Hedgehog artificial macrophage with atomic-catalytic centers to combat Drug-resistant bacteria

Pathogenic drug-resistant bacteria represent a threat to human health, for instance, the methicillin-resistant Staphylococcus aureus (MRSA). There is an ever-growing need to develop non-antibiotic strategies to fight bacteria without triggering drug resistance. Here, we design a hedgehog artificial macrophage with atomic-catalytic centers to combat MRSA by mimicking the “capture and killing” process of macrophages. The experimental studies and theoretical calculations reveal that the synthesized materials can efficiently capture and kill MRSA by the hedgehog topography and substantial generation of •O2− and HClO with its Fe2N6O catalytic centers. The synthesized artificial macrophage exhibits a low minimal inhibition concentration (8 μg/mL Fe-Art M with H2O2 (100 μM)) to combat MRSA and rapidly promote the healing of bacteria-infected wounds on rabbit skin. We suggest that the application of this hedgehog artificial macrophage with “capture and killing” capability and high ROS-catalytic activity will open up a promising pathway to develop antibacterial materials for bionic and non-antibiotic disinfection strategies.

A ccompanied by the overuse of antibiotics, a series of pathogenic bacteria are becoming the so-called drugresistant bacteria, which puts numerous lives in danger and becomes an ever-increasing crisis [1][2][3][4][5] . As a representative pathogenic drug-resistant bacteria, the methicillin-resistant Staphylococcus aureus (MRSA) can resist many types of common antibiotics 6 . Currently, there is an ever-growing need to develop non-antibiotic strategies to fight bacteria without triggering drug resistance, and numerous efforts have been paid to this field with the advancements of nanoparticles (NPs), for instance, loading multiple antibacterial molecules into NPs to reduce the formation of drug-resistance 7 , and tuning geometry of NPs to enhance the enzyme inhibition and antibacterial activities 8 . Recently, generating reactive oxygen species (ROS) via chemodynamic, photodynamic, sonodynamic, and enzyme-mimetic materials to combat bacteria has also been explored as promising alternative antibacterial strategies [9][10][11] . Among diverse ROS generation materials, the enzyme-mimetic catalysts, including nanoliposome 12 , metalorganic frameworks [13][14][15] , and inorganic materials 16,17 , have gained tremendous popularity by catalyzing oxygen and hydrogen peroxide (H 2 O 2 ) to generate ROS via the existing natural pathway. Nevertheless, when treated at a very low concentration, such enzyme-mimetic catalysts present insufficient bacterial killing activities, while utilizing antibacterial materials at high concentrations will always lead to a severe concern on biocompatibility. Thus, developing more effective ROS-based enzymemimetic strategies to fight drug-resistant bacteria is of great importance in current material and biomedical sciences. The human immune system has provided us with an ideal example to overcome this bionic challenge; it has been clarified that macrophages play a critical role in disinfection and wound healing [18][19][20] . The macrophage can efficiently eradicate pathogenic bacteria by successively capturing, phagocytosing, and killing actions. The macrophage exhibits a rough surface with microfold, prominences, and recognition receptors, which facilitate the detection and capture of bacteria. After phagocytosis, various peroxidases (POD) in the lysosome can produce abundant localized ROS to kill phagocytosed bacteria rapidly 21,22 . Recently molecular biology reveals that these ROS-generation enzymes mainly consist of heme-containing catalytic sites, which can generate highly toxic hypochlorous acid (HClO) and other types of ROS in the presence of Cl − and H 2 O 2 [23][24][25] . Though V 2 O 5 nanorods have been reported to show excellent haloperoxidase (HPO)-mimetic production activity for HClO, the potential cellular toxicity and poor bacterial capture ability of V 2 O 5 may limit their application in clinical treatment 26 . Thus, it is a crucial challenge to develop ROS catalytic materials that can simultaneously mimic the "capture and killing" process of macrophages to combat bacteria, especially integrated with heme-mimetic catalytic sites.
To address this challenge, we present here the design of a hedgehog artificial macrophage with heme-mimetic atomic-catalytic centers to combat MRSA. The main motivation of this work originates from two aspects: (1) designing physiologically stable, micron-sized, and hedgehog carbonaceous particles to mimic the topography of macrophage for localized bacterial capture, (2) precisely engineering metal-N x -C y based single-atom catalytic sites to mimic the heme-structure (Fig. 1a); here the Fe-N x -C y is the primary candidate, while the V-N x -C y is synthesized for comparison due to the natural widely existence of Fe-HPO (Fig. 1b) and V-HPO (Fig. 1c) for HClO production. The experimental studies and theoretical calculations reveal that the synthesized artificial macrophage can efficiently capture and kill MRSA by the hedgehog topography and localized massive generation of •O 2 − and HClO with its Fe 2 N 6 O catalytic centers. The synthesized artificial macrophage exhibits a low minimal inhibition concentration (8 μg/mL Fe-Art M with H 2 O 2 (100 μM)) to combat MRSA and can rapidly promote the healing of bacteria-infected wounds on rabbits skin, which is comparable to the treatment efficiency of vancomycin.

Results
Design and analysis of hedgehog artificial macrophage. To mimic the "capture and killing" function of macrophages both from the surface topography and ROS generation capability, we designed a micron-sized hedgehog particle doped with Fe-N x or V-N x based atomic catalytic centers (Fig. 1d). First, a hedgehog metal-organic framework precursor (Supplementary Figs. and 2) was synthesized by a hydrothermal transformation method with urea and Zn-MOF-74 ([Zn 2 (DOBDC), DOBDC = 2,5-dihydroxybenzenedicarboxylic acid]. After carbonization and sublimation of Zn, a highly stable micron-sized hedgehog carbonaceous particle can be obtained and served as the pristine artificial macrophage (Art M) ( Supplementary Fig. 3) 27,28 ; subsequently, the Fe-N x or V-N x catalytic centers are synthesized by the in situ coating and carbonization of metal-phenanthroline complex (Supplementary Fig. 4) 29 . Then, the synthesized Art M and Fe-doped or V-doped-Art M (named as Fe-Art M or V-Art M) are applied to act as artificial macrophages to capture and kill MRSA by the hedgehog topography and massive generation of ROS with its POD-/HPO-mimetic catalytic activities (Fig. 1e). Interestingly, it is also noticed that the carbonaceous Fe-Art M exhibits good dispersibility in diverse aqueous media than Fe-doped carbonous micro-particles (Fe-C-MPs, smooth surface) (Supplementary Figs. 5-7). An earlier report has suggested that the hedgehog structure can efficiently enhance the aqueous dispersibility of particles due to the decreased contact area, trapping of air, autoionization of water, etc. 30 , which may also facilitate the bacterial capture capability.
The representative hedgehog morphology and the spikes of different types of Art Ms are verified by scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HR-TEM) (Fig. 2a, b and Supplementary Figs. [8][9]. In the meantime, the surface area and pore size distribution of the carbonaceous particle are analyzed by the Brunauer-Emmett-Teller (BET) method. All the Art Ms show high surface areas (up to 1895.8 m 2 /g, Fig. 2c) with abundant pores ranged from 1.01 to 4.84 nm ( Supplementary Fig. 10). The carbon structure and metal phases of the Art Ms are detected by X-ray diffraction (XRD) pattern, and the existing weak characteristic peaks of carbon (002) and (101) diffractions suggest that all of these porous Art Ms display low graphitic degree with abundant defects and no trace of metal-related phases as observed in the HRTEM image (Fig. 2d). This unique porous carbonaceous structure may greatly facilitate the transportation of substrates during the ROS catalytic reactions. Furthermore, the high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images and corresponding energy dispersive spectroscopy (EDS) elemental mapping images show that atomic level Fe and V atoms are homogeneously distributed on the hedgehog particles ( Fig. 2e and Supplementary  Fig. 11).
To directly demonstrate the concept that these hedgehog artificial macrophages can be applied to capture the bacteria via its spiky topography and bioadhesive carbonaceous structure 32 , we have incubated the MRSA with different Art Ms and other comparison samples (Fe-C-MPs, smooth surface, Supplementary  Fig. 12, polystyrene microspheres (smooth surface), SiO 2 microsphere (slightly rough surface)). It can be observed that all types of Art Ms can efficiently capture abundant bacteria with no significant difference, as demonstrated by the SEM and fluorescence microscopy images (Fig. 2f-j Supplementary Fig. 16). These data indicate that the Fe-Art M and V-Art M exhibit contact-killing activity to bacteria. As presented in Supplementary Fig. 24, we have analyzed the intrinsic oxidase-mimetic activity of these Art Ms, which suggests that the Fe-Art M exhibits very high catalytic transformation of oxygen into ROS via activation and cleavage of the O-O bond 33 . This good contact killing activity of Fe-Art M preliminarily supports the proposed concept for the bacterial "capture and killing" function of hedgehog artificial macrophage.
Atomic catalytic centers analysis of Fe-Art M. As verified that the Fe-Art M exhibits the best contact killing activity to MRSA, in the following section, we will mainly focus on exploring the detailed characteristics of Fe-Art M. First, we use X-ray photoelectron spectroscopy (XPS) to reveal the atomic catalytic centers  Fig. 3b, the Fe 2p spectra can be divided into two peaks belonging to Fe 2+ /Fe 3+ ionic species, and no Fe 0 and other valence states can be detected 34,35 , suggesting the Fe-N x coordination structure. Besides, the V-Art M (0.10 at.% of V) also reveals similar metal-N x coordination structures as shown in Supplementary Figs. 25-29. Then, to reveal the precise atomic coordination Fe-N x structure, the Fe K-edge X-ray absorption near-edge structure (XANES) was tested for Fe-Art M with Fe foil, Fe 2 O 3 , and Fephthalocyanine (FePc) as references (Fig. 3c, d and Supplementary Fig. 30), the pre-edge peak in Fe-Art M positioned at around 7114.6 eV is approximately located between FePc and Fe 2 O 3 , suggesting a different Fe-N structure with higher valency compared to the representative Fe-N 4 . Thus it is believed that the Fe atoms may carry positive charges and are partially oxidized compared with the Fe-N 4 structure in FePc 36 . According to the Fourier-transformed k 3 -weighted extended X-ray absorption fine structure (FT-EXAFS) spectra, the EXAFS fitting results and corresponding wavelet-transform (WT) images are summarized in   The high-resolution HAADF-STEM images are presented in Fig. 3i, j and Supplementary Fig. 33; there are abundant paired bright atom dots, thus further verifying the Fe 2 N x(x=6-8) O structure 37,38 . The distance shown in the intensity profiles is around 2.20-2.30 Å (Fig. 3k), which is consistent with the major Fe-Fe bond (2.47 Å) detected from EXAFS. Furthermore, the Fe 2 N 6 O configuration also fits well with the corresponding Fe K-edge EXAFS for Fe-Art M in R-space (Fig. 3f). Therefore, we conclude that the simulated Fe 2 N 6 O configuration with a bond distance of~2.32 Å ( Evaluation of enzyme-mimetic ROS production activity. After validation of the enzyme-mimetic catalytic centers of Fe-Art M, we further tested and compared their POD-like and HPO-like activities for ROS production. By using the classic 3,3,5,5-tetramethylbenzidine (TMB) and o-phenylenediamine (OPD) colorimetric methods, the Fe-Art M shows the best POD-like performance in a pH-dependent manner (  Table 5). Although not possessing the highest V max , the Fe-Art M exhibits the highest TON values among these current established POD-mimics (Fig. 4d). The electron paramagnetic resonance (EPR) and fluorescence analysis reveal that the generated ROS is •O 2 −40 , and the Fe-Art M produces a much higher amount of •O 2 − than the V-Art M and pristine Art M ( Fig. 4e and Supplementary Fig. 36), which is similar to the natural heme-containing POD 41 . Finally, the production of •O 2 − has also been confirmed by the free radical quenching experiments (Fig. 4f); meanwhile, no obvious generation of •OH has been detected, and the 1 O 2 is originated from pristine Art M particle due to the generation of 1 42 . Thus, we utilize the CB as a probe to measure the HPO-mimetic activity of Fe-Art M 43 . As can be seen in Fig. 4g, h and Supplementary Figs. 41 and 42, the Fe-Art M displays much better catalytic activity than V-Art M and pristine Art M; besides, the Fe-Art M also reveal the best activity under acid condition. Taken together, we have demonstrated that Fe-Art M presents the best POD-like and HPO-like activities for ROS production than V-Art M and pristine Art M, which further proves that Fe-N catalytic centers exhibit the best activity to function as artificial macrophage.
Theoretical calculation of enzyme-mimetic ROS production pathways. To elucidate the detailed POD-like and HPO-like ROS catalytic pathways of Fe-Art M, we use the DFT calculations to simulate the •O 2 − and HClO-generating processes. First, we compare the Fe 2 N 6 O with Fe 2 N 6 O adsorbing H 2 O 2 molecules through the differential charge density, reduced density gradient, and projected density of states (PDOS). At a position of −2.5 eV, the O 2p and H 1s orbitals of H 2 O 2 have a certain weak interaction with the N 2p orbital, and charge at the OH group transfers to the N direction; besides, the acting force between Fe and O is Van der Waals action ( Fig. 5a and Supplementary Figs. [43][44][45]. Then, we propose three potential POD-like reaction paths and elucidate the Gibbs free energy of all intermediates for each reaction step, resulting in the theoretical free energy for different   For reaction path 1, the rate-determining step happens at the third step, the desorption of H + and formation of O* (iv) exhibit with a relatively low free energy barrier (0.86 eV) compared to the high free energy of OH* desorption (1.51 eV) in reaction path 3, which also theoretically explain why there is no generation of •OH in Supplementary Fig. 37. Then, there are two potential paths, in path 1, an OH* move to O* site and form OOH* (v) with a free energy of 0.71 eV, while in path 2, the formation of two O* will need free energy of 1.70 eV, this should be very difficult to happen compared with path 1. Therefore, we demonstrate that reaction path 1 shows the most reasonable POD-like catalytic processes of Fe-Art.
For the HPO-like catalytic process, we propose two potential reaction paths (Fig. 5d, e and Supplementary Fig. 47). Both paths share the same process of forming two OH*; subsequently, one OH* can easily combine with Cl* species to produce a molecule of HClO with a free energy of −1.67 eV. Then, there are two potential paths; one is to remove the OH* directly with a very high free energy of 2.71 eV, which is supposed to be not reasonable. Another path is to adsorb H 2 O 2 first and then remove an H* to the OH* to form H 2 O* + OOH*. The OOH* will desorb and become •O 2 − and H + . Therefore, the Fe-Art M could generate •O 2 − and HClO simultaneously in the condition of H 2 O 2 and Cl − , which agrees well with our experimental results shown in Supplementary Fig. 48.
Sterilization assessment of drug-resistant bacteria. After validating that the Fe-Art M can function as an artificial macrophage   . g, h Detection of HPO-mimetic activity using CB as agents, the relative activity is referred to the rate of Intensity 512 nm /Intensity 648 nm . (In h, n = 3 independent experiments per group, data are presented as mean ± SD). In a, h, p values are assessed by unpaired Student's two-sided t-tests. V is the initial velocity, V max is the maximal reaction velocity, K m is the Michaelis constant, TON is turnover number. In a, g, h, a.u. indicates the arbitrary units. Source data are provided as a Source Data file.
to simultaneously capture the bacteria by its hedgehog topography and catalytic generation of ROS with its POD-/HPOmimetic activities, we then carefully examine its activity in combating drug-resistant bacteria (Fig. 6a). First, we evaluate the cellular compatibility of samples with L929s cells according to the animal wound disinfection process, which shows limited shortterm toxicity, and the cells can soon recover after the removal of   that of vancomycin (4 μg/mL) and much lower than that of 32 μg/ mL V-Art M with H 2 O 2 (100 μM). Then, to assess the animal wound disinfection activity, MRSA was introduced to infect rabbit skin wounds. The photographs and change of wound sizes were recorded and shown in Fig. 6d, e    (heart, liver, spleen, lung, and kidney), the data are displayed in Fig. 6f-k and Supplementary Figs. 58-60. The infected wound shows extensive areas of necrotic polymorphonuclear leukocytes and a high number of neutrophils (1011 ± 116) 44 ; the Fe-Art M + H 2 O 2 group displays nearly no neutrophils as the vancomycin groups, indicating that MRSA has been eradicated, which is much better than that of the H 2 O 2 treated group (912 ± 139). Besides, no damage or abnormalities are found in the major organs compared to the healthy group, thus suggesting that the skin disinfection treatment of Fe-Art M + H 2 O 2 is very safe (Supplementary Figs. 58 and 59). Second, as displayed in Fig. 6g, j, for collagen regeneration, we found collagen takes large areas of the tissue slices in the Fe-Art M + H 2 O 2 (92 ± 5%) and vancomycin (96 ± 2%) groups, and the collagen fibers are uniformly and regularly arranged, thus indicating positive wound-healing effects. Furthermore, as shown in Fig. 6h, k and Supplementary Fig. 60, through the CD31 staining, regenerated endothelial cells in Fe-Art M + H 2 O 2 (109 ± 12) and vancomycin group (106 ± 25) can be labeled to reflect newly formed blood vessels after revascularization in the wound area, which is much higher than that of the control (22 ± 5) and H 2 O 2 treated groups (32 ± 5). Overall, the above in vivo disinfection results demonstrate that the synthesized Fe-Art M can serve as an antibioticfree agent to kill MRSA efficiently and promote wound healing rapidly, which has comparable treatment effects as vancomycin.

Discussion
In summary, our finding demonstrates that the synthesized hedgehog Art M with atomic-catalytic centers is an ideal material to combat MRSA and can mimic the "capture and killing" process of natural macrophage. The electron and fluorescent microscope analyses reveal that the synthesized Art Ms can efficiently capture MRSA by the hedgehog topography. The experimental analyses and theoretical calculations prove that the high catalytic production activity of •O 2 − and HClO are originated from the Fe 2 N 6 O catalytic centers, and the Fe-Art M exhibits much higher ROS catalytic kinetics than the V-Art M and pristine Art M. The synthesized Fe-Art M exhibits a low minimal inhibition concentration (8 μg/mL Fe-Art M with H 2 O 2 (100 μM)) to MRSA and can rapidly promote the wound healing by in vivo disinfection of rabbit skin. We are convinced that this highly ROScatalytic active Fe-Art M opens up a promising avenue to develop antibacterial materials for bionic and non-antibiotic disinfection strategies.
Synthesis of M-Art M (M = V, Fe). The synthetic process of M-Art M was divided into two steps. Firstly, 50 mL of methanol containing zinc acetate dehydrate (0.22 g, 1 mmol) was mixed with 20 mL of methanol containing 2,5-dihydroxyterephthalic acid (60 mg, 0.3 mmol), and the mixture was further ultrasonicated for about 30 min. After washing with water three times, urea (25/ 100 mg, 0.42/1.68 mmol) in 75 mL deionized water was added in a 175°C-Teflonlined autoclave for 24 h. The precipitate was then washed with water and ethanol and dried under vacuum. Afterward, with a heating ramp rate of 5°C/min, the asprepared material was transferred into a ceramic boat and placed in a tube furnace, and then heated to 1100°C for 1 h. Then the products was gathered at room temperature and washed with HCl (1 M) solution at 80°C for further use. Subsequently, 0.02 g Art M was dispersed in 2.5 mL of ethanol solution, and then a 30 μL mixture (metal acetylacetonate and 1,10 phenanthroline with a molar ratio of 2:1) was added. Finally, the dispersion was dried overnight and heated to 600°C in an N 2 atmosphere for 2 h.
Characterization of V-Art M and Fe-Art M. N 2 adsorption analysis was conducted on a Quantachrome Autosorb IQ instrument. All powder samples were degassed at 150°C overnight prior to actual measurement. The surface area was calculated by using BET calculations. The pore size distribution (PSD) plot was recorded from the adsorption branch of the isotherm based on the QSDFT model for slit/cylinder pores. XPS were obtained by using an XPS device (ESCAL 250) to detect the compositions of the M-Art M and analyzed by Avantage to confirm the successful introduction of the metals. SEM images were obtained by using an Apreo S HiVoc (Thermo Fisher Scientific, FEI). Cryo-SEM images were obtained by using Quanta 450FEG (FEI Ltd., USA). TEM images and EDS mapping were obtained via a Talos F200x TEM microscope (FEI Ltd., USA) operated at 200 kV and analyzed by GMS-freeanalysis. XRD pattern presented the crystal phase state via a Bruker D8 Focus X-ray diffractometer and analyzed by MDI Jade and Origin. The X-ray absorption spectra were collected on the beamline BL07A1 in NSRRC and the radiation was monochromatized by a Si (111) double-crystal monochromator. XANES and EXAFS data reduction and analysis were handled via Athena software. Cs-corrected (S)TEM (FEI Titan Cubed Themis G2 300) was used for high-resolution HAADF-STEM. Oxidase-like performance of Fe-Art M. As for the oxidase-like activity test, apart from being the absence of H 2 O 2 , the detection procedure of oxidase performance is similar to that of peroxidase.
Furthermore, the haloperoxidase-mimetic activity test was determined by CB.   6 Sterilization assessment of drug-resistant bacteria of Fe-Art M. a Schematic illustration for the "capture and killing" function of Fe-Art M by the hedgehog topography and massive generation of ROS on the infective wound skin. b Real-time OD 600 values for different samples at 10 µg/mL against MRSA (n = 3 independent experiments per group, data are presented as mean ± SD). c Killing ratio of MRSA treated with V-Art M and Fe-Art M at different concentrations in the presence of H 2 O 2 (n = 3 independent experiments per group, data are presented as mean ± SD). d Digital photographs and e the change of wound sizes on different days (n = 3 independent experiments per group, data are presented as mean ± SD). f HE staining, g Masson and h CD31 staining images of the epidermal histological sections in different groups at day 15. i Number of neutrophils, j collagen index, and k number of blood vessels. In i-k, n = 3 independent experiments and data are represented as mean ± SD. In b, e, i-k, p values are assessed by unpaired Student's two-sided ttests compared to the control groups. Experiments were repeated independently (f-h) three times with similar results. Source data are provided as a Source Data file.
The •OH generation activity was detected by DMPO. Five hundred microliter of catalysts (0.1 mg mL −1 , buffer) and 25 μL of H 2 O 2 (1 M) were added into a 2 mL buffer, and then, 10 μL DMPO was added.
Antibacterial experiments. In vitro. MRSA (ATCC 43300, Gram-positive) and extended-spectrum β-lactamase-producing Escherichia coli (ATCC 53104, Gramnegative) were used as the representative pathogenic bacteria to investigate the bacterial capture and eradication abilities of the enzyme-mimetic catalytic systems, the drug-resistance of bacteria is induced by standard protocols. The V-A M and Fe-Art M systems with H 2 O 2 containing NaCl were cultured with 2 mL of bacterial suspensions (10 7 CFU mL −1 ) for 14 h at 37°C. The final concentrations of materials and H 2 O 2 are 10 μg mL −1 and 100 μM, respectively. The OD 600 values of the treated suspensions were measured by a UV-vis spectroscopy (UV-3600, Shimadzu) to investigate the growth of bacteria at 3, 6, 9, 12, and 14 h. Meanwhile, the cultured suspensions were diluted 10 4 times and taken for agar plate counting. The antibacterial properties were estimated by comparing the CFUs of different systems with that of the control. Then, OD 600 values of the bacterial suspensions treated by the systems at different concentrations (4,8,16, and 32 μg mL −1 ) were recorded to calculate the MIC values.
The bacterial capture ability and bactericidal performance of these systems were visualized via SEM, cryo-SEM, fluorescence microscopy (DMIRE2, Leica), and CLSM images (N-STORM & A1, Nikon, equipped with NIS-Elements Viewer). After being treated with different samples for 30 min under continuous shaking, the bacterial suspensions were then fixed with 2.5 wt% glutaraldehyde and dehydrated with a gradient of ethanol/water solution. Then, the SEM images were obtained to observe the bacterial capture ability and their morphologies. The in situ cryo-SEM images were also taken directly without fixation, dehydration, and centrifugation steps. Moreover, the bacteria were also stained by LIVE/DEAD BacLight Viability Kits (SYTO-9 for live cells and propidium iodide for dead cells) for observation using fluorescence microscopy and CLSM.
Cytocompatibility experiments. For living/dead fluorescence staining, we first coculture cells with 10 μg/mL particles and H 2 O 2 (100 μM) for 30 min, then change the medium and wash with PBS twice to remove the particles, and then observe live/dead cells immediately or after 1 day. Subsequently, Calcein-AM and PI working solution were added directly to stain cells for 5 min at 37°C, and washed with PBS. Finally, the cells were immediately visualized by fluorescence microscopy.
For the viability of the L929 (ATCC, CCL-1), we first co-culture cells with 10 μg/mL particles and H 2 O 2 (100 μM) for 30 min, then change the medium and wash with PBS twice to remove the particles, and then observe cell viability immediately or after 1 day. 200 μL of CCK-8-based diluted solution was added to each well, respectively. After 1 h of incubation, the mixture was detected by a microplate reader (Model 550, Bio-Rad) at 450 nm and analyzed by a standard protocol to calculate the cell viability.
Wound disinfection experiments. In vivo: Rabbits (New Zealand White Rabbit, male, 8-12 weeks) were selected as models in animal experiments, and all the studies on these animals were performed by following the animal ethical standard from Animal Ethics Committee in West China Hospital, Sichuan University, Chengdu, China. The protocols of all the animal experiments were permitted and carried out as requested. The infected wound model was made by two main steps: i) removal of a hole around 1 cm in diameter on the epidermis; ii) introduction of 100 μL of MRSA suspension (1 × 10 8 CFU mL −1 ) and incubation for 1 day. After that, Fe-Art M, H 2 O 2 , and diluted saline were added to treat the infected wound. The final concentrations of materials and H 2 O 2 are 10 μg mL −1 and 100 μM, respectively. Normal saline, H 2 O 2 (100 μM), and vancomycin (16 μg mL −1 ) were also used as a contrast. The antibacterial treatment time is 30 min; after the disinfection process, the PBS solution was used to wash and remove the residual samples. To observe the residual bacteria in the wound after disinfection treatment, 5 μL of fluids were collected from different wounds, respectively, for agar plate counting. After 16 days, the treated wounds were cut off and fixed with 10% formaldehyde solution for histological HE, Masson, and CD31 staining analysis. The HE and Masson images were taken by optical microscope. The CD31 immunofluorescence images were taken by a CLSM.
DFT calculations methods. The DFT calculations were performed by Vienna Ab initio Simulation Package (VASP) 45 with the projector augmented wave (PAW) method 46 . The exchange-functional was treated with the generalized gradient approximation (GGA) of Perdew-Burke-Ernzerhof (PBE) 45 functional. The planewave basis's cut-off energy is set at 450 eV for optimizing cell and atoms calculations. The vacuum spacing being perpendicular to the plane of the catalyst is at least 15 Å. The Brillouin zone integration is performed on the primitive cell using 3 × 3 × 1 Monkhorst-Pack k-point 47 . The onvergence energy threshold of 10 −5 eV is applied for self-consistent calculations. The equilibrium lattice constants within 0.03 eV/Å are optimized for the maximum stress on each atom. As a key element in the DFT + U method, it is widely believed that the localized 3d electrons correlation of transition metal in the fourth period can be described by considering onsite coulomb (U) and exchange (J) interactions. Furthermore, we applied DFT + U through the rotationally invariant approach with the corresponding U -J values being 3.29 48 .
Gibbs free energy is acquired via adding corrections including entopic (TS) and zero-point energy (ZPE) to calculated DFT energy, hence Δ G = ΔE DFT + Δ ZPE -TΔ S , in which ΔE DFT means calculated DFT reaction energy, Δ ZPE means changes in ZPE calculated from the vibrational frequencie, Δ S means changes in the entropy regarding thermodynamics databases.
Statistical analysis. All data were expressed in this manuscript as mean ± SD. All the results have been performed three times by independent experiments. A twotailed Student's t-tests was used to analyze the statistical significance between the two groups by using GraphPad Prism 7.0 (GraphPad Software Inc.).