Novel Antimicrobial Peptides Based on Large-Scale Pretrained Models and Their Applications

Abstract

The present invention discloses novel antimicrobial peptides based on large pre-trained models and their applications, belonging to the field of bioengineering technology. The antimicrobial peptides AMP3, AMP6, AMP13, AMP19 and AMP23 provided by the present invention have broad-spectrum antibacterial properties and have great application value in the development and preparation of antibacterial agents. They can be applied in the aquaculture industry as feed additives, and can also be developed into antibacterial agents and antibacterial drugs, etc., with broad application prospects.

Description

Technical Field

[0001] The present invention relates to novel antimicrobial peptides based on large pre-trained models and their applications, belonging to the field of bioengineering technology. Background Art

[0002] Antimicrobial peptides are short-chain polypeptides (which can also be understood as short-chain proteins) with antibacterial effects. Compared with small-molecule antibiotics, antimicrobial peptides have multiple advantages, including more common broad-spectrum properties, more diverse sources, immunomodulatory effects, and so on. Therefore, antimicrobial peptides provide a new breakthrough in the fight against "superbugs" to meet the challenge of the global depletion of new antibiotics.

[0003] Currently, there are various methods proposed in the academic community for generating antimicrobial peptides through artificial intelligence. The most advanced and closest prior art to the present invention is the paper published by Das et al. in 2021 (Das, Payel, et al. "Accelerated antimicrobial discovery via deep generative models and molecular dynamics simulations." Nature Biomedical Engineering 5.6 (2021): 613-623.). However, the feature extraction of the polypeptide sequences by Das et al. is not based on a large model, so the effect is poor. Therefore, developing new screening methods and obtaining new antimicrobial peptides have become important issues that need to be solved urgently. Summary of the Invention

[0004] The present invention provides antimicrobial peptides with the amino acid sequences shown in (1) to (5) below:

[0005] (1) GLFKIFKKSVKHACKKLK (SEQ ID NO.1);

[0006] (2) KYLRRKGGRKWVKNAFRNLM (SEQ ID NO.2);

[0007] (3) YRGIRGKGWWKAAKAALKAVSAAAKH (SEQ ID NO. 3);

[0008] (4) KIKEKLRQKLGNIMEKHLHKLAHRILLKK (SEQ ID NO.4);

[0009] (5) VYRKTFNAKKIGLDIRRGWKGLKKIGGKI (SEQ ID NO.5).

[0010] The present invention also provides a composition with antibacterial function, and the composition contains one or more of the following antibacterial peptides: GLFKIFKKSVKHACKKLK, KYLRRKGGRKWVKNAFRNLM, YRGIRGKGWWKAAKAALKAVSAAAKH, KIKEKLRQKLGNIMEKHLHKLAHRILLKK, VYRKTFNAKKIGLDIRRGWKGLKKIGGKI.

[0011] In one embodiment, the composition is a pharmaceutical antibacterial agent or a non-pharmaceutical antibacterial agent.

[0012] In one embodiment, the non-pharmaceutical antibacterial agent includes a disinfectant, a cleaning agent, a feed additive or a textile additive.

[0013] In one embodiment, the concentration of the antibacterial peptide in the composition is ≥ 1.875 μM.

[0014] The present invention also provides the use of the antibacterial peptide in the preparation of an antibacterial agent.

[0015] In one embodiment, the application is as an antibacterial active substance or as an active ingredient in an antibacterial agent for inhibiting or killing Gram-positive bacteria, Gram-negative bacteria or fungi.

[0016] In one embodiment, the Gram-positive bacteria include, but are not limited to, Staphylococcus aureus ( Staphylococcus aureus ), Bacillus pumilus ( Bacillus pumilus ), Bacillus subtilis ( Bacillus subtilis ), Micrococcus luteus ( Kocuria rhizophila ).

[0017] In one embodiment, the Gram-negative bacteria include, but are not limited to, Escherichia coli ( Escherichia coli ).

[0018] In one embodiment, the antibacterial agent is a pharmaceutical antibacterial agent.

[0019] In one embodiment, the antibacterial agent is a non-pharmaceutical antibacterial agent.

[0020] In one embodiment, the non-pharmaceutical antibacterial agent includes, but is not limited to, disinfection products, food additives, feed additives or textile additives.

[0021] In one embodiment, the concentration of the antibacterial peptide in the antibacterial agent is ≥ 1.875 μM.

[0022] The present invention also claims the use of the antibacterial peptide or its salt in the preparation of products in the fields of medicine, daily chemicals, etc.

[0023] In one embodiment, the products include, but are not limited to, disinfectants, cleaners, preservatives, biological pesticides, etc.

[0024] In one embodiment, the application includes, but is not limited to, adding the antimicrobial peptide to the product or semi-finished product during the product production process.

[0025] Beneficial effects:

[0026] The antimicrobial peptides AMP3, AMP6, AMP13, AMP19 and AMP23 provided by the present invention have broad-spectrum antibacterial properties and have great application value in the development and preparation of antibacterial agents. They can be used as feed additives in the aquaculture industry, or developed into antibacterial agents and antibacterial drugs, etc. Therefore, they have broad application prospects. Description of the drawings

[0027] Figure 1 It is the method flow and discriminator architecture, where A is the method flow and B is the discriminator architecture.

[0028] Figure 2 It is the test results of the growth of the tested strains by the predicted and randomly generated antimicrobial peptides. The concentration of the tested drug is set to 60 μM, and an equal volume of DMSO is used as a negative control. All conditions are tested with three biological replicates. Among them, A is the tested strain Staphylococcus aureus S. GDMCC 1.221, B is the tested strain Bacillus pumilus B. pumilus GDMCC 1.225, C is the tested strain Bacillus subtilis B. subtilis GDMCC 1.222, D is the tested strain Micrococcus luteus K. rhizophila GDMCC 1.226, E is the tested strain E. coli GDMCC 1.335, and F is the tested strain E. coli DH5ɑ. Detailed implementation manners

[0029] The polypeptides used in the following examples were purchased from Nanjing Genscript Biotech Co., Ltd. The production method of the polypeptides adopted the solid-phase polypeptide synthesis method. The crude peptides with a synthesis purity of 40 - 70% were used for antibacterial experiments. For the convenience of calculation, the purity of all crude peptides was set to 50%. The crude peptides were further purified by high-performance liquid chromatography to obtain high-purity polypeptides with a purity exceeding 90% for MIC experiments. The molecular weights of all synthesized polypeptides were identified by mass spectrometry.

[0030] The tested strains used in the following examples, among which, Staphylococcus aureus ( Staphylococcus aureus ), GDMCC 1.221, Bacillus pumilus ( Bacillus pumilus ), GDMCC 1.225, Bacillus subtilis ( Bacillus subtilis) GDMCC 1.222, Micrococcus luteus ( Kocuria rhizophila ) GDMCC 1.226 and Escherichia coli ( Escherichia coli ) GDMCC 1.335 were purchased from Guangdong Microbial Culture Collection Center (GDMCC), which can be used for antibiotic potency testing. The strain E. coli DH5ɑ is a commonly used laboratory strain and is preserved by the laboratory.

[0031] Example 1 Method for generating and screening effective antimicrobial peptides

[0032] like Figure 1 As shown in the flow chart, the antimicrobial peptide design method includes the following steps:

[0033] (1) Configure the operating environment of Python and its toolkits: Use software including the Python environment management software Anaconda and the version control program git. Use Anaconda to install Python 3.8, create and activate a Python environment dedicated to this method. Use the Python toolkit management command pip to install toolkits related to machine learning, including flax, pandas, datasets, evaluate, scikit-learn, and natural language processing toolkits, including SentencePiece and transformers. Anaconda installs toolkits related to the deep learning framework Pytorch, including pytorch, torchvision, torchaudio, pytorch-cuda, and the learning frameworks keras and tensorflow-gpu.

[0034] (2) Obtain samples: Obtain known antibacterial peptide (hereinafter referred to as antimicrobial peptide) sequences (as positive samples) from the Antimicrobial Peptide Database (SATPdb). Obtain short-chain peptide sequences (as negative samples) and their annotations from the Protein Public Database (UniProt-SwissProt). Eliminate specific keywords contained in peptide annotations; exclude peptides containing non-natural protein amino acids in positive and negative samples, and limit positive and negative samples to within 50 amino acid residues; and ensure that the number of peptides containing the same number of amino acids in positive and negative samples is the same.

[0035] (3) Fine-tune the protein generation model ProtGPT2: Use git to obtain the transformers project directory on GitHub, and fine-tune the protein generation model ProtGPT2 using known antimicrobial peptide sequences ( Figure 1 A).

[0036] (4) Use a large pre-trained model to extract features from the sequences, and use the positive and negative sample sequences for which feature extraction has been performed to train the discriminator: The architecture of the discriminator is a deep convolutional neural network (as shown in Figure 1 Figure B), which is implemented by the Keras toolkit in Python. The optimizer is Adam (learning rate is 0.001), the loss function is binary crossentropy, the evaluation method is accuracy, the feature (X) is the feature extracted in the above step (embedding), and the label (y) is the yes / no antimicrobial peptide in the positive and negative samples.

[0037] (5) Use the fit function in the Keras model object to train the neural network for X and y: During the training process, the training set is split into a training set and a validation set in a 9:1 ratio. The maximum number of training epochs is set to 75, and the EarlyStopping function in Keras is used to set an early exit mechanism, with the patience value set to 6, that is, if the loss value does not decrease within six rounds of training, then the training will end early.

[0038] (6) Use the reserved test set to test the trained discriminator and evaluate the accuracy (using the evaluate function in the Keras model object). Subsequently, randomly selected negative controls and polypeptides generated without fine-tuning are also predicted (using the predict function in the Keras model object) to evaluate the changes before and after fine-tuning of the generation model for reference.

[0039] (7) Use the trained discriminator to predict these generated candidate polypeptides: Screen the candidate polypeptides according to the predicted value (antimicrobial peptide probability > 0.95), and randomly select 24 candidate sequences, numbered sequentially as "AMP1" - "AMP24" for experimental verification. In addition, according to the length distribution and amino acid distribution of "AMP1" - "AMP24", a set of polypeptide sequences are randomly generated as negative controls, numbered sequentially as "RP1" - "RP10".

[0040] Example 2 Bacterial culture

[0041] Single colonies of Staphylococcus aureus GDMCC 1.221, Bacillus pumilus GDMCC 1.225, Bacillus subtilis GDMCC 1.222, Micrococcus luteus GDMCC 1.226, and Escherichia coli GDMCC 1.335 were picked from fresh plates and inoculated into LB liquid medium respectively, placed on a shaker, and cultured overnight at 37 °C and 200 rpm. The overnight bacterial solution was inoculated into fresh LB medium at a ratio of 1:100 and cultured for 4 - 5 h until the logarithmic growth phase (OD600 0.4 - 0.6). The bacterial solution in the logarithmic phase was diluted with LB medium to OD 600 = 0.1, then further diluted 1000 times, and aliquoted into 96-well plates (the order of magnitude of the diluted bacterial solution was about 1×10 6 CFU / mL), 100 μL per well, for antibacterial experiments and MIC experiments.

[0042] Example 3 Antibacterial Experiment

[0043] The polypeptide screened in Example 1 was synthesized by chemical synthesis. The synthesized crude peptide was dissolved in dimethyl sulfoxide (DMSO) to a protein concentration of 3 mM to prepare a polypeptide stock solution. In the 96-well plate containing the diluted bacterial solution (prepared according to the method of Example 1), 4 μL of the crude peptide stock solution was added to each well, and the total volume was supplemented to 200 μL with fresh LB medium, so that the concentration of the polypeptide to be tested was 60 μM. An equal volume of DMSO was used as a negative control, and 200 μL of fresh LB medium was used as a blank control. After adding the samples, the lid of the 96-well plate was sealed with a sealing film and placed in a high-speed shaking incubator, and cultured at 37 °C and 500 rpm for 11 - 12 h. Three biological replicates were performed for each condition. After the culture was completed, the OD 600 of each well was detected with a microplate reader, and the OD 600 of the blank control was subtracted.

[0044] Verified (such as Figure 2 , Table 1), under the test conditions, AMP3, AMP6, AMP13, AMP14, AMP16, AMP19, AMP23, and RP4 could completely inhibit the growth of Staphylococcus aureus GDMCC 1.221, AMP12 and AMP15 could partially inhibit the growth of this bacterium, and the positive rate of the predicted polypeptide was 37.5% (9 / 24), higher than the positive rate of 10% (1 / 10) of the randomly generated polypeptide.

[0045] AMP3, AMP6, AMP8, AMP13, AMP14, AMP16, AMP19, AMP23 and RP4 can completely inhibit the growth of Bacillus pumilus GDMCC 1.225. AMP4, 5, 7, 12, 15 and 18 can partially inhibit the growth of this bacterium. The predicted positive rate of polypeptides is 58.3% (14 / 24), which is higher than the positive rate of 10% (1 / 10) of randomly generated polypeptides.

[0046] AMP3, AMP6, AMP8, AMP13, AMP14, AMP16, AMP19 and AMP23 can completely inhibit the growth of Bacillus subtilis GDMCC 1.222. AMP1, AMP4, AMP5, AMP10, AMP12, AMP15, AMP17, AMP18, AMP20, AMP21, RP1, RP2, RP3, RP4 and RP6 can partially inhibit the growth of this bacterium. The predicted positive rate of polypeptides is 78.3% (18 / 24), which is higher than the positive rate of 50% (5 / 10) of randomly generated polypeptides.

[0047] AMP3, AMP6, AMP8, AMP13, AMP16, AMP19, AMP23 and RP4 can completely inhibit the growth of Micrococcus luteus GDMCC 1.226. AMP24, RP6, RP7 and RP8 can partially inhibit the growth of this bacterium. The predicted positive rate of polypeptides is 29.2% (7 / 24), which is close to the positive rate of 30% (3 / 10) of randomly generated polypeptides.

[0048] AMP3, AMP6, AMP13, AMP19 and AMP23 can completely inhibit the growth of Escherichia coli GDMCC 1.335. All randomly generated polypeptides have no inhibitory effect on the strain GDMCC 1.335. The predicted positive rate of polypeptides is 20.8% (5 / 24), which is much higher than the positive rate of 0% (5 / 10) of randomly generated polypeptides; AMP3, AMP6, AMP13, AMP19, AMP23, AMP24 and RP4 can completely inhibit the growth of the strain DH5ɑ. The predicted positive rate of polypeptides is 25% (6 / 24), which is higher than the positive rate of 10% (1 / 10) of randomly generated polypeptides.

[0049] Table 1 shows the information of the antibacterial peptides predicted and randomly generated in Example 1 and their antibacterial effects

[0050]

[0051] Note: The concentration of the tested drug was set at 60 μM, and an equal volume of DMSO was used as a negative control. All conditions were tested with three biological replicates. Among them, AMP represents the polypeptide predicted by this method, and RP represents the randomly generated polypeptide. "+++" indicates complete growth inhibition, "+" indicates inhibitory effect compared with the control but not complete inhibition, and "-" indicates no inhibition.

[0052] Example 4 MIC Experiment

[0053] The high-purity polypeptide sample was dissolved in DMSO to 6 mM to prepare a stock solution, and then diluted into 8 gradients with DMSO by the method of two-fold serial dilution. In the 96-well plate containing the diluted bacterial solution (prepared with reference to Example 2), 4 μL of the sample was added to each well, and the total volume was supplemented to 200 μL with fresh LB medium, so that the final concentrations of the tested polypeptides were 120 μM, 60 μM, 30 μM, 15 μM, 7.5 μM, 3.75 μM, 1.875 μM, and 0.9375 μM respectively, and the bacterial solution concentration was xx CFU / mL. After adding the samples, the 96-well plate was placed in a high-speed shaking incubator and cultured at 37 °C and 500 r.p.m. for 16 - 18 h. Each condition was replicated three times biologically. After the culture, the growth of bacteria in the wells was observed, and the lowest antibacterial drug concentration in the well without bacterial growth was the minimum inhibitory concentration. After verification, except that the MIC of AMP19 for some strains was greater than 120 μM, the MICs of AMP3, AMP6, AMP13, and AMP23 were all between 3.75 μM and 120 μM, proving that these antibacterial peptides have broad-spectrum properties, great application value, and good application in the development and preparation of antibacterial agents.

[0054] Table 2 shows the results of the MIC experiment

[0055]

[0056] Note: All conditions were tested with three biological replicates.

[0057] Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person familiar with this technology can make various modifications and decorations without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be defined by the claims.

Claims

1. Antimicrobial peptide, characterized in that, The amino acid sequence is as shown in any one of (1) to (5) below: (1) GLFKIFKKSVKHACKKLK; (2) KYLRRKGGRKWVKNAFRNLM; (3) YRGIRGKGWWKAAKAALKAVSAAAKH; (4) KIKEKLRQKLGNIMEKHLHKLAHRILLKK; (5) VYRKTFNAKKIGLDIRRGWKGLKKIGGKI.

2. A composition with antibacterial function, characterized in that, Containing one or more of the following antimicrobial peptides: the amino acid sequence of the antimicrobial peptide is as shown in any one of (1) to (5): (1) GLFKIFKKSVKHACKKLK; (2) KYLRRKGGRKWVKNAFRNLM; (3) YRGIRGKGWWKAAKAALKAVSAAAKH; (4) KIKEKLRQKLGNIMEKHLHKLAHRILLKK; (5) VYRKTFNAKKIGLDIRRGWKGLKKIGGKI.

3. The composition according to claim 2, characterized in that, The composition is a pharmaceutical antimicrobial agent or a non - pharmaceutical antimicrobial agent.

4. The composition according to claim 3, characterized in that, The non - pharmaceutical antimicrobial agent includes a disinfectant, a cleaning agent, a feed additive or a textile additive.

5. The composition according to any one of claims 2 to 4, characterized in that, The concentration of the antimicrobial peptide in the composition is ≥ 1.875 μM.

6. Use of the antimicrobial peptide according to claim 1 in the preparation of an antimicrobial product, characterized in that, The application is as an antibacterial active substance, or as an active ingredient in an antimicrobial agent, for inhibiting or killing Gram - positive bacteria or Gram - negative bacteria; the Gram - positive bacteria are Staphylococcus aureus, Bacillus pumilus, Bacillus subtilis and / or Kocuria rhizophila; the Gram - negative bacteria are Escherichia coli.

7. Use of the antimicrobial peptide or a salt thereof according to claim 1 in the preparation of products in the pharmaceutical or daily chemical fields, characterized in that, The product is a disinfectant, a cleaning agent, a preservative or a biopesticide.

Images