1. Introduction
Antimicrobial peptides (AMPs) have unique antibacterial activity and multifaceted mechanisms of action, and are considered ideal alternatives to traditional antibiotics, especially at present, as public health safety is challenged by drug-resistant pathogens [
1], [
2]. The discovery and structural design of novel AMPs have been the focus of researchers. For example, by exploiting physical and structural vulnerabilities within the pathogen’s cellular envelope, Simonson et al. [
3] identified a peptide that undergoes instructed tryptophan-zippered assembly within the mycolic acid-rich outer membrane of
Mycobacterium tuberculosis to specifically kill the pathogen. Narayana et al. [
4] developed two peptides that can eliminate resistant pathogens and preformed biofilms by combining database-guided discovery with structure-based design. Furthermore, with the leapfrog development of nanotechnology, a wide variety of nanometals have achieved unprecedented success when applied to antibacterial infection, showing promise for antibiotic-free strategies [
5], [
6]. Crucially, certain nanometals—including silver (Au) [
7], gold (Ag) [
8], and copper (Cu) [
9]—have been revealed to possess multifaceted bactericidal mechanisms involving non-specific targets, similar to those of AMPs. However, novel antibacterial agents seamlessly reconstituted from AMPs and nanometals have been lacking.
As an example of how AMPs and nanomaterials can function in tandem, some hosts can limit the availability of zinc, manganese, and iron in regions containing bacteria by generating high-affinity metal-binding peptides [
10]. Termed
nutritional immunity, this inhibitory mechanism starves pathogens of these essential micronutrients. Alternatively, some hosts can leverage the inherent toxicity of nickel or copper and deliberately pool these metals into areas containing pathogenic bacteria [
11]. The amino-terminal copper-nickel binding motif (ATCUN), which has been noted in AMPs due to its importance in immune regulation, is composed of the sequence H
2N-XXH and is found in the N-terminus; the XX element of the motif can be any amino acid other than proline (Pro) [
12]. The Cu group can be rapidly bonded to the motif in a distorted square planar geometry through the backbone of the two deprotonated N-amide atoms and the imidazole ring of the histidine (His) [
13]. Subsequently, the reconstructed complex utilizes a Fenton-like mechanism involving Cu
1+/Cu
2+ redox pairs and hydrogen peroxide (H
2O
2) to form reactive oxygen species (ROS) that can destroy any molecule surrounding the generator [
11], [
14]. Noticeably, the interaction between the Cu sites and H
2O
2 is an extremely important link in determining the efficiency of ROS generation in this reaction [
15].
Inspired by such natural host defense mechanisms, we constructed an ATCUN-motif AMP complex (Cu@G-AMPs) incorporating single-atom Cu catalysts in this proof-of-concept study. Functionally, Cu@G-AMPs can be localized to the cell membrane or interfere with internal targets (e.g., DNA, RNA, and proteins), generating and transmitting its deadly ROS cargo
in situ when approaching pathogens (
Fig. 1). Finally, Cu@G-AMPs, which inherits the immunomodulatory properties of its AMPs, successfully performs the functions of pulling edge closure, stabilizing granulation tissue, promoting collagen fiber proliferation, alleviating inflammation, and promoting neovascularization in wound areas infected by drug-resistant bacteria (
Fig. 1). The emergence of Cu@G-AMPs provides a new perspective for the field of public health security to address the dilemma of drug-resistant bacteria having developed precise regulatory mechanisms to inactivate antibiotics.
2. Materials and methods
2.1. Materials
Piscidins-3 (FIHHIFRGIVHAGRSIGRFLTG) was manufactured using solid-phase peptide synthesis by GL Biochem Co., Ltd. (China). N-Hydroxy succinimide (NHS), guanine, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC), 3,3′,5,5′-tetramethylbenzidine (TMB), and vancomycin (VAN) were purchased from Macklin Biochemical Co., Ltd. (China). Propidium iodide (PI) and ROS assay kits were obtained from Beyotime Biotechnology Co., Ltd. (China). Phosphate-buffered saline (PBS; pH = 7.4) and SYTO9 were ordered from Thermo Fisher Scientific (USA). N,N-Dimethylformamide (DMF), Methylthiazolyldiphenyl-tetrazolium bromide (MTT), H2O2, glutaric dialdehyde, and dimethyl sulfoxide (DMSO) were purchased from Sinopharm Chemical Reagent Co., Ltd. (China). Unless otherwise indicated, the methicillin-resistant Staphylococcus aureus (MRSA; 43300) applied in this study was acquired from ATCC (USA). Mueller-Hinton broth (MHB) and Luria-Bertani (LB) culture were ordered from Hope Biotechnology Co., Ltd. (China).
2.2. Preparation of Cu@G
Single-atom metals were anchored to a sheet substrate derived from guanine (G) by means of coordination pyrolysis [
16]. In general, 2.5 mL of Cu(NO
3)
2 (5.3 mmol∙L
−1) was added dropwise to 60 mL of guanine solution (441 mmol∙L
−1) and stirred at 80 °C for 12 h. Afterward, the dried and ground reaction samples were placed in a tubular furnace at 800 °C for 2 h. The reaction process occurred under a nitrogen (N
2) atmosphere at a heating rate of 5 °C∙min
−1. Lastly, the new reaction products were reduced for 2 h in a hydrogen (H
2) atmosphere at 800 °C.
2.3. Preparation of Cu@G-AMPs
The Cu@G were modified with ATCUN-motif AMPs via an amide reaction [
17]. To be specific, EDC (15 mL, 0.4 mmol∙L
−1) and NHS (9 mL, 0.4 mmol∙L
−1) were first added sequentially to the suspension obtained from 100 mg of Cu@G dissolved in 200 mL of DMF and stirred for 1 h. Then, 30 mg of AMPs (piscidins-3) was added to the above suspension and stirred for 2 d. Finally, the new suspensions were placed in DMF and deionized water for dialysis.
2.4. Antibacterial experiment
The antibacterial activity profile of the nanomaterials
in vitro was evaluated by conducting plate bacteria-killing experiments [
18]. In detail, the concentration of the bacterial suspension (MRSA during the logarithmic growth phase) was adjusted to 1 × 10
8 colony forming units (CFU)∙mL
−1 by means of PBS (pH = 7.4, 6.7, and 5.5). PBS, H
2O
2 (final concentration, 1 mmol∙L
−1), and Cu@G-AMPs + H
2O
2 (final concentration, 128 μg∙mL
−1 + 1 mmol∙L
−1) were added to centrifuge tubes containing 100 μL of diluted bacterial suspension and incubated at 37 °C for 30 min. Finally, 100 μL of the above reaction solution diluted 1000-fold was evenly applied to a plate containing LB medium and incubated at 37 °C for 18 h. Each trial was repeated at least three times.
2.5. Live/dead bacterial fluorescent imaging
The concentration of the bacterial suspension (MRSA during the logarithmic growth phase) was adjusted to 1 × 10
8 CFU∙mL
−1 with PBS (pH = 6.7). PBS, H
2O
2 (final concentration, 1 mmol∙L
−1), and Cu@G-AMPs + H
2O
2 (final concentration, 128 μg∙mL
−1 + 1 mmol∙L
−1) were added to centrifuge tubes containing 100 μL of diluted bacterial suspension and incubated at 37 °C for 1 h. Next, bacterial cells washed three times with PBS (pH = 7.4) were incubated for 15 min using PI and SYTO9 dye [
18]. Finally, these bacterial cells were imaged using confocal laser scanning microscopy (A1; Nikon, Japan).
2.6. Hemolysis assay
Blood cells isolated from fresh mouse blood (Institute of Cancer Research, male, 30-35 g) were treated with different concentrations of nanomaterials (256, 128, 64, 32, and 16 μg∙mL
−1) at 37 °C for 1 h. Here, deionized water and PBS (pH = 7.4) were used as positive and negative controls, respectively. After incubation, the optical density (OD)
576 value of each tube sample was measured and the hemolysis rate was calculated according to the formula below [
19]. OD
576 sample was the optical density of the sample. OD
576 blank was the optical density of the blank control. OD
576 water was the optical density of the positive control.
2.7. Cell viability assay
MC3T3-E1 cells were cultured in plates containing alpha minimum essential medium (α-MEM) supplemented with 1% penicillin and 10% fetal bovine serum. The culture was maintained in a humidified incubator at 37 °C under an atmosphere of 5% CO
2. The cell viability of the MC3T3-E1 cells was measured via MTT assay [
20], [
21]. In detail, the cells were seeded at a density of 1 × 10
5 cells and then incubated to form a confluent monolayer. The cell culture medium was removed and added to 96-well plates filled with different drug concentrations (256, 128, 64, 32, 16, and 8 μg∙mL
−1). After 24 h of incubation, 20 μL of MTT solution was added to each well. Subsequently, the mixture solution was displaced with DMSO to dissolve the formazan crystals. Finally, the OD
595 value of each tube sample was measured.
2.8. Enzyme-like activity test
Different concentrations of nanoparticles (256, 128, 64, and 32 μg∙mL
−1) were completely dissolved in acetate buffer (pH = 7.4, 6.7, 5.5, and 4.7). Then, 10 μL of TMB (33 mmol∙L
−1) and 10 μL of H
2O
2 (10 mmol∙L
−1) were added to the buffer solution and stirred evenly. Finally, the OD
652 values of the prepared samples were monitored at different times (once per minute, total 30 min) [
22]. Each trial was implemented in triplicate, and the reported results are the averages of independent trials.
2.9. In situ detection of ROS production
A kit was used to check the ROS content of the bacterial cells treated with nanomaterials [
19], [
23]. In brief, the concentration of the bacterial suspension (MRSA during the logarithmic growth phase) was adjusted to 1 × 10
8 CFU∙mL
−1 with PBS (pH = 6.7). Fluorescent probes (DCFH-DA) were added to the bacterial suspension and incubated for 20 min. Subsequently, the bacterial cells were washed three times with PBS (pH = 6.7) to remove fluorescent probes that had not entered the cells. Immediately, different concentrations of nanoparticles (128, 64, and 32 μg∙mL
−1) and H
2O
2 (final concentration, 1 mmol∙L
−1) were added to the probe-loaded bacterial cells and incubated for 1 h. AMPs (piscidins-3; final concentration, 128 μg∙mL
−1) were used here as control drugs. Finally, the fluorescence intensity of these samples (525-nm emission wavelength and 488-nm excitation wavelength) was detected.
2.10. Proteomics experiment
The concentration of the bacterial suspension (MRSA during the logarithmic growth phase) was adjusted to 1 × 108 CFU∙mL−1 with PBS (pH = 6.7). Then, PBS (pH = 6.7), AMPs (piscidins-3; final concentration, 128 μg∙mL−1), and Cu@G-AMPs + H2O2 (final concentration, 128 μg∙mL−1 + 1 CFU∙mL−1) were added to centrifuge tubes containing 100 μL of diluted bacterial suspension and incubated at 37 °C for 1 h. The bacterial precipitate purified by centrifugation was washed with PBS three times and stored at −80 °C.
Bacterial cells were lysed in the presence of 400 μL of lysis buffer supplemented with protease inhibitors, 200 μg∙mL
−1 DNAse I, 200 μg∙mL
−1 lysostaphin, and PBS for 20 min at 37 °C, and sonicated ten times with 30-s cycles [
23], [
24], [
25]. The cell protein extract was clarified by centrifugation at 15 000
g for 10 min. Samples were mixed with Laemmli sample buffer (SB) and analyzed by means of sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis followed by Western blot or Coomassie blue staining. Furthermore, label-free quantitative mass spectrometry was done at Iproteome Biotech Co., Ltd. (China). Detailed protein extraction, peptide isolation, and bioinformatics are described in Appendix A.
2.11. Parallel reaction monitoring (PRM) verification
PRM is a proteomic result-verification technique with high sensitivity, resolution, and strong specificity. The detailed steps used here for bacterial cell processing and protein extraction were similar to those used in proteomics experiments [
25]. Hence, the following content mainly describes the identification process of peptide fragments. Specifically, peptide fragments identified in the data-dependent acquisition (DDA) library were separated using high-performance liquid chromatography (HPLC) with nanoliter flow rates. Peptide fragment information that met the analysis conditions was normalized (Xcalibur, Thermo Fisher Scientific, USA) and analyzed (Skyline, University of Washington, USA). This and the related proteomics experiments were particularly supported by the Human Phenome Institute (Fudan University, China).
2.12. MRSA-infected skin model
All the studies on these animals were performed following animal ethical standards from the Animal Ethics Committee in Fuzhou University, China. The mice (ICR, male, 30-35 g) were adaptively fed in cages at a constant temperature (26 °C) for 7 d. The mice were allowed to drink and eat freely throughout the experiment. First, 1% sodium pentobarbital was injected into the abdominal cavity of each mouse at a dose of 50 mg∙kg
−1. The hair was removed from the same spot on the dorsal skin of each anesthetized mouse. A circular wound with a diameter of 6 mm was constructed on the disinfected skin of each mouse. Then, these wound areas were dripped with 20 μL of MRSA suspension (during the logarithmic growth phase, 1 × 10
8 CFU∙mL
−1). Notably, once the bacterial suspension had completely immersed the skin wound, these areas were wrapped and fixed with a clear dressing (Tegaderm, 3M, USA) [
26], [
27], [
28]. Model mice with bacterial-infected skin were randomly divided into six groups and given different treatment regimens. The detailed respective treatment regimen for each model group was as follows: ① 20 μL of PBS (pH = 6.7); ② 20 μL of H
2O
2 (1 mmol∙L
−1); ③ 20 μL of piscidins-3 (128 μg∙mL
−1); ④ 10 μL of Cu@G (128 μg∙mL
−1) + 10 μL of H
2O
2 (1 mmol∙L
−1); ⑤ 10 μL of Cu@G-AMPs (128 μg∙mL
−1) + 10 μL of H
2O
2 (1 mmol∙L
−1); ⑥ 20 μL of VAN (16 μg∙mL
−1). Finally, the wound area and body weight of the groups of treated mice were recorded daily, with the entire course of treatment lasting for 7 d. Blood and skin tissue from the mice were collected and analyzed on days 3 and 7.
2.13. Closure evaluation and colony density testing
Changes in the size of the mouse skin wounds were determined and monitored through software (Image J). Meanwhile, wound closure was measured as a proportion of the initial wound size occupied by the latest wound size. After 1 day, the treated mouse wounds were thoroughly wiped with cotton swabs and soaked in normal saline (1 mL). Then, the diluted saline (100 μL) was transferred to the medium (MHB, 10 mL) for incubation (37 °C). Finally, the OD
600 value of the medium was determined after completing the incubation (18 h) [
29].
2.14. Hematoxylin-eosin (H&E), Masson, and toluidine blue staining
The wound tissue on day 3 and day 7 was carefully observed by making slices [
26], [
27], [
28]. In brief, the wound tissues were collected with scissors, washed with saline, dried with gauze, and fixed with 4% paraformaldehyde. Next, sections were stained with H&E, Masson, and toluidine blue, respectively. Lastly, the dehydrated sections were encapsulated with neutral gum. All the stained sections were examined by microscope and collected. At least five images of each sample were made available for statistics and analysis.
2.15. Immune-related experiment
The immune status of the mice treated in the groups was detected by means of immunofluorescence and enzyme-linked immunosorbent assay (ELISA) [
26], [
27], [
28]. The immunofluorescent sections involving platelet endothelial cell adhesion molecule-1 (CD31) and goat anti-rabbit immunoglobulin G (IgG) H&L were utilized to assess the vascular status of the wound tissue; the immunofluorescent sections involving proliferating cell nuclear antigen (PCNA) and Cy3-conjugated goat anti-mouse IgG were also applied to examine the vascular condition of the wound tissue. The tumor necrosis factor-α (TNF-α), interleukin-1 beta (IL-1β), and interleukin-10 (IL-10) levels in the blood of the groups of treated mice (at days 3 and 7) were accurately determined by means of ELISA.
3. Results and discussion
3.1. Design, preparation, and characterization
The main process of preparing the new antibacterial agent is illustrated in
Fig. 2(a). In brief, we first obtained single Cu atoms confined to carbon-based nanomaterials doped with heteroatoms (Cu@G) by means of coordination pyrolysis reactions from a mixture of guanine and metal salts. Subsequently, we created antibiotic-free nanoreactors (Cu@G-AMPs) with bacteriostatic activity involving non-specific targets by modifying ATCUN-motif AMPs onto carbon-based scaffolds through amide reactions. Piscidins-3, which has the advantages of a low hemolysis rate to blood cells, strong binding to intracellular DNA, and severe disruption of bacterial biofilms, is regarded as a promising candidate antimicrobial agent [
31]. Furthermore, single-atom catalysts possess the advantages of accurate active-site identification, tunable activity, and high metal utilization [
32]; thus, the addition of copper provided a valuable opportunity to replicate or even amplify the redox properties of AMPs. That is, single Cu atoms, which differ from traditional catalysts from the sub-nanoscale to microscale, bridge the gap between homogeneous catalysis and heterogeneous catalysis in Fenton-like reactions [
33]. In practical terms, single-atom metals as “catalysts” are ubiquitous in nature and play a key role in vital activities, such as molybdenum in nitrogenase, iron in heme, and magnesium in chlorophylls [
11].
Achieving single-atom Cu dispersion was a key technical obstacle to the preparation of Cu@G-AMPs. This involved the generation of thermodynamic structural motifs or un/metastable phases, which require the selection of suitable substrate materials to bind to metal precursors in order to prevent the metal atoms from aggregation, agglomeration, and nanocrystal/cluster growth [
34]. The direct application of carbon-based materials with a high surface area as the substrate to incorporate the metal precursors thus became an ideal solution for the scalable preparation of single Cu atoms [
35]. Guanine, which is a nucleic acid base component consisting of a pyrimidine-imidazole ring system with conjugated double bonds [
16], was designated here as the raw material for the synthesis of carbon skeletons. The substrate material formed by guanine doped with abundant heteroatoms is conducive to the stable dispersion of single Cu atoms [
16]. Certainly, guanine’s inherent biosafety properties were an important reason why it was adopted here.
As expected, the Cu@G-AMPs presented a high-surface-area sheet morphology similar to that of graphene-like materials [
35], which facilitated the full dispersion of the single Cu atoms (
Figs. 2(b) and (c)). Furthermore, the similar appearance of Cu@G and Cu@G-AMPs suggested that the modification process of the AMPs did not change the morphology of the carbon substrate materials (Figs. S1 and S2 in Appendix A). Crucially, this surface structure enhances the ability of Cu@G-AMPs and Cu@G to effectively trap pathogens [
22]. To directly observe the structure of the Cu at the atomic level, we applied aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) to examine the Cu@G-AMPs in detail. In HAADF-STEM, the Cu in Cu@G-AMPs was expected to be observed as independent highlights on a carbon substrate doped with heteroatoms, due to the greater mass of Cu in comparison with those of C, N, and O [
36]. In the 2-nm field of view, we unsurprisingly observed highly dispersed single Cu atoms with brighter white spot characteristics (
Fig. 2(d)). Importantly, the Cu elements observed in the larger field of view did not appear in clusters or particles, except for random dispersion in the state of single atoms. The element Cu was uniformly distributed over the surface of Cu@G-AMPs and Cu@G, respectively (
Fig. 2(e); Fig. S3 in Appendix A).
The results from ultraviolet and visible spectrophotometry (UV-Vis) showed that aqueous solutions of Cu@G-AMPs and Cu@G had no characteristic absorption peaks in the wavelength range of 200-600 nm but were heterogeneous states (Fig. S4 in Appendix A). The reason for the poor ability of the substrate material formed by guanine to capture wide-spectrum light may be the strong quantum size effect [
32], [
34]. The results from Fourier-transform infrared spectroscopy (FTIR) revealed that the absorption peaks of the N
H, C
N, and C
N bonds involving N-doped carbon-based materials appeared in the wavelength range of 500-1210 and 1580 cm
−1 for Cu@G and Cu@G-AMPs, respectively (
Fig. 2(f)) [
37], [
38]. Compared with Cu@G, the FTIR spectra of Cu@G-AMPs exhibited enhanced absorption peaks associated with modified AMPs in the ranges of 1110-1210 and 1400-1700 cm
−1. The Zeta potential test results verified that Cu@G-AMPs, which contains AMPs, had a higher positive charge than Cu@G (
Fig. 2(g)). Antimicrobials with an appropriate positive charge will make a preferable choice when attacking pathogenic bacteria or mammalian cells [
39]. The X-ray diffraction (XRD) patterns of Cu@G-AMPs and Cu@G exhibited a wide peak near 25°, which may be caused by the amorphous carbon (002) in the nanomaterials (Fig. S5 in Appendix A) [
40]. In addition, the XRD patterns of Cu@G-AMPs showed a slight shift to low intensity compared with those of Cu@G; this shift may be a result of the AMPs’ modification, which caused a certain degree of deformation in the original sheet nanostructure. In addition, the results of inductively coupled plasma optical emission spectrometer (ICP-OES) showed that Cu@G-AMPs contained 0.41% Cu, which is in line with the characteristics of single-atom metal catalysts’ low content and rich active sites (Fig. S6 in Appendix A) [
41]. In summary, we have preliminarily confirmed through multiple molecular-structure-identification techniques that Cu@G-AMPs is an antibiotic-free nanomaterial containing ATCUN-motif AMPs and Cu groups.
According to the basic theory of catalytic chemistry, the catalytic antibacterial activity of Cu@G-AMPs can be attributed to the strong electronic interaction between the d-orbitals of the Cu atoms and the p-orbitals of the coordinated nonmetal atoms (e.g., C, N, or O) [
42]. As an alternative explanation, the antibacterial activity of Cu@G-AMPs can be viewed as being closely related to the discrete quantum states of the single Cu atoms and their local chemical coordination environments, which include the types of the nearest-neighbor atoms and the coordination number [
35]. Therefore, we further characterized the single-atom metal dispersion state and substrate single-atom Cu catalysts by determining the photothermal performance and enhanced nanozyme coordination structure of Cu@G-AMPs. In the Raman spectra of Cu@G-AMPs and Cu@G (Fig. S7 in Appendix A), peak D (1360 cm
−1) and peak G (1580 cm
−1), representing amorphous carbon, were detected [
43]. The corresponding
ID/
IG values of the Cu@G-AMPs and Cu@G samples were very similar. Indeed, prominent peaks representing pyridinic N, pyrrolic N, and graphitic N were identified via X-ray photoelectron spectroscopy (XPS) for both Cu@G-AMPs and Cu@G. Relative contents of 47.63% pyridinic N (397.5 eV) and 44.12% pyrrolic N (401.3 eV) were calculated in the N spectrum of Cu@G-AMPs (Fig. S8 in Appendix A). However, the N spectrum of Cu@G showed a relative content of pyridinic N (399.3 eV) of up to 89.83% (Fig. S9 in Appendix A). Interestingly, the N atoms in pyridinic N and pyrrolic N belong to electron donors near the Fermi level, which can provide extremely important sites for the stability of the anchored single Cu atom on the substrate [
41].
Using CuO and Cu foil as the controls, we further characterized the basis of the graphene-based metal single-atom catalysts’ stability and catalytic performance by determining the structure of the single Cu atoms in Cu@G-AMPs via X-ray absorption fine structure (XAFS) spectroscopy. XAFS includes extended X-ray absorption fine-structure (EXAFS) and X-ray absorption near-edge-structure (XANES) spectroscopy [
33]. Specifically, the results from the XANES of Cu showed that the energy absorption threshold of Cu@G-AMPs was between those of Cu foil and CuO (
Figs. 3(a) and (b)). This situation implied that the charge range of the single Cu single in Cu@G-AMPs was between 0 and +2. Correspondingly, the Cu in Cu@G-AMPs exhibited a prominent peak involving Cu-N/O at 1.5 Å and had no prominent peak involving Cu
Cu at 2.2 Å, confirming that the single metal atoms are highly dispersed (
Fig. 3(c)).
In addition, by performing a wavelet-transform (WT) analysis on the EXAFS parameters of Cu@G-AMPs, we confirmed the true coordination environment of the single Cu atoms with neighboring atoms in the k and
R space. The contour extreme values of Cu@G-AMPs (
R = 1.4 Å,
K = 6.7 Å
−1) (
Fig. 3(d)) were close to those of CuO (
R = 1.5 Å,
K = 6.8 Å
−1) (Fig. S10 in Appendix A) but had a large gap with those of Cu foil (
R = 2.3 Å,
K = 7.6 Å
−1) (Fig. S11 in Appendix A), indicating that Cu@G-AMPs has no Cu
Cu interaction but does have Cu
N/O coordination. This result is consistent with the XRD of Cu@G-AMPs (Fig. S5), which lacks the characteristic peaks involving Cu
Cu crystals. These two characterization methods revealed that the Cu in Cu@G-AMPs is dispersed at the atomic level and does not polymerize into a structure similar to metal crystals.
A quantitative analysis of the EXAFS on Cu@G-AMPs further revealed that the coordination number of the single Cu atoms in the first shell is approximately 2 and the average bond length is 1.91 Å (
Figs. 3(e) and (f), where
Fig. 3(f) depicts the interaction details of single Cu atoms in Cu@G-AMPs with substrates doped with N/O). Frankly, the intrinsic chemical properties of N and O constitute a great obstacle to the accurate identification of the heteroatomic coordination environment within the same bond length range of single Cu atoms [
34]. On the whole, we successfully prepared and characterized a novel nanoreactor (Cu@G-AMPs) with potential Fenton-like catalytic activity, which is formed by fusing AMPs containing ATCUN motifs and single Cu atoms confined to a carbon-based substrate doped with heteroatoms.
3.2. Antibacterial properties and mechanisms
Setting MRSA (ATCC 43300) as the main test strain, we verified whether the Cu@G-AMPs prepared above exhibited
in vitro antibacterial activity that met the intended targets. To specify, the results of the spread plate method showed that Cu@G-AMPs exhibited antibacterial activity against MRSA that was dependent on the pH of the treated environment (
Fig. 4(a)) [
26]. The antibacterial action of Cu@G-AMPs was generally more fully exercised under acidic conditions. The results of SYTO9/PI staining also confirmed that Cu@G-AMPs exhibited ideal
in vitro antibacterial activity against MRSA at pH 6.7 (
Fig. 4(b)), with the bacteria community treated with Cu@G-AMPs exhibiting intense red fluorescence and slight green fluorescence, compared with those treated with PBS and H
2O
2. Here, SYTO9 is a dye that can penetrate the cell membrane and causes any live bacteria to show green fluorescence, while PI is a dye that cannot penetrate the cell membrane and only allows membrane-damaged bacteria to show red fluorescence [
44]. To summarize, Cu@G-AMPs demonstrated bacteriostatic activity against MRSA under acidic conditions, which is consistent with the classical view that catalytic metals perform antimicrobial activity through Fenton-like mechanisms. It is known that the activity and efficiency of Fenton/Fenton-like reactions depend on the concentrations of H
+ and H
2O
2, which are usually energetic within the weakly acidic range [
33]. In addition, the concentration of H
2O
2 in the Cu@G-AMPs catalyzed reaction system was much lower than the current clinical practice guidelines (166 mmol∙L
−1), which practically avoids the occurrence of high concentrations of H
2O
2 encroachment on normal tissues [
33].
The toxicity of Cu@G-AMPs against the MC3T3-E1 cells was determined by means of an MTT colorimetric assay (Fig. S12 in Appendix A). In the concentration range of 0-128 μg∙mL−1, the viability of the Cu@G-AMPs-treated cells was higher than 75.83% ± 5.67%. The results of the MTT colorimetric assay revealed that Cu@G-AMPs has a low level of cytotoxicity in the concentration range of antibacterial action. Notably, Cu@G-AMPs did not cause lysis and rupture of red blood cells in the concentration range of 0-128 μg∙mL−1 (Fig. S13 in Appendix A). Thus, the antibiotic-free nanomaterial we have painstakingly constructed has ideal biocompatibility.
As a novel antimicrobial agent that integrates AMPs containing ATCUN motifs and single Cu atoms, Cu@G-AMPs may have one or more bacteriostatic mechanisms. Therefore, we examined the inherent enzyme-like activity of Cu@G-AMPs using TMB as the substrate. Theoretically, Cu@G-AMPs with enzyme-like activity convert H
2O
2 to hydroxyl radicals (·OH), which would oxidize colorless TMB to blue oxTMB (
Fig. 5(a)) [
45]. Here, we investigated the correlation between Cu@G-AMPs concentration and its enzyme-like activity and found that the absorption value of the whole reaction system at 652 nm increased in a concentration-dependent manner (
Fig. 5(b)). This finding indicated that, as the concentration of Cu@G-AMPs increased, so did the oxidizing species (i.e., oxTMB) in the reaction system. Furthermore, the enzyme-like activity of Cu@G-AMPs increased with a decrease in the pH of the reaction system, which was consistent with the Fenton-like reaction mechanism of artificial nanomaterials (
Fig. 5(c)). Subsequently, the results of a 2',7'-dichlorodihydrofluorescein diacetate (DCFH-DA) fluorescence assay confirmed that MRSA cells treated with any concentration of Cu@G-AMPs had richer ROS than bacterial cells treated with 32 μg∙mL
−1 AMPs (
Fig. 5(d)). Crucially, with an increase in the Cu@G-AMPs concentration, the amount of ROS in the treated MRSA cells also increased. Here, “ROS” refer to the byproducts formed by aerobic cells during metabolism, including singlet oxygen, ozone, superoxide anion, H
2O
2, and ·OH, where ·OH is recognized to have the highest toxicity and reactivity against bacterial cells among all ROS [
46]. Based on these results, Cu@G-AMPs, which has intrinsic enzyme-like activity, can catalyze the formation of ROS represented by ·OH from H
2O
2.
In the external world where ROS are present, MRSA employs various protective mechanisms to counter their inactivation, including producing adaptive mutations, activating the regulation of stress-resistant genes, and changing to a dormant status [
26]. As proteins are the ultimate executors of the response mechanisms of MRSA, their situation can perfectly reflect the transient biological state of the organism itself [
26]. For this reason, we further explored the potential bacteriostatic mechanism of Cu@G-AMPs from the perspective of bacterial stress response by applying label-free quantitative proteomics (often denoted as just “label-free”) techniques. Specifically, 106 up-regulated and 183 down-regulated differentially expressed proteins (DEPs) were identified from piscidins-3-treated samples, compared with PBS-treated bacteria (AMPs/PBS, fold changes ≥ 2 or ≤ 0.5,
P-adjusted value ≤ 0.05) (
Fig. 6(a)); 6 up-regulated and 15 down-regulated DEPs were identified from Cu@G-AMPs-treated samples, compared with PBS-treated bacteria (Cu@G-AMPs/PBS) (
Fig. 6(b)); and 223 up-regulated and 118 down-regulated DEPs were identified from Cu@G-AMPs-treated samples, compared with AMPs-treated bacteria (Cu@G-AMPs/AMPs) (
Fig. 6(c)). In brief, the AMPs altered the expression levels of a large number of proteins in MRSA, whereas Cu@G-AMPs affected only a small number of key proteins. Moreover, the important DEPs that were sensitive to Cu@G-AMPs were mainly involved in the sensing, regulatory, and repair and recombination systems of MRSA (
Fig. 6(d)).
To better understand the mechanism of the inhibitory activity of Cu@G-AMPs against MRSA, we assigned these DEPs to gene ontology (GO) categories. The statistical results of these DEPs in MRSA showed that they were mainly assigned to biological processes that included the response to external stimuli, horizontal gene transfer, DNA-mediated transformation, the establishment of competence for transformation, and negative regulation of cell communication (
Fig. 6(e)). However, DEPs within MRSA assigned to CC and MF were rare; that is, the bacterial cells respond to ROS stimulation by expressing multiple gene products involved in a wide range of BP. Simultaneously, MRSA is able to prevent further invasion of ROS by adjusting protein expression levels such as its antioxidant system [
26]. When the external oxidative pressure exceeds the limit pressure that bacterial cells can bear, it will undoubtedly cause irreversible negative effects on the cells’ antioxidant system [
47]. Hence, the DEPs identified from Cu@G-AMPs-treated bacteria, including lytM, msaA, isaB, mecA, and msrA, were significantly down-regulated compared with those identified from the PBS-treated bacteria (
Fig. 6(d)). Moreover, when exogenous ROS cause DNA or RNA damage within MRSA cells, the associated repair system may be activated to maintain the integrity of the bacterial genome. In conclusion, Cu@G-AMPs achieves antimicrobial activity by disrupting the stress response systems—including the quorum sensing regulation, antioxidant enzymes, and gene repair and recombination—within MRSA.
PRM is a proteomic result-verification technique with high sensitivity, resolution, and strong specificity [
48]. We selected three target proteins (aaa, mecA, and msrA) related to the oxidative stress function of MRSA for PRM. We also selected 29 DEPs containing cysteine (Cys) residues for the PRM tests, mainly because it is very easy for ROS to cause the oxidation of Cys residues in peptides and convert them into chemical bonds such as S-glutathione, S-sulfinic acid, S-sulfonic acid, disulfide bonds, and S-hyposulfonic acid [
49]. In general, the quantification of these selected DEPs in PRM was similar to the results from the label-free techniques (
Fig. 6(f)). That is, the detection results of the label-free technology were verified to be accurate and reliable.
3.3. Healing therapy for MRSA-infected wounds
AMPs applied in the body often encounter complex microenvironmental constraints such as wound fluid, protease, and pH [
26]. Accordingly, we further evaluated the application prospects of Cu@G-AMPs
in vivo through mouse wound-resistant bacterial infection models and drug grouping treatment strategies (
Fig. 7(a)). The total treatment period for each mouse in the wound infection model was 7 d. The wound area of each group showed a concrete trend of healing and reduction with the extension of treatment time (
Fig. 7(b)). The wound area treated with Cu@G-AMPs and VAN was close to closure after 5 days, in comparison with the other treatment groups (Fig. S14 in Appendix A). It should be noted that the stability of the
in vivo environment of the AMP piscidins-3 could not be ensured [
2], so the contribution of piscidins-3 in promoting wound healing in mice did not show theoretical advantages compared with that of PBS. Importantly, we quickly assessed the bacterial density of the wounds treated with antimicrobials on the first day and found that Cu@G-AMPs and VAN had more ideal antibacterial activity than the other drugs (Fig. S15 in Appendix A). Coagulation, inflammation, proliferation, and remodeling are the necessary stages of wound healing in mice, and this process mainly involves fibroblasts, epithelial cells, inflammatory cells, mast cells, and mesenchymal cells [
50]. Correspondingly, H&E staining images of wounds treated with Cu@G-AMPs and VAN for 3 d revealed a small number of inflammatory cells and loose connective tissue compared with the other treatment groups (
Fig. 7(c)). Subsequently, at 7 d, the wound profile treated with Cu@G-AMPs manifested thicker granulation tissue and nearly intact surface cells, indicating that Cu@G-AMPs carries out multiple functions, including pulling the closure of wound edges and stabilizing granulation tissue (
Fig. 7(c)).
We further evaluated the actual progress of wound healing in the groups of treated mice by means of Masson staining, which can turn collagen fibers blue and myofibers red [
26]. Large blue areas appeared in the wound profiles treated with Cu@G-AMPs at both 3 and 7 d, indicating that Cu@G-AMPs inhibited the growth of undesirable granulation and promoted the proliferation of collagen fibers (
Fig. 7(d)). The fact that the AMP piscidins-3 is an immune factor derived from the secretion of fish skin mast cells motivated us to examine the infiltration of mouse wound mast cells by means of toluidine blue staining [
10]. Mast cell infiltration is a universal signal of skin tissue injury and allergic reactions [
12]. The absence of significantly infiltrated mast cells in the wound profile treated with Cu@G-AMPs at 3 and 7 d indicated that Cu@G-AMPs had anti-inflammatory functions similar to those of VAN (
Fig. 7(e)). To investigate whether Cu@G-AMPs has adverse side effects on the animals, the weight fluctuations of the mice were monitored throughout the experiment. A weight change of more than 20% is usually used as a key proxy of morbidity, distress, and overall toxicity [
30]. Fortunately, no side effects (e.g., swelling, redness, or itchiness) or
in vivo toxicity were detected during the healing process of the skin wound (
Figs. 7(b) and (c); Fig. S16 in Appendix A).
The contents of platelet endothelial cell adhesion molecule-1 (CD31) and PCNA in the wounds of mice treated with different drugs were analyzed by means of immunofluorescence staining. Green-fluorescent CD31 and red-fluorescent PCNA are a pair of common markers representing angiogenesis [
26]. More CD31 and PCNA appeared in the wound area treated with Cu@G-AMPs and VAN for 3 d compared with other treatment groups (Figs. S17 and S18 in Appendix A). When the wound was in the remodeling phase on day 7, less CD31 and PCNA was reasonably present in the samples treated with Cu@G-AMPs and VAN compared with the other treatment groups (Fig. S17 and Fig. S18). That is, Cu@G-AMPs continued the pro-angiogenic properties of the AMPs. In fact, the entire cycle of bacterial wound healing is accompanied by an inflammatory response, which is the result of the regulatory balance of various positive and negative immune factors [
51]. Specifically, as a multifunctional negative regulator, IL-10 plays a role in antagonizing inflammatory mediators and downregulating inflammatory responses in the body [
50]. The sample group treated with Cu@G-AMPs and VAN showed lower levels of IL-10 than the other treatment groups after 3 d, which was reversed after 7 d (
Fig. 8(a)). IL-1β and TNF-α are typical multifunctional positive regulators of the body’s immune response [
50]. The sample group treated with Cu@G-AMPs and VAN detected higher levels of IL-1β and TNF-α than the other treatment groups after 3 d, which was reversed after 7 d (
Figs. 8(b) and (c)). Based on these results, there was a gap in the real-time release of immune factors from wounds in the different healing cycles caused by the drug therapies.
4. Conclusions
Inspired by host defense mechanisms involving ATCUN-motif AMPs, we were fortunate enough to develop an artificial AMP complex (Cu@G-AMPs) incorporating single-atom Cu catalysts for antibacterial therapy. Single Cu atoms were anchored with a coordination number of 2 and an average bond length of 1.91 Å on a substrate prepared from guanine doped with a rich heteroatom. Showing Fenton-like catalytic activity, Cu@G-AMPs generated and delivered its deadly ROS cargo when encountering MRSA. These ROS caused irreversible damage to the stress response system of MRSA, including the quorum sensing regulation, antioxidant enzymes, and gene repair and recombination within MRSA. Notably, Cu@G-AMPs was capable of adapting to the complex microenvironment in vivo and showed indications of pulling edge closure, stabilizing granulation tissue, promoting collagen fiber proliferation, alleviating inflammation, and promoting neovascularization in wound areas suffering from drug-resistant bacterial infection. It is expected that the Cu@G-AMPs designed here will provide a new perspective on disrupting the precise regulatory laws that resistant bacteria have developed to inactivate antibiotics.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (32272311), Fujian Province Science & Technology Project, China (2023N3008), and Fujian Major Project of Provincial Science & Technology Hall, China (2020NZ010008).
Compliance with ethics guidelines
Xuan Chen, Wei Luo, Qun Gao, Congrong Chen, Lichan Li, Dongbo Liu, and Shaoyun Wang declare that they have no conflict of interest or financial conflicts to disclose.
PDF
(5084KB)
