Introduction

The emergence and spread of bacterial resistance in the environment dramatically increase the prevalence of drug-resistant bacteria1, particularly multidrug-resistant (MDR) bacteria. The prevalence of MDR strains represents a significant challenge in modern medicine because of their capacity to induce infections that are difficult to treat or even untreatable2,3. These infections are often associated with the formation of biofilms, which are complex structures that can protect bacteria from the effects of antimicrobial agents. The formation of biofilms results in bacteria exhibiting approximately tenfold greater resistance to antibiotics than their nonbiofilm-forming counterparts. This resistance is because bacteria within biofilms are in a dormant and subcellular state, rendering them highly insensitive to antibiotics, which is the primary reason for biofilm resistance and infection4. S. aureus, an ESKAPE (Enterococcus faecalis, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) pathogen, is a zoonotic bacterium with a strong biofilm-forming ability that can cause skin infections, mastitis and endometritis. As one of the few representative drugs for the treatment of MDR S. aureus and its biofilm infections, the antimicrobial vancomycin has shown a tendency to diminish in efficacy as the scope of application and dosage increases5, and drug-resistant S. aureus strains have been discovered one after another. The development of a new antimicrobial drug is needed to eradicate the biofilm formed by MDR S. aureus and the persisters in the biofilm2.

Low-cost, high-efficiency antimicrobial peptides (AMPs) are now the preferred alternative to antibiotics for dealing with infections caused by MDR bacteria, addressing the ineffectiveness and inefficiency of antibiotics in treating such bacteria5,6. AMPs are small and short peptides with cationic, amphiphilic, and hydrophobic properties and perform important functions with other molecules at low levels7,8,9. Furthermore, AMPs can also eradicate bacteria rapidly through mechanisms different from those of traditional antibiotics, such as disrupting cell membranes, inhibiting cell wall synthesis, and interfering with nucleic acids and metabolism, which makes them less susceptible to the development of drug resistance10. A framework for the future development of AMPs for use in medicine and other applications was provided. The defensin family of AMPs has been the subject of extensive research, and the LL-37-derived peptide SAPP-148 has been shown to have potent inhibitory effects on the growth and formation of biofilms of multidrug-resistant bacteria, as well as exerting a strong bactericidal effect on such bacteria11. The first fungal defensin, plectasin, has been extensively studied because to the characteristics similar to those of vancomycin and daptomycin against gram-positive (G+) bacteria with narrow-spectrum bactericidal effects, however, plectasin has led to failure in clinical phase II trials due to the high toxicity and poor stability12,13. AP138, a representative second-generation peptide derivative of plectasin, has inhibitory and killing effects on the G+ bacteria Staphylococcus and Streptococcus, and has potential as an anti-G+ drug candidate14,15.

Chemical synthesis is widely used for the synthesis of AMPs, particularly for the synthesis of peptides consisting of 10 amino acids and the nonribosomal synthesis of AMPs. However, the heterologous expression of AMPs with long natural amino acid (>20 aa) compositions is more cost effective16,17,18,19. Our recent study revealed that the natural AP138 sequence was expressed in high yield in the Pichia pastoris expression system. Furthermore, it has inhibitory effects on G+ bacteria and interacts with bacterial cell membranes, but its inhibitory effect on drug-resistant and multidrug-resistant bacteria is weaker than that on sensitive strains, and it might have weaker inhibitory and scavenging effects on biofilms15. In addition to achieving high-yield preparations of derived AMPs, enhancing the antimicrobial activity of a newly derived AMP against drug-resistant (DR) bacteria and their biofilms in in vivo environments, such as in the presence of serum and different ions, is equally important. Given the inherent limitations of AP138 in terms of its activity against multidrug-resistant bacteria and its capacity to inhibit biofilm formation, we sought to leverage bioinformatics to engineer novel AMPs that would not only retain the fundamental structural and functional relationships of the original peptide but also enhance its antimicrobial activity and inhibit the formation of biofilms in an in vivo-mimicking environment20,21.

The AP138-derived peptide was designed by point mutation through the bioinformatic random forest (RF) algorithm (accuracies of 93.2%) in the CAMP database, which not only maintains the original basic structure-function relationship but also improves the “AMP probability”, including the positive charge, hydrophobicity and α-helicity of AMPs. Finally, a high yield of AP138-derived peptide was attained via a yeast expression system. Its antimicrobial activity and stability in a physiological environment were then evaluated. The optimally derived peptide A24 was assessed in vitro, including activity against multidrug-resistant bacteria and internal persistent bacteria of biofilms, an exploration of bactericidal mechanisms, and an assessment of the development of resistance by a resistance induction assay at subinhibitory concentrations for 30 days. Finally, the therapeutic efficacy and safety were evaluated in a mouse peritonitis model and endometritis model.

Results

Screening of highly active peptides derived from AP138

AP138-derived peptides were designed using the bioinformatics tool “Rational Design of Antimicrobial Peptides” in CAMP22. The CAMP database has been modeled by the RF algorithm for the prediction and design of amino acid sequences based on parameters such as positive charge, hydrophobicity and secondary structure. The outcome of the prediction analysis is in the form of AMP probability scores. The “AMP probability” represents the potential for antibacterial activity; the higher the score (max = 1), the greater the likelihood of the sequence having antimicrobial activity23. The selection pipeline and protocol of the 12 candidate peptides used in this experiment are described below (Supplementary Fig. S1) (1) Focusing on the best AMP candidates reported to specifically target gram-positive bacteria via searches of key literature databases online and mainly from our previous works and papers in our hands, a total of 12 candidates were initially selected—AP138, AP114, Py4, NZX, MP1102, MP1106, DLP4, ID13, P2, etc. Finally, AP138 was selected from among them in terms of key antimicrobial properties such as activity, stability, toxicity, cost and other druggability factors at the initial stage of this work. (2) The sequence of AP138 as a mother template was input into the CAMP design window to generate a series of variants (100) via the rational design under the random forest algorithm online (Supplementary Fig. S1), and then these 100 derived peptide sequences were listed for evaluation, comparison and screening. (3) The top 43 candidates (AMP probability >0.9945) were preferably selected from the above 100 sequences (Supplementary Fig. S1, Supplementary Table S1). (4) The 43 sequences were compared with the MEGA software to construct an evolutionary tree, and the online prediction software was used to analyze the positive charge and hydrophobicity. The family lines of these 43 sequences were divided into various groups according to a comprehensive consideration of positive charge (+3.5, +4.25, +4.5, +5.5) and genetic distance—(I) family 1, variants with a net charge of +4.25 (11 sequences), AP138 and A11 were selected; (II) family 2-1, variants with a net charge of +4.5 (15 sequences), A42, A2, A36, A40, and A19 were selected; family 2-2, variants with a net charge of +4.5 (7 sequences), A15 was selected; family 2-3, variants with a net charge of +4.5 (3 sequences), A24 was selected; (III) family 3-1, variants with a net charge of +5.5 (4 sequences), A27 was selected; family 3-2, variants with a net charge of +5.5 (3 sequences), A24 was selected; and (IV) family 4, variants with a net charge of +3.5 (1 sequence), A25 was selected after which a total of 12 representative sequences were chosen for the above evaluation (Supplementary Fig. S1). The pPICZαA peptide plasmid was constructed and introduced into P. pastoris X-33 cells via linearization and electrotransfer. The AP138-derived peptides AP138-1, AP138-2, AP138-3, etc., are abbreviated as A1, A2, A3, etc. The 12 candidate sequences were screened by the inhibition zone assay; among them, A2, A4, A15, A24, and A36 had potent antimicrobial activity against MRSA ATCC 43300. A24, in which serine is replaced with alanine at position 9 in the sequence of AP138, was more hydrophobic, and the 3D predicted structure revealed a significant change in the electron cloud at position 9, which had a clearer and larger inhibition zone than the other derived peptides (A2, A4, A15 and A36) (Fig. 1a–d and Supplementary Fig. S2). Compared with AP138 and other clinical drugs, A24 presented the highest level of activity against MRSA ATCC 43300, with a minimum inhibitory concentration (MIC) of 2 μg/mL (Table 1), and the minimum bactericidal concentration (MBC) value remained unchanged in the presence of 25% plasma (Table 2). These findings indicate that A24 maintains high stability and antimicrobial activity in serum and different saltion environments. The A24-positive transformants were fermented in a 5 L fermenter at high density for 120 h. The total protein concentration in the fermentation supernatant reached 4.1 mg/mL, the wet weight of the bacterium reached 0.38 g/mL, and the inhibition zone diameter was the largest (Fig. 2a, b). The peptide was purified on a cation exchange column, and only one target band of approximately 4.6 kDa was identified by Tricine-SDS-PAGE. The molecular weight of the peptide was identified by mass spectrometry to be 4438.5 Da, which is consistent with the theoretical values (Fig. 2c–e). A24 has a typical α-helix and β-fold structure in aqueous solution according to predictions and circular dichroism (CD) assays, and the α-helix structure changed significantly (3.4–6.49%) in a 20 mM sodium dodecyl sulfate (SDS) solution. This change may be associated with the observed antibacterial activity. Conversely, the structure essentially remained unaltered in a 50% trifluoroethanol (TFE) solution, which was designed to simulate the eukaryotic cell membrane environment24 (Fig. 1e, f). This stability may be related to the low toxicity.

Fig. 1: The secondary structure of antimicrobial peptide A24.
figure 1

The 3D structure of the antimicrobial peptide A24. a, c are the α-helix and β-folding predicted by the AP138 and A24 ball-and-stick models, respectively, which do not change significantly, and at 9 as the electron cloud changes; b and d are the predictions of the AP138 and A24 stacking models, respectively, with the blue color representing the hydrophilicity, and the reddish-orange representing the hydrophobicity, which shows a significant change of the structure, with an increase in the reddish-orange color, with the model. e and f are the one-to-one counterparts of the AP138 and A24 secondary structure changes in ddH2O, 20 mM SDS and 50% TFE.

Full size image
Table 1 The MIC of AP138-derived peptides against MRSA ATCC 43300 in different environments
Full size table
Table 2 Determination of MBC of A24 in PBS and 25% plasma against Gram-positive bacteria
Full size table
Fig. 2: Expression of A24 in high-density fermentation in a 5 L fermenter.
figure 2

a Total protein and bacterial wet weight time curves at 0–120 h, b Inhibition of MRSA ATCC 43300 by antimicrobial peptide A24 at fermentation level was verified by inhibition zone test. c Tricine-SDS-PAGE analysis of A24 expression at the fermentation level, lanes 1–6 are fermentation supernatants from 0 to 120 h, respectively. d Tricine-SDS-PAGE analysis of fermentation supernatant purification, lane 4 is the elution peak. e MALDI-TOF MS analysis of the purified A24.

Full size image

The MBCs against selected G+ bacteria were determined to further evaluate the antibacterial activity of A2425, with nisin, vancomycin, and ceftiofur sodium serving as the controls. The results revealed that the MBCs of A24 in phosphate buffered saline (PBS) against S. aureus, Streptococcus agalactiae, Streptococcus dysgalactiae and Staphylococcus epidermidis were 2–8 μg/mL, whereas in 25% plasma, the MBCs were 2–16 μg/mL. The activity essentially remained unchanged. While AP138 and nisin in PBS demonstrated anti-G+ bacterial (S. aureus, S. agalactiae, S. dysgalactiae and S. epidermidis) activities of 8–32 μg/mL and 16–64 μg/mL, respectively, in 25% plasma, the activities were 8–64 μg/mL and 32–128 μg/mL, respectively. The MBCs of vancomycin and ceftiofur sodium in 25% plasma were 2–8 μg/mL (Table 2). The results indicate that A24 is highly stable in plasma and has superior antimicrobial activity against the aforementioned bacteria, with a more pronounced effect than that observed with AP138 and nisin. These findings suggest that A24 has strong antimicrobial activity against S. aureus, S. agalactiae, and S. epidermidis. These findings imply that A24 may have greater potential for application in treating drug-resistant bacteria.

Killing MDR strains without resistance selection by A24

Eight strains of drug-resistant S. aureus, which are stored in our laboratories, were used as test organisms to further investigate the bactericidal activity of A24 against resistant S. aureus. As shown in Supplementary Table S5, the MBC values of A24 against these S. aureus strains in PBS were 2–16 μg/mL; those of AP138, ceftiofur sodium, and vancomycin were 8–64 μg/mL, 2–32 μg/mL, and 2–4 μg/mL, respectively, which indicated that the activity of A24 was greater than those of AP138 and ceftiofur sodium. Moreover, in 25% plasma, the activity of A24 remained unchanged, with MBC values of 8–32 μg/mL, AP138 activity was greater than 128 μg/mL, vancomycin activity ranged from 4–8 μg/mL, and ceftiofur sodium activity ranged from 8–64 μg/mL, which indicated that A24 maintained high stability and bactericidal activity in plasma, which was superior to those of AP138, ceftiofur sodium and vancomycin. Overall, A24 has a potent bactericidal effect on drug-resistant S. aureus, indicating its potential as a promising candidate drug. A24 was subjected to further investigation to ascertain its efficacy against drug-resistant bacteria in vitro, and the killing curve, postantibiotic effects, and synergistic effects were determined. The results of the killing curve (Fig. 3a, b) revealed that different concentrations of A24 resulted in greater bactericidal effects than vancomycin, in which 2× MIC and 4 × MIC of A24 were able to kill all the methicillin-resistant Staphylococcus aureus (MRSA) ATCC 4300 bacteria at 1 h and 0.5 h, respectively, but 2× MIC of Van required 6 h to achieve the same outcome. When S. aureus CVCC 546 was treated, 2× MIC and 4× MIC of A24 were observed to eradicate the bacteria within 1.5 h and 1 h, respectively, and 2× MIC of Van killed all the bacteria within 6 h. However, 1× MIC of AP138 was unable to eradicate all the bacteria completely, and a rebound phenomenon was observed. The results of the time-killing curve indicate that A24 exhibited a more rapid bactericidal effect than Van. The results of the postantibiotic effect (PAE) assay (Fig. 3c, d) showed that the 2× MIC and 4× MIC of A24 against MRSA ATCC 43300 had durations of 1.5 h and 2.1 h, respectively, which were longer than those of the antibiotic and AP138. The 2× MIC and 4× MIC of A24 against S. aureus CVCC 546 were 2 h and 4 h, respectively, which were longer than those observed for the other treatments. The results of the analysis of the synergistic effect showed that the Fractional inhibitory concentration index (FICI) of A24 and ceftiofur sodium was 0.3125, indicating a superior combination effect (Supplementary Table S2).

Fig. 3: Time killing curve, PAE and 30 d resistance induction of A24 against MRSA ATCC 43300 and S. aureus CVCC 546 respectively.
figure 3

a Time killing curve of A24 against MRSA ATCC 43300, Van and AP138 as control; b Time killing curve of A24 against S. aureus CVCC 546, Cef and AP138 as control; c The post-antibiotic effect of A24 against MRSA ATCC 43300, Van and AP138 as control; d The post-antibiotic effect of A24 against S. aureus CVCC 546, Cef and AP138 as control; e Resistance of A24 against S. aureus ATCC 43300; f Resistance of A24 against S. aureus CVCC 546.

Full size image

The issue of drug resistance has consistently been a matter of significant concern within the scientific community. The ability of the drug-resistant S. aureus strains MRSA ATCC 43300 and CVCC 546 to develop resistance to A24 was evaluated. The results of the 30-day serial subinhibitory concentration-induced resistance of S. aureus (Fig. 3e, f) revealed that the MIC value of A24 against MRSA ATCC 43300 increased 4-fold, approaching that of vancomycin, whereas the MIC value of AP138 increased 8-fold at the 30-day passaging stage. In addition, the MIC of A24 against S. aureus CVCC 546 also increased 4-fold, reaching the same level as vancomycin, and the MICs of AP138 and ceftiofur sodium increased 8-fold at the 30-day passaging stage. The resistance genes of S. aureus CVCC 546 induced by A24, AP138, Cef, and Van at subinhibitory concentrations were identified, and the results are shown in Supplementary Fig. S3. No new resistance genes were produced following induction with A24, and the strain was observed to exhibit a reduction or even complete elimination of the drug resistance genes blaZ and mecA. A consideration of the use of A24 in combination with antibiotics as a means of shifting bacterial resistance to antibiotics would be prudent.

Eradication of biofilms and persistent bacteria by A24

The effects of A24 on MRSA ATCC 43300 and S. aureus CVCC 546 and their biofilms were observed26. The inhibitory effect of A24 on biofilms was dose-dependent (Supplementary Fig. S4). It inhibited 88% and 53% of the early and mature biofilms of MRSA ATCC 43300 at 8× MIC, respectively, which was close to the inhibition level at 16× MIC. In addition, A24 inhibited the early biofilm of S. aureus CVCC 546 by 92% and the mature biofilm by 60% at 4× MIC, which was greater than the effect of AP138 on the biofilm. These findings indicate that A24 has a pronounced inhibitory effect on early-stage biofilms and mature biofilms.

The inhibitory effect of A24 on biofilm endobacteria revealed that the bactericidal rates of 4× MIC A24 on the two strains of MRSA ATCC43300 and S. aureus CVCC 546 were 99.9% and 99.9%, respectively, which were greater than those of the 4 × MICs of AP138 (99.6% and 99.9%) and vancomycin (33.33% and 96%) on the same bacteria. Furthermore, the bactericidal rate of either 8× MIC of A24 or ceftiofur sodium reached 99.9% (Fig. 4b, d). This finding indicates that A24 has a notable inhibitory effect on the formation of early biofilms, with a more pronounced effect than that observed with AP138 and vancomycin. Further investigation into the effect of A24 on the retention of bacteria in the biofilm revealed a notable decrease in the number of bacteria (Fig. 4c, e). The results revealed that the number of bacteria decreased by 9 and 9.56 orders of magnitude after A24 treatment and then decreased by 4.2 and 6.9 orders of magnitude compared with that after vancomycin treatment against MRSA ATCC 43300 and S. aureus CVCC 546, respectively, and the number of bacteria changed to 2 and 7.1 orders of magnitude after AP138 treatment and decreased to 9 and 4.46 orders of magnitude compared with that after vancomycin treatment (4.2 and 1.8 orders of magnitude, respectively). These results indicate that A24 has a pronounced bactericidal effect on the bacteria within the biofilm, consistent with previous results. The biofilm was treated with 8 × MIC of A24 and stained with SYTO9/PI, which resulted in almost red fluorescence, indicating that A24 killed all the bacteria within the biofilm (Fig. 4a). An analysis of previous tests revealed that A24 has a strong ability to kill resistant bacteria in biofilms and inhibit the formation of biofilms. However, the capacity of A24 to eradicate biofilms is relatively limited.

Fig. 4: Effect of antimicrobial peptide A24 on biofilm and viability of drug-resistant bacteria.
figure 4

a Untreated and effect of A24 on MRSA ATCC 43300 and S. aureus CVCC 546 biofilm and activity; b Activity of different concentrations of A24 on MRSA ATCC 43300 biofilm endobacteria; c Killing of biofilm-persistent bacteria by 4× MIC A24; d Activity of different concentrations of A24 on S. aureus CVCC546 biofilm endobacteria; e Killing of biofilm-persistent bacteria by 4× MIC A24.

Full size image

Effects of A24 on bacterial morphology and membrane permeability

The surface morphology and internal ultrastructure of A24-treated S. aureus were observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM), respectively, to investigate the bactericidal mechanism of A24 against S. aureus. The results (Fig. 5a, b) revealed that the surface of untreated S. aureus was clear, glossy, and uniformly regular and globular; in contrast, the surface of S. aureus was rough, crumpled, collapsed, and fragmented after treatment with A24, and a large number of granular substances adhered to the surface. However, this phenomenon was not observed following vancomycin treatment, leading to the hypothesis that the bactericidal mechanism might be inconsistent. Further observation of the internal ultrastructure of S. aureus CVCC 546 revealed a clear and uniform cell wall and cytoplasm in the untreated group. In contrast, the cell wall and membrane of S. aureus treated with A24 exhibited structural damage and blurred boundaries, leading to leakage of the contents. The permeation and depolarization of the cell membrane by A24 were observed by staining with the fluorescent dyes PI and DiSC3(5), respectively, to further explore the potential impact of A24 on the cell membrane. Additionally, protein leakage, potassium ion leakage, and membrane mobility changes were examined after treatment with A24. The results (Fig. 5c–e) showed that 1) the untreated group of S. aureus CVCC 546 exhibited exclusively green fluorescence (SYTO9), nearly 100%, with no red fluorescence (PI) after SYTO9/PI staining (Fig. 5c). Conversely, the majority of S. aureus bacteria displayed red fluorescence, over 90%, after treatment with A24, which indicated that A24 destroyed the cell membrane, leading to the entry of PI into S. aureus CVCC 546, where it bound to the nucleus. 2) The membrane potential test (Fig. 5d) revealed that the untreated group remained in a stable state at approximately RFU = 2000. Following treatment with 1× MIC, 2× MIC or 4× MIC of A24, the membrane potential exhibited a notable change for 10 min, reaching the highest value observed. The change in the membrane potential in the 1× MIC and 2× MIC A24 treatment groups was slightly greater than the value of 1000 RFUs in the CK group, and the effect of membrane penetration was lower than that of nisin. The membrane potential increased by 3000 RFUs compared with that of the CK group after being treated with 4× MIC of A24 and was greater than that of the nisin group. 3) Potassium ion leakage (Fig. 5e) notably increased and reached its highest value after A24 treatment for 15 min, approaching the level observed with nisin. The above tests investigating the effect of A24 on the cell membrane indicated that A24 can destroy the bacterial membrane within a short time, increasing the membrane permeability of S. aureus.

Fig. 5: Antibacterial peptide A24 destroys the bacterial membrane, inserts into the membrane, and interferes with metabolism to kill bacteria quickly.
figure 5

a Surface structure morphology of A24-treated S. aureus by SEM. The scalebar represents 1 μm; b Ultrastructure of A24-treated S. aureus by TEM; c SYTO9/PI was used to detect the disruption of bacterial membranes by A24, and the images were viewed by fluorescence microscopy; d Increasing in membrane potential of A24-treated S. aureus CVCC 546; e K+ leakage after A24 treatment of bacteria; f A24-treated S. aureus CVCC 546 increased the bacterial metabolic toxic product ROS; g A24-treated S. aureus CVCC 546 increased intracellular ATP.

Full size image

The levels of reactive oxygen species (ROS) and ATP were determined to further investigate the effects of A24 on bacterial metabolism27. The results (Fig. 5f, g) revealed that A24 dose-dependently induced the accumulation of toxic ROS in S. aureus. The ROS level of the untreated strain was 641, and the ROS level increased to 936.5, 980, and 1432 after treatment with 1×, 2×, and 4× MIC of AP138, respectively. The ROS level increased to 980, 991, and 1098 after treatment with 1×, 2×, and 4× MIC of A24, respectively. The accumulation of ROS prompted oxidative damage to bacteria, which ultimately resulted in bacterial death. The intracellular ATP assay revealed that the intracellular ATP level of the untreated strain was intracellular luminescence intensity (RLU) = 1376, and the RLU increased to 10,674, 11,618, and 11906 after treatment with 1×, 2×, and 4× MIC of AP138, respectively. The RLU increased to 3676, 6585, and 8422 after treatment with 1×, 2×, and 4× MIC of A24, respectively. The increase in the intracellular ATP level may enhance bacterial metabolism, thereby facilitating the accumulation of toxic products and promoting the transition of drug-resistant bacteria from a dormant state to an active state. All the results demonstrated that A24 treatment caused metabolic disorders and ultimately led to bacterial death.

Evaluation of A24 safety in vitro and in vivo

The toxicity of A24 was verified in vitro and in vivo to further assess its safety28. First, the results of the hemolysis test (Fig. 6a) revealed that the twofold gradient concentrations of A24 (2–256 μg/mL) resulted in almost no hemolysis (<1%) at the highest concentration of 256 μg/mL, and the results of the cell cytotoxicity assay (Fig. 6b, c) revealed that mouse macrophages RAW264.7 and uterine epithelial cells BNCC 359233 had high survival rates of up to 78% and 85%, respectively, after treatment with a high concentration of 256 μg/mL, indicating better biocompatibility in vitro. In vivo acute toxicity tests were conducted via intraperitoneal and uterine administration (10 mg/kg) for seven consecutive days to further evaluate the safety of A24 in vivo29 (Supplementary Fig. S5). The body weights of the mice after intraperitoneal injection tended to increase in line with those of the control group, and no significant differences in leukocytes (neutrophils and lymphocytes) or biochemical indices (AST and ALT) were observed. Additionally, the lung, heart, liver, kidney, lung and spleen had clear and regular tissue structures; clear glands; no obvious hemorrhage or inflammatory cell infiltration; and no significant difference from those of the control group. Moreover, the results (Fig. 6d–g) obtained after uterine injection revealed that the body weight of the mice was not affected compared with that of the control group and tended to increase. The white blood cell count indicated that the whole-cell, neutrophil, and lymphocyte counts were normal; the biochemical indices AST, ALT, and total protein were within the normal range (p > 0.05). The uterine tissue structure was normal, with no obvious damage, such as congestion or redness, evident. The results of hematoxylin-eosin (HE)-stained sections revealed no hemorrhage, epithelial cell detachment, or inflammatory cell infiltration in the tissue, indicating the safety of uterine administration. The results of the above tests showed that A24 exhibited favorable biocompatibility in vitro and in vivo, thereby indicating the safety of the medication in vivo.

Fig. 6: The safety of A24 in vitro and in vivo.
figure 6

a Hemolysis of A24 against mouse erythrocytes; b, c Cytotoxicity of the A24 against RAW264.7 and BNCC 359233 cells; In vivo, ICR female mice (n = 4) were uterus administered with A24 (10 mg/kg) daily for a week: d Body weight; e Histology images (H&E) stained; f Whole-blood cell profiles; g Serum biochemical index. CK, untreated group; scale bar, 100 µm.

Full size image

Evaluation of A24 efficacy against systemic and local infections

Mouse models of peritonitis and endometritis were established to evaluate the therapeutic efficacy of A24 and ensure its application in vivo30,31. The white blood cell (WBC) counts decreased to 11.61 ± 0.92 (×109/L) and 8.64 ± 0.67 (×109/L) after treatment with 5 and 10 mg/kg A24, respectively (Supplementary Table S3). The neutrophil count was markedly elevated in the model group (1.22 ± 0.18–15.11 ± 2.03 ×109/L), reflecting the presence of a bacterial infection. Following treatment with 5 mg/kg and 10 mg/kg A24, the neutrophil count decreased to 4.31 ± 0.91 ×(109/L) and 2.30 ± 2.33 × (109/L), respectively.

The survival rates of the mice are shown in Fig. 7a. The survival rate was 100% after treatment with 10 mg/kg AP138 or 5 mg/kg A24, whereas treatment with 5 or 10 mg/kg Cef resulted in a survival rate of only 75%. These findings indicated that A24 had greater therapeutic efficacy at low doses. The results of the bacterial load in each organ are shown in Fig. 7b. A24 had a significant bactericidal effect on bacteria in the spleen, lung, kidney, and liver, with an average reduction of 4–5-log, and the bacterial inhibition rate was greater than 99%, indicating that A24 has a potent bactericidal capacity in vivo. The gross anatomical images (Supplementary Fig. S6) revealed that the positive control group presented suppurative foci in the liver and notable enlargement of the spleen, but the suppurative foci disappeared after treatment with 10 mg/kg A24. The results of the analysis of inflammatory cytokines are presented in Supplementary Fig. S6. The levels of the proinflammatory cytokines TNF-α, IL-1β, and IL-6 were significantly reduced to normal levels after treatment with 10 mg/kg A24. HE staining of tissue (Fig. 7c) and pathological sections revealed notable alterations in the structure and glands of the spleen, lung, kidney, and liver, accompanied by many inflammatory cells and hemorrhages in the model group. However, the structure of the tissues in the A24-treated group was markedly restored, with a clear delineation of the tissue structure and glands, and a minimal presence of inflammatory cells. In conclusion, A24 exerted a superior protective effect on peritonitis, increased survival rates, and reduced inflammatory cytokine and tissue inflammatory levels.

Fig. 7: A24 treatment of peritonitis with the drug-resistant bacterium S. aureus CVCC546.
figure 7

The therapeutic groups were treated with A24 (5, 10 mg/kg of body weight, 0.2 mL, n = 4), and the cef (5, 10 mg/kg, 0.2 mL, n = 4) after post-infection (2 h and 6 h) (a) A24 improves survival in mice; (b) A24 removes organ loads of bacteria; (c) A24 alleviates inflammation in various tissues.

Full size image

Moreover, the therapeutic effect of A24 on endometritis is shown in Fig. 8a and Supplementary Table S4. The WBC count in the control group was 14.04 ± 1.08 (×109/L), and the WBC count increased to 18.38 ± 0.81 (×109/L) after bacterial infection, whereas it decreased to 11 (×109/L) after treatment with 5 mg/kg, 10 mg/kg, and 15 mg/kg A24, which was comparable to the results of the negative control group. Compared with the negative control group, the bacterial infection group presented a 2-fold increase (12.18 ± 1.04 ×109/L). Following A24 treatment, the bacterial load was restored to the normal level of 6.62 ± 1.07 ×109/L, which was comparable to that observed in the ceftiofur sodium group. Moreover, the bacterial load was significantly lower than that in the untreated group (Fig. 8b). The inflammatory cytokine levels are shown in Supplementary Fig. S7. Compared with those in the CK group, the levels of the proinflammatory cytokines TNF-α, IL-6, and IL-1β were significantly increased, whereas the level of the anti-inflammatory cytokine IL-10 was significantly decreased in the infection model group. After treatment with A24, the levels of inflammatory cytokines returned to normal levels. Among them, the levels of the proinflammatory cytokines TNF-α, IL-6, and IL-1β were 560, 670, 571, 562, and 565 ng/L; 85, 145, 90, 93, and 83 pg/mL; and 40, 80, 55, 50, and 45 ng/L in the CK, positive, 15 mg/kg AP138, 15 mg/kg A24, and 10 mg/kg Cef groups, respectively. The levels of the anti-inflammatory cytokine IL-10 were 915, 650, 930, 940, and 850 pg/mL in the above groups. The results demonstrated that A24 treatment was superior to the same dose of A138 and comparable to ceftiofur sodium. The results of the pathological analysis of uterine tissue sections are shown in Fig. 8c. The normal uterine tissue was found to have a clear structure, no inflammatory cell infiltration, a clear gland structure, and no shedding of uterine epithelial cells; however, the uterus of the infected group was found to have a large number of infiltrating inflammatory cells (lymphocytes and neutrophils), shedding of uterine epithelial cells, blurred irregularity of the gland structure and intertissue congestion. In severe cases, these changes resulted in penetrating inflammation, characterized by the loss of the organized cellular structure, hyperplasia, the absence of glands, and the loss of fertility. However, the uterine tissues returned to normal levels after the intraperitoneal injection of 10 mg/kg A24. These findings indicated that A24 had an excellent therapeutic effect in vivo, exerting bactericidal and anti-inflammatory effects.

Fig. 8: A24 Treatment of endometritis with S. aureus CVCC 546.
figure 8

Therapeutic groups were treated with A24 (5, 10, 15 mg/kg, n = 4) and the cef (5, 10 mg/kg, 0.2 mL, n = 4) by intraperitoneal injection for 3 consecutive days (a) Uterine morphology; (b) A24 removes bacteria in the uterus; (c) A24 alleviates inflammation in the uterus.

Full size image

Discussion

MDR bacteria are a serious threat to human health, especially in local and systemic infections caused by G+ organisms of the ESKAPE group, which have the potential to become untreatable32,33,34. AMPs have the potential to be the most promising antibiotic alternatives35, but they also face challenges, including weak activity against resistant bacteria, poor stability in the in vivo environment, and high costs associated with in vitro preparation36,37. RF algorithms are integrated into AMP design, including mutation and modification, with the aim of maintaining a balance of hydrophobicity, positive charge, and amphiphilicity to improve antimicrobial activity and reduce toxicity38,39. In this work, we designed 12 derived AMPs with improved hydrophobicity and positive charge via the RF algorithm in the CAMP database. The most promising of these peptides, designated A24, exhibited superior antimicrobial activity against S. aureus, along with two other peptides, A2 and A4, which also demonstrated notable efficacy against this bacterial strain. The overall success rate for the design process was 25%, implying that this methodology can be employed to develop AMPs. 1) Hydrophobicity and positive charge were identified as the two primary design criteria, with the objective of increasing hydrophobicity and charge while limiting the risk of toxicity40,41,42. Through the design of the AP138-derived peptide, the RF algorithm achieved a better balance between hydrophobicity and charge, with a hydrophobicity of up to 35% and a charge of up to +5.5. These thresholds should not be exceeded, as they may induce toxicity. 2) Simple amino acids A, V, and L were used as the amino acids for the substitution of mutation sites 3G, 5N, 6G, 7P, 9S, 10E, 12D, 16H, 21S, and 25Y, respectively, and only the activities of three sites (9S, 10E, and 25Y) increased after the substitution. G is important in the sequence and should be maintained as much as possible. AMPs possess a distinctive spatial conformation, and the replacement of short-chain amino acids, such as alanine, can effectively preserve the original spatial conformation while increasing hydrophobicity43. 3) The amino acid cysteine (C) has the advantage of maintaining the spatial structure and activity of AMP sequences44. The results of this experiment demonstrated that C-substituted mutations in the sequences (S9C, D12C, and S21C) resulted in a complete loss of sequence activity. We postulated that the insertion of the C amino acid may have disrupted the normal S‒S bond in the original sequence, thereby causing the loss of activity due to the disordered spatial conformations. Consequently, the C amino acid was not changed and was inserted into AMP sequences as a potential modification. 4) The amino acids glycine (G), lysine (K), arginine (R), histidine (H), and proline (P) are the most frequently occurring amino acids in AMPs and are therefore essential for maintaining their activity. In this experiment, R14, R17, and R26 in the sequence of AP138 were replaced by the plectasin sequences Q14, N17, and K26, respectively. The original sequences, G, K, and R, were maintained in their original form. 5) Both ends of the sequence maintain the amphiphilic character of C-terminal hydrophilicity and N-terminal hydrophobicity, and the amino acids on both sides of the cysteine neighborhood are homophilic45. This analysis, in conjunction with the antimicrobial activity and membrane-acting structure of A24, led to the identification of A24 (GFGCNGPWAEDDLRCHRHCKSIKGYRGGGYCAKGGFVCKCY) as the optimal candidate for pharmacological studies, following the derivation of peptides from AP138 (S9A) and plectasin (D9A, M13L, Q14R, N17R, and K26R).

Heterologous expression represents a superior approach to the production and preparation of drugs, vaccines, and food additives in a variety of fields. It provides a means of circumventing the challenges of high cost and technical difficulties that are inherent to chemical synthesis46,47. However, the low expression of AMPs has been an important bottleneck limiting the industrialization of AMPs. The main reasons for low expression are the instability and toxicity of basic amino acids, as well as the low molecular weight of AMPs48,49. Furthermore, different families of AMPs may have different preferences for different expression systems50; therefore, a deeper understanding of the evolutionary progress of AMPs is essential to identify the most suitable expression system for their production51. Our laboratory has previously achieved successful high-yield expression of AMPs belonging to the plectasin family, which is transferable19,52,53. The hydrophobicity and positive charge were 35% and +4.5, respectively, which may indicate the potential low toxicity to P. pastoris, thereby increasing the likelihood of successful expression. A24 was successfully expressed, with the amount of the target protein in the 5 L fermenter reaching a level of grams (total supernatant protein concentration, 4.1 g/L). The product was found to be >95% pure, which greatly reduces the cost of chemical synthesis and paves the way for the clinical use of the drug.

Biofilms, which contain resistant bacteria and represent a resistant barrier that escapes killing by conventional antibiotics, are important in triggering persistence and reinfection54. A24 was able to kill different strains of resistant G+ bacteria, such as S. aureus, S. agalactiae and S. epidermidis, and the MICs of A24 (4–8 µg/mL) against resistant S. aureus were significantly lower than those of AP138 and nisin (8–32 µg/mL and 16–64 µg/mL). Moreover, A24 inhibited biofilm formation, and A24 inhibited early biofilm formation of MDR S. aureus, including both standard and clinical drug-resistant S. aureus, within 12 h by 99.9% at a dose of 4× MIC. Additionally, it eradicated up to 68% of mature biofilms, while most traditional antibiotics are incapable of traversing the biofilm and exert minimal or no bactericidal effect on the bacteria within the biofilm55. AMPs are capable of traversing biofilms and killing bacteria within biofilms, efficiently mitigating biofilm-associated infections11,28. Furthermore, the bacteria that form biofilms remain in a dormant state for an extended period, with a low energy demand and inherent resistance to antibiotics, which presents a significant challenge for the treatment of biofilm-associated chronic infections. Additionally, many drug-resistant bacteria have developed resistance to traditional antibiotics, such as penicillin and sulfonamide, which have limited penetration into biofilms and are not lethal to the bacteria that persist within them56,57. The antibiotic ceftiofur sodium was observed to have a weak lethal effect on the bacteria present in the biofilm at high doses. In contrast, A24 was able to remove up to 99.9% of bacteria at low concentrations.

AMPs have cationic and hydrophobic properties and are capable of interacting with bacterial membranes to exert bactericidal functions. The AMPs LI14 and L0069 have been reported to increase the penetration of multidrug-resistant S. aureus membranes, causing metabolic disorders and leading to rapid bacterial death10,31. In our study, the structure (wall and membrane) of S. aureus was significantly altered after treatment with A24 for 1 h, with wrinkles (approximately 100%) and leakage (approximately 25%) observed via SEM and TEM, respectively. A24 was found to disrupt the cell membrane of S. aureus, causing a series of alterations in physicochemical properties, including increased membrane permeability, leakage of the contents and metabolic disturbances, leading to rapid bacterial death. The RLU of ATP level increased from 1376 to 8422, causing S. aureus to revert from a dormant state to an active state and later leading to the accumulation of toxic ROS (641–1098) after treatment with 4× MIC of A24, resulting in rapid bacterial death10,31,58. Although the unique bactericidal mechanism of AMPs is not susceptible to resistance, the risk of resistance to some AMPs has also been reported59,60. Furthermore, after 30 days of resistance induction, the MIC value of A24 against S. aureus increased only 4-fold, which was consistent with that of vancomycin and better than that of the parent peptide and ceftiofur sodium.

Local and systemic infections caused by multidrug-resistant bacteria increase the difficulty and cost of treatment, seriously endangering human health and affecting the economic development of livestock and poultry61. Endometritis occurs when pathogenic microorganisms invade the exposed uterus to colonize and multiply, causing inflammation and even necrosis of the uterus and affecting normal health and the ability to reproduce62,63. Most cases of chronic endometritis, which has weakened the potency of antibiotics, are associated with AMPs as therapeutic agents64. The hemolysis rate was less than 2%, and no cytotoxicity (survival rate >85%) was observed in vitro. An additional seven consecutive days of toxicity testing revealed no abnormal changes in biochemical indices or histopathology in vivo, suggesting drug safety.

Treatment with the antimicrobial peptide A24 (10 mg/kg) in a postpartum mouse model of endometritis effectively reduced bacterial levels by 2–3 log10 CFU/0.1g in the uterus, significantly reduced inflammation levels and restored the structures and glands of the uterine tissue to normal. Moreover, systemic infection was modeled by inducing peritonitis in mice. A24 significantly inhibited S. aureus 546 (4–5 log10 CFU/0.1g reduction) growth in the liver, kidney, and spleen tissues, and a concentration of 10 mg/kg A24 had the most pronounced inhibitory effect. The survival rate of the mice increased to 100%, with a reduction in the levels of inflammatory cytokines to levels comparable to those observed at the control group. Additionally, the tissues were restored to normal, with a well-defined structure, clearly defined glands, and no accumulation of inflammatory cells. A24 has been shown to be safe in vitro and in vivo and exerts a significant therapeutic effect on systemic infection in a peritonitis model and localized infection in an endometritis model. A24 is capable of significantly clearing pathogens, which provides compelling evidence that A24 has substantial pharmaceutical research value for the treatment of diseases caused by multidrug-resistant S. aureus, such as endometritis.

Overall, in this study, the RF algorithm was employed for the rational design approach to increase the positive charge and hydrophobicity of the derived AMPs40,65,66,67, which were recombinantly expressed in P. pastoris to achieve a high yield of AP138 and the best-derived peptide A24 with elevated hydrophobicity. A24 was able to efficiently inhibit multidrug-resistant G+ bacteria (2–8 μg/mL) and markedly inhibit clinical isolates of MRSA (2–16 μg/mL). Moreover, A24 exhibited enhanced antimicrobial activity and superior stability in physiological environments (25% plasma, 150 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl2, and 2.5 mM MgCl2) compared with AP138 and other drugs. AMPs have the capacity to alter antibiotic susceptibility in S. aureus and eradicate MDR S. aureus more rapidly than conventional antibiotics29,68. The ability of the screened A24 peptide to cure diseases caused by biofilms and their resistant bacteria supports its use as a new drug candidate. The ability of A24 to treat diseases resulting from biofilm-related infections and antibiotic-resistant bacteria lends credence to its potential as a prospective pharmaceutical agent.

Materials and methods

Strains, culture, and plasmids

Standard strains and clinical isolates (S. aureus, S. epidermidis, S. agalactiae, Streptococcus dysgalactiae) were used for this test. For the experiment, all strains need to be revived and activated first, after which they are prepared in Mueller-Hinton Broth (MHB; S. aureus, S. epidermidis), Tryptone Soy Broth (TSB; S. agalactiae, S. dysgalactiae), medium at 37 °C.

RAW 264.7 mice macrophages were obtained from Peking Union Medical College. Bovine endometrial epithelial cell line BNCC359233 was purchased from by BeNa Culture Collection (Beijing, China).

All recombinant plasmid including pPICZαA-A24 were synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). Plasmid extraction kits, antibiotics (Vancomycin and ceftiofur sodium), Dulbecco’s modified Eagle medium (DMEM), and fetal bovine serum (FBS) were purchased from Tiangen Co., Ltd, China MeilunBio (China) respectively. Other reagents were analytical grade.

Peptide design and preparation

The antimicrobial peptide AP138 was obtained from the antimicrobial peptide database DBAASP69. The RF algorithm of rational design was used to design derived peptides with a better AMP probability22. The tools APD, DBAASP, and ProtParam were used for the physical and chemical characterization of the protein sequences70. I-TASSER online software was used to predict the 3D structures of these peptides, and PyMOL software was used to determine their structures71.

The derived peptide genes, including A24 (GenBank: submission 2783807), were obtained via the reverse translation tool as previously described and inserted into the pPICZαA plasmid. The recombinant plasmid pPICZαA-A24 was digested with PmeI and transformed into competent P. pastoris X-33 cells via electroporation, with 120 h of high-density culture in a 5 L fermenter. The AKTAxpress system was used to purify the peptide from the fermentation broth. The results of the peptide expression and purification were analyzed by Tricine-SDS–PAGE and MALDI-TOF/TOF MS (Ultraflextreme, Bruker, Germany), respectively15. The concentration and purity of the peptides were assayed using Bradford assay kits and HPLC, respectively.

CD spectroscopy and high-resolution MS

A24 was analyzed in ddH2O, 50% TFE (simulated eukaryotic cell environment), and 20 mM SDS (simulated bacterial membrane hydrophobic environment) via circular dichroism spectroscopy (Bio-Logic MOS450 spectropolarimeter, France).

Antimicrobial and bactericidal activities

The MIC and MBC values were determined to evaluate antimicrobial and bacterial activities, and these assessments were performed according to CLSI 202172. Briefly, logarithmic growth-stage bacteria (1× 105 CFU/mL in MHB, 90 µL/well) were co-incubated with twofold dilutions of the peptide (1280 to 2.5 µg/mL, 10 µL/well) in 0.1 M PBS and subjected to physical conditions (25% plasma (v/v), 150 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl2, and 2.5 mM MgCl2) in a 96-cell plate. After 18 h of incubation at 37 °C, the MICs were determined as no visible growth of bacteria, and MBCs were determined as no bacteria growing on the MHA plate (killed ≥99.9% of bacteria).

Time‒kill curve experiments of A24 against S. aureus CVCC 546 and MRSA ATCC 43300 were performed as described previously18. Briefly, the bacteria (1× 105 CFU/mL) were coincubated with 1×, 2×, or 4× MIC of A24 at 37 °C and centrifuged at 200 rpm in a 10 mL volume. The samples were collected and plated on MHA plates for 0–24 h. The number of viable bacteria was determined at each time point.

The PAEs against S. aureus CVCC 546 and MRSA ATCC 43300 were detected using a previously described method53. Briefly, the bacteria (1× 106 CFU/mL) were coincubated with 1×, 2×, or 4× MIC of A24 at 37 °C and 200 rpm in a 50 mL volume. (Formula: PAE = T-C, where “C” and “T” represent the time at which the bacterial number increased 10-fold in the control and treatment groups, respectively.

The synergistic effect of A24 (with antibiotics) on bacteria was detected using a checkerboard assay and a growth curve. All methods were described previously in our laboratory. (Formula: FICI = FICpeptides + FICantibiotic, FIC = MICc/MICa, where MICa and MICc are the MICs of the peptide alone and in combination with antibiotics. FICI ≤ 0.5 synergy, 0.5 < FICI ≤ 1 addition, 1 < FICI ≤ 4 no effect, and FICI > 4 antagonism.)

Serial passage resistance induction studies

An MIC assay was performed to evaluate the bactericidal effect of the peptides against bacteria, as described previously73. Briefly, mid-log phase MRSA ATCC 43300 and S. aureus CVCC 546 (1× 105 CFUs/mL) were coincubated with A24 or antibiotics (1280 to 2.5 µg/mL) and added to 96-well plates. After a 16–18 h incubation at 37 °C, the sub-MIC bacteria were used to inoculate the subsequent cultures. After 30 passages, the mixture was further cultured for 3 days without the peptide or antibiotics to stabilize the bacterial state. All MICs were recorded.

Biofilm inhibition and eradication assays

Biofilm inhibition was analyzed as described previously11,74. Briefly, mid-logarithmic growth-phase cultures of MRSA ATCC 43300 and S. aureus CVCC 546 (1× 108 CFU/mL in TSB) were added to 96-cell plates, and different concentrations of peptides (128–0.25 µg/mL) were subsequently added to the above system. The untreated bacteria in TBS were served as controls. Following a 24-h incubation at 37 °C, the plate was stained with 1% crystal violet for 30 min, washed and solubilized in DMSO. The optical density at 600 nm was determined as a measure of the biofilm biomass using a microplate spectrophotometer.

The biofilm eradication assay was different from the biofilm inhibition assay. The bacteria in the mid-logarithmic growth phase above were cultured at 37 °C for 24 h without any peptides or antibiotics. The planktonic bacteria were subsequently removed with PBS. The biofilms were subsequently coincubated with peptides (128–0.25 µg/mL) at 37 °C for 24 h, after which they were treated as described above.

The L7007 LIVE/DEAD BacLight™ Bacterial Viability Kit SYTO9/PI was used for staining, and the images were observed with a confocal scanning laser microscope (Nikon A1) to visualize the disruption of the biofilm by the peptides. Briefly, the mid-logarithmic growth-phase bacteria were cultured in confocal glass-bottom microwell dishes (BeyoGold™ 35 mm) at 37 °C for 24 h. A final concentration of 4× MIC of A24 was added to the dish and incubated for another 24 h at 37 °C; planktonic bacteria were removed with PBS, and biofilms were stained with a SYTO9/PI kit for 15 min. Samples were visualized via confocal scanning laser microscopy.

Bactericidal activity against persisters

The mid-logarithmic growth-phase MDR bacteria MRSA ATCC 43300 and S. aureus CVCC 546 were diluted in TSB to 1× 108 CFU/mL in 24-well plates and incubated at 37 °C for 24 h to obtain persisters from the biofilm. A concentration of 100 × MIC of Van was added to the plates after they were washed with PBS and incubated for 24 h at 37 °C. Finally, 16× MIC of A24 was added after washing and incubated for another 24 h at 37 °C, and the bacteria were counted as described above. Bacteria were coincubated with Van or PBS as a control.

SEM and TEM

S. aureus CVCC 546 in the mid-logarithmic growth phase were treated with 2× MIC of A24 for 0.5 h, 1 h, or 2 h at 37 °C, washed twice and fixed with 2.5% glutaraldehyde overnight. The bacterial morphology was observed via SEM (Hitachi SU 8000, Japan). Furthermore, the ultrastructural changes in the above bacteria (control and treatment for 1 h) were examined using a transmission electron microscope (TEM, Hitachi H7650B, Japan). The methods used for SEM and TEM were described previously59.

Interactions with the membrane

The SYTO9/PI assay and K+ leakage assays were used to evaluate the integrity of the cell membrane59. Briefly, mid-log phase S. aureus CVCC 546 (1× 108 CFU/mL) was coincubated with 2× MIC of A24 at 37 °C (15/30/60/90/120 min), and the untreated and nisin treatment groups served as controls. Next, the supernatants were analyzed for K+ leakage using inductively coupled plasma‒mass spectrometry (ICP‒MS) (Santa Clara, CA, USA), and the treated bacteria were stained with SYTO9/PI and observed under a fluorescence microscope (Nikon 80i).

Bacterial membrane permeability assay

The membrane-permeabilizing potential of A24 was detected with the membrane probe DiSC3(5), and the method was described previously29. Briefly, S. aureus CVCC 546 in the mid-logarithmic growth phase were first coincubated with 0.5 mM DiSC3(5)) at 37 °C for 1 h in the dark, after which the stained bacteria and A24 (4×, 2×, and 1× MIC) were mixed in black 96-well plates. The results were determined using a spectrophotometer (excitation, 622 nm; emission, 670 nm; Infinite M200).

Bacterial metabolic interference

Intracellular ATP and ROS levels were detected with kits to further study the effects of A24 on the bacterial metabolism mechanism15,29. Briefly, 1) for the analysis of intracellular ATP levels, the mid-log phase S. aureus CVCC 546 was diluted to 1× 108 CFU/mL and incubated with different concentrations of A24 (4×, 2×, or 1× MIC). The intracellular supernatant was detected with an Infinite M200 microplate reader (Tecan, Luminescence signal). 2) For the ROS assay, the probe DCFH-DA (2’,7’-dichlorodihydrofluorescein) was used to stain bacteria to measure the levels of ROS. Briefly, the mid-logarithmic growth phase S. aureus CVCC 546 was coincubated with 10 µM DCFH-DA at 37 °C for 0.5 h and treated with A24 (4×, 2×, or 1× MIC) for 1 h. The results were detected with a microplate reader (488/525 nm, Tecan, Männedorf, Switzerland).

Safety in vivo and in vitro

For the hemolysis assay, 8% fresh mouse red blood cells were coincubated with a 2-fold gradient (0.5–256 µg/mL) of the peptide in equal volumes for 1 h at 37 °C. PBS and 0.1% Triton X-100 served as the blank and positive controls, respectively.

For the cytotoxicity assay, RAW264.7 mouse macrophages and the uterine epithelial cell line BNCC 359233 were coincubated with A24 (0.5–256 µg/mL) for 24 h at 37 °C in 5% CO2. The results were detected via the MTT (thiazolyl blue tetrazolium bromide) method, as described previously.

For the assessment of acute toxicity in mice, A24 (10 mg/kg, ICR mouse body weight) was injected intraperitoneally injected and into the uterus (n = 4 mice per group)10. Whole blood and tissues (heart, liver, spleen, lung, kidney, and uterus) were collected on the seventh day for the whole-blood cell analysis, biochemical analysis, and HE staining, respectively.

Infection and treatment of murine peritonitis

Six- to eight-week-old female mice weighing approximately 20 g were intraperitoneally infected with mid-log phase S. aureus CVCC 546 (5× 108 CFU/mL, 1 mL). The therapeutic groups were treated with A24 (5 or 10 mg/kg body weight, 0.2 mL, n = 4) or cef (5 or 10 mg/kg, 0.2 mL, n = 4) after infection (2 h and 6 h). The Cef groups were the antibiotic control groups, the uninfected group was the blank control group (CK), and the untreated groups were the negtive controls. The survival of the mice was recorded daily for 7 days.

Moreover, the model of peritonitis described above was used75, and the dose of 5 mg/kg and 10 mg/kg A24 were administered. Serum was collected at 2 and 6 h after treatment for the determination of inflammatory cytokine levels using ELISA kits. Organs were collected for assessments of bacterial translocation and HE staining at 24 h posttreatment.

Infection and treatment of murine endometritis

Postpartum mice were used for endometritis model establishment and studies of the effects of drug treatment to better simulate the infection-prone state in the postpartum period76. Eight-week-old female ICR mice were caged with male mice (3:1) for approximately 21 days. During the postnatal period, 2× 108 CFU/mL S. aureus CVCC 546 were injected into the uterus twice consecutively via an instillation needle. The therapeutic groups were treated with AP138 (5, 10, or 15 mg/kg, n = 4), A24 (5, 10, or 15 mg/kg, n = 4) or Cef (5 or 10 mg/kg, 0.2 mL, n = 4) by intraperitoneal injection for 3 consecutive days. Serum was collected after 3 days of treatment for detection of inflammatory cytokine levels using ELISA kits (R&D Systems, USA), and uterine tissues were subjected to HE staining.

Statistics and reproducibility

Unless otherwise noted, experiments were repeated at least three times. All data are presented as means ± standard deviation (SD). The software GraphPad Prism (version 8, USA) was used to analyze all data and ANOVA was as the method to determine the statistical significance. The results are presented as means ± standard deviation (SD). A P value of <0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.