Dual-functional self-assembled nanosystems with enhanced protease resistance: Promoting bacterial aggregation and immune activation for multidrug-resistant bacterial infection

Abstract

Multidrug-resistant bacterial infections are increasing globally and posing a greater threat to human health. The application of direct bactericidal agents can induce secondary infections and treatment failures. The antibacterial strategy of the innate immune system brings inspiration. Here, we developed highly stable bacterial-aggregating peptides with immunoregulatory function. These peptides were designed to capture multidrug-resistant bacteria, prevent their dissemination, and activate the antibacterial immune response of the host. Among these peptides, the central-bola amphiphile R₂F₄R₂ highly captured bacteria without directly killing them. R₂F₄R₂ was believed to self-assemble through the lateral connection of peptide chains. The tetra-Phe segments formed a hydrophobic core of nanoparticle, with Arg residues appearing on the surface. Notably, R₂F₄R₂ enhanced chemotactic response and phagocytic ability of macrophages, supported a transition to M2-macrophage phenotype to combat bacterial infection. Transcriptome sequencing and molecular docking analyses revealed that R₂F₄R₂ regulated the gene expression associated with immunoregulatory functions and modulated calcium-Rap1 signaling pathways. Finally, R₂F₄R₂ exhibited exceptional stability against proteolytic degradation and effectively entrapped invading pathogenic bacteria Escherichia coli to alleviate skin infections and intestinal inflammation. Overall, the bacterial-aggregating peptides represent a novel and effective strategy to combat multidrug-resistant infections.

Graphical abstract

1. Introduction

Antimicrobial resistance (AMR) has turned into a global public health challenge in the 21st century [1]. Without successful interventions, projections suggest that this number can rise to approximately 10 million annually by 2050 [2]. Traditional approaches that focus on directly killing bacteria may not be sufficient to control the emergence of antibiotic resistance [3]. The unique antibacterial mechanism of human α-defensin 6 (HD6), expressed and secreted by Paneth cells in the intestinal lumen, has attracted our attention. HD6 self-assembles into nanonets that entangle and agglutinate invading pathogens to prevent bacterial invasion [4]. Unlike conventional antimicrobial peptides (AMPs), HD6 employs a novel antibacterial mechanism of "entangling without attacking", which could decrease the risk of microbial resistance [5]. A similar mechanism is observed in the reduced form of human β-defensin 1 (hBD1), which entraps bacteria in extracellular net structures to prevent bacterial invasion without directly killing them [6]. Consequently, the application of antibacterial strategy of HD6 could combat multidrug-resistant bacterial infections [7].

Multiple lines of defense within the innate immune system collaborate to offer antibacterial defense [8]. This system relies on professional phagocytes, such as neutrophils, dendritic cells and macrophages, to recognize and phagocytose invading microbes and dead cells, thus maintaining host homeostasis [9]. While existing researches have focused on the bacterial-aggregating capacity of HD6 and mimics, the critical interplay with host immune defenses has not been extensively explored [10,11]. We proposed a molecular design strategy involves integrating bacteria entrapment and immunomodulatory activities, which enhances antimicrobial efficiency and minimizes the microbial resistance risk [7].

Despite the considerable promise of bacterial-aggregating peptides in combating drug-resistant bacterial infections, their proteolytic degradation and physiological instability fundamentally restrict their clinical translation [12,13]. Accordingly, there is an urgent need to develop clinically viable bacterial-aggregating peptides. Several methods for forming bacterial-aggregating peptides often require complex modifications. For example, incorporating highly hydrophobic groups or applying hydrophobic coatings can substantially increase synthesis costs and complicate process control, thereby restricting clinical applicability [10,14]. The remarkable properties of simple and naturally unmodified short peptides as nanomaterials are particularly compelling [15,16]. Furthermore, their self-assembly into structured nanomaterials enhances stability under physiological conditions by shielding cleavage sites and reducing protease binding affinity [15,17]. Consequently, we hypothesized that integrating supramolecular self-assembly with optimized peptide sequences resistant to protease could enhance the bioavailability and therapeutic efficacy of peptide-based antimicrobial materials (Fig. 1).

Fig. 1.

Schematic illustration of the dual-functional nanoparticle R₂F₄R₂ with enhanced protease resistance, promoting bacterial aggregation and immune activation for combating multidrug-resistant bacterial infections.

Based on the designed strategy above mentioned, we employed a straightforward surfactant-like peptide motif to guide supramolecular assembly, favoring low manufacturing costs and easy quality control (Fig. 2A) [18]. The choice of Arginine (Arg) and Phenylalanine (Phe) was based on their unique side-chain characteristics. The aromatic side chain of Phe enables robust π-π stacking interactions [19,20], whereas the guanidinium group of Arg serves as an effective hydrogen bond donor [21]. At physiological pH, the positively charged guanidinium group of Arg further promotes cation-π interactions with the phenyl ring of Phe [22]. Subsequently, conventional surfactant-like, central- and end-bola amphiphiles, and alternating hydrophobic and hydrophilic amino acids patterns were constructed. Proline (Pro) contains a pyrrole ring with a unique spatial configuration obstructing protease cleavage through steric hindrance, particularly when positioned after amino acids susceptible to trypsin and chymotrypsin [23]. This strategic placement enhances the protease stability of bacterial-aggregating peptides. Emerging evidence has illustrated that MTB Rv2626c may improve bacterial clearance via M2 macrophage polarization and recruitment. Finally, considering the regulation of the C-terminal 123–131 amino acid Rv2626c motif (₁₂₃LPEHAIVQF₁₃₁) with immune function of macropphages [24], the motif was linked to construct an upgraded multidimensional antibacterial system (Table S1, Figs. S1-S2).

Fig. 2.

Construction of peptide libarary and molecular structural basis of bacteria-aggregating activity. (A) Construction of bacterial-aggregating peptide systerm. Amino acid and peptides composition of the system; (B) SEM images of R₂F₄R₂ at a concentration of 2 µM, 32 µM and 512 µM. Scale bar: 1 µm and 500 nm (magnification); (C) AFM images of R₂F₄R₂ at a concentration of 2 µM, 32 µM and 512 µM. Scale bar: 400 nm; (D) CD spectra of R₂F₄R₂ in water; (E) FTIR spectra of R₂F₄R₂; (F) XPS spectra of C1s of R₂F₄R₂; (G) Size distribution of R₂F₄R₂ measured by DLS; (H) Zeta-potential of R₂F₄R₂. The differences between the groups were determined using one-way ANOVA followed by Tukey’s post hoc analysis. Values with a, b, and c indicate significant differences (P < 0.05).

Subsequently, we screened the potential bacterial-aggregating peptides against pathogens and observed that the central-bola amphiphile, R₂F₄R₂, exhibited excellent bacterial agglutination capability (Fig. 3A). R₂F₄R₂ was stabilized by hydrogen bonding and π-π interaction, and self-assembled into nanoparticles. Owing to the dual reinforcement from the self-assembly system and the Pro residue arrangement, R₂F₄R₂ maintains high integrity against common gastrointestinal proteases, thereby enhancing its oral bioavailability. Notably, R₂F₄R₂ exhibited immunomodulatory effects by promoting macrophage recruitment, phagocytosis, and M2 macrophage polarization. Comprehensive in vivo assessments indicated that R₂F₄R₂ effectively mitigated Escherichia coli-infected wound infection and enteritis. Collectively, this peptide can serve as a source of inspiration for the advancement of antimicrobial therapies, introducing innovative possibilities for future development (Fig. 1).

Fig. 3.

Bacterial-entrapping and aggregating activity of peptides. (A) E. coli 25922 agglutination index of peptide system within 1 h across different concentrations; (B) E. coli 25922 and (C) S. aureus 29213 agglutination after exposure to R₂F₄R₂ for 6 h; (D) Live/dead fluorescence image of E. coli 25922 and S. aureus 29213 treated with 32 µM R₂F₄R₂ and melittin for 2 h. Scale bar: 10 µm.

2. Materials and methods

2.1. Materials

All reagents and solvents employed in the organic synthesis were obtained from commercial suppliers and used without further purification. These custom-designed peptides were synthesized by GL Biochem (Shanghai) Ltd. and subsequently purified via RP-HPLC, achieving a purity level of > 95% (Fig. S1). The fidelity and molecular masses of the peptides were confirmed using MALDI-TOF-MS (Fig. S2). The powder was dissolved in deionized water at a concentration of 1.28 mM and stored individually at –40 °C.

2.2. CD, DLS and zeta potential

Circular dichroism (CD) measurements were conducted on R₂F₄R₂ at a concentration of 64 µM using a Jasco J-820 spectropolarimeter (Jasco, Japan) maintained at 25 °C. Baseline corrections were applied by subtracting the spectra of water.

The peptide nanoparticles were incubated at 37 °C overnight before analysis. The hydrodynamic diameter and zeta potential were determined using a Zetasizer Nano Z90 (Malvern Instruments, Worcestershire, U.K.) with samples placed in low-volume quartz cuvettes. The viscosity was set to that of pure water, and the refractive index was matched to the proteins. The series measurement mode was employed to track changes in hydrodynamic radius over time at 37 °C.

2.3. Peptide self-assembly observations

The morphologies of the peptide nanoparticles were visualized using scanning electron microscopy (SEM) and atomic force microscopy (AFM). The peptide samples were incubated overnight at 37 °C to promote self-assembly. A 10 µl aliquot of each peptide in sterile water was applied to coverslips, air-dried, and subsequently coated with gold using a Polaron SC7640 Sputter Coater. Imaging was performed using a Hitachi S-4800 SEM (Hitachi, Japan) operated at 5 kV. For AFM, 10 µl of the same peptide sample was evenly spread on a mica sheet, allowed to dry completely in air, and then examined using a Bruker Fastscan AFM (Bruker, USA).

To examine lipopolysaccharide (LPS) and lipoteichoic acid (LTA) effects on peptide conformation, R₂F₄R₂ was incubated overnight at 37 °C with 100 µg/ml LPS and 10 µg/ml LTA. The solution was then drop-cast onto a copper-coated carbon grid, air-dried, and stained with 0.1% phosphotungstic acid for 3 s. Representative images were acquired using a TEM (HT7800, Hitachi High-Tech Corp.). Subsequently, the critical aggregation concentration of R₂F₄R₂ in LPS/LTA solution was determined using the fluorescent probe ANS (Shanghai Aladdin Biochemical Technology Co., Ltd.), following our established protocols [13].

2.4. FTIR and XPS analysis

Fourier transform infrared (FTIR) spectras were recorded using a Nicolet 6700 spectrometer (Thermo Nicolet Ltd., USA). Dried peptide samples (2 mg) were loaded into a KBr cuvette, and spectra were collected in the range of 4000–400 cm⁻¹ with a resolution of 4 cm⁻¹, averaging 32 scans per sample.

Dried peptide samples (20 mg) were loaded into a high-resolution X-ray photoelectron spectroscopy (XPS) to obtain individual spectra for carbon, nitrogen and oxygen using an Escalab 250Xi system (Thermo Fisher) equipped with Al Kα X-rays (1486.6 eV). The elemental ratios were determined from the spectra, with the C1s peak at 284.8 eV used as a calibration reference.

2.5. Bacterial negative stain TEM

E. coli 25922 and Staphylococcus aureus (S. aureus) 29213 were rinsed three times with 10 mM PBS and adjusted to an OD₆₀₀ of 0.4 during their logarithmic growth phase. The resulting bacterial suspension was then combined with peptide nanoparticles at a final concentration of 32 µM and incubated at 37 °C for 2 h. After incubation, 20 µl of the bacterial mixture was deposited onto a copper grid for 2 min and stained with 0.1% phosphotungstic acid for 3 s. The samples were air-dried for 18 h at room temperature before being observed using a Hitachi H-7650 TEM (Hitachi, Japan).

2.6. TEM and SEM characterization

E. coli 25922 and S. aureus 29213 morphologies exposed to the peptide were characterized using SEM and TEM. Bacteria in the logarithmic phase were washed three times with 10 mM PBS and resuspended to an OD₆₀₀ of 0.4. The bacterial suspension was then treated with R₂F₄R₂ nanoparticles at a final concentration of 32 µM and incubated at 37 °C for 2 h. After incubation, the bacterial cells were centrifuged, and the supernatant was removed. The pellet was subsequently fixed overnight at 4 °C using 2.5% (w/v) glutaraldehyde. Sample preparation followed standardized protocols, as previously outlined. Finally, the external and internal morphologies of E. coli 25922 and S. aureus 29213 were examined using an S-4800 SEM (Hitachi, Japan) and a Hitachi H-7650 TEM (Hitachi, Japan).

2.7. Cytotoxicity assay

Peptide cytotoxicity was evaluated using three mammalian cell lines [murine macrophage RAW 264.7; human embryonic kidney (HEK) 293T; intestinal porcine epithelial (IPEC-J2) cells], using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye reduction assay. RAW 264.7, HEK 293T, and IPEC-J2 cells were seeded into 96-well plates at a density of 3 × 10⁴ cells/well in RPMI 1640, high-glucose DMEM and DMEM/F12 (with 10% fetal bovine serum, v/v), respectively, and incubated overnight at 37 °C in 5% CO₂ atmosphere. After medium removal, cells were exposed to peptides at varying concentrations for 4 h. Subsequently, 50 µl of 5 mg/ml MTT (prepared in sterile PBS) was added and incubated for 4 h. The supernatant was carefully removed, and 150 µl dimethyl sulfoxide was added. The plates were then gently agitated for 15 min. The controls included wells without cells (negative control) and wells without peptide treatment (positive control). The cytotoxicity was calculated using the formula: (measured OD − negative control OD)/(positive control OD − negative control OD) × 100%.

2.8. Live/dead staining assay

To assess bacterial cell toxicity in response to peptide treatment, a live/dead staining assay was performed. E. coli 25922 and S. aureus 29213 cells, adjusted to an OD₆₀₀ of 0.4 in PBS, were treated with 32 µM peptide nanoparticles and 32 µM melittin for 2 h at 37 °C. After incubation, 5 µg/ml SYTO9 and 20 µg/ml PI were added, followed by an additional 30 min incubation at 37 °C. The resuspended bacterial sample was transferred to a glass slide for visualization using a super-resolution microscope (DeltaVision OMX SR, GE Healthcare UK Ltd.).

2.9. Bactericidal activity assay

Bacterial suspensions of E. coli 25922 and S. aureus 29213 in logarithmic growth phase were prepared in PBS (10 mM, pH 7.0) and adjusted to an OD₆₀₀ of 0.4, followed by a 1000-fold dilution in PBS. Aliquots of the diluted bacterial suspension were mixed with varying concentrations of R₂F₄R₂ in PBS. After 3 h of incubation at 37 °C, serial dilutions were plated on Mueller-Hinton agar and cultured overnight at 37 °C for colony counting.

2.10. Hemolytic properties

The hemolytic activity of the peptides was evaluated by quantifying hemoglobin release from lysed erythrocytes. Human red blood cells (hRBCs) were isolated from whole blood and washed three times with PBS to remove white blood cells and platelets. A 50 µl aliquot of 5% hRBCs was mixed with 50 µl peptide solutions at concentrations ranging from 2 to 128 µM in a 96-well plate. Triton X-100 (0.1%) was used as a positive control, and PBS served as the negative control. The peptide-hRBC mixtures were incubated at 37 °C for 1 h, followed by centrifugation at 1000 g for 5 min at 4 °C. Hemoglobin release was quantified by measuring the supernatant absorbance at 570 nm using an Infinite M200 Microplate Reader (Tecan). Hemolysis percentage was calculated using the formula: (measured OD − negative control OD)/(positive control OD − negative control OD) × 100%.

2.11. In vitro bacterial entanglement and sedimentation

Bacterial cells cultured to the exponential phase at 37 °C with shaking at 200 rpm were centrifuged at 1000 g for 5 min, resuspended in PBS to an OD₆₀₀ of 0.4, and mixed with peptide nanoparticles at concentrations ranging from 32 to 512 µM. The mixtures were transferred to ethanol-sterilized cuvettes and incubated at room temperature for 6 h. At designated time points, 50 µl aliquots from the supernatant were diluted with 450 µl PBS, gently vortexed, and serially diluted ten-fold from 10⁻² to 10⁻⁴, then incubated overnight at 37 °C. The solution turbidity and the bacterial colony-forming units (CFU) in the supernatant were recorded. The bacteria-entrapping efficiency was computed as a weighted average of the proportion of E. coli 25922 agglutination within 1 h at different peptide concentrations.

2.12. In vivo detection of bacterial entanglement

Under anesthesia, a midline laparotomy was performed, and a 4-cm segment of the distal ileum was surgically exposed and ligated using vascular clips to preserve the mesenteric vascular arcades. The isolated intestinal segment was injected with 100 µl PBS containing 4 × 10⁸ CFU enterotoxigenic E. coli (ETEC) K88 mixed with 32 µM R₂F₄R₂ nanoparticles. After 2 h, glutaraldehyde fixative was injected into the loop, and the entire ileal segment was fixed overnight. The samples were then processed for SEM, as previously described [25].

2.13. Proteolytic resistance assays

Peptide resistance to proteolysis was evaluated using Tricine-SDS-PAGE, RP-HPLC, and CD spectroscopy. Peptide solutions (2.56 mM) were mixed in equal volumes with a protease solution (20 mg/ml, final concentration: 10 mg/ml) or with simulated intestinal fluid (SIF) and simulated gastric fluid (SGF) [26]. The protease concentration used was comparable to the concentrations (10 mg/ml) present in SIF and SGF, as specified by the U.S. Pharmacopoeia [26]. The control groups consisted of the peptide alone and the enzyme alone. The mixtures were incubated at 37 °C, and samples were collected at predetermined time points (0, 1, 2, 4 and 8 h). After heating at 100 °C for 10 min, the mixtures were centrifuged at 13,000 g for 10 min to remove precipitated proteases. The supernatants were analyzed using 16.5% Tricine-SDS-PAGE, RP-HPLC and CD spectroscopy, following our established protocols [27,28].

2.14. Phagocytosis assay

RAW 264.7 cells were seeded into 24-well plates at a density of 2 × 10⁵ cells per well and incubated overnight at 37 °C in a 5% CO₂ atmosphere. E. coli ATCC 25922 (OD₆₀₀ = 0.4 in PBS) was pretreated with 32 µM R₂F₄R₂ nanoparticles and Rv2626c_123–131 for 2 h at 37 °C. Pretreated E. coli was then added to the adherent RAW 264.7 cells at a multiplicity of infection (MOI) of 50. To assess phagocytosis, cells were incubated with the bacterial particles for 2 h, followed by treatment with 5 µg/ml gentamicin for 2 h at 37 °C. The cells were then lysed using 1% Triton buffer, and the lysate was plated on agar plates to determine the CFU.

For visualizing macrophage phagocytosis of peptide-pretreated E. coli, RAW 264.7 cells were seeded in confocal Petri dishes (2 × 10⁵) and incubated overnight. E. coli strain BL21, harboring the PET-28a-EGFP plasmid for green fluorescence expression, was prepared according to a previously described method [12]. Sample preparation was performed as described above. After removing non-internalized E. coli, the infected macrophages were fixed and permeabilized using immunostaining fixation and permeabilization solutions containing Triton X-100 (Beyotime, China) for 10 min. The nuclei were stained with DAPI. The samples were visualized using a Leica TCS SP8 confocal laser scanning microscope (Germany).

2.15. RNA extraction and quantitative qRT-PCR

E. coli ATCC 25922 (OD₆₀₀ = 0.4 in PBS) was pretreated with 32 µM R₂F₄R₂ nanoparticles for 2 h at 37 °C, followed by RAW 264.7 cell infection at an MOI of 50 for 2 h. Total RNA was isolated from intestinal tissues or cell lines using RNAisoPlus reagent. After quantification, cDNA was synthesized using the PrimeScript RT Master Mix according to manufacturer specifications. Quantitative real-time PCR was performed using SYBR Green mix with primers listed in Table S2. Gene expression was quantified using the 2^−ΔΔCT method, normalized to β-actin.

2.16. Western blot analysis

RAW 264.7 cells were infected with R₂F₄R₂-pretreated E. coli (MOI 50; 2 h). Protein extracts were obtained from cells using RIPA buffer supplemented with 1% PMSF (IP0280, Solarbio). Following denaturation at 95 °C for 10 min, protein concentrations were determined using the BCA assay. Equal amounts of protein were separated by 10%–12% SDS-PAGE and transferred to PVDF membranes. After blocking with 5% non-fat milk for 2 h, membranes were incubated with primary antibodies against TLR4 (19811, Proteintech), phospho-NF-kB p65 (82335-1-RR, Proteintech), MYD88 (67969, Proteintech), and β-actin (GB11001, Servicebio) at 4 °C overnight, followed by HRP-conjugated secondary antibodies for 1 h at room temperature. Signals were detected using ECL substrate (KGC4601–100, Keygen BioTECH), and band intensities were quantified with ImageJ software with β-actin for normalization.

2.17. ELISA for cytokine detection

Macrophage samples were prepared as described in Section 2.16. Subsequently, cell culture supernatants were collected and analyzed for IL-6, TNF-α, IL-10 and TGF-β secretion using commercial ELISA kits (JM-02446M2, JM-02415M2, JM-02459M2 and JM-02418M2; Jingmei Biotechnology Co., Ltd., Jiangsu, China), according to the protocols of the manufacturer. The absorbance was measured at 450 nm using a microplate reader.

2.18. Transwell chemotaxis assay

A chemotaxis assay was performed using a 24-well Transwell plate with 8 µm pores (Corning). The bottom chamber was filled with 600 µl RPMI 1640 medium (serum-free) containing 32 µM R₂F₄R₂ nanoparticles and Rv2626c_123–131. RAW 264.7 cells (2 × 10⁵) suspended in 200 µl serum-free RPMI 1640 medium were seeded into the upper chamber and incubated for 24 h. Following incubation, the chambers were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet to visualize migrated macrophages. The stained cells were observed and imaged under a microscope to quantify chemotactic activity.

2.19. Macrophage phenotype polarization assay

Flow cytometry was used to evaluate the macrophage phenotype. Macrophage samples were prepared as described in Section 2.14. Subsequently, cells were exposed to fluorescence-labeled antibodies from Elabscience, specifically F4/80 (E-AB-F0995C), CD206 (E-AB-F1135E) and CD86 (E-AB-F0994D), following the guidelines of the manufacturer. Cell fluorescence was analyzed by using flow cytometry and FlowJo 10.0 software.

2.20. RNA isolation and transcriptome analysis

To evaluate the transcriptional changes in gene expression in macrophages in response to E. coli infection and peptide treatment, the macrophage samples were prepared as described in Section 2.14. After removing the culture medium, the cells were washed thrice with PBS. Total RNA was extracted using TRIzol, and the samples were immediately frozen in liquid nitrogen. RNA sequencing (RNA-seq) was performed on an Illumina NovaSeq 6000 platform at Shanghai Personal Biotechnology Co., Ltd. Difference expression of genes was analyzed by DESeq (version 1.30.0). Calculate P-value by hypergeometric distribution method (P < 0.05) to find the Kyoto encyclopedia of genes and genomes (KEGG) term with significantly enriched differential genes. GSEA based on the KEGG term applies a widely used method, a permutation test.

2.21. Molecular docking

To further study the binding region and interaction modes of R₂F₄R₂ with calcium signaling molecule (PDB ID: 5FBH, resolution = 2.7 Å), the AlphaFold 3 server was employed to perform molecular docking [29]. The calcium signaling molecule was defined as a receptor, and R₂F₄R₂ as a ligand. The model with interface predicted TM-score (ipTM) and a pTM score of 0.93, and were considered high-quality predictions (> 0.8) and were downloaded. The files from the server were visualized using Pymol.

2.22. Ethics statement

All experimental procedures involving human blood samples were reviewed and approved by the Ethics Committee of Northeast Agricultural University Hospital, Harbin, China (Approval No. NEAUEC20230269). Animal studies were conducted in compliance with the animal welfare guidelines approved by the Animal Welfare Committee of Northeast Agricultural University (Protocol No. #NEAU-[2013]−9).

2.23. Wound-healing mouse model

Female ICR mice aged 8–10 weeks and weighing ∼20 g were obtained from Liaoning Changsheng Biotechnology Company (Benxi, China). A full-thickness wound was created in the dorsum center using an 8 mm biopsy punch to remove the epidermis and superficial parts of the dermis. Two animal infection models based on the cutaneous wound model were established and tested.

In the initial experiment, twenty-four mice were randomly allocated to two groups. One group was inoculated with E. coli alone, whereas the other group received E. coli pre-incubated with R₂F₄R₂ nanoparticles. Full-thick skin incisions (∼8 mm) were made on the dorsal surface of each mouse, and the treatments were applied directly to the wounds. This experimental setup allowed for a more accurate assessment of the therapeutic effects of R₂F₄R₂ in a model that closely mimics human wound healing dynamics. The mice were assigned to a blinded protocol as follows: E. coli group (1 × 10⁸ CFU/ml) and R₂F₄R₂ group (1 × 10⁸ CFU/ml; R₂F₄R₂ pre-treatment, 32 µM). For the R₂F₄R₂ treatment group, E. coli (1 × 10⁸ CFU/ml) was incubated with R₂F₄R₂ (32 µM) in 1 ml 0.9% saline for 2 h. For the E. coli alone group, the bacteria were processed without any pre-incubation. Following incubation (for the R₂F₄R₂ group) or direct addition (for the E. coli group), both samples were centrifuged. The resulting bacterial pellets were resuspended in 50 µl of 0.9% saline and applied to the wounds. Both groups were monitored for 48 h to evaluate infection progression. Wound tissues were harvested at the 48 h time point for histological and immunofluorescence analyses.

In a follow-up experiment, thirty-six mice were randomly assigned to three groups. The control group received 50 µl 0.9% saline. E. coli and R₂F₄R₂ groups were treated with 50 µl 0.9% saline containing E. coli (4 × 10⁸ CFU/ml). Mice in the R₂F₄R₂ group were administered 200 µL R₂F₄R₂ nanoparticles (32 µM) on Day 1 and 4 post-infection. This experimental design allowed for a comprehensive evaluation of the therapeutic efficacy of R₂F₄R₂ in modulating wound healing and resolving infection.

2.24. Enteritis mouse model

Experiment 1: The cumulative toxicity assay for R₂F₄R₂ was first investigated in vivo. Eight ICR mice in each group were orally administered with saline, 10, 20 and 30 mg/kg R₂F₄R₂ for 3 d at 24 h intervals. Throughout the 3 d, changes in body weight were monitored. On Day 4, all animals were euthanized via cervical dislocation. The liver and kidneys were excised and weighed, and the intestinal tissues were fixed in 4% paraformaldehyde for hematoxylin and eosin (H&E) staining. Serum samples were collected from the orbital vein and analyzed using an automatic biochemical analyzer to measure BUN, ALT, AST, ALP and TBIL levels. These assessments provided insights into potential hepatotoxicity and nephrotoxicity.

Experiment 2: To disrupt colonization resistance and facilitate infection with enterotoxigenic ETEC K88, 36 healthy female ICR mice (6 weeks old) were divided into 3 groups and administered an antibiotic cocktail (0.4 mg/ml kanamycin, 0.035 mg/ml gentamicin, 850 U/ml colistin, 0.215 mg/ml metronidazole, and 0.045 mg/ml vancomycin) in their drinking water from Day 1 to 2. Subsequently, the mice were switched to normal water. On Day 3–4, the mice in E. coli and R₂F₄R₂ groups were challenged with ETEC K88 inoculation (2.0 × 10⁹ CFU/ml) twice daily. At 48 h post-infection, mice in the R₂F₄R₂ treatment and the control groups received oral gavage of 200 µl either 30 mg/kg R₂F₄R₂ nanoparticles (optimal concentration determined in pre-experiments) or saline twice daily from Day 5–6. On Day 7, mice were euthanized by cervical dislocation under full anesthesia induced by isoflurane. Throughout the experimental period, changes in body weight were monitored. The livers, kidneys, lungs, and spleens were excised and weighed. Colon samples were collected and fixed in 4% paraformaldehyde for subsequent H&E staining, PAS staining, and immunofluorescence analysis.

2.25. 16S rRNA amplicon sequencing

Total microbial genomic DNA was extracted from cecal contents using the Qubit dsDNA HS Assay Kit (Q32854, Invitrogen). The hypervariable V3-V4 regions of 16S rDNA were amplified using specific primers (forward: ACTCCTACGGGAGGCAGCA; reverse: GGACTACCAGGGTATCTAATCCTGTT). The amplified products were sequenced using the Illumina platform. The DADA2 pipeline was used for data processing, including primer removal, quality filtering, denoising, splicing, and clustering. Species annotation was performed using the GreenGenes database (Release 13.8, http://greengenes.secondgenome.com/). Sequencing was conducted by Shanghai Personal Biotechnology Co., Ltd., China, and data analysis was performed using the free online platform Personalbio Gene Cloud (https://www.genescloud.cn/). The Kruskal–Wallis test was used to determine differential OTU abundance analyses among these three groups.

2.26. Histological and immunofluorescent staining

After fixation in 4% paraformaldehyde, the tissues were embedded in paraffin and sectioned at 4 µm. Following dewaxing in xylene and rehydrating using ethanol gradients, paraffin sections were stained. H&E and PAS staining were performed for histological analysis and goblet cells, respectively. Masson staining of wound tissues was determined for collagen deposition in the wound remodeling phase. For TUNEL analysis, paraffin sections of wound tissues were stained using the TUNEL assay kit (Servicebio) following the protocol of the manufacturer. For immunofluorescent staining, the paraffin sections were incubated with primary antibodies, including KI67, F4/80, CD86, CD206, IL-6, IL-10 and nuclear factor κB (NF-κB). The corresponding secondary antibody was marked with appropriate species-specific antibodies in 5% BSA in PBS in the dark at room temperature for 1 h. The nuclei were counterstained with DAPI. All images for H&E and PAS staining were obtained using a digital slide scanner (SN101240120, Winmedic Tech Co., Ltd.), and representative images were captured. Other images were acquired using a fluorescence microscope (Nikon, Eclipse C1). Histopathological analysis of H&E-stained sections was conducted by a single blinded evaluator for the treatment groups. Intestinal lesions were scored according to established criteria [30], accounting for inflammatory cell infiltration and tissue damage severity. Parallel wound healing assessment employed a 0–5 scale evaluating epithelization, granulation tissue formation, inflammatory cell presence, abscess development, and tissue architecture integrity [31].

2.27. Statistical analysis

Data are presented as mean ± SD. Statistical analyses were performed using an unpaired t-test for comparisons between two groups and a one-way ANOVA followed by Tukey's post hoc test for comparisons among multiple groups. Differences were considered statistically significant at *P<0.05, **P<0.01, ***P<0.001.

3. Results and discussion

3.1. Deciphering the supramolecular self-assembly architecture of R₂F₄R₂

Structure-function studies of HD6 could reveal that the molecular structure is the determining factor for its distinctive bacteria-agglutinating activity [32]. Consequently, the R₂F₄R₂ microstructure was first characterized using SEM and AFM. The images in Fig. 2B-C and S8-S9 reveal that R₂F₄R₂ could transform the morphology from loose nanoparticles with an uneven particle size distribution to regularly well-proportioned, uniform and tight nanoparticles with diameters of 132.8 ± 5.68 nm to 53.45 ± 1.76 nm as the peptide concentration increased (2–512 µM). Briefly, this implies that R₂F₄R₂ formed supramolecular structures driven by concentration, which was closely related to the bacterial aggregation activity.

The secondary structure of R₂F₄R₂ was further analyzed using CD spectroscopy (Fig. 2D). In aqueous buffer, the R₂F₄R₂ CD spectrum exhibited a minimum at ∼198 nm and near-zero signals beyond 220 nm, indicating a random coil conformation. This contrasts sharply with alternating [RF] sequences that self-assemble into fibrillar β-sheet structures with strong amyloidogenic characteristics [33]. The presence of only two Arg residues near each terminal is insufficient to maintain the alignment required for strong hydrogen bonding and β-sheet formation [34]. Additionally, introducing a large amount of Pro could disrupt the secondary structure of potential β-sheet and decrease the propensity to aggregate into amyloid-like fibrils [35].

R₂F₄R₂ chemical bonding state was analyzed using XPS and FTIR spectroscopy. The presence of an absorbance peak at 1630 cm^-1 in the FTIR spectra indicates the hydrogen bonding interactions (Fig. 2E). The XPS spectra confirmed that the R₂F₄R₂ was stabilized primarily by π-π stacking forces with a peak located at 292.16 eV (Fig. 2F). Furthermore, full XPS survey, N1s, and O1s spectra are illustrated in Figs. S10-S11. Collectively, these results confirm that the R₂F₄R₂ nanoparticle structure was stabilized by hydrogen bonding and π-π interactions. The hydrodynamic sizes of R₂F₄R₂ nanoparticles in aqueous buffer were analyzed using dynamic light scattering (DLS) with a Zetasizer Nano ZS90. The results exhibited a decrease in hydrodynamic diameter as the concentration increased, aligning with the more compact nanoparticle structures observed in SEM (Fig. 2G). The zeta potential results indicated a high number of positive surface charges on the peptides, facilitating their binding and entrapment of bacteria by utilizing the abundant negative charges of the bacteria (Fig. 2H) [36,37].

3.2. Unraveling the bacterial aggregation activity and entrapment mechanism of R₂F₄R₂

The bacterial entrapment efficiency results depicted in Fig. 3A and S3 demonstrated that central-bola amphiphile R₂F₄R₂ exhibited the highest comprehensive bacterial agglutination efficiency compared to other bola-amphiphilic peptides, reaching 88.76% against E. coli 25922. Notably, R₂F₄R₂ almost sedimented all E. coli 25922 and S. aureus 29213 clumps at high concentrations, while exhibiting a time- and concentration-dependent relationship (Fig. 3B and 3C). Except for E. coli 25922 and S. aureus 29213, the sedimentation phenomenon following R₂F₄R₂ treatment was also observed in Gram-negative bacteria, ETEC K88, Salmonella typhimurium 7731 and 14028, and Gram-positive bacteria, S. aureus 25923 (Fig. S4). Importantly, R₂F₄R₂ sedimented three clinically isolated drug-resistant bacterial strains, including tetracycline-resistant E. coli 2566, polymyxin E-resistant E. coli 1515, and methicillin-resistant S. aureus 43300 clumps, particularly at high concentrations (Fig. S5). In these dose-response studies, we found that compared to the highest tested concentration of 512 µM, the peptide R₂F₄R₂ at 32 µM already exhibited effective aggregation activity against the pathogenic bacteria. To visually investigate the bacterial agglutination behavior induced by the central-bola amphiphile, an E. coli strain (BL21) consisting of the PET-28a-EGFP plasmid was constructed. Interestingly, R₂F₄R₂ could agglutinate E. coli at a low concentration of 32 µM, whereas no cell agglutination was observed when bacteria were incubated with other peptides at the same concentration (Fig. S6). To further investigate the relationship between the antimicrobial and agglutination activities of R₂F₄R₂, bacterial cultures were analyzed using confocal microscopy with SYTO9/propidium iodide (PI) nucleic acid fluorescent labels, enabling simultaneous monitoring of cell agglutination and viability over time. After 2h-incubation with R₂F₄R₂ and melittin (as a control), E. coli and S. aureus in the R₂F₄R₂-treated group exhibited aggregated live (green) and almost no dead (red) cells, but the melittin group emitted strong red fluorescence (Fig. 3D). Furthermore, the bactericidal activity assay exhibited that R₂F₄R₂ exhibited no bactericidal activity against E. coli and S. aureus within the tested range (Fig. S7). Based on these results, we selected 32 µM as the working concentration for subsequent experiments. The use of the minimum effective concentration (32 µM) aligns with strategies to reduce evolutionary pressure and delay potential resistance. Compared to traditional AMPs, such as melittin, with direct bacterial-killing activity, R₂F₄R₂ possesses "entangling without attacking" activity. The mechanism of action and this fundamental difference suggest that this strategy is not subject to severe bacterial resistance limitations.

Next, we incubated bacterial cultures with R₂F₄R₂ and visualized the samples using TEM and SEM to understand the mechanisms of bacterial aggregation visually. Under negative staining TEM, R₂F₄R₂ bound to the E. coli and S. aureus cell surface and formed amorphous aggregates at the bacterial cell surface without disrupting cell membrane integrity (Fig. 4B and 4C). Subsequently, the changes in the internal and external morphologies of E. coli and S. aureus cells before and after treatment with R₂F₄R₂ were observed. As anticipated, TEM and SEM analysis illustrated that the untreated bacterial cell surface appeared smooth and intact. Following R₂F₄R₂ treatment, a slight shrinkage of the bacterial cell membrane was observed without intracellular content leakage (Fig. 4D and 4E), and the bacteria became surrounded by clearly visible aggregates (Fig. 4F and 4G). This phenomenon is like the in situ self-assembled fibrous networks formed by natural HD6 in trapping bacteria (Fig. 4A) [4].

Fig. 4.

Bacterial aggregation induced by R₂F₄R₂. (A) Schematic illustration of bacterial aggregation; (B-C) Negative-stained TEM images; (D-E) TEM images and (F-G) SEM images of E. coli 25922 and S. aureus 29213 after exposure to R₂F₄R₂.

Recent studies suggest that cell agglutination can be consistently driven by specific interactions between peptides and the bacterial surfaces. In a study by Petrlova et al., thrombin-derived C-terminal peptides were found to bind to bacterial LPS, forming amorphous amyloid-like aggregates [38]. Similarly, the BTT1 analog exhibited amyloid-nucleating properties when exposed to LTA and LPS on the bacterial surface, allowing precise control over microbial entrapment and elimination [39]. Therefore, R₂F₄R₂ was incubated with LPS and LTA extracted from Gram-negative and Gram-positive bacteria, respectively. Notably, a significant increase in ANS fluorescence intensity was observed upon LPS and LTA addition to R₂F₄R₂ (Figs. S12–S13). These results suggested that the aggregation of Gram-negative and Gram-positive bacteria by R₂F₄R₂ can be enhanced through interaction with LPS and LTA [40]. We then used negative-stained TEM to observe peptide assembly in LPS or LTA presence. Fig. S14 illustrates that in a water environment, R₂F₄R₂ exhibited limited self-assembly, forming small particle-like structures. However, these structures aggregated into larger and more uniform particles when LPS or LTA was introduced. Given the high LPS and LTA concentration in the outer membranes of Gram-negative and Gram-positive bacteria, respectively, serving as primary interaction sites for foreign substances [41], these findings are highly significant. Collectively, these results highlight the successful design of a trap-only HD6 peptide, offering potential for addressing drug resistance and microbiota dysbiosis.

3.3. Evaluating the proteolytic resistance profile of R₂F₄R₂ in simulated physiological environments

Peptide instability in systemic environments remains a significant hurdle in developing the bacterial-aggregating peptides. As depicted in Fig. S15A, R₂F₄R₂ demonstrated substantial resistance to digestion by pepsin, trypsin, chymotrypsin, SIF and SGF, with bands resembling those of the controls remaining visible even after 8 h of incubation. Similarly, the RP-HPLC profile exhibited that the peak shapes and profiles of the R₂F₄R₂ after incubation with various proteases were well maintained compared to the peptide alone (Fig. S15B). R₂F₄R₂ was slightly degraded after incubation with pepsin, SIF and SGF, but there was still a mass of intact peptides (above 54.97%). R₂F₄R₂ secondary structure after protease treatment was also evaluated using CD spectroscopy (Fig. S15C). The spectra illustrated that the peptide retained a structural profile similar to the controls across all treatments. Notably, the [θ]_{198 nm} values for R₂F₄R₂ exposed to proteases displayed only slight deviations from the untreated peptide, suggesting minimal alterations in the secondary structure due to these proteases. In comparison, melittin was degraded entirely after incubation with 1 mg/ml trypsin for 1 h [42].

A key advantage of this study lies in the strategic incorporation of proline into the R₂F₄R₂ sequence, a rational design approach that effectively evades protease cleavage sites and fundamentally enhances peptide stability [27,43]. Compared to previously reported bacterial-aggregating peptides, R₂F₄R₂ exhibits remarkable resistance to high protease concentrations (10 mg/ml), superior to that reported for bacterial aggregating peptides to date, such as BTT2 (15 nM) [39] and SAP (peptide: trypsin = 1:2, v/v) [12]. Importantly, R₂F₄R₂ exhibited potential resistance against SIF and SGF. These findings not only validate the effectiveness of the rational design strategy but also highlight the distinctive advantages of R₂F₄R₂ as a promising drug candidate, thereby facilitating its further development.

3.4. Elucidating the immunomodulatory function and mechanism of R₂F₄R₂ on RAW 264.7 macrophages

Macrophages provide the first line of defense against infections. Upon infection, blood-derived macrophages migrate to infected sites, where they become activated to identify, engulf and kill invasive bacteria [44]. In this study, we proposed that the bacterial clumps aggregated by R₂F₄R₂ can re-rouse the antibacterial and anti-inflammatory functions of macrophages (Fig. 5I). First, the chemotactic function of peptides on murine macrophages (RAW 264.7) was detected using transwell assays. The assays revealed that compared with Rv2626c_123–131 treatment and control, R₂F₄R₂ treatment resulted in more macrophages in the same field of view, indicating that the self-assembled nanosystem could enhance macrophage migration (Fig. 5A). A confocal laser scanning microscope was used to explore whether bacterial agglutination facilitated the efficient clearance of bacterial clusters by phagocytes at infection sites. Compared with the control and Rv2626c_123–131 groups, a significantly higher number of E. coli aggregates (indicated by green fluorescence) were engulfed by macrophages treated with R₂F₄R₂ (Fig. 5C). Furthermore, intracellular live bacteria quantification across different groups using the spread plate method reinforced the finding that R₂F₄R₂ reactivated macrophages, enhancing their phagocytic and antibacterial infection (Fig. 5B).

Fig. 5.

Immunoregulatory effect of R₂F₄R₂ on macrophages. (A) Typical images of RAW 264.7 cells cultured on transwell stimulated with R₂F₄R₂ and Rv2626c_123–131. Scale bar: 150 and 75 µm (magnification); (B) Quantification of intracellular live bacteria in different groups using spread plate method. Data are presented as mean ± SD; n = 6; (C) Confocal laser scanning microscope of phagocytosis of E. coli (green) clumps pretreated with 32 µM R₂F₄R₂ and Rv2626c_123–131 for 2 h by RAW 264.7 cells (DAPI, blue) or not; (D) Typical scatter plots of macrophage phenotype markers CD86 (M1 macrophage marker) and CD206 (M2 macrophage marker) as detected using a flow cytometry; (E) Proportions of CD86⁺CD206⁻ and CD86⁻CD206⁺ cells. Data are presented as mean ± SD, n = 4; (F) Levels of inflammatory cytokines mRNA IL-6, TNF-α, IL-10 and TGF-β in RAW 264.7 cells determined by qPCR (n = 6); (G) Quantification of inflammatory cytokines in the supernatant by ELISA (n = 6); (H) Protein level of inflammatory cytokines in RAW 264.7 cells (n = 3); (I) Schematic illustration of macrophages immunoregulatory function activated by R₂F₄R₂; (J) Quantification of inflammatory cytokines in RAW 264.7 cells (n = 3). The difference in B was determined using one-way ANOVA followed by Tukey’s post hoc analysis. In E-G and J, the differences between the groups were determined using an unpaired t-test. *P < 0.05, **P < 0.05 and ***P < 0.001.

Macrophages exhibit remarkable adaptability, rapidly changing their functional states in response to the dynamic microenvironment created by tissue infection or inflammation, a process referred to as "polarization" [45]. This polarization is crucial for resolving inflammation and restoring tissue homeostasis after infection or injury [45]. Macrophages are classified into M1, characterized by a proinflammatory phenotype, whereas M2 are recognized for their anti-inflammatory properties [46]. Next, we investigated whether R₂F₄R₂ treatment altered the macrophage phenotype during E. coli infection. Flow cytometry assay revealed that the M1 macrophage percentage in the E. coli-infected group decreased from 18.20% to 1.56%, whereas the M2 macrophage percentage increased from 17.00% to 63.20% after R₂F₄R₂ treatments (Fig. 5D and 5E). Furthermore, R₂F₄R₂ alone affected the macrophage phenotype, particularly the M2-like phenotype (Fig. S16). Next, we investigated the potential anti-inflammatory properties of R₂F₄R₂. The qPCR and ELISA results demonstrated that R₂F₄R₂ treatment effectively downregulated the proinflammatory cytokine expression and secretion (IL-6 and TNF-α), while upregulating the anti-inflammatory cytokine expression and secretion (IL-10 and TGF-β) (Fig. 5F and 5G). The TLR/MyD88/NF-κB inflammatory signaling is an important pathway in the body's inflammatory system. It participates in the occurrence and regulation of multiple diseases. Our results showed that R₂F₄R₂ treatment suppressed the TLR4, MyD88 and NF-κB (p65 phosphorylation) activation in RAW 264.7 cells infected with E. coli (Fig. 5H and 5J).

These findings indicate that bacterial clumps aggregated by R₂F₄R₂ enabled macrophages to swiftly migrate to infection sites, facilitating the engulfment and eradication of bacteria. Its effect on the macrophage antibacterial function may be a consequence of gene expression [47]. Thus, RNA-seq analysis was performed to examine the transcriptomes of RAW 264.7 macrophages. Principal component analysis revealed distinct clustering of the control and peptide transcriptomes (Fig. 6A). There were 82 significantly upregulated and 171 significantly downregulated genes in peptide versus control transcriptomes, as depicted in the volcano plot of the 1400 most varying genes (Fig. 6B).

Fig. 6.

Immunoregulatory mechanism of R₂F₄R₂ on macrophages. (A) Principal coordinate analysis. n = 5 for each group; (B) Volcano plot depicting DEGs identified by DESeq2 [69], defined as those with a |log₂ (fold change)| > 1.5 and an adjusted P < 0.05 (after Benjamini-Hochberg correction); (C) GSEA of the total genes associated with "lysosome", "IL-17 signaling pathway", and "viral protein interaction with cytokine and cytokine receptor" in macrophages using GSEA v3.0 based on the KEGG database by a permutation test; (D) Top 20 KEGG pathways based on the Rich factor using hypergeometric distribution method; (E) Hierarchical cluster analysis of genes associated with immune response in the bacteria clearance process; (F) Complex structure of R₂F₄R₂ and calcium signaling molecules; (G) 3D mode of action between R₂F₄R₂ and calcium signaling molecules. Yellow dashed line: hydrogen bond; Red dotted line: salt bridge.

Gene Set Enrichment Analysis (GSEA) of RNA-seq data based on the KEGG database revealed that gene sets related to lysosomes were significantly enriched in R₂F₄R₂-treated macrophages. Lysosomes merge with phagosomes to form phagolysosomes, which are characterized by an acidic, hydrolytic lumen that effectively destroys pathogens. Enhanced lysosomal function and bactericidal capacity may manage successive rounds of phagocytosis [48]. Moreover, pathways related to acute and chronic inflammatory responses, such as the IL-17 signaling pathway and viral protein interactions with cytokine and its receptor pathways, were notably suppressed (Figs. 6C and S17).

The differentially expressed genes (DEGs, P < 0.05, |log₂ (fold change)| > 1.5) were analyzed using KEGG, with the top 20 pathways regulated by R₂F₄R₂ treatment, mainly involving calcium and Rap1 signaling activation, represented as bubble plots (Fig. 6D). In this study, the interaction modes between R₂F₄R₂ and calcium signaling molecules were simulated using molecular docking. As illustrated in Fig. 6F and 6G, hydrogen bonds formed between R₂F₄R₂ and the calcium signaling molecule at SER20-ASP30, GLU23-GLY173, THR172 and GLN28-SER330, and a salt bridge was formed between PHE29-ARG93. Accumulating evidence suggests that calcium signaling, a ubiquitous second messenger, is crucial for mediating various macrophage functions, including migration, survival, phagocytosis, and anti-inflammatory response during bacterial infections [49,50]. Previous studies have also demonstrated that calcium influx can facilitate the small GTPase Rap1 activation following TLR stimulation, thereby promoting macrophage activation through a feedforward mechanism [51]. Furthermore, complement cascade pathways are involved in macrophage polarization [52], phagocytosis [53], lysosome and reactive oxygen species (ROS) [54], directly affecting the bactericidal ability of macrophages, and were significantly enriched [55].

In the immune response, macrophages perform phagocytosis and inflammation regulation by migrating to the site of pathogenic infection, inflammation or injury [46]. Hierarchical cluster analysis of immune response-related genes in the bacterial clearance process demonstrated that genes positively regulating cell migration (Ccn4, Madcam1, Pdgfrb and Ptpru) were significantly upregulated in the R₂F₄R₂ group compared with the control group (Fig. 6E). We have conducted qPCR analysis to verify the expression changes of several critical genes identified in the RNA-seq experiment (Fig. S18). Opsonin binding plays a crucial role in augmenting the antibacterial activity of macrophages by facilitating bacterial recognition and phagocytosis, leading to the release of antimicrobial molecules, such as ROS [56]. Moreover, opsonized bacteria are readily internalized by macrophages, leading to their degradation in phagolysosomes [57]. Notably, in the R₂F₄R₂ group, the genes encoding complement component C1q receptor (Cd93) and complement factor H (Cfh), which negatively regulate opsonin binding, were markedly downregulated by 1.95- and 2.48-fold, respectively. Genes negatively regulating the response to ROS (Met and Ngb) were markedly downregulated by 4.08- and 1.78-fold, respectively. Surfactant proteins are critical for defending the host against infectious microorganisms and modulating the innate immune response to various pathogen-associated molecular patterns. These proteins act as pattern recognition receptors by selectively binding to bacterial, viral, and fungal surfaces, thereby enhancing phagocytosis and promoting intracellular killing [58]. In this study, the surfactant protein family, including SFTPB, SFTPC and SFTPD, the key gene Calml4 in the C-type lectin receptor signaling pathway, and Unc13d, were significantly upregulated in RAW 264.7 macrophages exposed to E. coli pretreated with R₂F₄R₂.

Altogether, these results demonstrate that R₂F₄R₂, based on the combined strategy of bacteria-entrapping peptides with immunomodulatory peptides, not only captures bacteria efficiently but also actively activates macrophage immune functions, owing to the immunoregulatory function of Rv2626c and bacterial clusters entrapped by R₂F₄R₂. Our study addresses a significant gap in the field, as previous studies have primarily focused on the trapping ability of such peptides while paying little attention to their potential role in regulating the innate immune system of the host [39,59]. The R₂F₄R₂ immunomodulatory capacity provides a secondary defense, allowing host cells to eliminate the bacteria that escape aggregation [12,60]. This dual-functional mechanism not only improves antibacterial efficacy but also lowers bacterial resistance risk. To ensure long-term efficacy, future efforts should explore optimized dosing regimens and combination therapies with other antimicrobial agents to mitigate potential adaptive evolution.

3.5. In vivo assessment of the therapeutic potential and biocompatibility of R₂F₄R₂

Inspired by the in vitro evaluation of bacterial aggregation ability of R₂F₄R₂ and immune-modulatory effects and its excellent biocompatibility demonstrated by cell viability (Fig. S19) and hemolysis assays (Fig. S20), the in vivo efficacy and therapeutic potential of the peptide were further investigated. Given the notable prevalence of E. coli in soft-tissue and intestinal infections [61], cutaneous wound and intestinal infection models were established to assess the effectiveness of the peptide.

In mice, excisional wound healing is primarily driven by wound contraction, whereas in humans, epithelialization plays a dominant role in wound closure [62]. To better model human wound-healing mechanisms and reduce contraction, full-thickness wounds created with an 8 mm biopsy punch were stabilized using a splint and adhesive tape (3 M, USA) (Fig. S21). Twenty-four mice were randomly divided into two groups: one group was inoculated with E. coli alone, whereas the other group received E. coli pre-incubated with R₂F₄R₂ (Fig. 7A). Forty-eight hours post-injection, R₂F₄R₂-incubated E. coli group exhibited no signs of infection, whereas the E. coli-only group exhibited clear abscess formation. These results demonstrate that R₂F₄R₂ effectively entangles and agglutinates E. coli, thereby reducing bacterial invasion and alleviating E. coli-induced skin lesions. Further pathological analysis of wound samples was conducted to evaluate the healing process, which includes four overlapping phases: hemostasis, inflammation, proliferation, and remodeling with scar tissue formation. H&E staining revealed that the R₂F₄R₂-treated skin tissue maintained relatively normal morphology, with a great number of blood vessels (black circle) and hair follicles (red circle) (Figs. 7B). Conversely, the skin tissue of the E. coli group exhibited a significant presence of inflammatory cells (black arrows). In normal skin, cell proliferation and apoptosis are tightly regulated [63]. To assess these processes, KI67 and TUNEL staining were performed to highlight the cellular activities essential for wound healing. TUNEL assay images revealed a large number of apoptotic cells in the E. coli group, whereas no apoptotic cells were observed in the R₂F₄R₂-incubated E. coli group (Fig. 7C). There were more KI67⁺fluorescent cells in the R₂F₄R₂ group than in the E. coli group, indicating that R₂F₄R₂ could inhibit E. coli-induced skin apoptosis (Fig. S23). Masson staining was conducted to detect collagen deposition, and the results indicated that wounds treated with R₂F₄R₂ exhibited much higher levels of collagen deposition (black square) than those untreated with R₂F₄R₂ (Fig. 7D). Notably, a clear epidermal structure (black dashed line) was observed in the wounds treated with R₂F₄R₂, which was not observed in the E. coli group.

Fig. 7.

Evaluation of the inhibition effect of R₂F₄R₂ on E. coli invasion in vivo. (A) Schematic diagram of the construction of infected wound. After incubating E. coli with R₂F₄R₂ for 2 h or not, the culture system was agitated and the 50 µl resuspension was smeared on the incisional skin wounds. The wound tissues after 48 h were harvested for histological analyses to evaluate the infection; (B) H&E-stained images of the skin incisions (scale bar: 50 µm). The red arrow: epidermis. the blue arrows: ermis; black dashed line: the boundary of epidermis and dermis. the black arrow: the infiltration of inflammatory cells; red circle: hair follicles; black circle: new blood vessels; (C) Images of the wound tissues labeled with TUNEL (green) to evaluate skin cell apoptosis (scale bar: 100 µm); (D) Masson staining images of wound tissues to evaluate the remodeling stage (scale bar: 100 µm). Black square: collagen deposition; black dashed line: the boundary of epidermis and dermis.

A cutaneous wound model was established in thirty-six mice divided into three groups (n = 12). The groups received (1) E. coli inoculation alone (control), (2) E. coli inoculation followed by R₂F₄R₂ treatment, or (3) saline solution only (blank control) in Fig. 8A. The in vivo topical dose was guided by the effective in vitro concentration of 32 µM, following a standard rationale in early-stage AMP researches for wound healing [64]. Wound images indicated that mice treated with R₂F₄R₂ exhibited a significantly accelerated healing process compared to other groups. Furthermore, the wound sites in the R₂F₄R₂-treated group exhibited greater contraction than those in the E. coli-infected group (Fig. 8B). To visualize the bacterial capture and agglutination effects of R₂F₄R₂ on bacteria in vivo, the wound surface was collected for bio-SEM (Fig. 8C). E. coli in the R₂F₄R₂ group was entangled and aggregated by extensive fibrous network-like structures. Further magnification revealed a meshwork originating from the bacterial surface that effectively ensnared the E. coli (Figs. 8C and S24). These observations, reminiscent of naturally occurring self-assembled fibrous networks involved in HD6-mediated bacterial entrapment, suggest that R₂F₄R₂ can efficiently capture bacteria in vivo. H&E staining revealed that on Day 3, the skin tissues from the E. coli group displayed acute inflammatory responses (Figs. 8D and S25), whereas those treated with R₂F₄R₂ exhibited significant inflammatory infiltration. This reduction highlights the anti-inflammatory effects enhanced by bacterial entrapment and the aggregating properties of R₂F₄R₂. By Day 6, tissues infected with E. coli displayed a sustained hyperinflammatory state, characterized by the inflammatory cell infiltration (black arrow), incomplete epithelialization (black dashed line), and vasodilation and congestion (black circle). Conversely, tissues treated with R₂F₄R₂ exhibited complete epithelialization (black dashed line) and a significant reduction in inflammatory cell infiltration. Additionally, KI67 expression, a well-established marker of cell proliferation, was markedly upregulated in R₂F₄R₂-treated wounds during the proliferative phase (Fig. 8E). In the early wound healing stage, the epidermis and dermis in the E. coli group exhibited significant disruption, accompanied by a reduction in collagen fibers. During the recovery period from Day 3 to 6, the R₂F₄R₂-treated tissues demonstrated a more organized arrangement of collagen fibers (black square), characterized by robust fibroplasia and hair follicles (black circle), resembling the remodeling phase of normal skin (Fig. 8F). Fig. 8E illustrates that CD86, a marker associated with proinflammatory macrophages, was predominantly expressed in untreated and E. coli-treated wounds. Conversely, CD206, a marker of anti-inflammatory macrophages, was predominantly observed in R₂F₄R₂-treated wounds. These findings suggest that R₂F₄R₂ promotes a shift toward a pro-healing immune microenvironment, facilitating tissue repair and regeneration. Notably, immunofluorescence staining revealed that IL-6 (M1 marker) expression decreased, whereas IL-10 (M2 marker) expression increased in macrophages treated with R₂F₄R₂ (Fig. 8E). These results collectively indicate that R₂F₄R₂ can re-rouse the macrophage immune response, further eliminating the pathogen.

Fig. 8.

Experimental design for cutaneous wound model and treatment. (A) Scheme of R₂F₄R₂ for treating cutaneous wound; (B) Schematic images of wound contraction during 6 d; (C) SEM images of E. coli entrapped by R₂F₄R₂in vivo. Scale bar: 5 µm and 2 µm (magnification); (D) Histopathological H&E staining of skin tissues on Day 3 and 6. The red arrow: epidermis. Blue arrows: ermis; black dashed line: the boundary of epidermis and dermis. Black arrow: the infiltration of inflammatory cells; red circle: hair follicles; black circle: the vasodilation and congestion. Scale bar: 100 µm and 50 µm (magnification); (E) Immunofluorescence analysis of KI67, M1 (CD86-F4/80 and IL-6) and M2 (CD206-F4/80 and IL-10) macrophages markers. Scale bar: 100 µm; (F) MASSON staining of skin tissues on Day 3 and 6. Black square: collagen deposition; black dashed line: the boundary of epidermis and dermis; black circle: hair follicles. Scale bar: 200 µm and 100 µm (magnification).

The high resistance of peptides to proteases ensures their potential application as oral drugs. Before investigating whether R₂F₄R₂, an oral drug, could mitigate E. coli-challenged enteritis, the cumulative toxicity of R₂F₄R₂ was determined in vivo. Previous studies have indicated that nanoparticle-based peptides with similar bacterial-aggregating mechanisms exhibit efficacy within the 10–30 mg/kg range [65]. We therefore first evaluated the safety of R₂F₄R₂ within this concentration range. Eight mice in each group were orally administered with saline, 10, 20 and 30 mg/kg R₂F₄R₂ for 3 d at 24 h intervals. As illustrated in Figs. S26–S28, there were no notable differences in body weight, organ weight ratios, or serum markers, indicating that R₂F₄R₂ did not cause hepatotoxicity or nephrotoxicity. H&E staining revealed no significant tissue abnormalities across the treatment groups (Fig. S29). Collectively, these findings suggest that R₂F₄R₂ exhibits good biocompatibility in vivo, establishing a foundation for dosage determination in preclinical testing.

ETEC K88, a major pathogen linked to diarrhea in both human infants and young animals. To assess the potential of R₂F₄R₂ in the treatment of enteritis, we conducted a preliminary dose experiment using 10, 20 and 30 mg/kg of R₂F₄R₂ in an ETEC K88-infected enteritis model, according to the procedure (Fig. 9A). The results showed that while all tested doses alleviated body weight loss by Day 7, the 30 mg/kg dose exhibited obvious protective effect (Fig. S30). Histopathological examination of intestinal H&E-stained sections further supported this finding. ETEC infection induced severe inflammatory infiltration, lymphoid hyperplasia (red arrow), and a reduction in goblet cells, whereas all R₂F₄R₂ treatments reduced inflammatory cell infiltration (black arrow), with the 30 mg/kg dose again showing the most significant improvement (Fig. S31). Based on these preliminary efficacy and safety data, we selected 30 mg/kg as the optimal dose for the formal therapeutic experiment. Compared with the ETEC group, the R₂F₄R₂ group exhibited improvement in body weights, relative liver and lung weight changes (Fig. 9B and 9C). Importantly, SEM analysis revealed that ETEC incubated with R₂F₄R₂ was enveloped and immobilized (Fig. 9D), mirroring the bacterial entrapment observed with natural HD6 [4]. ETEC aggregation by the intertwined R₂F₄R₂ likely contributes to bacterial clumping and prevents invasion. To further explore how R₂F₄R₂ protects mice from intestinal injury, we conducted a comprehensive assessment of colonic pathology. H&E staining revealed that treatment with R₂F₄R₂ reduced the number of inflammatory cells and improved shedding of the epithelial cells, villus necrosis, and edema triggered by the ETEC challenge (Figs. 9E and S32). Furthermore, periodic acid-Schiff (PAS) staining indicated that R₂F₄R₂ treatment preserved the integrity of the intestinal mucosal barrier in ETEC-challenged mice by enhancing goblet cell numbers (Fig. 9E). Immunofluorescence analysis (Fig. 9F) revealed that the R₂F₄R₂-treated group exhibited decreased CD86 and NF-κB expression relative to the infected group, alongside elevated CD206 and KI67 levels, indicating R₂F₄R₂ treatment diminished inflammation-related signaling pathway activation and potentially promoted macrophage differentiation toward the anti-inflammatory M2 phenotype to alleviate inflammatory response.

Fig. 9.

Experimental design for intestinal infection model and treatment. (A) Scheme of R₂F₄R₂ for treating intestinal infection; (B-C) Changes in the body weight curves and liver, kidney, lung and spleen weights of mice during experimental period (n = 8). Data are presented as mean ± SD. The differences between the groups were determined using one-way ANOVA followed by Tukey’s post hoc analysis. *P < 0.05 and ***P < 0.001; (D) SEM images of ETEC entrapped by R₂F₄R₂in vivo. Scale bar: 5 µm and 2 µm (magnification); (E) Histopathological H&E staining and PAS staining of colon tissues. Scale bar: 100 µm; (F) Immunofluorescence analysis of KI67, M1 (CD86) and M2 (CD206) macrophages markers, and inflammation-related signaling pathway NF-κB. Scale bar: 100 µm; (G) α-Diversity was analyzed by the Chao1 and Shannon indexes. The Kruskal-Wallis test was applied for statistical analysis; (H) PCoA analysis of the cecal microbiota structure measured by Jaccard distances; (I) Average relative abundances of taxa at the phylum and family level; (J) Hierarchically heatmap clustered with top 20 family. The abundance difference is shown as a color key, n = 5 per group.

Subsequently, 16S rRNA sequencing was performed to compare the intestinal microbiota composition and abundance between the infection and treatment groups. Mice treated with R₂F₄R₂ exhibited higher α-diversity (Chao1 and Shannon indices) and β-diversity based on Jaccard distances than infected mice, indicating that R₂F₄R₂ facilitated the disrupted microflora restoration (Fig. 9G and 9H). Differences in community composition and species at the phylum and family levels were also analyzed. The results in Fig. 9I suggest that R₂F₄R₂ enhanced the probiotics abundance, such as Muribaculaceae, Bacteroidaceae, and Coprobacillaceae in the Bacteroidota phylum, to maintain and protect the function of the intestinal mucus layer barrier [66] and decreased the harmful bacteria abundance in the Proteobacteria phylum. Besides, R₂F₄R₂ enhanced the abundance of dietary fiber decomposers, Ruminococcaceae, in Firmicutes, to produce short-chain fatty acids involved in regulating energy metabolism and immune function, and to inhibit the harmful bacteria growth (Fig. 9J) [67]. Meanwhile, it decreased harmful bacteria that exacerbate intestinal inflammation, such as Enterobacteriaceae in the phylum Proteobacteria [68]. Together, R₂F₄R₂ regulates bacteria-challenged intestinal homeostasis in two main ways: (1) Increasing the abundance of probiotics to maintain the intestinal barrier and enhance immune function and (2) decreasing the abundance of pathogen-related intestinal inflammation.

3.6. Limitations

This study has certain limitations that warrant consideration. Firstly, the absence of negative control peptides or vehicle controls could constitute a critical limitation, as it introduces ambiguity in distinguishing whether the observed effects originate from the specific function of R₂F₄R₂ or from experimental system biases. However, this limitation does not materially affect the main efficacy conclusion of the R₂F₄R₂. Secondly, although there were no obvious signs of acute systemic toxicity caused by R₂F₄R₂ as an antibiotic at these doses (30 mg/kg), a full long term toxicological evaluation still needs to be performed before considering any possible application.

4. Conclusion

In the present study, we designed a highly simplified self-assembled motif consisting of only three amino acids (Arg, Phe and Pro) in the structure, following the pattern of conventional surfactant-like amphiphiles, bola-amphiphiles (central and end), and alternating hydrophilic and hydrophobic residues. The arrangement of these peptides ensured a comprehensive investigation of hydrophobic effects and π-π stacking on bacterial aggregation activity. Among these peptides, R₂F₄R₂ exhibited potential resistance to SIF and SGF, thereby enhancing its oral bioavailability. The proteolysis-resistant ability of R₂F₄R₂ guarantees its effective bacterial-capturing capability in vivo. As expected, the excellent peptide R₂F₄R₂ can boost antibacterial activity of macrophages by modulating calcium/Rap1 signaling pathway and antibacterial immune response-related gene expression, leading to macrophage migration to the infectious site, the phagocytic uptake of E. coli, and the switch of macrophages to an anti-inflammatory M2 phenotype, due to the immunoregulatory function of Rv2626c and bacterial clusters entrapped by R₂F₄R₂. This synergistic effect contributes to the outstanding therapeutic efficacy of R₂F₄R₂ against E. coli-induced cutaneous infections and intestinal inflammation. Collectively, the design strategies and technical insights utilized in this study have the potential to expand the scope of AMPs design.

CRediT authorship contribution statement

Nan Gao: Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Yikuan Bian: Investigation, Data curation, Conceptualization. Shasha Wang: Methodology, Investigation, Data curation, Conceptualization. Pengfei Bai: Investigation, Data curation, Conceptualization. Chunyang Fang: Data curation, Conceptualization. Jiaqi Sun: Data curation, Conceptualization. Yihan Li: Data curation, Conceptualization. Na Dong: Writing – review & editing, Validation, Conceptualization. Anshan Shan: Writing – review & editing, Funding acquisition, Formal analysis, Data curation, Conceptualization. Jiajun Wang: Writing – review & editing, Validation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Conflicts of interest

The authors declare that there is no conflicts of interest.