The bacterial strains Staphylococcus aureus (S. aureus) CMCC(B) 26003 and Pseudomonas aeruginosa (P. aeruginosa) CMCC(B)10104, and the fungal strain Candida albicans (C. albicans) CMCC(F) 98001, were purchased from Beijing Kezhan Biotechnology Co., Ltd. 2019, China. The bacterial strains Bacillus thuringiensis and Escherichia coli were obtained from our laboratory. Mouse RAW 264.7 macrophages were purchased from Shanghai Yaji Biotechnology Co., Ltd. 2022, China.

Mueller–Hinton broth (MHB), Mueller–Hinton agar (MHA), nutrient agar (NA), and Sabouraud dextrose agar (SDA) were obtained from BEIJING AOBOXING BIO-TECH Co., Ltd. 2019, China. Trypsin was obtained from Nanjing Beyotime Biotechnology Co., Ltd., 2021. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was purchased from Sigma-Aldrich, 2021 (China).

Staphylococcus aureus CMCC(B) 26003, B. thuringiensis, E. coli, and P. aeruginosa CMCC(B)10104 were cultured in NA medium and C. albicans CMCC(F) 98001 was cultured in SDA medium. They were incubated in a constant-temperature (bacteria: 37°C; fungi: 28°C) incubator for 16–20 h. Thereafter, single colonies were picked and incubated in 50 mL MHB at 160 rpm/min at the optimal growth temperature (bacteria: 37°C; fungi: 28°C). Microbial suspensions of 1 × 10⁶ CFU/mL were then prepared (Chen et al., 2016).

The whole-genome sequence of Protaetia brevitarsis Lewis larvae (GenBank accession number: CM014285.1) was compared to the AMP sequences in APD3.¹ Aligned sequences were used as template peptide sequences (Houyvet et al., 2018).

The template peptide was used to screen out the truncated sequences, and then the antimicrobial region within the predicted peptide was screened using the random forest classifier calculation method in CAMP² to predict the possibility of becoming AMPs. Thereafter, the obtained truncated candidate peptides were inputted into Antimicrobial Peptide Predictor³ in APD3 for prediction of the probability of being the truncated sequence of an AMP. The physicochemical properties of the peptides were analyzed by using the ExPASy website software ProtParam⁴ in order to select peptides with a high electrostatic charge, high hydrophobic amino acid ratio, high stability, CAMP prediction of high probability of being an AMP, and APD3 prediction of high probability of being the truncated sequence of an AMP (Yang et al., 2018).

In addition, depending on the structure–function relationships relevant to each AMP, the peptide was selectively modified by amino acid substitution or N-terminal acetylation. For example, negatively charged amino acid residues could be replaced with positively charged amino acid residues (ANWDKVIR changed to ANWKKVIR) and -NH₂ could be added to the end of the peptide (STLHLVLRLR-NH₂; STLHLVLRLR) to preserve its functional activity (Chen et al., 2016; Yang et al., 2018). The 16 designed peptides were sent to Beijing Coolaber Technology Ltd. (2020). for chemical synthesis.

100 μL of 1 × 10⁶ CFU/mL microbial suspension (E. coli, P. aeruginosa CMCC(B)10104, B. thuringiensis, S. aureus CMCC(B) 26003, or C. albicans CMCC(F) 98001) was evenly applied to dishes containing 20 mL NA (bacteria) or SDA (fungi). After the microbial solution solidified, six holes per dish were punched with a sterile pipette tip of 6 mm in diameter, with each hole spaced >20 mm apart and > 10 mm apart from the edge of the dish. The holes were marked with (+), (−), and (1–16), representing the positive control (ampicillin), negative control (deionized water), and treatment groups (16 candidate peptides dissolved in sterile water), respectively. Next, 55 μL of 1 mg/mL control/treatment was added to each hole and incubated in a constant-temperature (bacteria: 37°C; fungi: 28°C) incubator for 16–18 h (Chen et al., 2016; Gong et al., 2022), with three replicates for each treatment group. The diameter of the inhibition circle was measured using ImageJ software (National Institutes of Health, United States) in order to preliminarily determine whether the candidate peptides had inhibitory activity.

The MIC values of five candidate peptides (FD4, FD10, FD12, FD14, and FD15), which were selected based on physicochemical properties and antimicrobial effects (determined by the agar punch method), were evaluated using a micro-broth dilution method (Zhou and Zhou, 2018). Serial two-fold dilutions of the candidate peptides were prepared in 96-well plates containing the same microbial inoculum (E. coli, P. aeruginosa CMCC(B)10104, B. thuringiensis, S. aureus CMCC(B) 26003, or C. albicans CMCC(F) 98001). The plates were incubated (bacteria: 37°C; fungi: 28°C) for 16–18 h. Ampicillin was used as the positive control. The MIC was defined as the lowest peptide concentration where no visible growth occurred (Diao et al., 2014). The experiment was conducted in triplicate.

Six AMPs (FD10, FD12, FD13, FD14, FD15, and FD16), selected based on antimicrobial activity and the peptide modifications, were processed by Heliquest⁵ (Gautier et al., 2008) in order to generate helical wheel projection diagrams.

During the antimicrobial activity screening it was found that the five antimicrobial AMPs (FD4, FD10, FD12, FD14, and FD15) had inhibitory effects against E. coli and S. aureus CMCC(B) 26003. Therefore, the follow-up experiments involved these two bacteria, along with the three AMPs with the best antimicrobial effects (FD10, FD12, and FD15). The dynamic antimicrobial effects were analyzed by measuring the effects of the AMPs on bacterial growth over 10 h. First, 100 μL microbial solution (1.0 × 10⁶ CFU/mL) was added to the wells of a 96-well culture plate, and then 100 μL AMP (at 1× and 1/4× MIC) was added and the mixtures were cultured at 37°C (Mishra et al., 2013; Memariani et al., 2016). 100 μL MHB with 100 μL microbial solution was used as the control group. The optical density at 600 nm (OD₆₀₀) was measured every hour for 10 h using a microplate reader (Hou et al., 2007). The experiment was conducted in triplicate.

The kill-time curve assay method was used to investigate the bactericidal effects of the essential oil according to the technique described by Joray et al. (2011). A time–kill assay was used to examine the rate at which three AMPs (FD10, FD12, and FD15) killed two microbes (S. aureus CMCC(B) 26003 and E. coli) in vitro over time, based on the number of viable bacteria left at various time points after AMP exposure. Briefly, AMP at 2× MIC was added to MHB containing 10⁶ CFU/mL S. aureus CMCC(B) 26003 or E. coli and the mixture was then incubated at 37°C and 100 rpm. At various time points, a sample was taken, serially diluted, and plated on MHA for colony counting (Flamm et al., 2019).

AMP at 2× MIC was incubated with the same volume of 1 mg/mL trypsin solution at 37°C for 1 h, followed by a 5 min 95°C exposure to inactive the trypsin. Thereafter, 100 μL of the resulting solution was added to 96-well plates, followed by 100 μL microbial suspension (1 × 10⁶ CFU/mL). Sterile water with 1 × 10⁶ CFU/mL microbial suspension was used as the blank control. After culturing at 37°C for 16–18 h, colony counting was conducted (Tresnak and Hackel, 2020). The experiment was conducted in triplicate.

E. coli and S. aureus CMCC(B) 26003 (1 × 10⁶ CFU/mL) were incubated in MHB with various final physiological concentrations of serum salts (150 mM NaCl, 4.5 mM KCl, 1 mM MgCl₂, and 4 mM FeCl₃). 50 μL of these E. coli or S. aureus CMCC(B) 26003 suspensions and 50 μL of AMP (serially diluted two-fold to a final concentration of 256 μg/mL) were added to a 96-well plate and incubated at 37°C for 16–18 h (Chou et al., 2016). The MIC values were then determined. The experiment was conducted in triplicate.

The hemolytic activity of AMPs on mammalian erythrocytes was determined by optical absorption (Lee et al., 2011). Briefly, fresh mouse erythrocytes were collected, washed, and resuspended in phosphate-buffered saline PBS at a concentration of 8% (v/v). Thereafter, 100 μL of this solution was added to a 96-well plate containing an equal volume of AMP (1–256 μg/mL) and incubated at 37°C for 1 h. The positive control was 100 μL 1% Triton X-100 and 100 μL 8% erythrocytes, and the negative control was 100 μL phosphate-buffered saline PBS and 100 μL 8% erythrocytes. Next, each mixture was centrifuged at 1,200 g for 15 min, and 100 μL of the supernatant was transferred to a clean 96-well plate to measure the light absorption value at 576 nm with a microplate reader. The percent hemolysis was calculated using the following formula: Percent hemolysis = [(A–A₀)/(A₊–A₀)] × 100%, where A, A₊, and A₀ represent the absorbance of the peptide sample and positive and negative controls, respectively (Li et al., 2020).

The cytotoxicity of AMPs toward mouse RAW 264.7 macrophages was determined by MTT assays (Meng and Kumar, 2007). Cells were added into 96-well microtiter plates (2.5 × 10⁴ cells/well) and incubated for 24 h. After another 24 h incubation with Amp and FD10, FD12, FD15 (1 to 256 μg/mL), MTT solution was added and incubated for 4 h. Followed the removing of the MTT, dimethyl sulfoxide (DMSO) (150 μL/well) was added and the absorbance was measured at 570 nm. Untreated cells were used as a control. The cell survival rate was calculated using the following formula: survival rate (%) = (Abs_{570 nm} of treated sample/Abs_{570 nm} of control) × 100% (Wang et al., 2018). The cytotoxicity detection was repeated in triplicate.

All data are presented as mean ± standard deviation (SD), with “n” represents the number of samples. One-way analysis or two-way analysis of variance (ANOVA) in SPSS software was used to analyze significant differences between groups. p < 0.05 indicated that the differences were statistically significant. GraphPad Prism 8.0 (GraphPad Software, United States) was used for statistical computing and plotting.

After comparing the Protaetia brevitarsis Lewis genomic sequence to known AMP sequences in a database, nine aligned sequences were identified and used as template peptides (AP02030, AP02257, AP02096, AP00489, AP01575, AP02128, AP01540, AP02012, and AP00208) (Supplementary Figure S1).

PHYRE²⁶ was used to predict the secondary structure of the nine template peptides to clearly and intuitively observe their structures, and to help to predict their functions. They all have typical secondary structures, i.e., α-helical structures (green spirals) and β-pleated structures (blue arrows) (other lines are random coil regions) (Supplementary Figure S2). The figure, which indicates the confidence of predictions (decreasing from red to purple), shows that the α-helical region prediction confidence is high (mostly in red). AMPs mainly depend on their α-helical structures, while the β-pleated structures were treated as auxiliary structures in terms of the subsequent related design work.

Based on the nine template peptides, key truncated sequences were identified (Supplementary Tables S1–S9), and 16 candidate peptides, designated FD1–16, were designed (Table 1). FD13(ANWDKVIR) contains a negatively charged amino acid residue, which was detrimental to the interaction of the antimicrobial peptide with bacterial membranes, so it could be considered to be replaced with Lys, which becomes FD14 (ANWKKVIR), thereby increasing the net charge from +1 to +3 without changing the hydrophobic amino acid ratio. The polypeptide sequence of FD16 (STLHLVLRRR) was the C-terminal sequence of the template peptide, and the control peptide FD15 (STLHLVLRR-NH₂) could be set up, and -NH₂ could be added to the FD16 terminal in order to preserve the integrity of functional activity.

The physicochemical properties of the 16 candidate peptides were clearly and concisely compared (Table 2). The net charge of each candidate peptide was 2–4, which was favorable for interacting with the negatively charged bacterial membrane. The high ratio of hydrophobic residues indicated easy insertion into the membrane interior to destroy the bacterial cells. The small molecular weight lowered chemical synthesis costs. The instability index was not high except for FD1, FD15, and FD16, which showed that most of the candidate peptides were relatively stable.

The antimicrobial effects of the 16 candidate peptides against five microbes (E. coli, P. aeruginosa CMCC(B)10104, B. thuringiensis, S. aureus CMCC(B) 26003, and C. albicans CMCC(F) 98001) were compared to the effects of ampicillin (positive control) and sterile water (negative control). The candidate peptides differed in their inhibitory action against the five microbes, based on the size of the inhibition circles formed (Supplementary Figure S3).

Five peptides (FD4, FD10, FD12, FD14, and FD15) generated inhibition circles, with the growth of the five microbes being prevented in these zones (the remaining candidate peptides did not form inhibition circles). FD4 was effective against E. coli and S. aureus CMCC(B) 26003, with inhibition circle diameters of 10.12 and 9.47 mm, respectively (Table 3). However, they were much smaller than those of ampicillin (29.37 and 19.98 mm), indicating a lower antimicrobial effect. FD10 was more effective (larger inhibition circle diameter) than ampicillin against P. aeruginosa CMCC(B)10104, B. thuringiensis, and S. aureus CMCC(B) 26003, and it was the only candidate peptide with activity against fungi (C. albicans CMCC(F) 98001), indicating excellent inhibitory effects. FD12 was comparable to ampicillin against B. thuringiensis and S. aureus CMCC(B) 26003, indicating antimicrobial activity against Gram-positive bacteria. FD14 and FD15 were both much less effective than ampicillin.

Based on the physicochemical properties and antimicrobial effects of the candidate peptides determined using the agar punch method, five (FD4, FD10, FD12, FD14, and FD15) were selected for further antimicrobial activity analysis. FD10 had an antimicrobial effect on all five microbes at a low MIC; the MIC against Gram-negative bacteria E. coli and P. aeruginosa CMCC(B)10104 was as low as 8 μg/mL, while those against Gram-positive B. thuringiensis and S. aureus CMCC(B) 26003 were 8 and 16 μg/mL, respectively, and that against C. albicans CMCC(F) 98001 was 16 μg/mL. FD12 and FD15 also had antimicrobial activity against S. aureus CMCC(B) 26003 with a MIC of 16 μg/mL each and against E. coli with a MIC of 32 μg/mL each (Table 4).

Heliquest was used to calculate the helix properties of six AMPs (FD10, FD12, FD13, FD14, FD15, and FD16). Supplementary Figure S4 shows the helical wheel projection diagrams of the AMPs. In this conformation, they did not have different physicochemical properties from those predicted by ExPASy-ProtParam. Both FD10 and FD12 (especially FD10) had a high proportion of electrostatically charged and hydrophobic amino acids, which are favorable for adsorption to the bacterial cell membrane and insertion into the membrane, causing bacterial death. Despite FD14 having a positively charged Lys in place of the negatively charged Asp in FD13, neither FD13 nor FD14 were antimicrobial, with both having a low hydrophilic index. The C-terminal acetylation modification gave FD15 antimicrobial activity. Helical wheel projection diagrams could not be constructed for FD13 and FD14 because the sequences were considered excessively short (only eight amino acid residues).

In the dynamic antimicrobial analysis, the OD₆₀₀ of the microbial solutions were essentially unchanged until 1 h, indicating slow microbe proliferation (Figure 2). At 1× MIC, for both AMPs and ampicillin, E. coli and S. aureus CMCC(B) 26003 growth was almost inhibited at 10 h (relative to the control [MHB with microbial solution]).

Ampicillin at 1/4× MIC led to slow E. coli growth after 1 h and increased growth after 2 h, while the AMPs at 1/4× MIC led to E. coli growth only after 3–4 h, and after 4 h, E. coli continued to grow. FD10 had a significantly higher inhibitory effect against E. coli than the other AMPs (Figure 2A). Similarly, ampicillin at 1/4× MIC led to S. aureus CMCC(B) 26003 growth at around 2 h, while AMPs at 1/4× MIC led to slow growth after 3 h (Figure 2B), and after 4 h, the number of S. aureus CMCC(B) 26003 began to increase more rapidly.

This indicates that, compared to ampicillin, the AMPs exhibited good microbial inhibition and broad antimicrobial activity even at concentrations below the MIC.

All three tested AMPs (FD10, FD12, and FD15) were bactericidal immediately after E. coli was exposed to them, with FD10 having the steepest curve and the fastest decline at any given time point, indicating that it was the most effective at killing E. coli, followed by FD12 and then FD15 (Figure 3A). The bactericidal effect of the three AMPs on both bacteria (E. coli and S. aureus CMCC(B) 26003) was most rapid before 0.5 h, with a large reduction in the number of colonies, and the rate of bactericidal action gradually decreased with time until the curves plateaued. Similarly, all three AMPs were bactericidal immediately after S. aureus CMCC(B) 26003 was exposed to them, with FD10 and FD12 causing a more rapid decline in S. aureus CMCC(B) 26003 than FD15, indicating that they were better than FD15 at killing S. aureus CMCC(B) 26003 (Figure 3B). By 10 h, the inhibitory effect of all three AMPs on S. aureus CMCC(B) 26003 decreased and the number of colonies began to increase, but the bacteria grew more slowly in the FD10 group.

After being treated with trypsin for 1 h, the antimicrobial activity of the three AMPs was greatly reduced. The antimicrobial activity was less than half of the original activity, with >60% of the bacteria surviving (Figure 4). This shows that the three AMPs were easily decomposed by trypsin and exhibited reduced activity.

The salt sensitivity of antimicrobial agents is generally tested by the addition of physiological concentrations of various salts (Table 5). The antimicrobial activities of the three AMPs (FD10, FD12, and FD15) against E. coli and S. aureus CMCC(B) 26003 were minimally affected, or even promoted, by the presence of certain monovalent (K⁺) and trivalent (Fe³⁺) cations, while they were slightly decreased in the presence of Na⁺ or Mg²⁺. Among FD10, FD12, and FD15, FD10 was the most stable in the presence of salt; when certain salt ions were present, it maintained or increased its antimicrobial activity.

The percentage of mouse erythrocyte hemolysis after FD10 and FD12 treatment (at 256 μg/mL) were 0.31 and 0.40%, respectively, which were lower than that of ampicillin (0.52%), while the rate after FD15 treatment was 0.55%, which was higher than that of ampicillin (Figure 5A). The survival rates of mouse RAW 264.7 macrophages after FD10, FD12, and FD15 treatment (at 256 μg/mL) were 75.37, 67.56, and 60.50%, respectively, which were lower than that of ampicillin (84.90%) (Figure 5B). The results showed that the hemolytic effects and cytotoxicity of FD10 and FD12 were lower than those of ampicillin, while the hemolytic effect of FD15 was higher than that of ampicillin.