Broad-spectrum antibacterial peptide with low hemolytic activity, application thereof and broad-spectrum antibacterial drug

Abstract

This invention belongs to the field of antimicrobial peptide technology, and particularly relates to a broad-spectrum antimicrobial peptide with low hemolytic activity and its applications, as well as a broad-spectrum antibacterial drug. The antimicrobial peptide is named AY6, and its amino acid sequence is shown in SEQ ID NO.1. Antimicrobial peptide AY6 exhibits significant antibacterial effects against both Gram-positive and Gram-negative bacteria, with a bactericidal rate far exceeding that of traditional antibiotics, enabling highly effective responses to bacterial infections. AY6 has low hemolytic activity and good biocompatibility, making it suitable for medical anti-infective applications. The antimicrobial peptide provided in this application has a short sequence, is easy to synthesize chemically, can save on large-scale production costs, and has broad application prospects.

Description

Technical Field

[0001] This invention belongs to the field of antimicrobial peptide technology, and particularly relates to a broad-spectrum antimicrobial peptide with low hemolytic activity and its applications. Background Technology

[0002] With the long-term and widespread use of antibiotics, bacterial resistance has become an increasingly serious problem, and the continuous emergence of multidrug-resistant bacteria has posed a significant challenge to clinical treatment and public health safety. Traditional antibiotics have relatively simple mechanisms of action and are prone to inducing drug resistance. The development of new antibiotics is time-consuming and costly, making it urgent to develop alternative antibacterial strategies with novel mechanisms of action.

[0003] Antimicrobial peptides are a class of short peptide molecules that are naturally produced or artificially synthesized by organisms. They are typically composed of multiple amino acid residues and possess broad-spectrum antimicrobial activity. Antimicrobial peptides exert their bacteriostatic or bactericidal effects primarily by disrupting bacterial cell membrane structure, interfering with cell metabolism, or inhibiting microbial biofilm formation. Their mechanisms of action differ significantly from those of traditional antibiotics, making them less likely to induce drug resistance. They show promising applications in medical anti-infection, biomedical materials, and food safety.

[0004] However, antimicrobial peptides still face many challenges in their practical development and application. On the one hand, natural antimicrobial peptide resources are limited, and different antimicrobial peptides vary greatly in terms of antimicrobial activity, stability, and biosafety. On the other hand, some antimicrobial peptides have problems such as high hemolytic or cytotoxicity and insufficient in vivo stability, which limit their further application.

[0005] Developing an antimicrobial peptide with excellent comprehensive performance, which not only has significant antibacterial or bactericidal activity against bacteria and other microorganisms, but also has good biocompatibility, such as low hemolysis and low cytotoxicity, is of great significance for overcoming the problems of insufficient safety or limited application of existing antimicrobial peptides in practical applications. Summary of the Invention

[0006] To address the problems existing in the prior art, the present invention first provides a broad-spectrum antimicrobial peptide with low hemolytic activity, the amino acid sequence of which is shown in SEQ ID NO.1.

[0007] The present invention further provides the application of the antimicrobial peptide described above in the preparation of antimicrobial drugs.

[0008] Preferably, the bacteria inhibited by the antibacterial drug include Gram-positive bacteria and Gram-negative bacteria.

[0009] Preferably, the bacteria inhibited by the antibacterial drug are any one or a combination of Escherichia coli, Bacillus subtilis, Pseudomonas aeruginosa, Acinetobacter baumannii, Staphylococcus aureus, and Enterococcus faecalis.

[0010] Preferably, when the inhibited bacteria is *Escherichia coli*, the minimum effective concentration of the antimicrobial peptide in the antimicrobial drug is 4 μg / mL; when the inhibited bacteria is *Bacillus subtilis*, the minimum effective concentration of the antimicrobial peptide in the antimicrobial drug is 8 μg / mL; when the inhibited bacteria is *Pseudomonas aeruginosa*, the minimum effective concentration of the antimicrobial peptide in the antimicrobial drug is 8 μg / mL; when the inhibited bacteria is *Acinetobacter baumannii*, the minimum effective concentration of the antimicrobial peptide in the antimicrobial drug is 26.7 μg / mL; and when the inhibited bacteria is *Enterococcus faecalis*, the minimum effective concentration of the antimicrobial peptide in the antimicrobial drug is 32 μg / mL.

[0011] Finally, the present invention also provides a broad-spectrum antibacterial drug containing a pharmaceutically effective dose of the broad-spectrum antibacterial peptide as described above.

[0012] The beneficial effects of this invention are as follows: This application uses a sequence generation algorithm based on a diffusion model combined with parallel computing on a large computing cluster to generate a massive number of virtual candidate antimicrobial peptide sequences. Then, a deep learning model is used to predict and finally screen representative antimicrobial peptides AY1-6, among which AY6 has superior broad-spectrum antimicrobial properties.

[0013] Experiments show that the antimicrobial peptide AY6 has significant antibacterial effects against both Gram-positive and Gram-negative bacteria, especially against Escherichia coli and Pseudomonas aeruginosa, with a minimum inhibitory concentration as low as 4 μg / mL. Moreover, its bactericidal speed is far superior to that of traditional antibiotics (it can kill the target strain in 1-4 hours), making it highly effective in combating bacterial infections.

[0014] Experiments show that AY6 has no hemolytic activity in the concentration range of 4~512 μg / mL. Its low hemolytic activity and good biosafety avoid the problem of insufficient safety of traditional antimicrobial peptides, making it more suitable for medical anti-infection and other scenarios.

[0015] The antimicrobial peptides provided by the invention have short sequences, are easy to synthesize chemically, can save on the cost of large-scale production, and have broad application prospects. Attached Figure Description

[0016] Figure 1 This is a diagram showing the amino acid composition of 154 sequences.

[0017] Figure 2 The image shows the MIC heatmaps of six candidate antimicrobial peptides against three bacteria, with the colors of the heatmaps corresponding to the MIC values.

[0018] Figure 3 The results show the bactericidal kinetics of the antimicrobial peptide AY6 against Escherichia coli.

[0019] Figure 4 The results show the bactericidal kinetics of the antimicrobial peptide AY6 against Bacillus subtilis.

[0020] Figure 5 The results show the bactericidal kinetics of the antimicrobial peptide AY6 against Pseudomonas aeruginosa. Detailed Implementation

[0021] To facilitate understanding, the technical solution of the present invention will be described in more detail below with reference to experiments.

[0022] Example 1

[0023] Acquisition of antimicrobial peptides

[0024] 1. Formation of antimicrobial peptides

[0025] This study employs a diffusion-based generative model for antimicrobial peptide sequence design, reproducing the previously reported antimicrobial peptide diffusion generation architecture (Science Advances, 2025, 11, eadp7171) to systematically explore the virtual peptide sequence space in a high-performance computing environment. The diffusion model learns the probability distribution characteristics of known antimicrobial peptide sequences during the training phase. During the generation phase, it uses random noise sequences as initial input and gradually recovers amino acid sequences conforming to the characteristics of antimicrobial peptides through multi-step denoising iterations, thereby generating novel antimicrobial peptide sequences. Basic constraints are imposed on the length and amino acid composition of candidate peptide sequences during sequence generation to ensure the biological rationality of the generated sequences. The sequence length is limited to the common range of typical antimicrobial peptides (approximately 8-50 amino acid residues), and only 20 natural amino acids are included. In the implementation, batch processing job scripts are constructed in a high-performance computing cluster environment, and tasks are submitted to a GPU computing queue for parallel execution through a job scheduling system. The job scripts set multi-core CPU and GPU resource requests and call the diffusion model main program to execute sequence sampling and generation tasks. The model performs parallel inference in a GPU-accelerated environment, generating a large number of candidate antimicrobial peptide sequences through random seed initialization and multiple sampling strategies to achieve high-throughput exploration of the potential antimicrobial peptide sequence space. After sequence generation, a self-developed Python script is used to uniformly process and standardize the format of the raw sequence data. This script automatically reads the valid sequence information in the generated results, removes blank lines and redundant characters, and renumbers and organizes the sequences according to the standard FASTA format to generate standardized sequence files. Simultaneously, basic quality control processing is performed on the generated sequences, including removing completely repetitive sequences and sequences containing non-standard amino acid characters or abnormal symbols. Through these steps, the large-scale raw sequence data generated by the diffusion model can be transformed into standardized FASTA input files, providing a unified data format foundation for subsequent deep learning predictions and bioinformatics analysis.

[0026] 2. Deep learning screening

[0027] To accurately assess the antimicrobial potential of candidate sequences, this application integrates two advanced deep learning models for parallel prediction of minimum inhibitory concentration (MIC) values. First, an antimicrobial peptide prediction model based on the EvoGradient framework is used for preliminary screening. The EvoGradient model originates from a recently reported framework for antimicrobial peptide design and optimization (Nature Microbiology, 2025, 10: 332–347), which comprises an antimicrobial peptide classification model and an MIC regression prediction model. It uses a deep neural network to extract features from amino acid sequences and determine antimicrobial activity, while simultaneously quantitatively predicting the MIC values ​​of candidate sequences. The model constructs a training dataset based on a publicly available antimicrobial peptide database. The training data mainly comes from reported antimicrobial peptides and their corresponding MIC experimental data, establishing a mapping relationship between sequence features and antimicrobial activity through supervised learning. In the prediction stage, generated FASTA sequence files are imported into the model in batches as input, yielding antimicrobial activity classification results and predicted MIC values ​​for different strains. Secondly, to further improve the reliability and robustness of the prediction results, this study introduces the BERT-AmPEP60 protein language model based on the Transformer architecture for independent prediction. This model is built upon a ProtBERT pre-trained protein language model and fine-tuned under supervision on an antimicrobial peptide dataset to achieve prediction of antimicrobial peptide activity and MIC values ​​(Journal of Chemical Information and Modeling, 2025, 65:3186–3202). The ProtBERT model, through pre-training on a large-scale protein sequence database, can learn contextual dependencies and potential structural features in amino acid sequences, thereby improving the characterization ability of antimicrobial peptide functional properties. Since there is a lack of training weight files suitable for this research task in publicly available resources, the model was retrained based on existing antimicrobial peptide datasets in this study to obtain optimized model parameters for the MIC prediction task of *Escherichia coli* and *Staphylococcus aureus*. Finally, based on the prediction results of the two models mentioned above, the predicted MIC values ​​for both *Escherichia coli* (EC) and *Staphylococcus aureus* (SA) were set to ≤10 μM. From tens of thousands of generated sequences, 154 highly active candidate peptides were initially screened. The characteristics of the screened sequences are as follows: Figure 1 As shown in Table 1, the horizontal axis represents the types of amino acids in the antimicrobial peptides, using one-letter abbreviations for amino acids, with their specific meanings as shown in Table 1.

[0028] Table 1. Comparison of Amino Acid Names

[0029] 3. Obtaining candidate peptides

[0030] To ensure the clinical application potential of candidate antimicrobial peptides, rigorous biosafety assessments were conducted on the highly active sequences obtained from the initial screening. The HemoPI2 online platform was used to predict the hemolytic potential of candidate sequences in batches, screening for sequences with "non-hemolytic" characteristics. Simultaneously, the ToxinPred tool was used to assess the toxicity of all sequences, excluding molecules with cytotoxic risks. At this stage, the activity threshold was further tightened (requiring predicted MIC values ​​≤ 5 μM), and cross-referenced with the "non-hemolytic" and "non-toxic" screening results.

[0031] Based on the above screening, in order to further verify the model prediction results and conduct experimental optimization, from the 154 highly active candidate antimicrobial peptides, taking into account the predicted MIC value, amino acid composition characteristics and sequence differences, 6 representative antimicrobial peptides were selected for chemical synthesis and numbered AY1 to AY6, as shown in Table 2.

[0032] Table 2 Results of 6 representative antimicrobial peptides

[0033] In the table, the SVM score is the support vector machine (SVM) score output by the ToxinPred model; a higher value indicates that the peptide is more likely to be toxic. HemoPI is a model for predicting hemolytic activity, with the half-maximal concentration (HCG) as the median. 50 The lower the value, the stronger the hemolytic activity and the more dangerous it is.

[0034] All six antimicrobial peptides met the basic screening criteria of low predicted MIC values, non-hemolytic activity, and non-toxicity, and were used for subsequent comparative studies of their in vitro antibacterial activity and bactericidal effects. A biotechnology company was commissioned to synthesize the six candidate antimicrobial peptides using a solid-phase synthesis process.

[0035] Example 2

[0036] Minimum inhibitory concentration (MIC) comparison experiment of candidate antimicrobial peptides

[0037] The antimicrobial peptides AY1–AY6 of six candidate antimicrobial peptides were tested against Escherichia coli using the microbroth dilution method. Escherichia coli MG1655, Bacillus subtilis ( Bacillus subtilis )168 and Pseudomonas aeruginosa ( Pseudomonas aeruginosaThe minimum inhibitory concentration (MIC) of PAO1 was determined. *Escherichia coli* MG1655 is a classic Gram-negative bacterium strain commonly used for evaluating the in vitro antimicrobial activity of antimicrobial peptides; *Bacillus subtilis* 168 is a Gram-positive bacterium whose cell wall structure differs significantly from that of Gram-negative bacteria, and can be used to assess the broad-spectrum antimicrobial properties of antimicrobial peptides; *Pseudomonas aeruginosa*... PAO1 These are common opportunistic pathogens in clinical practice, exhibiting strong resistance to multiple antibiotics. They serve as an important model for evaluating the inhibitory ability of antimicrobial peptides against drug-resistant Gram-negative pathogens. Selecting the above three strains as test subjects, covering Gram-negative / positive and clinically drug-resistant bacteria, can comprehensively reflect the antimicrobial spectrum and potential application value of candidate antimicrobial peptides.

[0038] The experimental methods were performed according to ISO 20776-1:2019. Specifically, before the experiment, each experimental strain was inoculated into a suitable liquid culture medium and cultured with shaking at 37°C until the logarithmic growth phase. Subsequently, the bacterial suspension was diluted with sterile Mueller-Hinton Broth (MHB) to adjust each strain to the same initial inoculation concentration, which was approximately 1 × 10⁻⁶. 5 CFU / mL. Candidate antimicrobial peptides AY1–AY6 were dissolved in sterile PBS (phosphate-buffered saline) buffer and diluted using a two-fold serial dilution method to prepare antimicrobial peptide solutions of different mass concentrations, ranging from 0.5 to 256 μg / mL. Each concentration gradient of antimicrobial peptide solutions was added to a predetermined volume into a 96-well plate, followed by inoculation of an equal volume of bacterial suspension into each well to ensure consistent final volume and bacterial concentration. Control groups were also included: a bacterial growth control well containing only bacterial suspension without antimicrobial peptides; and a blank control well containing only sterile culture medium, without bacterial suspension or antimicrobial peptides. The 96-well plates were incubated at 37 ℃ for 16–20 h. After incubation, the turbidity of the bacterial suspension in each well was visually observed to assess bacterial growth. The MIC value was defined as the lowest concentration of antimicrobial peptide at which no bacterial growth was observed.

[0039] Six candidate antimicrobial peptides, AY1 to AY6, were tested in parallel under the same experimental conditions to compare the differences in antimicrobial activity of different antimicrobial peptides against various bacterial strains.

[0040] Experimental results are as follows Figure 2As shown in the figure. The results indicate that the antimicrobial activities of different candidate antimicrobial peptides differed significantly against the tested strains. Among them, antimicrobial peptide AY6 showed lower MIC values ​​in Escherichia coli, Bacillus subtilis, and Pseudomonas aeruginosa, and its overall antimicrobial activity was superior to the other five candidate antimicrobial peptides, demonstrating more stable and broad-spectrum antimicrobial performance.

[0041] Further comprehensive comparison of the MIC results of various candidate antimicrobial peptides revealed that antimicrobial peptide AY6 exhibited a more balanced antimicrobial effect across different strains and maintained a high level of antimicrobial activity in multiple tested strains. Based on the above MIC comparison results, antimicrobial peptide AY6, with the best overall antimicrobial activity, was selected for subsequent hemolytic activity evaluation and bactericidal kinetic experiments.

[0042] Example 3

[0043] Evaluation of the antimicrobial activity of antimicrobial peptide AY6 against clinical pathogenic bacteria

[0044] The minimum inhibitory concentration (MIC) of antimicrobial peptide AY6 against a variety of Gram-positive and Gram-negative bacteria was determined using the micro-broth dilution method, with the experimental method being the same as in Example 2.

[0045] The test strains included Gram-positive bacteria: Staphylococcus aureus O269 and Enterococcus faecalis 22023; and Gram-negative bacteria: Acinetobacter baumannii O351 and Klebsiella pneumoniae O515. The experimental results are shown in Table 3.

[0046] Table 3. MIC of antimicrobial peptide AY6 against clinical pathogenic bacteria

[0047] It can be seen that the antimicrobial peptide AY6 has antibacterial activity against a variety of Gram-positive and Gram-negative bacteria, and shows good antibacterial effect against Acinetobacter baumannii and Enterococcus faecalis, indicating that the antimicrobial peptide has certain broad-spectrum antibacterial potential.

[0048] Example 4

[0049] Hemolytic assessment

[0050] The hemolytic activity of the antimicrobial peptide AY6 was evaluated using a 2% sheep erythrocyte suspension. Before use, sheep erythrocytes were thoroughly resuspended in sterile PBS, centrifuged, resuspended, and washed three times until the supernatant was clear and free of visible red color, thus removing free hemoglobin and plasma residues. The washed erythrocytes were then resuspended in sterile PBS to prepare a 2% (v / v) erythrocyte suspension. Antimicrobial peptide AY6 was prepared into sample solutions of different concentrations (4–512 μg / mL) using sterile PBS as the solvent through a two-fold serial dilution method. Control groups were included in the experiment: the negative control group consisted of an equal volume of sterile PBS and a 2% sheep erythrocyte suspension; the positive control group consisted of an equal volume of 1% Triton X-100 and a 2% sheep erythrocyte suspension, representing complete hemolysis. AY6 solutions of various concentrations were mixed with an equal volume of 2% sheep erythrocyte suspension and incubated at 37 ℃ for 30 min to simulate acute hemolysis. After incubation, the reaction system was centrifuged at 1000 g for 3 min to allow non-hemolyzed erythrocytes to settle. The supernatant from each sample was then carefully aspirated and transferred to 96-well plates, with 100 μL added to each well. The absorbance (OD) of hemoglobin in the supernatant of each well was measured at 540 nm using a microplate reader. 540 The hemolysis rate was calculated using the following formula, and the results are shown in Table 4.

[0051]

[0052] Table 4. Hemolysis rate of antimicrobial peptide AY6 at different concentrations

[0053] As shown in Table 4, AY6 has no hemolytic activity and good biosafety (the hemolysis rate in the table is negative, which is due to experimental systematic error and there is actually no hemolytic activity).

[0054] Example 5

[0055] bactericidal kinetics

[0056] Escherichia coli MG1655, Bacillus subtilis 168, and Pseudomonas aeruginosa PAO1 were inoculated onto LB solid agar plates and incubated upside down at 37°C until colonies appeared. Single colonies were then picked up with an inoculation loop and inoculated into LB liquid agar, incubated at 37°C and 180 rpm until the logarithmic growth phase. The bacterial suspension was then adjusted to OD500 with sterile PBS. 600=0.01 for later use. Set up a negative control with PBS, adding 900 μL of sterile PBS and 100 μL of bacterial suspension to each tube; for the experimental group, add 800 μL of sterile PBS and 100 μL of bacterial suspension to each tube, and finally add 100 μL of peptide solution diluted with sterile PBS to bring the final concentration to 2. MIC. Positive control is 2 corresponding volumes. MIC ciprofloxacin CIP. Incubate at 37°C and 180 rpm. Take 10 μL of bacterial suspension at time points of 0, 0.25, 0.5, 1, 4, 8, 12, and 24 h, and serially dilute to 1000-fold at 10-fold intervals (96-well plate operation, 90 μL of PBS added to each well before dilution). Drop each cell suspension onto LB agar, incubate the plate upside down at 37°C overnight, and count the colonies. Plot a time-kill curve with time on the x-axis and the logarithm of the number of surviving bacteria on the y-axis. The results are shown below. Figures 3-5 As shown.

[0057] Depend on Figures 3-5 It can be seen that although the antimicrobial peptide AY6 has a similar bactericidal effect to ciprofloxacin CIP against Escherichia coli MG1655, it has a rapid bactericidal effect against Bacillus subtilis 168 and Pseudomonas aeruginosa PAO1. At the MIC concentration, AY6 can kill Bacillus subtilis 168 in 4 hours and Pseudomonas aeruginosa PAO1 in 1 hour, while the positive control ciprofloxacin CIP requires 12 hours.

[0058] The above tests demonstrate that the antimicrobial peptide AY6 has broad-spectrum antibacterial activity, high biosafety, and rapid bactericidal effect against clinical pathogens such as Bacillus subtilis and Pseudomonas aeruginosa, with better efficacy than ciprofloxacin, and has good application potential.

[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A broad-spectrum antimicrobial peptide with low hemolytic activity, characterized in that, The antimicrobial peptide is named AY6, and its amino acid sequence is shown in SEQ ID NO.

1.

2. The use of the antimicrobial peptide as described in claim 1 in the preparation of antimicrobial drugs.

3. The application as described in claim 2, characterized in that, The bacteria inhibited by the antimicrobial drug include Gram-positive and Gram-negative bacteria.

4. The application as described in claim 2, characterized in that, The bacteria inhibited by the antibacterial drug are any one or a combination of Escherichia coli, Bacillus subtilis, Pseudomonas aeruginosa, Acinetobacter baumannii, Staphylococcus aureus, and Enterococcus faecalis.

5. A broad-spectrum antibacterial drug, characterized in that, The drug contains a pharmaceutically effective dose of the broad-spectrum antimicrobial peptide as described in claim 1.

Images