Introduction

Bacterial resistance refers to a bacterium’s natural or acquired ability to survive exposure to antibiotics designed to eliminate or slow its growth. Although this phenomenon is natural, it has escalated into a public health crisis due to its widespread increase. The rise in antibiotic resistance has led to the emergence of multidrug-resistant strains, complicating the treatment of common infections and presenting significant challenges for essential medical procedures, such as chemotherapy, organ transplants, cesarean sections, and various surgeries1,2,3. According to the CDC’s 2019 Antimicrobial Resistance Threats Report, over 2.8 million antimicrobial-resistant infections occur annually in the United States, resulting in more than 35,000 deaths4.

The primary driver of bacterial resistance is the overuse and misuse of antibiotics. Additionally, other factors contribute to its growing prevalence, including the pharmaceutical industry’s inadequate development of new antibiotic drugs. Therefore, the pursuit of new antibiotics becomes essential. There is an urgent need for new antibiotics to address infections caused by carbapenem-resistant Gram-negative bacteria, which are classified as critical-priority antimicrobial-resistant pathogens by the World Health Organization5. Carbapenems, a group of β-lactam antibiotics, are commonly employed in initial empiric treatment for various infections. However, their overuse has increased bacterial resistance, particularly in pathogens such as Pseudomonas spp., Acinetobacter baumannii, and Enterobacterales, including E. coli. Infections from these bacteria can lead to severe and often fatal conditions, such as bacteremia and pneumonia6.

Animal venoms represent a primary source of bioactive molecules. Their toxicity and diverse biological effects have attracted significant scientific interest7. In particular, the toxins found in snake venom demonstrate remarkable potential for discovering new drugs, whether through direct application or as structural frameworks for synthesizing novel pharmacological agents. The snake Bothrops asper, belonging to the Viperidae family, produces venom rich in a cytotoxic protein called Myotoxin II. This myotoxin is classified within the phospholipase A2 (PLA2) enzyme family. Snake venom phospholipases are divided into two types: PLA2 Asp49, which serves a catalytic function, and PLA2 Lys49, which does not exhibit catalytic activity. Myotoxin II, found in the venom of Bothrops Asper, is a Lys49 type PLA2 with antibiotic properties and may offer an encouraging basis for developing new medications8,9.

The region of Myotoxin II that enhances its bactericidal effect is a slight stretch of amino acids (115–129) situated near the C-terminal loop, represented by the sequence KKYRYYLKPLCKK7. This sequence, recognized for its high concentration of cationic amino acids surrounding hydrophobic ones, gives the peptide amphipathic properties, allowing it to interact effectively with the membranes of various Gram-negative bacteria7. When the peptide attaches to the membrane surface in lipopolysaccharide (LPS), the amino acids align to form a layer that outlines a pore. In this arrangement, the side chains of non-polar amino acids interact with the aliphatic tails of fatty acids within the lipid bilayer. In contrast, the hydrophilic chains remain exposed to the surrounding solvent. Consequently, as additional peptides bind, the pore size expands, leading to cell death due to the escape of intracellular components and osmotic disruption10,11. However, the peptide exhibits high toxicity levels7.

A productive method for improving peptide selectivity is through trial and error12,32. Alternatively, rational design, also known as structure-based drug design (SBDD) or molecular docking, is a drug design technique that simulates molecular interactions to inform the development of new drugs. This approach facilitates predictions regarding the binding modes and affinities of receptors and ligands13. Molecular docking is a crucial method for designing antibiotic peptides, as it enables the modeling of specific interactions between the peptide’s amino acids and components of the bacterial membrane14. This ability allows for the rational modification of the peptide sequence to enhance its affinity and selectivity for pathogenic bacteria, while minimizing potential side effects on human cells. Previous research has demonstrated that modifying the amino acid sequence of peptides 115–129 can produce homologs with enhanced antibiotic effectiveness against both antibiotic-sensitive and resistant strains of Escherichia coli, as well as lower toxicity to eukaryotic cells15,33.

This study utilizes SBDD techniques to develop a peptide with selective antibiotic properties derived from Myotoxin II of Bothrops Asper. The peptide’s effectiveness against E. coli, P. aeruginosa, and K. pneumoniae was assessed, and its safety was evaluated in Caco-2 and L929 cell lines, as well as human red blood cells.

Materials and methods

Bioinformatic

The crystallized protein Myotoxin II was identified using the Protein Data Bank (PDB) under the code 1CLP16. This protein, obtained from the organism Bothrops Asper, consists of 121 amino acids. The crystal was acquired without ligands and has a resolution of 2.80 Å. Previous research has shown that Myotoxin II exhibits a strong interaction with the lipopolysaccharide of several gram-negative bacteria; therefore, this component was selected as the receptor for the current study17. The lipopolysaccharide (LPS) was obtained from the ChemSpider chemical structures database, identified by ID 10,143,54718. Using the PyMOL software, chain A was removed from the 1CLP model sequence. Water molecules were then eliminated, and the energy was optimized in the Avogadro program19 using the MMFF94s force field through four procedural steps.

Clarification of the active site (LPS) and creation of the Grid box

We selected six amino acids from the original Myotoxin II sequence (positions 115–129) to act as ligands (KYYLPC), while lipopolysaccharides (LPS) served as the receptor. The LPS was partitioned into quadrants for molecular docking using the Webina software. This methodology enabled us to pinpoint the binding site of the designed peptide sequences on LPS, which yielded the most negative affinity energy. The grid box dimensions were established at 50 × 46 × 54 points, with a spacing of 0.375 Å, and the center coordinates were set at 25.694, − 34.227, and 2.796.

Modifications and molecular docking

The original sequence from 115 to 129 was analyzed to assess affinity energies, static binding configurations, interactions, and interatomic distances. Using this sequence as a base, 20 modifications were generated via ChemDraw 20.0 and converted into PDB format through Avogadro. Each modified sequence underwent molecular docking with AutoDock Vina. A detailed examination of molecular interactions and interatomic distances was conducted using PyMOL or Discovery Studio software.

Peptide synthesis

The SurGenoma S.A.S. laboratory in Bogotá, Colombia, located at Carrera 79D #35B-03 Sur, received the peptide sequence for synthesis. We obtained 100 mg of the lyophilized peptide, which has a molecular mass of 1354.54 g/mol and a purity of 99.08%.

Antimicrobial assay

The antimicrobial activity of the peptide was assessed using the plate microdilution test, following the guidelines of the Clinical and Laboratory Standards Institute (CLSI) as outlined in CLSI M07 and M100. The test focused on Gram-negative strains, particularly Escherichia coli (ATCC 25922), Pseudomonas aeruginosa (ATCC 27853), and two antibiotic-resistant strains, Klebsiella pneumoniae (ATCC BAA-2146) and Escherichia coli (CTX-M).

Mueller–Hinton Broth (MHB) was used to prepare the inoculum for the strains, and the cultures were incubated at 35 °C for 18 h. The bacterial suspensions were adjusted to an absorbance of 0.08 to 0.13 at 625 nm, corresponding to a 0.5 McFarland standard, which is approximately equivalent to 100 to 200 million colony-forming units (CFU) per milliliter (mL). Then, the 0.5 McFarland suspension was diluted at a 1:20 ratio, resulting in a concentration of 5 million colony-forming units (CFU) per milliliter (mL). For each bacterial suspension, 100 µL was added to each well of a 96-well plate, resulting in a final concentration of fifty thousand CFU/well. Additionally, 100 µL of peptide serial dilution (from 10 to 0.15 mg/mL) previously prepared in saline solution was added to each well. The plates were then incubated for 24 h at 35 °C. MHB without bacterial inoculum served as a blank control, MHB with bacterial inoculum functioned as the bacterial growth control, and ciprofloxacin in MHB (0.5 μg/mL) served as a positive control for antimicrobial activity. To evaluate the antimicrobial activity of the samples, the minimal inhibitory concentration (MIC) was determined by measuring the optical density using a BIO-RAD iMark Microplate Reader, comparing the absorbance at 625 nm of all samples with that of the bacterial growth control. All experiments were conducted in triplicate.

Toxicity and security

Hemolysis assay

A hemolysis assay was conducted to evaluate the effect of the peptide candidate on the cell membrane of human erythrocytes. A 10 mL blood sample was collected using anticoagulants from a healthy volunteer who gave prior informed consent. The blood was centrifuged at 1500 rpm for 5 min, and then the supernatant was discarded. The erythrocyte pellet was washed and diluted in a 1:10 ratio with 1 × PBS. Next, 100 μL of the diluted erythrocytes were seeded into each well of a 96-well plate. Peptide solutions were added at serial dilutions from 10 to 0.07 mg/mL, with a volume of 100 μL per well. The samples were incubated at 37 °C for 1 h. After incubation, the plates were centrifuged at 3500 rpm for 15 min using a SIRENA AFI-C080R-E centrifuge. Following centrifugation, 100 μL of the supernatant was transferred to a new plate, and the absorbance was measured at 595 nm using a BIO-RAD iMark™ Microplate Reader. Triton X-100 (10% v/v) served as a positive lysis control, while PBS (1x) was used as the negative lysis control. All experiments were performed in triplicate.

Security cell assay

The various concentrations of peptides evaluated were used to assess toxicity and safety across two cell lines: a healthy L929 line (ATCC CCL-1) and a Caco-2 (ATCC HTB-37) intestinal cell model. Both cell lines were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and incubated at 37 °C with 5% CO2. They were kept in culture until they reached confluency. The cells were seeded into 96-well plates at a density of 10,000 cells per well. The next day, the medium was replaced with dilutions of the peptide in complete medium, and the cells were maintained in culture for 24 h. To assess toxicity, the treatments were replaced with 100 µL of Resazurin at 440 nm, followed by a four-hour incubation period. The impact of the peptides on the cell lines was evaluated using a plate reader with an excitation wavelength of 530 nm and an emission wavelength of 590 nm. Untreated cells and cells treated with DMSO served as negative and positive controls, respectively.

Results and discussion

Rational design and molecular docking

Lipopolysaccharide (LPS) is present in many, though not all, gram-negative bacteria20, where it serves various biological functions. LPS is located solely in the outer leaflet of the gram-negative outer membrane. In Escherichia coli and Salmonella enterica serovar Typhimurium, the LPS-to-glycerophospholipid ratio is approximately 0.15 to 121. Around 2 × 106 LPS molecules, covering roughly 75% of the cell surface22 Molecular docking analysis identified the O-antigenic polysaccharide (O-PS), core oligosaccharide (Core OS), and lipid A of the LPS site10 and then attempted to divide it into quadrants, as illustrated in Fig. 1. The first three quadrants revealed unusual affinity energies, indicating a lack of significant binding sites in those regions (10, 7, and 6 kcal/mol, respectively). Conversely, the fourth quadrant, corresponding to the lipid region of LPS, exhibited intense affinity energy along with six distinct binding modes (− 3.3 kcal/mol). Therefore, the fourth region of LPS was selected as the target for binding the designed peptide compounds.

Fig. 1
figure 1

2D representation of lipopolysaccharide (LPS) segmented into quadrants.

Full size image

Designing peptides based on sequences 115–129 of Myotoxin II

Based on the original sequence 115–129 of Myotoxin II, the peptide design aimed to enhance binding affinity to LPS while reducing toxicity in eukaryotic cells. The initial sequence showed an affinity energy of -6.0 kcal/mol, serving as a reference for further modifications. Table 1 identifies the best peptide interaction site with LPS through molecular docking.

Table 1 Molecular docking was performed on the complete sequence (115–129) and the fragments 115–120 and 121–129 of Myotoxin II.
Full size table

To determine the most promising area for developing peptide variants, we divided sequences 115–129 into two segments: 115–120 and 121–129. We then performed molecular docking on both segments. The results showed that the 115–120 sequence exhibited lower affinity energy than the 121–129 sequence, indicating a stronger affinity with LPS, as illustrated in Table 1. From this data, we selected the 115–120 segment as the basis for structural modifications to enhance the peptide’s affinity selectivity.

The initial modification involved substituting tyrosines with lysines and arginine with tryptophan, resulting in peptide 1 (KKKWK), as shown in Fig. 2 and Table 2. This change caused a slight reduction in binding affinity with LPS, yielding an energy measurement of − 5.5 kcal/mol. As a cationic amino acid, lysine contributes a positive charge to the sequence due to the amino group in its side chain. The positive charge from lysine could enhance interactions with the phosphate groups of LPS through electrostatic forces. Conversely, substituting tryptophan might disrupt the balance of hydrophobic interactions within the peptide, reducing affinity energy toward LPS. The indole ring of tryptophan also promotes π–π interactions with neighboring amino acids, stabilizing the peptide’s three-dimensional structure. This structural stabilization is essential for antibiotic efficacy, as it ensures that the peptide is correctly oriented for optimal insertion and interaction with cell membranes.

Fig. 2
figure 2

synthetic peptide variants derived from the original C-terminal sequence 115–129 of Myotoxin II.

Full size image
Table 2 Molecular docking and interatomic distances of twenty synthetic peptide variants derived from the original C-terminal sequence 115–129 of Myotoxin II.
Full size table

The proposed mechanism indicates that the first interaction between Gram-negative E. coli and antimicrobial peptides (AMPs) is mediated by lipopolysaccharides (LPS). These LPSs act as kinetic barriers that hinder AMPs from binding to the bacterial membrane while also interacting with the LPS’s negative charges23. This interaction ultimately leads to pore formation, which induces osmotic shock and results in cell lysis. Therefore, while the presence of lysine imparts a positive charge in the sequence, the absence of tryptophan diminishes the peptide’s capacity for effective hydrophobic interactions, thus reducing its ability to penetrate the plasma membrane efficiently.

In contrast, peptide 2 (KHPHYRY) was designed from the original sequence 115–120 to achieve a balance between cationic and aromatic, hydrophobic amino acids. The tyrosine phenolic groups enable peptide 2 to engage in π-π interactions and facilitate hydrogen bonding. Furthermore, having lysine at position one and histidine at position two enhances lysine’s interaction with LPS, as illustrated in Table 2. This improvement results from the proximity of histidine to lysine. The imidazole group in the histidine side chain promotes a cation-π interaction with the protonated amino group of lysine, which is critical for increasing affinity for LPS and enhancing the stability of the protein-LPS complex at physiological temperatures24.

In the third modification, the histidine at position four of peptide two was removed, forming peptide 3 (KHPYRY). This change enhanced the affinity energy to − 6.7 kcal/mol. Substituting histidine with tyrosine in the peptide sequence altered its hydrophobic properties, thereby strengthening the hydrophobic interactions with the receptor binding site. Moreover, arginine at position five was found to make several interactions with LPS, especially when surrounded by tyrosines, with bond distances of less than 4 Å.

Peptide 4 (KHLYRYH) was developed from peptide 3 (KHPYRY) by substituting proline with leucine at position three and adding histidine at position eight, achieving an affinity energy of − 6.7 kcal/mol. This substitution likely enhances the peptide’s conformational flexibility, allowing it to interact more effectively with its environment, particularly with hydrophobic binding sites on the receptor25. Previous research has demonstrated that short peptides rich in leucine can inhibit proinflammatory responses by interacting with LPS. They accomplish this by neutralizing LPS and inducing the dissociation of LPS micelles, which prevents the activation of the proinflammatory signaling pathway26. Furthermore, incorporating a cationic amino acid, such as histidine, at the peptide’s end facilitates its interaction with the negatively charged regions of LPS.

The fifth modification reduced the sequence length to 5 amino acids, yielding peptide 5 (KHPRY). This sequence preserved lysine in the first position and histidine in the second, while a cationic amino acid flanked by hydrophobic amino acids was added at the C-terminus. Consequently, this modification resulted in an affinity energy of − 6.8 kcal/mol.

Next, the best affinity energies were combined, and some peptides with promising affinity were obtained, ranging from 6 to 16. However, peptide 4 (KHLYRYH-6.7 kcal/mol) and peptide 5 (KHPRY-6.8 kcal/mol) formed peptide 17 (KHLYRYRH). This new sequence retained positive charges at both ends of the peptide, ensuring the stability of the center while maintaining the optimal arrangement of arginine surrounded by hydrophobic amino acids. This setup enabled the arginine at position five to engage in multiple interactions with LPS atoms, particularly forming hydrogen bonds with bond distances around 2 Å.

The final step involved transforming peptide 17 into peptides 18 through 20. Specifically, peptide 20 (KHWYKHYRH) is detailed in Scheme 1. This new variant includes a substitution at position 5, replacing arginine with lysine. The lysines at positions 1 and 5 may facilitate initial interaction with the bacterial membrane, while the lysine at position seven potentially supports deeper peptide insertion. Furthermore, the H–W–Y–K sequence establishes a charge-hydrophobicity pattern crucial for enhancing affinity. There is also the introduction of histidine at position six and the substitution of leucine with tryptophan at position 3. Molecular docking results reveal that peptide 20 exhibits the strongest affinity, with an energy of − 7.6 kcal/mol.

Scheme 1.
scheme 1

A 2D Structure of Peptide 20 KHWYKHYRH.

Full size image

Adding tryptophan at the 3-position is crucial, as it enhances the peptide’s structural stability through hydrophobic interactions15,33. Research conducted by Hui-Yuan Yu et al. shows that tryptophan residues are more effectively situated within a hydrophobic environment by interacting with dodecyl phosphocholine micelles and lipid vesicles, highlighting tryptophan’s affinity for hydrophobic membrane areas15,33. Furthermore, Paramo et al. note that tryptophan-rich peptides demonstrate enhanced antibacterial effectiveness against both Gram-positive and Gram-negative bacteria, while exhibiting lower toxicity to eukaryotic cells17.

Conversely, placing histidine in position six and replacing arginine with lysine affects the cation-π interactions, thereby stabilizing the three-dimensional structure of the sequence. Lysine in position 1 and histidine in position two can interact with the indole ring of tryptophan in position 3. Likewise, the aromatic character of tyrosine in position 4 enables it to participate in cation-π interactions with the positively charged lysine groups in position five and histidine in position 6, thereby enhancing structural stabilization and proper protein folding27. Among the amino acids that demonstrated stronger interactions with lipopolysaccharide (LPS), lysine at positions one and five was notably significant. This suggests the formation of hydrogen bonds with short distances not exceeding 2.94 Å, as shown in Table 2.

Based on the rational design of the Myotoxin II-derived peptide 20, 100 mg of the KHWYKHYRH sequence was obtained as an external service and reconstituted in PBS for biological analysis assays. The inhibitory effects of the KHWYKHYRH peptide were shown in the E. coli strain (ATCC 25922), which had an IC50 value of 0.27 mg/mL, thereby confirming the proposed rational design illustrated in Fig. 3. These effects were noted to a lesser degree in P. aeruginosa (ATCC 27853), which had an IC50 of 2.93 mg/mL. Conversely, the peptide was ineffective against resistant strains, such as the NDM-1 positive K. pneumoniae (ATCC BAA-2146) and the ESBL clinical isolate E. coli (CTX-M), as shown in Fig. 4. This lack of effectiveness may stem from variations in lipopolysaccharide (LPS) structures between Gram-negative bacteria. It is well-known that lipid A and 3-deoxy-D-manno-oct-2-ulosonic acid (Kdo), components of the core oligosaccharides linked to lipid A, can undergo adaptive changes. These modifications can impact LPS synthesis and the core structure, regardless of the type of bacteria or the site of infection. Such changes can generally lead to immune evasion, persistent inflammation, antibiotic resistance, and increased antimicrobial resistance28,29

Fig. 3
figure 3

IC50 tdetermination of pepide 20.

Full size image
Fig. 4
figure 4

Evaluation of the antibacterial effect of peptide 20.

Full size image

Investigating the in vitro cytotoxicity of peptides is critical before they can be considered for in vivo testing. To assess the safety of peptide 20, we examined its hemolytic activity in human red blood cells (Fig. 5). Our results showed that none of the tested concentrations exhibited hemolytic activity, reinforcing our hypothesis regarding rational design. Utilizing the active sequence of Myotoxin II and applying a structure-based design, we created a peptide with an IC50 value of 0.27 mg/ml against E. coli and an IC50 of 2.93 mg/mL against P. aeruginosa (ATCC 27853) while avoiding hemolytic activity.

Fig. 5
figure 5

Hemolytic activity of peptide 20.

Full size image

As part of the peptide 20 safety evaluation, tests were conducted on the reference cell line L929 (Fig. 6b). Since many antimicrobial agents are taken orally and can be absorbed in the intestines, they may have cytotoxic effects on epithelial cells30. Consequently, the safety assessment included the Caco-2 cell line (Fig. 6a)31, revealing that none of the tested concentrations exhibited toxicity affecting cell viability in the L929 and Caco-2 lines.

Fig. 6
figure 6

Resazurin assay for assessing cell viability.

Full size image

This further demonstrates that rational design rooted in structure facilitates the creation of safer peptides with potential biological efficacy against E. coli and P. aeruginosa.

Conclusions

The original sequence 115–120 of Myotoxin II was modified 20 times via molecular docking, revealing that peptide sequence 20 (KHWYKHYRH) demonstrated the highest affinity energy of -7.6 kcal/mol for lipopolysaccharide (LPS). The in vitro potency was evaluated against E. coli, yielding an IC50 of 0.27 mg/mL, while Pseudomonas aeruginosa (ATCC 27853) exhibited an IC50 of 2.93 mg/mL. In contrast, th e peptide proved ineffective against resistant strains, including the NDM-1-positive Klebsiella pneumoniae (ATCC BAA-2146) and the ESBL clinical isolate E. coli (CTX-M). Furthermore, the safety profile of peptide 20 was assessed, indicating that none of the concentrations tested resulted in hemolytic activity or loss of cellular viability in L929 and Caco-2 cells. This finding highlights that rational, structure-based design is an effective strategy for developing safe peptides.