Method for designing and screening antibacterial peptide for inhibiting fusarium graminearum

Abstract

The invention provides a design and screening method of an antibacterial peptide for inhibiting fusarium graminearum, which comprises the following steps: screening the antibacterial peptide targeting myosin of the fusarium graminearum from a peptide chain with inhibition on the fusarium graminearum, and carrying out saturated mutation to obtain an optimal triple mutant for remarkably improving the affinity and stability of the myosin; by combining with the catalytic center of myosin to inhibit ATP hydrolytic activity of myosin and prevent growth and diffusion of fungi, the effect of preventing and treating wheat scab caused by fusarium graminearum is superior to that of micromolecule cyanoacrylate bactericides, the influence on other non-target organisms can be reduced, and the micromolecule cyanoacrylate bactericides are environment-friendly, and the application of the micromolecule cyanoacrylate bactericides to the wheat scab caused by fusarium graminearum to the prevention and treatment of wheat scab caused by fusarium graminearum.

Description

Technical Field

[0001] The present invention relates to the technical field of antimicrobial peptides, and in particular to a design and screening method for antimicrobial peptides that inhibit Fusarium graminearum. Antimicrobial peptides that target myosin of Fusarium graminearum are screened by computer-aided drug design and machine learning and other technologies, and saturation mutation is performed to obtain an optimal triple mutant with significantly improved affinity and stability. The optimal triple mutant is applied to the prevention and treatment of wheat fusarium rust caused by Fusarium graminearum, and the effect is better than that of a small molecule cyanoacrylate fungicide, phenamacril. Background Art

[0002] Fusarium graminearum is a widespread soil-borne fungal pathogen. The fusarium headblight (Fusarium headblight) caused by it can cause premature fading of spikelets and lead to seed abortion, wilt or size reduction, resulting in reduced grain yield and quality, which has a devastating impact on wheat production. Specifically, the conidia of Fusarium graminearum attach to the surface of wheat spikelets. With the help of hydrophobic proteins, the spores germinate on the wheat flower tissue and use creeping hyphae to grow on the surface of the plant to form multicellular attachment cells, leading to tissue death at the infected site. During the infection process, a variety of trichothecenes harmful to humans and animals are produced, such as deoxyniverin and zearalenone. my country stipulates that the maximum residue limits of deoxyniverin and zearalenone in wheat for human consumption are 1 mg / kg and 60 μg / kg, respectively.

[0003] At present, the main means to solve Fusarium head blight include disease-resistant breeding and fungicides. Wheat Fusarium head blight resistance is controlled by multiple genes. The Fhb1 gene is the main gene for wheat resistance to Fusarium head blight. It is widely used in wheat breeding around the world and improves resistance to Fusarium head blight. However, due to the limited disease-resistant germplasm, fungicides are still the main strategy for the prevention and control of Fusarium head blight. The benzimidazole fungicide carbendazim was introduced in the late 1960s and became one of the effective agents for the prevention and control of Fusarium head blight during the heading and flowering period of wheat. It binds to the β-tubulin of pathogenic fungi and inhibits the assembly of microtubules, thereby preventing the growth of fungal hyphae. However, long-term use has led to resistance to carbendazim in many plant pathogenic fungi. In 1992, isolates of Fusarium graminearum resistant to carbendazim were first discovered in Zhejiang Province, China, and this resistance has now been widely reported in other wheat-producing areas in China. Studies have shown that the mechanism of resistance to benzimidazole fungicides is due to point mutations in the β-tubulin gene in pathogenic fungi or mutations in different amino acid sites of the β2-tubulin gene of Fusarium graminearum at codons 167, 198 and 200. Currently, the widely used small molecule cyanoacrylate fungicide, phenamacril, prevents and controls Fusarium graminearum by inhibiting the myosin activity of Fusarium graminearum, but Fusarium graminearum has developed resistance and the prevention effect has gradually decreased.

[0004] In summary, wheat breeding cycle is long, and there are no highly resistant varieties at present. In harsh field environments, the antifungal efficiency of biocontrol agents will decrease. Long-term use of chemical disinfectants will increase the resistance of fungal pathogens and cause environmental pollution. Therefore, it is urgent to explore new alternative methods to inactivate Fusarium graminearum. Among them, antimicrobial peptides have diverse mechanisms of action, a wide antibacterial spectrum, and low toxicity to host cells compared with genetically modified crops and chemical pesticides. They are more in line with the requirements of green environmental protection and are the most potential technical direction for solving the problem of wheat fusarium rust. However, the research on antimicrobial peptides in the prevention and control of Fusarium graminearum is still in its infancy. The research mainly focuses on the synthesis and structural modification of antimicrobial peptides, the inhibitory effect on pathogens and their in-depth study of their mechanism of action. The research on the application of antimicrobial peptides in the prevention and control of Fusarium graminearum is relatively small, and lacks in-depth systematic research. Although antimicrobial peptides with inhibitory effects on Fusarium graminearum have been screened, most of the inhibitory mechanisms are unclear, the peptide chain stability is poor, the application conditions are harsh, and the inhibitory effect on Fusarium graminearum in the field environment is limited. Summary of the invention

[0005] In view of the above problems, the present invention provides a method for designing and screening antimicrobial peptides that inhibit Fusarium graminearum. Antimicrobial peptides that target myosin of Fusarium graminearum are screened from peptide chains that are inhibitory to Fusarium graminearum, and saturation mutation is performed to obtain an optimal triple mutant that significantly improves the affinity and stability with myosin. The peptides inhibit the ATP hydrolysis activity of myosin by binding outside the catalytic center of myosin, thereby preventing the growth and spread of the fungus. This not only achieves the purpose of inhibiting Fusarium graminearum, but also reduces the impact on other non-target organisms, thereby being environmentally friendly.

[0006] A method for designing and screening antimicrobial peptides that inhibit Fusarium graminearum, comprising the following steps:

[0007] Step 1: Molecular docking of the peptide chain that has an inhibitory effect on Fusarium graminearum with the myosin of Fusarium graminearum to screen out the peptide chain that targets the myosin of Fusarium graminearum;

[0008] Step 2: Mutate the selected peptide chains and select mutants for the purpose of improving the affinity and stability with myosin.

[0009] Furthermore, the method for designing and screening antimicrobial peptides that inhibit Fusarium graminearum also includes step three: feature engineering to select amino acid composition, dipeptide composition, tripeptide composition, pseudo-amino acid composition and composition transition distribution, establish machine learning models and deep learning models, and predict the antibacterial ability of antimicrobial peptides.

[0010] Preferably, the machine learning model in step three includes a support vector machine, a random forest and a gradient boosting decision tree model.

[0011] Preferably, the deep learning model in step three includes a convolutional neural network, a recurrent neural network and a bidirectional recursive neural network model.

[0012] Preferably, the step 1 aims to screen out a peptide chain having a stronger binding free energy to myosin of Fusarium graminearum than the small molecule inhibitor phenamacril.

[0013] Furthermore, the peptide chain amino acid sequence screened in step 1 is NH2-RGFRRR-COOH.

[0014] Preferably, the mutation sites determined by single mutation in step 2 for the purpose of improving affinity and stability with myosin include 1, 3, 4, 5 and 6.

[0015] Preferably, the double mutation site combinations in step 2 for the purpose of improving affinity and stability with myosin include four: (1, 3) (3, 4) (3, 5) and (3, 6).

[0016] Preferably, in step 2, the triple mutation site combination for the purpose of improving affinity and stability with myosin is (1, 3, 4).

[0017] Furthermore, the amino acid sequence of the mutant obtained by triple mutation is NH2-KGPPRR-COOH.

[0018] Compared with the prior art, the present invention has the following technical advances:

[0019] 1) The present invention selects myosin as the target protein of Fusarium graminearum. Myosin is a key motor protein in fungi, involved in cell movement and infection, and its conservation in fungi and lack of homologs in plants make targeting myosin have high specificity and safety. The selection of myosin as a target is different from the idea that traditional pesticides mainly target fungal metabolic pathways or cell wall synthesis, which reduces the impact on the environment and non-target organisms.

[0020] 2) The present invention conducts molecular docking of antimicrobial peptides and myosin, and through accurate prediction of binding modes, deeply understands how antimicrobial peptides interact with myosin. The present invention performs saturation mutation on antimicrobial peptides, systematically replaces amino acid residues, evaluates their effects on antimicrobial peptide stability and binding affinity, and combines computational simulation with experimental verification to improve the optimization efficiency and accuracy of antimicrobial peptides. The antimicrobial peptides designed and screened by the present invention bind to myosin more freely than the small molecule inhibitor Phenamacril.

[0021] 3) The present invention combines multiple features such as amino acid composition, dipeptide composition, and tripeptide composition to establish multiple machine learning and deep learning models for the prediction of antimicrobial peptide activity. The strategy of comprehensive application of multiple models not only improves the prediction accuracy, but also provides a rich perspective for the design and optimization of antimicrobial peptides. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 The present invention is a flow chart of the design and screening method of antimicrobial peptides for inhibiting Fusarium graminearum.

[0023] Figure 2 This is a diagram of the binding pattern of the small molecule inhibitor Phenamacril to Fusarium graminearum myosin.

[0024] Figure 3 This is a diagram showing the binding pattern of the antimicrobial peptide AMP_07 and Fusarium graminearum myosin in the example.

[0025] Figure 4 Schematic diagram of the change in binding energy (Kcal / mol) of the antimicrobial peptide AMP_07 saturation mutation, where: (a) The amino acid sequence of the wild-type AMP_07 polypeptide is RGFRRR. (b) Change in free energy of AMP_07 single mutation (binding ability to myosin). (c) Change in binding energy of AMP_07 double mutation. (d) Change in binding energy of AMP_07 triple mutation.

[0026] Figure 5 Flowchart for predicting the MIC value of antimicrobial peptides against Fusarium graminearum by training machine learning and deep learning models.

[0027] Figure 6 This is a bar graph comparing the antibacterial effects of AMP_07 and triple mutant TP in the biological experiment to verify the antibacterial concentration test of antimicrobial peptides against Fusarium graminearum. DETAILED DESCRIPTION

[0028] The following is a further detailed description of the design and screening method of an antimicrobial peptide for inhibiting Fusarium graminearum according to the present invention in conjunction with the accompanying drawings and specific embodiments:

[0029] Example:

[0030] The inhibitory mechanism of the antimicrobial peptide provided by the present invention in inhibiting Fusarium graminearum is as follows:

[0031] This example uses myosin (PDB ID: 6UI4) of Fusarium graminearum as the research object. Myosin includes a superfamily of molecular motors driven by ATP and is involved in a variety of cellular processes, including muscle contraction, vesicle transport, cytokinesis, organelle movement and sensory transduction. Myosin has a highly conserved ATP and myosin binding motor domain, a force-transducing lever arm containing one to several isoleucine / glutamine (IQ) motifs, and a variable cargo-interaction tail domain. The IQ motif can serve as a binding domain for calmodulin and calmodulin-like myosin light chains, providing rigidity to the lever arm and increasing motor activity. ATP hydrolysis and the release of hydrolysis products are coupled to actin binding and large "powerful stroke" lever arm movements on actin.

[0032] The inhibitor inhibits the ATP hydrolysis activity of myosin by binding outside of its catalytic center. This binding does not prevent ATP binding, and ATP, Mg2+, and water molecules are still correctly positioned for ATP hydrolysis. In the presence of actin, ATP is rapidly hydrolyzed at the catalytic center, but the release of the phosphate after hydrolysis is slow because this process requires rearrangement of the phosphate-binding Switch1 and Switch2 loops and requires actin binding to accelerate. Structural alignments show that inhibitor binding sterically interferes with cleft closure in the absence of further major rearrangements in inhibitor-bound myosin (from Fusarium graminearum) compared to the nucleotide-free, actin-bound cleft-closed state of myosin V (from chicken). The inhibitor occupies one of the open cleft structures of myosin. The corresponding pocket in the myosin V closed cleft structure has only Too small to accommodate inhibitors. Although the extent of opening and closing of the myosin binding cleft varies in structures with and without nucleotides and nucleotide analogs, structural alignments suggest that different types of myosins are sterically perturbed by inhibitor binding when the cleft is closed. Thus, inhibitors may affect multiple steps in the myosin / actin cycle, but their binding is likely to keep myosin in an “open,” pre-power stroke state.

[0033] 1. Establishment and docking screening of antimicrobial peptide library:

[0034] 52 antimicrobial peptides with experimental inhibitory activity against Fusarium graminearum were collected. Their structures were predicted using Alphafold2. By modeling the structures of these antimicrobial peptides, an antimicrobial peptide library specifically targeting Fusarium graminearum was established. ADCP was used to molecularly dock the antimicrobial peptides with myosin of Fusarium graminearum, and the best antimicrobial peptide was AMP_07 (binding free energy of -19.2 kcal / mol).

[0035] Table 1. Calculation of binding free energy of antimicrobial peptides and myosin (“-” means no suitable conformation can be found)

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043] 2. Comparison of binding free energy between AMP_07 and the small molecule inhibitor phenamacril:

[0044] The binding free energy of antimicrobial peptides with small molecules and myosin was calculated by molecular mechanics-Boisson integral method. It was found that the binding free energy of AMP_07 (-107.1786 kcal / mol) is much greater than that of phenamacril (-44.1280 kcal / mol). In other words, the affinity of AMP_07 with myosin is indeed greater than that of the small molecule inhibitor phenamacril. Through molecular docking, the binding mode of the inhibitor with myosin was explored ( Figure 2 and 3 ). It was found that the antimicrobial peptide AMP_07 had hydrogen bond interactions with multiple key amino acids of myosin (L214, K216, S379, D383, G533, D539). And the amino acids that had hydrogen bond interactions were far more than those of small molecule inhibitors. This shows that the antimicrobial peptide AMP_07 has a better binding ability with Fusarium graminearum myosin than small molecules.

[0045] 3. Saturation mutation of AMP_07 to improve stability and affinity with myosin:

[0046] like Figure 4 As shown, in order to improve the binding force and stability of the antimicrobial peptide AMP_07 to myosin, AMP_07 was subjected to saturation mutation. In the single mutation, mutation sites 1, 3, 4, 5, and 6 all improved the affinity and stability of AMP_07. Based on the single mutation, double mutation was performed, and it was found that the optimal sites for double mutation included (1, 3) (3, 4) (3, 5) and (3, 6). Based on the two optimal results of the double mutation, triple mutation was performed to obtain the best triple mutant TP of AMP_07, whose amino acid sequence is KGPPRR.

[0047] 4. Establish an artificial intelligence model to predict the antibacterial ability MIC value of antimicrobial peptides:

[0048] like Figure 5 As shown, based on the established database of antimicrobial peptides against Fusarium graminearum, machine learning and deep learning models are trained to predict the MIC values ​​of antimicrobial peptides to accelerate the research and development of antimicrobial peptides. The amino acid sequence of the antimicrobial peptide is taken as the starting point, and the MIC value is the target. Feature engineering selects 5 features, namely amino acid composition, dipeptide composition, tripeptide composition, pseudo-amino acid composition and composition transition distribution. Model training selection, support vector machine, random forest and gradient boosting decision tree model (XGBoosting) in machine learning models, convolutional neural network, recurrent neural network and bidirectional recursive neural network in deep learning models. After adjusting the hyperparameters and related parameters of the model optimization, the best model is the gradient boosting decision tree model. The gradient boosting decision tree model mainly uses RMSE (root mean square error) as the main evaluation criterion, R 2 As an auxiliary evaluation criterion, the RMSE of the model is 1.8, which means that the average error between the MIC value predicted by the model and the actual MIC value is 1.8. As a regression model with a large MIC range, it is relatively excellent, and R 2 The fitting result is 0.76, which is close to the actual situation. In summary, the proposed gradient boosting decision tree is a relatively accurate model for predicting the MIC value of antimicrobial peptides against Fusarium graminearum.

[0049] 5. Biological experiments verify the inhibitory activity of antimicrobial peptide AMP_07 and its mutants against Fusarium graminearum:

[0050] In order to verify the antibacterial activity of antimicrobial peptides predicted by the machine learning model, biological experiments were carried out to verify the antibacterial ability of the antimicrobial peptide AMP_07 and its triple mutant TP against Fusarium graminearum.

[0051] like Figure 6 As shown in the figure, the experimental results are roughly the same as the trend of the model prediction results. The blue bar graph is the AMP_07 wild type, the red bar graph is the triple mutant TP, and the blank group is 0.09. When the concentration of the triple mutant TP is 8.89μM, the fungus basically did not grow. When the concentration of the AMP_07 wild type is 20.00μM, the fungus basically did not grow. Therefore, the model is basically reliable and can be used to accelerate the drug discovery of Fusarium graminearum. The obtained mutant has better antibacterial ability than the wild type.

[0052] The examples of the present invention are described in detail above in conjunction with the embodiments, but the present invention is not limited to the above examples. Various changes can be made within the knowledge scope of ordinary technicians in the field without departing from the purpose of the present invention, and should also be regarded as the protection scope of the present invention.

Claims

1. A method for designing and screening antimicrobial peptides that inhibit Fusarium graminearum, characterized in that: The following steps are involved: Step 1: Molecular docking of the peptide chain that has an inhibitory effect on Fusarium graminearum with the myosin of Fusarium graminearum to screen out the peptide chain that targets the myosin of Fusarium graminearum; Step 2: Mutate the selected peptide chains and select mutants for the purpose of improving the affinity and stability with myosin.

2. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 1, characterized in that: It also includes step three: feature engineering to select amino acid composition, dipeptide composition, tripeptide composition, pseudo-amino acid composition and composition transition distribution, establish machine learning models and deep learning models, and predict the antibacterial ability of antimicrobial peptides.

3. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 2, characterized in that: The machine learning models in step three include support vector machine, random forest and gradient boosting decision tree models.

4. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 2, characterized in that: The deep learning model in step three includes convolutional neural network, recurrent neural network and bidirectional recursive neural network models.

5. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 1, characterized in that: The goal of step 1 is to screen out a peptide chain with a stronger binding free energy to myosin of Fusarium graminearum than the small molecule inhibitor thiostrobin.

6. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 5, characterized in that: The peptide chain amino acid sequence screened in step 1 is NH2-RGFRRR-COOH.

7. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 6, characterized in that: In the step 2, the mutation sites determined by single mutation to improve the affinity and stability with myosin include 1, 3, 4, 5 and 6.

8. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 6, characterized in that: In the step 2, there are four double mutation site combinations for improving affinity and stability with myosin through double mutation: (1, 3) (3, 4) (3, 5) and (3, 6).

9. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 6, characterized in that: In the step 2, the three mutation sites for the purpose of improving the affinity and stability with myosin through double mutation are combined as (1, 3, 4).

10. The method for designing and screening antimicrobial peptides for inhibiting Fusarium graminearum according to claim 9, characterized in that: The amino acid sequence of the mutant obtained by triple mutation is NH2-KGPPRR-COOH.

Images