Male and female adult M. pulchra frogs (n = 3), which are not protected or endangered species, were acquired in the countryside around Guangzhou City, Guangdong Province of China, and then were identified according to the species record website (https://amphibiaweb.org/) and the book “Atlas of Amphibians of China”. The mice used in the living animal experiment were bought from the experimental animal center of Southern Medical University (Guangdong, China). The mice were raised under standard acclimatization environments of 12-h light/dark cycle at 25°C and fed normal rodent feed procedure. All experimental processes involving living animals were permitted by the animal care and use ethics committee of Southern Medical University and implemented in strict accordance with the guidelines of the Committee for Animal Care and Use.

The molecular cloning of cDNA encoding brevinin-2MP was carried out as reported previously by us with minor modification (Zeng et al., 2018). In short, the frogs were humanely euthanized with carbon dioxide, and their skins were stripped and cut into small pieces to extract total RNA using Trizol reagent (Life Technologies, Carlsbad, CA) in light of the manual of the manufacturer. The total cDNA in the skin was synthesized with the extracted RNA with a SMART™ cDNA Library Construction Kit (Takara biotechnology, Dalian, China) based on the protocol of the manufacturer and was used as the template for amplification of the cDNAs encoding brevinin-2MP with a combination of two oligonucleotide primers, HG (5′-AGA​TGA​AGA​AAT​CCC​TGT​TAC​TTC-3′) in the sense direction designed in light of the signal peptide of the reported precursor of brevinin-2GHa from Hylarana guentheri frog (Zhou et al., 2006) and CDS III (5′-ATT​CTA​GAG​GCC​GAG​GCG​GCC​GAC-3′) in the antisense direction in the PCR reaction. The PCR procedure was carried out in a mixture with Gene Tap polymerase (TianGen, Beijing, China), skin cDNA, and primers using a thermal cycler (Applied Biosystems, Foster, CA, United States). The PCR program was as follows: 4 min at 95°C; 30 cycles of 15 s at 94°C, 20 s at 46°C, and 20 s at 72°C; and, finally, at 72°C for 10 min for extension. The purified PCR product (about 250 bp) was cloned into pMD18-T vector (Takara biotechnology, Dalian, China) and sequenced by Applied Biosystems DNA analyzer, model 3730XL (Applied Biosystems, Foster, CA, United States).

The chemical and physical parameters of the polypeptide were obtained via the Bioinformatics Resource Portal (http://www.expasy.org/tools/). The structure prediction of brevinin-2MP was performed via the trRosetta algorithm (https://yanglab.nankai.edu.cn/trRosetta/) with the NMR structure of gaegurin 4 (PDB entry 2G9L) as a template on the basis of maximum similarity (Yang et al., 2020). The comparative 3D structure model of brevinin-2MP was generated and optimized with Modeller 9.12 (University of California San Francisco, San Francisco, CA, United States). The generated 3D structural model was visualized using PyMOL software (Schrödinger) without any refinement. As further confirmation, the secondary structure of brevinin-2MP was further predicted with the Jpred4 and SOPMA secondary structure prediction servers provided by the Division of Computational Biology, School of Life Sciences, University of Dundee and PBIL-IBCP of the Institute of Biology and Protein Chemistry.

Brevinin-2MP and brevinin-2MP labeled by FITC at the N terminal were synthesized by GL Biochem Ltd. (Shanghai, China). The crude brevinin-2MP peptide was purified by reversed-phase high-performance liquid chromatography on an inert silica gel ODS-SP column (Shimazu, Sumi, Japan) equilibrated with acetonitrile/water/trifluoroacetic acid at a flow rate of 1 ml/min. The peptide was eluted with 43% acetonitrile solution at 18 min. When the purity of the peptide, computed on the basis of the ratio of different front areas, was more than 95%, the peak was collected and lyophilized and further confirmed by MALDI-TOF mass spectrometry (Voyager-DE-PRO MALDI-TOF, AB SCIEX, Supplementary Figures S1, S2).

Circular dichroism (CD) was used to study the secondary structure of brevinin-2MP and to detect the changes of the secondary structure of brevinin-2MP in different solutions and temperatures. CD spectra were measured on Chirascan plus ACD spectropolarimeter (Applied Photophysics Ltd., Leatherhead, United Kingdom), which has a cell with 1 mm optical path length. The spectrum at 190–260 nm was determined at 25°C by a 0.1-cm-long cell with a width of 1 nm, scanning speed of 100 nm/min, and response time of 1 s. The samples were obtained by dissolving the brevinin-2MP powder in sodium dodecyl sulfate (SDS) solution (0, 30, 60, 90, and 120 mM) or 60 mM SDS solution containing a series of concentrations of sodium chloride (0, 100, 200, and 400 mM). For the temperature effect experiment, the peptide powder was dissolved in 60 mM SDS solution and incubated at different temperatures (25, 50, 70, and 90°C) for 1 h, and then the CD spectrum was measured. The concentration of peptide in all the above-mentioned samples was 100 μM. CD data are expressed as the mean residue ellipticity (θ) of three consecutive scans per sample in degree square centimeter per decimole. The following equation is used to calculate the mean residue ellipticity (θ, deg cm² dmol⁻¹): θ = ( θ obs × 1,000 ) ÷ ( c × l × n ) where θ_obs is the observed ellipticity (mdeg), c is the concentration (mM) of the peptide solution, l is the path length (mm), and n is the number of peptide residues.

The minimal inhibitory concentration (MIC) and the minimum bactericidal concentration (MBC) were measured to ensure the antimicrobial capacities of brevinin-2MP in 96-well plates, with a twofold dilution method as reported in our previous paper (Zeng et al., 2018). In brief, standard strains, including E. coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, S. aureus ATCC 25923, Propionibacterium acnes ATCC 6919, Bacillus subtilis CMCC 63501, and Candida albicans ATCC 10231, purchased from Guangdong Institute of Microbiology were grown to exponential phase at 37°C with Mueller–Hinton broth (MH broth) and diluted to 10⁶ CFUs/ml with fresh culture medium. Then, 50 μl/well of the resulting sample was then loaded to 96-well microplates which contained equal volumes of a serial dilution of brevinin-2MP with MH broth. After 16 h of gentle shaking while incubated at 37°C, the absorbance of the bacterial suspensions was read at 600 nm with a microplate reader (Tecan Trading AG, Männedorf, Switzerland). MIC was defined as the minimum concentration inhibiting visible growth. MBC was then determined following the MIC assay. Then, 10 μl of the sample which exhibited no evident growth after 16 h of incubation was inoculated onto MH agar plates. These plates were placed at 37°C for another 16 h. The MBC was defined as the peptide concentration at which there was no colony growth. In addition, the bacterial killing kinetics of brevinin-2MP against E. coli ATCC 25922 was carried out according to our previous method with minor modification (Zeng et al., 2018). Briefly, a final inoculum of 10⁶ CFUs/ml was grown in MH broth with brevinin-2MP (0×, 2×, and 4× MICs) at 37°C. Duplicate samples were withdrawn at 0, 15, 30, 60, 90, 120, 150, and 180 min before spreading on LB agar plates. Viable colonies were counted after the plates were placed for 16 h at 37°C. Ampicillin and sterile saline were used as the positive and negative controls, respectively. All experiments were repeated three times.

The cytotoxicity of brevinin-2MP on different mammal cells was examined by MTT. Three tumor cell lines (H460, M21, and MDA-MB-231) and two normal mammalian cells (RAW 264.7 macrophages and splenocytes) were inoculated in 96-well plates at a density of 5,000 cells per well and cultured in Dulbecco's modified Eagle’s medium (DMEM) or RPMI 1640 medium or in a medium containing continuous concentrations of brevinin-2MP (1.25, 2.5, 5, 10, and 20 μM) for 48 h at 37°C before the cell viabilities were measured with MTT methods as our reported methods (Chai et al., 2021). Hemolytic activity was also measured as described in our previous article with minor modification (Wu et al., 2021). In short, 2% of mouse erythrocyte suspension in Tris-buffered saline solution (v/v) was treated with different concentrations of brevinin-2MP (6.25, 12.5, 25, 50, and 100 μM) in a 96-V-well plate at room temperature for 2 h. Then, 1% Triton X-100 and phosphate-buffered saline (PBS) were applied as the positive and negative control, respectively. The absorbance of the supernatant at 540 nm was measured with a microplate spectrophotometer. The hemolysis ratio was counted with the following formula: percentage hemolysis = (OD_sample – OD_PBS) / (OD_Triton – OD_PBS) × 100%.

Both confocal laser scanning microscopy and scanning electron microscopy (SEM) experiments were performed to examine the potential antimicrobial mechanism of brevinin-2MP against the tested microorganisms according to our previous methods (Chai et al., 2021). In brief, E. coli and S. aureus at the late exponential phase were incubated with 1× MIC of brevinin-2MP at 37°C for 1.5 h. SYTO9 and propidium iodide (PI) from the LIVE/DEAD® BacLight kit (Invitrogen, Waltham, Massachusetts, United States) were incubated with the above-mentioned treated cells for 30 min at room temperature according to the instructions of the manufacturer. The stained cells were observed under confocal laser scanning microscopy (Leica TCS SP5; Leica Microsystems, Wetzlar, Germany) with emission/excitation wavelengths of 635/490 nm for PI and 500/480 nm for SYTO9, respectively. The living microbes with intact membrane were stained by SYTO9 emitting green fluorescence. However, the dead bacteria with broken membrane were stained by PI emitting red fluorescence. About 5–10 single-plane images per coverslip were obtained. For the SEM observation, bacterial suspensions at logarithmic growth phase (10⁶ CFUs/ml) were mixed with brevinin-2MP (2× MICs) for 30 min at 37°C. Then, the bacterial cells were harvested by centrifugation and sequentially fixed with 4% glutaraldehyde solution at room temperature for 4 h and 2.5% at 4°C overnight, respectively. After three times of washing with PBS, the bacteria were dehydrated sequentially with a series of concentration gradients of ethanol solution, followed by tert-butyl alcohol, and dried in a freeze-dryer (Quorum, UK). After gold coating, bacterial morphology was visualized by JSM-840 instrument (Hitachi, Japan) at a magnification of ×50,000; then, the bacterial morphology was captured at 5 kV voltage. A total of 5–10 single-plane images were obtained for each sample.

Membrane binding assays were undertaken with FITC-labeled brevinin-2MP according to our previous methods (Chai et al., 2021). In short, RAW 264.7 macrophages and bacterial cells in the logarithmic phase were prepared in PBS at a density of 1 × 10⁵ cells/ml and incubated with FITC-labeled brevinin-2MP (0, 2.5, 5, and 10 μM) for 15 min at 37°C. The unbound peptide was removed by washing with PBS containing 1% bovine serum albumin. Cell fluorescence intensity was detected with a FACscan flow cytometer (Becton Dickinson, United States). Cells without peptide treatment were regarded as the negative control.

RAW 264.7 macrophages were grown in DMEM medium with 10% FBS, 100 units per milliliter of ampicillin plus 100 μg/ml streptomycin, and seeded to 96-well plates at a density of 10⁵ cells per well. After cell adhesion, the cells were incubated with brevinin-2MP (0, 2.5, 5, and 10 μM) for 1 h before further incubation with LPS (100 ng/ml, E. coli O₅₅: B₅, L6529, Sigma Aldrich, St. Louis, Missouri, USA) in an incubator at 37°C for another 24 h. Then, the culture supernatants were collected to measure NO contents by Griess reagent (Beyotime Biotechnology, China), MCP-1, and IL-6 as well as TNF-α levels with enzyme-linked immunosorbent assay (Thermo Fisher Scientific, United States) in light of the manuals of the manufacturer. The cells without peptide and/or LPS incubation were considered as the negative control. Nitrite contents which reflect NO generation in medium were calculated from the standard curve obtained with NaNO₂ (Sigma-Aldrich, St. Louis, Missouri, United States). All experiments were repeated three times.

RAW 264.7 cells were seeded into six-well plates at 1 × 10⁶ cells per well and cultured in DMEM medium containing 10% FBS for 12 h. The cells were treated with different concentrations of brevinin-2MP (2.5, 5, and 10 μM) for 30 min and then stimulated with 100 ug/ml LPS for another 30 min. The cells lacking peptide and LPS treatment were considered as the negative control. All cells were collected by centrifugation and washed twice with pre-cooled PBS before the cytoplasmic or nuclear proteins were extracted using our previously published method for Western blot analysis (Nguyen et al., 2020). Primary antibodies against phospho-JNK/JNK, phosphor-ERK/ERK, phospho-p38/p38, NF-κB p65, Lamin A/C, and GAPDH (1: 1500; Cell Signaling Technology, Beverly, Massachusetts, United States) were applied in western blot analysis. All experiments were undertaken in triplicates.

The anti-inflammatory function of brevinin-2MP in vivo was measured with the carrageenan-induced mouse paws as reported previously by us (Deng et al., 2021). Briefly, male and female Kunming mice (20–22 g) were randomly divided into four groups, with six mice in each group. Before the experiment, the paw volume up to the ankle joint was measured with a plethysmometer (Taimeng PV-200 7500, China) before injecting them with 50 μl normal saline containing 1% carrageenan into the plantar side of the paw. The mice were intraperitoneally injected with brevinin-2MP (10 mg/kg), saline, or indomethacin (5 mg/kg) for 1 h. The paw volume up to the ankle joint was surveyed at 1, 2, 4, 8, and 24 h after the carrageenan administration. The increasing volume was calculated by the delta volume ( a − b ), where “a” and “b” are the volumes of the right hind paw after and before carrageenan administration, respectively. In the other experiments, the secondary batch of mice was injected sequentially with brevinin-2MP, indomethacin, saline, and carrageenan as mentioned above. The right hind paws of all mice were surgically cut at 4 h post-injection for histological examination and myeloperoxidase (MPO) activity analysis.

The statistical analysis of data was performed with Igor and GraphPad Prism 5.0 software (GraphPad Software Inc., La Jolla, CA, United States). One-way analysis of variance with Bonferroni’s multiple-Comparison test was carried out to judge the significance when comparing two or more groups with a control group and with a post-hoc Tukey test when making comparisons among three or more groups. The unpaired Student’s t-test was implemented to measure the significance between two experimental groups. Data are shown as mean ± SEM.

The cDNA sequence encoding AMP was obtained from M. pulchra frog skin gene library by PCR-based cloning technique. As displayed in Figure 1A, the predicted protein precursor consisted of 69 amino acid residues containing a 19-residue signal peptide, an acidic domain with 20 residues, and a mature peptide with 30 residues following the KR residues which form the classic protease cleavage site of frog defense peptide. NCBI Basic Local Alignment Search Tool (BLAST) analysis showed that the predicted precursor sequence highly resembled those of brevinin-2 peptides. Especially, they shared complete identity in the signal peptide and acidic domains of brevinin-2 peptides from Sylvirana guentheri . In addition, the sequence of mature peptide (GVITDTLKGVAKTVAAELLRKAHCKLTNSC) had only several residue differences with brevinin-2 peptides from S. guentheri ; for example, the former was only two amino acid different from brevinin-2GHb and brevinin-2GUb (Figure 1B) (Conlon et al., 2008a; Dong et al., 2017). Hence, we judged that this new peptide belonged to the brevinin-2 family and named it as brevinin-2MP. Brevinin-2MP possessed a theoretical PI of 9.31 with +3.01 net charge and aliphatic index of 107.33. The grand average of hydropathicity (GRAVY) of brevinin-2MP was predicted to be 0.210. Its relative mass without the intramolecular disulfide bond was predicted to be 3,141.71, which was nearly identical with the value shown in Supplementary Figure S2A from the mass spectrum result.

The secondary structure of brevinin-2MP was predicted with gaegurin 4 (PDB entry 2G9L) as the homologous template because the modeling credibility is very high, with a confidence of 96.3 and E-value of 0.00047 according to trRosetta algorithm search (Yang et al., 2020). Figure 2A suggests that the α-helix was the main conformation component in the predicted brevinin-2MP structure. In addition, residues G1-I3 and N28-C30 formed one coil at the N-terminal part and at the C-terminal part of the structure, respectively (Figure 2A). Furthermore, the major secondary structural components of brevinin-2MP predicted by both Jpred4 and SOPMA were also α-helix, which took 80.00 and 93.33% of the whole structure in the latter two methods, respectively (Figures 2B,C). In order to verify the prediction accuracy, the CD spectra of brevinin-2MP in different solutions were measured. As described in Figure 2D, the CD spectra of brevinin-2MP dissolved in H₂O presented a large negative peak at 205 nm and a small negative peak at 223 nm, which is the spectral characteristic of random coil. Subsequently, one positive peak at 195 nm and two negative peaks at 208 and 222 nm appeared in the CD spectra of brevinin-2MP dissolved in SDS solution, which suggested that the secondary conformation of brevinin-2MP was mainly composed of α-helix. The α-helix structure accounts for 96.42% of the mature peptide chain of brevinin-2MP, according to CDNN program calculation. In addition, the secondary structure of brevinin-2MP was hardly affected by SDS at the concentration range of 30–120 mM (Figure 2D). Similarly, the CD spectra of brevinin-2MP showed no significant difference after treatment at different temperatures, even at 90°C (Figure 2E). However, despite the fact that two negative peaks at 208 and 222 nm still existed with the increasing concentrations of sodium chloride in SDS solution, the positive peak at 195 nm disappeared, and some other peak appeared, which was mostly due to the severe signal distortion known to happen at high chloride concentrations (Figure 2F).

The antimicrobial properties of brevinin-2MP against Gram-negative and Gram-positive bacteria as well as fungi are shown in Table 1. Brevinin-2MP possessed antimicrobial activity against tested strains other than Pseudomonas aeruginosa. Among them, brevinin-2MP had the strongest activity against Bacillus subtilis CMCC 63501, with a MIC value of about 4.97 μM. What is more, the killing kinetics of brevinin-2MP against E. coli ATCC 25922 was studied by colony counting assay. At concentrations of 2× MICs, brevinin-2MP was able to completely eliminate all the microbes in 120 min, and E. coli ATCC 25922 could not restart growth on agar plates after treatment, which suggested that brevinin-2MP is bactericidal rather than bacteriostatic (Supplementary Table S1).

The attraction and attachment of AMPs to microbial cell surfaces are crucial to exert their antimicrobial activity (Vineeth and Sanil, 2017). Flow cytometry experiments were implemented to explore the binding of brevinin-2MP to bacteria. As described in Figure 3A, compared with the untreated group, the bacterial fluorescence intensities were enhanced in a concentration-dependent manner after E. coli ATCC 25922 and S. aureus ATCC 25923 were incubated with FITC-labeled brevinin-2MP for 15 min, indicating that brevinin-2MP could bind to them. It is well known that most AMPs exert their antibacterial activity by penetrating and destroying the integrity of the cell membrane. In order to determine the antibacterial mechanism of brevinin-2MP, both laser confocal scanning microscopy and scanning electron microscopy were applied. As shown in Figure 3B, only individual E. coli ATCC 25922 and S. aureus ATCC 25923 in the untreated control group were dyed red by PI, while most of the living bacteria were dyed green by STYO9. However, after treatment with brevinin-2MP of 1× MIC for 1 h, a large number of bacteria were dyed red, and the intensity ratio of PI/STYO9 was significantly increased, which indicated that brevinin-2MP, like most AMPs, obviously destroyed the integrity of the cell membrane and induced the inflow of PI into Gram-positive and Gram-negative bacteria. Consistently, the scanning electron microscopy images suggested that brevinin-2MP could damage the bacterial membrane. In comparison with the control group, the bacterial cell membrane treated with brevinin-2MP had an obvious expansion, deformation, and even a large amount of intracellular inclusion overflow (Figure 3C). All these results suggested that the cellular membrane permeabilization and disruption mechanism were responsible for the antimicrobial activity of brevinin-2MP.

As shown in Table 2, brevinin-2MP was a low hemolytic against mouse erythrocyte, and the hemolysis ratio was only 1.42 ± 0.16% even at the highest concentration of 100 uM. The effects of brevinin-2MP on the proliferation of splenocytes, RAW 264.7, H460, M21, and MDA-MB-231 cells were measured by the MTT assay. As reported in Table 3, H460 is the most sensitive to brevinin-2MP with a IC₅₀ value of about 5.77 ± 1.21 μM after 48 h of exposure. However, brevinin-2MP showed low cytotoxicity to other tested mammalian cell lines, with IC₅₀ values of more than 25 μM (Table 3).

The binding of AMPs to their targets on the surface of macrophages sets off cellular signaling pathway and regulates the secretion of pro-inflammatory factors. Therefore, the binding of brevinin-2MP to RAW 264.7 cells was firstly evaluated with flow cytometry. As shown in Figure 4A, similar with its binding with bacteria, after the FITC-labeled brevinin-2MP and RAW 264.7 cells were co-incubated for 15 min, obvious binding phenomenon occurred, and their binding effects were enhanced with the increased concentrations of brevinin-2MP. To define whether brevinin-2MP can affect the release of inflammatory cytokines in RAW 264.7 macrophages stimulated by LPS, we tested secondly the effect of brevinin-2MP on the viability of RAW 264.7 cells. As shown in Figure 4B, brevinin-2MP at the tested concentrations had no cytotoxicity toward RAW 264.7 cells. Thirdly, the contents of NO, TNF-α, and IL-6 in RAW 264.7 cells were investigated as illustrated in Figures 4C–F. Compared with RAW 264.7 cells without LPS stimulation, RAW 264.7 cells stimulated with 100 ng/ml LPS significantly increased the release of NO, MCP-1, IL-6, and TNF-α, while their increase trends were inhibited by pre-incubation with brevinin-2MP for 1 h before LPS stimulation. The inhibition rates of 2.5, 5, and 10 μM brevinin-2MP on these factors increased in RAW 264.7 cells to about 14–54%, 18–75%, 15–54%, and 10–57%, respectively.

It is well known that inflammatory response is closely related to the MARK/NF-κB signaling pathways (Akira et al., 2006). Thus, the expression of MARK/NF-κB signal proteins, like p38, JNK, ERK, and p65, was examined by western blot. As shown in Figures 5A,B, the contents of the p65 subunit translocated into the nucleus and phosphorylated p38 and ERK as well as JNK in 100 ng/ml LPS induced cells were significantly higher than those in normal cells, but these upregulations were inhibited in a concentration-dependent manner by brevinin-2MP which did not change the total protein levels of ERK, JNK, and p38. Specifically, after the administration of 10 μM brevinin-2MP, the contents of phosphorylated JNK, ERK, p38, and p65 in the nucleus were reduced by 11.54, 59.17, 88.14, and 53.26%, respectively. These results suggested that brevinin-2MP exerted its anti-inflammatory effect by inactivating the MAPK/NF-κB pathways in LPS-sensitized macrophage cells.

The effect of brevinin-2MP on acute inflammation was evaluated by a carrageenan-stimulated paw edema assay. As shown in Figures 6A,B, the swelling degree of the paw reached a maximum at 4 h after the carrageenan injection into the right paw. Compared with the model group, the swelling rates of the right paw injected with indomethacin and brevinin-2MP at 4 h were decreased by about 63.85 and 46.38%, respectively. At 24 h after injection, their swelling rates were reduced by 47.94 and 31.50%, respectively. The increasing MPO activity indicates neutrophil infiltration. As shown in Figure 6C, the MPO activity was 9.09 ± 0.19 units/mg protein in the control group, while it was significantly increased by carrageenan administration and reached 33.16 ± 3.77 units/mg protein in the model group only injected with carrageenan. This increase induced by carrageenan was obviously suppressed by indomethacin and brevinin-2MP, and the MPO activity in their treatment groups was 23.13 ± 0.53 and 23.15 ± 0.50 units/mg protein, respectively. Coincidentally, the histological results showed that subcutaneous edema and inflammatory cell infiltration appeared in all carrageenan-injected paws, while these were less intense in the groups injected with indomethacin or brevinin-2MP when compared to the group injected only by carrageenan (Figure 6D).