules in step (1) are used to verify the best pharmacophore model by Fisher random test, and the best pharmacophore model refers to the model with the highest matching degree of predicted activity value and activity of the test set molecules, which is intuitively reflected by the cross-validation correlation coefficient (q2) (the better the model, the larger the q2 value; the worse the model, the smaller the q2 value; generally, when q2>0.5, the model can be used for prediction and screening, and when q2>0.8, the accuracy of the model screening is better);

[0019] (5) Screening of molecular docking model

[0020] The crystal complex structure of sEH (PDB ID: 3WKE, resolution: ) is downloaded from the RCSB Protein Data Bank database, the Prepare Protein module in the DS software is used to remove unnecessary water molecules and heteroatoms in the docking process, remove the built-in ligand and complete the hydrogen atoms; the inhibitor molecule t-AUCB on the sEH complex is removed, and then the CDOCKER model and the LibDock model are selected to dock t-AUCB with sEH, and according to the docking results, the model that can dock t-AUCB with she to the closest original state is selected from the CDOCKER model and the LibDock model as the optimal molecular docking model;

[0021] (6) Construction of polypeptide set to be screened

[0022] The cod hydrophobic polypeptides meeting the following conditions are used as the set to be screened: tryptophan has inhibitory activity on sEH, and the polypeptide to be screened contains tryptophan; the sEH active site pocket size is small, and the molecular weight of the polypeptide to be screened is between 400-700 Da; the cod hydrophobic polypeptides meeting the above two conditions are subjected to energy minimization treatment to obtain the set to be screened;

[0023] (7) Screening of the best sEH inhibitory peptide

[0024] The best pharmacophore model obtained in step (4) and the optimal molecular docking model obtained in step (5) are used to screen the set to be screened obtained in step (6), and the best sEH inhibitory peptide is determined according to the results of the best pharmacophore model (Fit Value) and the molecular docking model (LibDock score);

[0025] Specifically, the closer the Fit Value and LibDock score is to the control or higher than the control, the better; the Fit Value is used to screen out inhibitors with the same matchable pharmacophore characteristic elements as the control, and when the structural characteristics of the sample are more matched to the pharmacophore model, the higher the Fit value is, but the Fit value cannot match the appropriate size required to enter the sEH inhibition site, so the molecular docking LibDock model is introduced for further screening to make up for the shortcomings of the pharmacophore model screening;

[0026] (8) Optimal sEH inhibitory peptide activity verification

[0027] The sEH optimal inhibitory peptide screened in step (7) is synthesized by using a solid-phase synthesis technique, and the sEH inhibitory activity thereof is determined.

[0028] The present application provides a polypeptide with sEH inhibitory activity, and the sequence thereof is: PLLW.

[0029] The present application starts from constructing a sEH active site pharmacophore, selects a reported active substance for pharmacophore evaluation, and selects the best model for rapid screening of polypeptides, and then combines the Flexible Docking algorithm to significantly increase the hit rate of virtual screening, and obtains a rapid and widely applicable virtual screening method.

[0030] [beneficial effects]

[0031] 1. The pharmacophore model is used as a primary screening method to quickly locate polypeptides with characteristic pharmacophores among numerous polypeptides to be screened, reduce the subsequent workload, and initially improve the hit rate.

[0032] 2. Flexible docking is used. On the basis of pharmacophore screening, the flexibility and spatial characteristics of the polypeptide are fully considered, which greatly increases the prediction hit rate.

[0033] 3. The pharmacophore model and flexible docking are used together to solve the problem that the speed and accuracy cannot be considered in the current polypeptide screening process. The former screens out the structure of the potential active inhibitor, and the latter screens out the structure that can enter the sEH inhibition site and play a role.

[0034] 4. China has abundant resources of cod, and cod protein is a good source of various bioactive peptides. The present application takes cod protein-derived polypeptides as an example to virtually screen polypeptides with sEH inhibitory activity, and the verification results show that the peptide has good binding capacity with sEH and can inhibit the hydrolysis of sEH to the substrate. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1Structure and activity data of sEH inhibitors for training set and test set.

[0036] Figure 2 Cost values (a) and correlation coefficients (R 2 ) Fisher validation

[0037] Figure 3 Schematic diagram of t-AUCB and different molecular docking models of sEH interaction.

[0038] Figure 4 Superimposition of PLLW and the first pharmacophore model.

[0039] Figure 5 Results of in vitro experiment to verify the in vitro activity of polypeptides screened.

[0040] Figure 6 Schematic diagram of PLLW (a) and t-AUCB (b) and molecular docking interaction with sEH. DETAILED DESCRIPTION

[0041] Terms:

[0042] Training set: In machine learning, samples are generally divided into three independent parts: training set, validation set and test set. Among them, the training set is used to establish the model, the validation set is used to determine the network structure or the parameter of controlling the complexity of the model, and the test set is used to test the performance of the finally selected optimal model.

[0043] Pharmacophore: It is a combination of characteristic three-dimensional structural elements, which can be divided into two types. One is a group of analogs with the same pharmacological effect, which has a certain basic structure, i.e. the same chemical structural part; the other is a group of molecules with completely different chemical structures, but they bond to the same receptor with the same mechanism and produce the same pharmacological effect.

[0044] Molecular docking: It is the process of mutual recognition between two or more molecules through geometric matching and energy matching, aiming to find the best binding mode of ligand molecules and receptor molecules. Molecular docking methods can be roughly divided into the following three categories according to different degrees of simplification: (1) rigid docking; (2) semi-flexible docking; (3) flexible docking. Flexible docking refers to the fact that the conformation of the research system can basically change freely during the docking process, which is generally used to accurately investigate the recognition between molecules. Since the conformation of the system can change during the calculation, flexible docking improves the accuracy of docking while consuming more computing time.

[0045] Discovery Studio 2017 R2 software: Discovery Studio TM (abbreviated as DS) is a professional life science molecular simulation software. The main functions of DS currently include protein characterization (including protein-protein interaction), homology modeling, molecular mechanics calculation and molecular dynamics simulation, structure-based drug design tools (including ligand-protein interaction, de novo drug design and molecular docking), small molecule-based drug design tools (including quantitative structure-activity relationship, pharmacophore, database screening, ADMET) and design and analysis of combinatorial libraries, etc.

[0046] Example 1

[0047] (1) Preparation of training set and test set molecular conformations

[0048] A total of 29 sEH inhibitors were collected from the PDB database (Protein Data Bank) and literature, and the specific physiological activity values were obtained. The inhibition activity data of sEH inhibitors were all determined by the same principle method, that is, by determining the amount of reaction product 6-methoxy-2-naphthaldehyde obtained by adding sEH inhibitor while sEH catalyzing hydrolysis of substrate PHOME, to judge the binding ability of inhibitor to sEH, and the inhibition activity is represented by IC 50 value. The inhibition activity IC 50 values of all 29 inhibitors cover 5 orders of magnitude, with a wide range (0.0018-2200 μM). When constructing the Hypogen pharmacophore model, the negative logarithm (pIC 50 ) of IC 50 value is used. Among all 29 inhibitor molecules, 18 are used as training set molecules and 11 are used as test set molecules. The chemical molecular structure formulas and activity data of all 29 inhibitors are shown in Table 1. Figure 1

[0049] Using ChemDraw 18.0 software, the chemical structures of all collected sEH inhibitors were drawn, and then the Minimize module in DS software was used to perform energy minimization of 29 inhibitor molecules by CHARMM force field, and the activity values of IC 50 values of the 29 inhibitors were edited and input into the Active column of the browser table, and the activity uncertainty value (Uncert value) was set to the default value 1.5.

[0050] (2) Extraction of pharmacophore characteristic elements

[0051] In the Discovery Studio 2017R2 software, the training set molecules were classified according to IC 50 ​The values are automatically sorted, and the top few compounds will be automatically defined as "active compounds" by the system, and the last few compounds will be defined as "inactive compounds". Here, Figure 1 Compounds 11, 24 and 25 in the table are all defined as active compounds. Then select the Feature Mapping program in the Edit and Cluster Pharmacophore Features module in the DS software, and select five feature elements required for matching extraction, namely hydrogen bond acceptor (Acceptor), hydrogen bond donor (Donor), hydrophobic center (Hydrophobe), ionizable positive center (Ionizable Positive) and aromatic ring center (Ring Aromatic), and the rest of the parameters are set to default values. After running the program, the pharmacophore feature elements required for generating Hypogen pharmacophore and their spatial arrangement form can be extracted from the training set molecules.

[0052] (3) Construction of Hypogen pharmacophore model

[0053] After obtaining the pharmacophore feature elements, select the 3D QSAR Pharmacophore Generation program in the Pharmacophore module in the DS software to generate the Hypogen pharmacophore model. Select hydrogen bond acceptor (Acceptor), hydrogen bond donor (Donor), hydrophobic center (Hydrophobe), ionizable positive center (Ionizable Positive) and aromatic ring center (Ring Aromatic) as feature elements, set the Minimum Interfeature Distance parameter to 1.5, select the Fast algorithm in the Conformation Generation module to generate the Hypogen pharmacophore model, and set the energy threshold to 10 kcal / mol, and the rest of the parameters are set to default values. After running the program, 10 pharmacophore models will be generated.

[0054] Table 1 shows 10 Hypogen pharmacophore models generated using the structure and activity information of the training set molecules, including the Cost value, correlation coefficient factor (r), root mean square deviation (RMSD) and pharmacophore feature elements of each model.

[0055] Table 1 Hypogen pharmacophore effect constructed by training set

[0056]

[0057]

[0058] Note: HBA, pharmacophore contains hydrogen bond acceptor characteristic elements; HBD, pharmacophore contains hydrogen bond donor characteristic elements; HY, pharmacophore contains hydrophobic center characteristic elements.

[0059] The Total cost value of the generated 10 pharmacophore models ranges from 58.0658 to 61.9782; the Null cost value is 121.698; the Fixed cost value is 51.3629; q2 is between 0.798 and 0.954; the RMSD ranges from 0.427 to 0.902. Among the 10 generated pharmacophore models, 3-4 pharmacophore characteristic elements are matched, most of which are hydrogen bond acceptors and hydrogen bond donors. The cost value is a key indicator for evaluating the pharmacophore model, which is a comprehensive evaluation of the deviation between the predicted activity and the actual activity of the model and the complexity of the pharmacophore composition. If the ΔCost value is > 60, it indicates that the confidence interval of the model is 75%-90%; if the value is < 40, the confidence interval of the pharmacophore model is less than 50%; the closer the cross-validation correlation coefficient (q2) is to 1 and the smaller the root mean square deviation value (RMSD) is, the better the pharmacophore effect is.

[0060] As can be seen from Table 1, the Total cost value of the first pharmacophore is the lowest (58.0658), close to the Fixed cost value (51.3629), and the Δcost value is 63.632 (> 60), which indicates that the confidence interval of the first pharmacophore model is 75%-90%, with good reliability. From the statistical indicators, compared with other pharmacophore models, the first pharmacophore has a certain predictive ability, with 4 matched pharmacophore characteristic elements, and the types are extensive, including 1 hydrogen bond acceptor characteristic, 1 hydrophobic center characteristic and 2 hydrogen bond donor characteristics. At the same time, compared with the indicators of the other 9 pharmacophores, the first pharmacophore has a larger q2 value (0.949) and a lower RMSD value (0.437), which also indicates that the first pharmacophore is better than other pharmacophore models. Therefore, the first pharmacophore is selected as the best pharmacophore and enters the next step of analysis.

[0061] (4) Verification of Hypogen pharmacophore model

[0062] In the process of generating the Hypogen pharmacophore model, the Fisher random test and the test set test are used to verify the generated Hypogen pharmacophore. First, the confidence coefficient of Fisher verification is set to 95%, and the generated pharmacophore is used for random cross test. If the total cost of the random hypothesis generated based on the compound hypothesis of the training set is greater than the value of the original hypothesis, it indicates that there is a correlation between the chemical structure and the biological activity in the compounds in the training set, and the generated pharmacophore suggestion model has a high confidence and is not accidentally generated. This method is mainly used to evaluate the statistical correlation of the generated pharmacophore model and determine whether the hypothesis generated by the training set is reasonable. Secondly, the molecules in the test set are used to determine the ability of all generated pharmacophore models to predict the activity values of all substances except the training set. The compounds in the test set are preprocessed in the same way as the compounds in the training set, and then the prediction of the activity of the test set compounds by the pharmacophore is observed. The correlation between the experimental and predicted values of the test set compounds is used to evaluate the prediction ability of the pharmacological model. The verified Hypogen pharmacophore model can evaluate the activity of the compounds.

[0063] In the process of Fisher random verification, all the activity values of the training set molecules are first shuffled and randomly assigned to each molecular structure, and then the pharmacophore model is established using the same parameters. If the random model obtained is not similar to the initial model generated, and does not have similar or better ΔCost value, RMSD value and q2 value, etc. Correlation coefficient, it can be shown that the generated model is not accidentally generated and has statistical significance. According to this verification method, 9 molecular tables with randomly assigned activity values are generated, and random pharmacophore models are reconstructed. The Cost value and RMSD value are evaluated at a confidence level of 95%. As shown in Table 1, the results show that the Cost value and RMSD value obtained by 9 random experiments are quite different from the values of the generated No. 1 pharmacophore model, indicating that the generated Hypogen pharmacophore model is not accidentally generated and has statistical significance. Figure 2

[0064] (5) Screening of sEH inhibitory activity peptides

[0065] It has been reported that tryptophan has inhibitory activity on sEH at a certain concentration, so the polypeptide sequences of the hydrophobic part of the cod peptides identified by mass spectrometry are screened, and short peptide sequences containing tryptophan are screened, and the molecular weight is not too large, controlled between 400-700 Da.

[0066] ​Molecular docking was performed using Discovery Studio 2017R2 Client software. The crystal complex structure of sEH (PDB ID: 3WKE, resolution: ) was downloaded from the RCSB Protein Data Bank database, and water molecules and heteroatoms that were not needed in the docking process were deleted using PrepareProtein, the built-in ligand was removed, and hydrogen atoms were completed. First, the docking model was selected, the ligand t-AUCB on sEH was removed, and the CDOCKER model and the LibDock model were selected to dock the ligand with sEH, respectively, and the docking mechanism of the two models was analyzed, and one of them closest to the original state was selected as the docking model. These polypeptides were docked with sEH (PDB: 3WKE) to screen polypeptide molecules that could match the size of the active center of sEH, and the best Hypogen pharmacophore was used to predict the physiological activity of these polypeptide molecules, and the Fit Value value was obtained for analysis.

[0067] Molecular docking was performed using Discovery Studio 2017R2 Client software, and the docking model used was first determined. As shown in Table 2 and Table 2, the original model of t-AUCB formed hydrogen bonds with Asp335, Tyr383 and Tyr466 of sEH, and the LibDock model also restored the formation of hydrogen bonds at the same amino acid residue position, but the CDOCKER model failed to restore the original model well. Therefore, by comparing with the original model, it can be concluded that the LibDock model is the best in the 3WKE structure, so the LibDock docking model is selected for simulation docking. Figure 3

[0068] Table 2 Comparison of LibDock and CDOCKER models for restoring ligands

[0069]

[0070] The polypeptide sequence of the cod peptide identified by mass spectrometry was screened using the first pharmacophore. The short peptide sequence containing tryptophan was screened, and t-AUCB, an sEH inhibitor, was used as a reference standard. Then, the first pharmacophore was used to screen the short peptide containing tryptophan, and the Fit value was generated. Part of the results are shown in Table 3, wherein the 4-peptide sequence of PLLW has the highest Fit value, the molecular weight is 527.31, the Fit value is in the range of 9.053, and the LibDock score value is 147.807, which is relatively close to t-AUCB. Figure 4 ​The superimposition result of PLLW and the first pharmacophore is shown in Figure 1. It can be seen that the characteristic elements of the first pharmacophore can be matched with PLLW.

[0071] Table 3 Results of the first pharmacophore predicting partial cod hydrophobic peptides

[0072]

[0073] (6) In vitro sEH inhibition activity determination of PLWW

[0074] The synthesized PLLW sequence was determined at multiple concentration gradients according to the following determination method. The concentration gradients were set to 10 mM, 100 mM, 500 mM, 550 mM, 750 mM, and 1000 mM. There must be at least two points on both sides of the half-inhibitory concentration. The IC 50 value was finally fitted.

[0075] The following high-performance liquid chromatography method was used for determination: mobile phase A: 100% water, mobile phase B: 100% acetonitrile, and both mobile phases A and B contained 0.01% formic acid.

[0076] Table 4 HPLC determination method of sEH enzyme activity

[0077]

[0078] The reaction group containing the inhibitor or sample was first incubated with sEH at 25°C for 30 min, and then the substrate PHOME was added. After 30 min of reaction, the sample was filtered using a 0.22 μm organic filter membrane, and then loaded. The loading amount was 10 μL, the PMT gain value was set to 15, and the fluorescence intensity of the reaction product 6-methoxy-2-naphthaldehyde was determined under the conditions of excitation wavelength 330 nm and emission wavelength 465 nm. The reaction system is shown in Table 5. The determination result is expressed by IC 50 value.

[0079] Table 5 Reaction system for determination of sEH enzyme activity

[0080]

[0081] The sEH in vitro inhibition activity of PLLW solutions with different concentration gradients was determined using high-performance liquid chromatography. The concentration gradients were set to 10 mM, 100 mM, 500 mM, 550 mM, 750 mM, and 1000 mM. The finally fitted curve is shown in Figure 5The curve equation is shown as y = A2 + (A1 - A2) / (1 + exp((x - x0) / dx)), where A1 = 67.699, A2 = 3.089, x0 = 536.676, dx = 30.584, r 2 = 0.94634. The IC 50 value fitted by the curve is 506.66 mM.

[0082] (7) Analysis of the mechanism of action of PLWW on sEH

[0083] After selecting the docking model, the 3D structure of the active peptide was constructed as a ligand, and energy minimization optimization was performed under the CHARMM force field. Then, the simulation docking of sEH and the active peptide sequence was performed, and the docking coordinates were x: -15.2371, y: -6.5367, and z: 61.5581. The number of generated docking conformations for each molecule was set to 10, and other parameters were default values. After docking, the mechanism of action of the active peptide and sEH was analyzed.

[0084] The inhibition mechanism of PLLW on sEH was studied using LibDock docking model, and the mechanism of action of cod active peptide PLLW and sEH was described from a molecular perspective. Figure 6 It can be seen that PLLW forms electrostatic interactions with Pro371 forms a carbon-hydrogen bond, and the amino acid residue Trp336 adjacent to the catalytic site also forms two electrostatic interactions While the sEH inhibitor mainly acts on three key amino acid residues Asp335, Tyr383 and Tyr466 in the active center of sEH.

[0085] The rapid virtual screening method of the polypeptide with sEH activity in the embodiment was applied to a total of 150 polypeptides to be screened, and polypeptides with potential inhibitory activity were obtained after screening. The polypeptide PLLW obtained by screening was synthesized in vitro and verified by combining with the chromogenic substrate experiment. The results show that the peptide has good binding capacity with sEH and inhibits its activity.

[0086] Compared with the prior art, the method of the present application can improve the speed while ensuring the hit rate of prediction. In the embodiment, the target polypeptide is finally obtained from 150 polypeptides, which only needs about 40 hours, while the traditional Flexible Docking needs more than 90 hours, and the efficiency is improved by more than 50%. Compared with the pharmaophore model alone, the present application combines the LibDock docking mode, and fully considers the flexibility and spatial conformation characteristics of the polypeptide, so as to avoid the possibility that the screened polypeptide cannot enter the enzyme active center pocket due to the size. Therefore, the method of the present application can play a positive and rapid role in the virtual screening of polypeptides, especially in the screening of a large number of polypeptides, and can provide a basis for the rapid progress of subsequent work.

[0087] Although the present application has been disclosed with the preferred embodiments as above, it is not intended to limit the present application, and any person skilled in the art can make various modifications and modifications without departing from the spirit and scope of the present application, therefore the protection scope of the present application should be defined by the claims.

Claims

1. A method for screening target protein inhibitors based on pharmacophore models and molecular docking, characterized in that, Comprising the following steps: (1) Training set and test set preparation Among them, the training set is used to construct the pharmacophore model, which is composed of molecules with target protein inhibitory activity; the test set is used to verify the matching of the pharmacophore model constructed based on the training set, which is composed of molecules not in the training set and having target protein inhibitory activity; (2) Pharmacophore model feature element extraction Using Discovery Studio 2017 R2 software, the pharmacophore feature elements are extracted from the training set molecules constructed in step (1); (3) Pharmacophore model construction The pharmacophore feature elements obtained in step (2) are used to construct the Hypogen pharmacophore model by selecting hydrogen bond acceptor, hydrogen bond donor, hydrophobic center, positive ion center and aromatic ring center as feature elements and using the Fast algorithm in the Conformation Generation module of the DS software Pharmacophore module; (4) Optimal pharmacophore model screening and verification From the ten pharmacophore models obtained in step (3), the best pharmacophore model is selected according to the statistical index, and at the same time, the test set molecules in step (1) are used to verify the best pharmacophore model by Fisher random test; (5) Molecule docking model screening The crystal complex structure of sEH is downloaded from the RCSB Protein Data Bank database, the Prepare Protein module of the DS software is used to delete water molecules and heteroatoms that are not needed in the docking process, remove the built-in ligand and complete the hydrogen atoms, remove the inhibitor molecule t-AUCB on the sEH complex, and then select the CDOCKER model and LibDock model to dock t-AUCB with sEH, and select the one closest to the original state as the optimal molecular docking model; (6) Construction of target protein inhibitor molecule set to be screened The target protein inhibitor molecules to be screened are subjected to energy minimization treatment to obtain the screening set; (7) Optimal target protein inhibitor screening The best pharmacophore model obtained in step (4) and the optimal molecular docking model obtained in step (5) are used to screen the screening set obtained in step (6), and the best target protein inhibitor molecule is determined according to the results of the best pharmacophore model and the molecular docking model.

2. The method of claim 1, wherein the method is characterized by, The inhibitor is a polypeptide.

3. The method according to claim 1 or 2, wherein the method is characterized in that, The target protein is soluble epoxide hydrolase.

4. The method according to claim 1 or 2, wherein the method is characterized in that, The target protein is soluble epoxide hydrolase, comprising the following steps: (1) Training set and test set preparation Collect 29 sEH-inhibiting compounds with clear chemical structures, 18 molecules as the training set to construct the pharmacophore model, and the remaining 11 compounds as the test set to verify the matching of the pharmacophore model constructed based on the training set; (2) Pharmacophore model feature element extraction The training set obtained in step (1) is used to extract the pharmacophore feature elements and their spatial arrangement forms required for generating the Hypogen pharmacophore by using the Edit and Cluster Pharmacophore Features module in the Discovery Studio 2017 R2 software, and selecting the Feature Mapping program. (3) Pharmacophore model construction According to the pharmacophore feature elements obtained in step (2), the 3D QSAR Pharmacophore Generation program in the Pharmacophore module of the DS software is used to select hydrogen bond acceptors, hydrogen bond donors, hydrophobic centers, positive ion centers and aromatic ring centers as feature elements, and the Fast algorithm in the Conformation Generation module is selected to generate the Hypogen pharmacophore model. (4) Optimal pharmacophore model screening and verification The statistical index cross-validation correlation coefficient (q 2 ) and root mean square error (RMS Error) of ten pharmacophore models obtained according to step (3) are screened to obtain the best pharmacophore model, and the best pharmacophore model is verified by using the test set molecules in step (1) and Fisher random test. (5) Molecular docking model screening The crystal complex structure of sEH is downloaded from the RCSB Protein Data Bank database, the Prepare Protein module in the DS software is used to delete unnecessary water molecules and heteroatoms in the docking process, remove the built-in ligand and complete the hydrogen atoms, remove the inhibitor molecule t-AUCB on the sEH crystal complex, and then select the CDOCKER model and LibDock model to dock t-AUCB with sEH, and select the model closest to the original state as the optimal molecular docking model. (6) Construction of the polypeptide set to be screened Polypeptides containing tryptophan and having a molecular weight of 400-700 Da are subjected to energy minimization processing to obtain the set to be screened. (7) Screening of the best sEH inhibitory peptide The best pharmacophore model obtained in step (4) and the optimal molecular docking model obtained in step (5) are used to screen the set to be screened obtained in step (6), and the best sEH inhibitory peptide is determined according to the results of the best pharmacophore model and the molecular docking model. (8) Activity verification of the best sEH inhibitory peptide The best sEH inhibitory peptide screened in step (7) is synthesized by solid-phase synthesis technology, and its sEH inhibitory activity is determined.

5. The method of claim 4, wherein the method is characterized by, In the pharmacophore model construction, screening and verification of steps (3) and (4), the Feature Mapping program in the Discovery studio 2017 R2 software is run, the Common Feature Pharmacophore Generation program is executed on the training set molecules after analyzing the results, and the correlation coefficient and root mean square error are cross-verified by analyzing and statistical results to screen the best pharmacophore model. At the same time, the test set molecules in step (1) are used to verify the best pharmacophore model by Fisher's random test.

6. A polypeptide having sEH inhibitory activity, which is obtained by the method according to any one of claims 1 to 5, characterized in that, The amino acid sequence is: PLLW.

7. Use of a polypeptide according to claim 6 for the manufacture of a medicament for the treatment of hypertension, inflammation, cardiovascular disease or diabetes.

8. A medicament for the treatment of hypertension, inflammation, cardiovascular disease or diabetes comprising a polypeptide according to claim 6.

Images