A non-bactericidal cathelicidin provides prophylactic efficacy against bacterial infection by driving phagocyte influx

The roles of bactericidal cathelicidins against bacterial infection have been extensively studied. However, the antibacterial property and mechanism of action of non-bactericidal cathelicidins are rarely known. Herein, a novel naturally occurring cathelicidin (PopuCATH) from tree frog (Polypedates puerensis) did not show any direct anti-bacterial activity in vitro. Intriguingly, intraperitoneal injection of PopuCATH before bacterial inoculation significantly reduced the bacterial load in tree frogs and mice, and reduced the inflammatory response induced by bacterial inoculation in mice. PopuCATH pretreatment also increased the survival rates of septic mice induced by a lethal dose of bacterial inoculation or cecal ligation and puncture (CLP). Intraperitoneal injection of PopuCATH significantly drove the leukocyte influx in both frogs and mice. In mice, PopuCATH rapidly drove neutrophil, monocyte/macrophage influx in mouse abdominal cavity and peripheral blood with a negligible impact on T and B lymphocytes, and neutrophils, monocytes/macrophages, but not T and B lymphocytes, were required for the preventive efficacy of PopuCATH. PopuCATH did not directly act as chemoattractant for phagocytes, but PopuCATH obviously drove phagocyte migration when it was cultured with macrophages. PopuCATH significantly elicited chemokine/cytokine production in macrophages through activating p38/ERK mitogen-activated protein kinases (MAPKs) and NF-κB p65. PopuCATH markedly enhanced neutrophil phagocytosis via promoting the release of neutrophil extracellular traps (NETs). Additionally, PopuCATH showed low side effects both in vitro and in vivo. Collectively, PopuCATH acts as a host-based immune defense regulator that provides prophylactic efficacy against bacterial infection without direct antimicrobial effects. Our findings reveal a non-bactericidal cathelicidin which possesses unique anti-bacterial action, and highlight the potential of PopuCATH to prevent bacterial infection.


Introduction
Antimicrobial peptides (AMPs) are a wide array of gene-encoded small defensive molecules that have been identified from prokaryotic to eukaryotic kingdoms, including bacteria, fungi, plantae, and animalia (Mygind et al., 2005;Radek and Gallo, 2007;Silva et al., 2014; Gallo, Yang

Results
A novel naturally occurring cathelicidin was identified from the skin of tree frog, P. puerensis To understand the peptidomics of P. puerensis skin, the skin secretions were firstly separated by molecular sieving fast pressure liquid chromatography (FPLC) as indicated in Figure 1-figure supplement 1A. The eluted peak containing the objective peptide in this study (marked by an arrow) was further purified by a reversed-phase high-performance liquid chromatography (RP-HPLC) C18 column for two times (Figure 1-figure supplement 1B and C, marked by an arrow). The purified peptide exhibited an observed molecular weight of 4295.9 Da (Figure 1-figure supplement 1D). Then, a total of 16 amino acids at N-terminus were determined as SRGGRGGRGGGGSRGG by automated Edman degradation. The N-terminus is enriched in glycine residues, which is possibly a novel member of cathelicidin antimicrobial peptides like those glycine-rich cathelicidins found in frog (Hao et al., 2012) and fish (Broekman et al., 2011).    According to this implication, we designed primer based on the conserved region of amphibian cathelicidins to clone the gene encoding the objective peptide. The nucleotide sequence (GenBank accession number: KY391886) encoding the precursor of the objective peptide was cloned from the skin cDNA library (Figure 1). The coding sequence of the precursor included 659 nucleotides that encodes a precursor containing 179 amino acid residues (Figure 1). The full-length amino acid sequence of the mature peptide (designated as PopuCATH) was determined as shown in Figure 1. BLAST comparison confirmed that the precursor of PopuCATH is definitely a novel member of cathelicidin antimicrobial peptide family, which shares a highly conserved signal peptide and cathelin domain at N-terminus with amphibian cathelicidins (Figure 1-figure supplement 2A). Phylogenetic tree analysis indicated that PopuCATH combined with amphibian and fish cathelicidins form the second cluster, showing close evolutionary relationship with amphibian cathelicidins and fish cathelicidins (Figure 1-figure supplement 3).
Primary structural analysis indicated that PopuCATH is composed of 46 amino acid residues, including 41 polar residues and 5 non-polar residues, which is a glycine-rich cathelicidin (21 glycine residues) like those found in frog and fish ( PopuCATH lacks direct antimicrobial activity but can prevent bacterial infection in tree frogs Cathelicidins were initially described for their direct antimicrobial activity (Gennaro et al., 1989). Therefore, we first detected the direct antimicrobial activity of PopuCATH in vitro by MIC assay. To our surprise, PopuCATH didn't show any antimicrobial activity against the tested bacteria (a total of 40 strains) at the concentration up to 200 µg/mL, including Gram-negative bacteria, Gram-positive bacteria, fungi, and aquatic pathogenic bacteria (Supplementary file 3). Similarly, in time-kill assays, 200 µg/mL of PopuCATH did not reduce the CFUs of E. coli, S. aureus, C. albicans, and A. hydrophila after incubation for 1, 2, 3, and 4 hr, respectively ( Figure 2A). Furthermore, 200 µg/mL of PopuCATH did not alter bacterial metabolic activity during the exponential growth phase of E. coli, S. aureus, C. albicans, and A. hydrophila after incubation for 1, 2, 3 and 4 hr, respectively ( Figure 2B). Cathelicidins are usually membrane-active agents which can alter the surface morphology of bacteria (Wei et al., 2013). As shown in Figure 2C, 200 µg/mL of PopuCATH did not alter the surface morphology of E. coli and S. aureus after PopuCATH treatment. While the positive control peptide PY (1× MIC, cathelicidin-PY), a previously described amphibian cathelicidin from P. yunnanensis (Wei et al., 2013) markedly inhibited bacterial growth (Supplementary file 3), showed bactericidal activity (Figure 2A), reduced bacterial metabolic activity ( Figure 2B), and altered bacterial surface morphology ( Figure 2C). These results indicated that PopuCATH lacks direct antimicrobial activity.
In order to understand whether PopuCATH has antimicrobial activity in vivo, PopuCATH (10 mg/ kg) was intraperitoneally injected into P. puerensis 8 hr, or 4 hr prior to (-8 hr or -4 hr), or 4 hr after ( + 4 hr) intraperitoneal bacterial inoculation, and the bacterial load was recorded. Compared to PBS treatment, PopuCATH (10 mg/kg) treatment at 8 hr or 4 hr before bacterial inoculation significantly reduced the bacterial load in tree frogs, but PopuCATH (10 mg/kg) treatment at 4 hr after bacterial inoculation did not significantly reduce the bacterial load ( Figure 2D), indicating that pretreatment with PopuCATH significantly prevented bacterial infection in tree frogs.

PopuCATH exhibits low toxic side effects to mammalian cells and mice
In order to further investigate the mechanism of action of PopuCATH against bacterial infection in vivo, it was necessary to move from a frog system to a mouse system. Thus, the toxicity of Popu-CATH to mammalian cells and mice were evaluated. At concentration up to 200 μg/mL, PopuCATH didn't show any cytotoxicity to mouse peritoneal macrophages and humane monocyte THP-1 cells ( Figure 3A), and didn't show any hemolytic activity to mouse erythrocytes and rabbit erythrocytes ( Figure 3B). Yang The immunogenicity of PopuCATH was evaluated by determination of its proliferative capacity on mouse lymphocytes isolated from mesenteric lymph node (MLN) and spleen. As shown in Figure 3C, PopuCATH did not induce proliferation of lymphocytes isolated from mouse MLN and spleen at any dose tested up to 200 μg/mL, unlike the positive control ConA.
Hypersensitivity to PopuCATH was assessed in a mast cell degranulation assay. Mast cells modulate immediate hypersensitivity reactions and nonspecific inflammatory reactions (Scott et al., 2007). PopuCATH did not induce mast cell degranulation at any dose tested up to 200 mg/mL ( Figure 3D)  aureus, and C. albicans were diluted in Mueller-Hinton broth, and A. hydrophila were diluted in nutrient broth at density of 10 5 CFU/mL. PopuCATH (200 μg/mL), PY (cathelicidin-PY, 1× MIC, positive control) or PBS was added and incubated at 37℃ or 25℃. At indicated time points, the CFUs were counted. (B) Microbial metabolic activity assay. E. coli, S. aureus, C. albicans, and A. hydrophila were diluted in DMEM at density of 10 5 CFU/mL, and PopuCATH (200 μg/mL), PY (cathelicidin-PY, 1× MIC, positive control) or PBS was added. Microbial dilution (100 μL/well) and WST-8 (10 μL/well) was added to 96-well plates, respectively, and incubated at 37℃ or 25℃. At indicated time points, absorbance at 255 nm was monitored. Metabolic activity was expressed as the percentage of the PBS-treated group. (C) SEM assay. E. coli ATCC25922 and S. aureus ATCC25923 were washed and diluted in PBS (10 5 CFU/mL). PopuCATH (200 μg/mL), PY (1× MIC, positive control) or PBS was added into the bacterial dilution and incubated at 37℃. After incubation for 30 min, bacteria were centrifuged at 1000 g for 10 min, and fixed for SEM assay. The bacterial surface morphology was observed using a Hitachi SU8010 SEM. (D) Anti-bacterial activity in tree frogs. PopuCATH (10 mg/kg) was intraperitoneally injected into P. puerensis (n = 5, 21-30 g) at 8 or 4 hr prior to (-8 or -4 hr), or 4 hr after ( + 4 hr) S. aureus ATCC25923 inoculation (10 8 CFU/frog, intraperitoneal injection). At 18 hr post bacterial challenge, peritoneal lavage was collected for bacterial load assay. **p < 0.01, ***p < 0.001, ns, not significant. while human cathelicidin LL-37 (positive control) markedly induced mast cell degranulation as described previously (Niyonsaba et al., 2001). The effect of PopuCATH on complement activation was determined by measurement of C3a after incubation of PopuCATH with mouse serum. PopuCATH didn't exhibit significant effect on complement activation at concentration up to 200 μg/mL, in contrast to the controls of ethylene glycol tetraacetic acid (EGTA) and zymosan ( Figure 3E).
The highest amount of administered substance that does not kill tested animals was recorded as the maximum tolerable dose (Scott et al., 2007). The maximum tolerable dose of PopuCATH was tested in C57BL/6 mice by intravenous and intraperitoneal delivery. The maximum tolerable dose of PopuCATH by intravenous delivery was between 75 and 100 mg/kg, and intraperitoneal delivery was between 125 and 150 mg/kg. Markedly, PopuCATH was not toxic at substantially higher concentrations via the intraperitoneal route, well above the doses (10, 20, and 40 mg/kg) used in the mouse models.

PopuCATH provides prophylactic efficacy against bacterial infection in mice
We next examined whether PopuCATH protects against bacterial infection in mice as observed in tree frogs. Mice were intraperitoneally injected with PopuCATH at 8 or 4 hr (-8 or -4 hr) prior to, or 4 hr after ( + 4 hr) intraperitoneal inoculation of bacteria. Compared to PBS injection, PopuCATH injection at -8 or -4 hr significantly reduced the bacterial loads in the abdominal cavity of mice post Gramnegative bacteria (E. coli, A. baumannii) and Gram-positive bacteria (S. aureus, methicillin-resistant S. aureus, MRSA) inoculation ( Figure 4A). Bacterial inoculation significantly elicited the production of pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) in mouse serum relative to control mice (sham), while PopuCATH injection at -8 or -4 hr significantly reduced the production of pro-inflammatory cytokines in mouse serum ( Figure 4B). Consistent with these findings, E. coli or S. aureus inoculation markedly induced inflammatory damage in the lung, and PopuCATH injection at -4 or -8 hr obviously rescued this inflammatory damage induced by bacterial inoculation ( Figure 4C), suggesting its prophylactic efficacy against bacterial infection, and PopuCATH (10 mg/kg) also showed prophylactic efficacy against bacterial infection via intravenous injection ( Figure 4-figure supplement 1). However, at the dose of 10 mg/kg, PopuCATH treatment at +4 hr did not significantly reduce the bacterial loads in abdominal cavity, the production of pro-inflammatory cytokines in serum, and the inflammatory damage in lung as compared to control mice (sham, Figure 4A-C).
In order to further evaluate the prophylactic efficacy of PopuCATH against bacterial infection, mice were intraperitoneally injected with PopuCATH before a lethal dose of E. coli or MRSA inoculation, and the survival rates of mice were monitored for up to 7 days. Compared to PBS treatment, Popu-CATH pretreatment markedly increased the survival rates of mice challenged by a lethal dose of E. coli or MRSA ( Figure 4D). Besides, PopuCATH exhibited a better prophylactic efficacy than those of LL-37 (human cathelicidin) and IDR-1 (bovine cathelicidin derivative) against a lethal dose of bacterial infection ( Figure 4D). We then evaluated the protective efficacy of PopuCATH in a CLP-induced sepsis model. We found that PopuCATH pretreatment markedly increased the survival rate of mice against CLP-induced sepsis ( Figure 4E). These data suggested that PopuCATH (10 mg/kg) pretreatment effectively provided prophylactic efficacy against bacterial infection and prevented sepsis induced by a lethal dose of bacterial inoculation or CLP in mice.
The online version of this article includes the following figure supplement(s) for figure 4:     4), suggesting that chemotaxis observed in mouse peritoneal cavity and peripheral blood were specifically due to PopuCATH rather than possible endotoxin contamination. In addition, we assayed if PopuCATH induce leukocyte influx in tree frogs. As shown in Figure 5-figure supplement 5, an intraperitoneal injection of PopuCATH obviously recruited leukocytes to the abdominal cavity of tree frogs, which is consistent with the data observed in mice.

Neutrophils and monocytes/macrophages, but not T and B cells, are required for the protective efficacy of PopuCATH in mice
PopuCATH was chemotactic to neutrophils and monocytes/macrophages in both mouse abdominal cavity and peripheral blood. To examine whether the protective capacity of PopuCATH depends on these phagocytic cell types, we evaluated the prophylactic efficacy of PopuCATH in mice after     the neutrophils or monocytes/macrophages were depleted by anti-Ly6G or anti-CSF1R antibody ( Figure 6-figure supplement 1). As shown in Figure 6, PopuCATH failed to provide prophylactic efficacy against E. coli ( Figure 6A) and S. aureus ( Figure 6B) infection in neutrophil-depleted mice, and PopuCATH was not efficacious against E. coli ( Figure 6C) and S. aureus ( Figure 6D) infection in monocyte/macrophage-depleted mice. As mentioned above, PopuCATH was primarily chemotactic to myeloid cells with a negligible impact on lymphoid cells. To confirm this finding, we next tested its prophylactic efficacy in Rag1 -/mice, which are T and B lymphocyte-deficient mice. In contrast to neutrophil and monocyte/macrophage depletion, PopuCATH still provided prophylactic efficacy Protective efficacy of PopuCATH against E. coli (A) or S. aureus (B) in neutrophil depletion mice. Anti-Ly6G antibody or rat IgG2a isotype antibody were intraperitoneally injected into C57BL/6 mice (18-20 g, n = 6) at doses of 500 µg/mouse on day 0 and day 2, respectively. (C, D) Protective efficacy of PopuCATH against E. coli (C) or S. aureus (D) in monocyte/macrophage depletion mice. Anti-CSF1R antibody or rat IgG2a isotype antibody were intraperitoneally injected into C57BL/6 mice (18-20 g, n = 6) at doses of 1 mg/mouse on day 0 followed by 0.3 mg/mouse on day 1 and day 2, respectively. (E, F) Protective efficacy of PopuCATH against E. coli (E) or S. aureus (F) in Rag1 -/mice (18-20 g, n = 6). At 4 hr before E. coli or S. aureus (2 × 10 7 CFUs/mouse) inoculation, PopuCATH (10 mg/kg) was intraperitoneally injected into neutrophil depletion mice (on day 3), monocyte/macrophage depletion mice (on day 3), and Rag1 -/mice. At 18 hr post bacterial inoculation, peritoneal lavage was collected for the bacterial load assay. ***p < 0.001, ns, not significant.
The online version of this article includes the following figure supplement(s) for figure 6: against E. coli ( Figure 6E) and S. aureus ( Figure 6F) infection in Rag1 -/mice. These data suggested that myeloid cells, but not lymphoid cells, are required for PopuCATH-mediated protection against bacterial infection in mice.

PopuCATH-induced phagocyte migration relies on its effect on macrophages
Given the increase in neutrophils and monocytes/macrophages in the abdominal cavity and peripheral blood, we were interested to investigate if PopuCATH acts as a chemoattractant for neutrophils and macrophages. As shown in Figure 7, PopuCATH (10 μM) did not directly induce neutrophil migration ( Figure 7A) and macrophage migration ( Figure 7B), suggesting that PopuCATH cannot act as a chemoattractant for neutrophils and macrophages. Macrophages have been shown to produce chemokines/ cytokines that recruit other cells, and macrophages are the major immune cells in mouse abdominal cavity (Scott et al., 2007;Yang et al., 2021). We next investigated whether PopuCATH induce phagocyte migration in the presence of macrophages. As shown in Figure 7, PopuCATH (10 μM) markedly induced neutrophil migration ( Figure 7A) and macrophage migration ( Figure 7B) in the presence of peritoneal macrophages. The addition of PopuCATH (10 μM) in the lower chamber elicited about 2.3 × 10 5 neutrophil migration ( Figure 7A) and 2.0 × 10 3 macrophage migration ( Figure 7B) when peritoneal macrophages were cultured in the lower chamber, implying that PopuCATH-induced phagocyte migration might rely on PopuCATH-triggered immune response in macrophages.

PopuCATH selectively induced the production of chemokines/cytokines in macrophages and mice
To confirm whether PopuCATH-induced phagocyte migration relies on PopuCATH-triggered immune response in macrophages, we stimulated mouse peritoneal macrophages with a single dose of PopuCATH (10 μM) for 4 hr and analysed the mRNA levels of chemokines/cytokines. As shown in Figure 8A, the mRNA levels of Cxcl1, Cxcl2, Cxcl3, Il1b, and Il6 were significantly increased by 60.9-, 74.2-, 17.0-, 15.5-, and 18.6-fold in peritoneal macrophages post PopuCATH treatment relative to PBS . PopuCATH-induced phagocyte migration relies on its effect on macrophages. For the direct chemotactic effect of PopuCATH to neutrophils or macrophages, 100 µL of neutrophil suspension (A) or macrophage suspension (B) (5 × 10 6 cells/mL) was added to the upper chamber, and 500 µL of PopuCATH (10 µM, dissolved in medium) or medium was added to the lower chamber. After neutrophils and macrophages were migrated at 37 ℃ for 8 hr, the increased cells in the lower chamber were collected and counted using a hemocytometer. For the co-cultured system, 500 µL of macrophage suspension (5 × 10 6 cells/mL) was seeded in the lower chamber. After macrophages were adherent to the lower chamber, 100 µL of neutrophil suspension (A) or macrophage suspension (B) (5 × 10 6 cells/mL) was added to the upper chamber. Then, the medium in the lower chamber was replaced with 500 µL of PopuCATH (10 µM, dissolved in medium) or fresh medium. Neutrophils and macrophages were migrated at 37 ℃ for 8 hr. The reduced cells in the upper chamber were counted using a hemocytometer. *p< 0.05, ***p < 0.001, ns, not significant.  Ccl5,Ccl6,Ccl8,Ccl9,Ccl10,Ccl14,Csf1,Il4,Il12, and Tnfa were slightly upregulated in peritoneal macrophages post PopuCATH treatment relative to PBS treatment, ranging from 1.5-to 3-fold (p > 0.05). We next stimulated mouse peritoneal macrophages with different dose of PopuCATH (5, 10, and 20 μM) to verify the results observed in Figure 8A. As shown in Figure 8B, mRNA levels of chemokines (Cxcl1, Cxcl2, and Cxcl3) and cytokines (Il1b and Il6) were significantly upregulated in a dose-dependent manner (p < 0.05). The others didn't generate a dose-dependent effect. To confirm the results observed by mRNA quantification, we detected the protein levels of the upregulated chemokines/cytokines by ELISA. PopuCATH significantly induced the protein production of CXCL1, CXCL2, and CXCL3 in a dose-dependent manner, whereas PopuCATH did not significantly induce the protein production of CXCL1, TNF-α, IL-1β, and IL-6 although their mRNA levels were upregulated ( Figure 8C). In vivo assay showed that an intraperitoneal injection of PopuCATH (10 mg/kg) significantly induced the production of the chemokines (CXCL1, CXCL2, and CXCL3) as well as the proinflammatory cytokines (IL-1β and IL-6) in mouse abdominal cavity ( Figure 8D). The results indicated that PopuCATH directly acted on macrophages and selectively induced the production of chemoattractant which are critical for the recruitment of phagocytes.
PopuCATH-induced chemokine production in macrophages were partially dependent on p38/ERK MAPKs and NF-κB signaling pathways To investigate the signaling pathways by which chemokines were induced by PopuCATH in macrophages, mouse peritoneal macrophages were pretreated with various inhibitors, including p38, ERK1/2, JNK1/2, PI3K and NF-κB, and responses induced by PopuCATH were analysed. As shown in Figure 9A, chemokines (CXCL1, CXCL2, and CXCL3) induced by PopuCATH were markedly attenuated after p38/ERK MAPKs, or NF-κB blockade, whereas inhibitors of JNK MAPK and PI3K pathway had no significant effect on PopuCATH-induced chemokine production in macrophages. Consistent with these results, PopuCATH (10 μM) significantly activated p38/ERK MAPKs and NF-κB p65 ( Figure 9B&C). But inhibition of p38/ERK MAPKs or NF-κB signaling pathways did not completely blocked PopuCATH-mediated chemokine production in macrophages, we cannot not exclude other possible signaling pathways were involved. The data suggested that PopuCATH-mediated chemokine production in macrophages partially depended on p38/ERK MAPKs and NF-κB signaling pathways.

PopuCATH promoted neutrophil phagocytosis through eliciting neutrophil extracellular traps
It is noted that PopuCATH primarily drove neutrophil influx in both peritoneal cavity and peripheral blood, and peaked at 4 hr post intraperitoneal injection of PopuCATH with an increment of approximately 1.54 × 10 5 neutrophils in mouse abdominal cavity ( Figure 5A&B) and 5.64 × 10 6 neutrophils in mouse peripheral blood ( Figure 5C&D) relative to control mice (sham), indicating that neutrophils exhibited a rapid response to PopuCATH. We herein tried to understand whether PopuCATH directly act on neutrophils to promote bacterial clearance. As illustrated in Figure 10A, PopuCATH significantly promoted phagocytic uptake of bacterial particles by mouse neutrophils. To investigate the mechanism by which PopuCATH promoted the phagocytic activity of neutrophils, the capacity of PopuCATH to induce neutrophil extracellular traps (NETs) were detected as indicated in Figure 10B. Single treatment of PopuCATH or PMA (positive control) markedly induced the formation of NETs as in PBS) (B), or PBS was added. After incubation at 37℃ for 4 hr, cells were collected, and the mRNA levels of chemokines/cytokines were detected by qPCR analysis, the protein levels of chemokines/cytokines were quantified by ELISA (C). (D) The protein levels of chemokine/cytokines in mice induced by PopuCATH (10 mg/kg). C57BL/6 mice (18-20 g, n = 6) were intraperitoneally injected with PopuCATH (10 mg/kg) dissolved in 0.2 mL PBS. Sham mice received the same volumes of PBS. At 4 and 8 hr post injection, peritoneal lavage was collected for quantification of the protein levels of chemokines/ cytokines by ELISA. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant.
The online version of this article includes the following figure supplement(s) for figure 8:
The online version of this article includes the following source data for figure 9: Source data 1. The original images of the unedited blots and images with the uncropped blots with the relevant bands clearly labelled.  inoculation significantly attenuated the bacterial load in tree frogs and mice, reduced inflammatory responses induced by bacterial inoculation in mice, and increased the survival rates of septic mice induced by lethal dose of bacterial inoculation and CLP. Except for the intraperitoneal injection, intravenous injection of PopuCATH also effectively provides prophylactic efficacy against bacterial infection ( Figure 4-figure supplement 1). The results indicates that PopuCATH can provide preventive capacity via both intraperitoneal and intravenous injection routes. While intramuscular injection had no significant preventive effects against bacterial infection (Figure 4-figure supplement 1). It is more likely that muscle is not rich in neutrophils, monocytes/macrophages, which are the key effector cells of PopuCATH. At the dose of 10 mg/kg, PopuCATH did not exhibit therapeutic efficacy against bacterial infection. In order to evaluate whether it has therapeutic efficacy at high doses, we increased the dose of PopuCATH. At doses of 20 and 40 mg/kg, PopuCATH significantly reduced the bacterial load when it was given at 4 hr after E. coli inoculation (Figure 4-figure supplement 2). It is possibly that bacteria have colonised in mice at 4 hr after bacterial inoculation, which need a higher dose of PopuCATH to drive more phagocytes for bacterial clearance. To the best of our knowledge, this is the first report of a non-bactericidal cathelicidin that can protect bacterial infection in vivo. The extent of microbial infection-mediated host damage largely depends on host's immune status. If the host can effectively initiate an immune defense, the invading microbes will be cleared, and host damage induced by microbial infection will be prevented or controlled. On the contrary, if the host cannot effectively initiate an immune defense, host will lose a balanced protection and microbial infection-mediated host damage will follow (Silva, 2010). Neutrophils and macrophages are two professional phagocytic cell types, which comprise a myeloid phagocyte system of host. Neutrophils and macrophages usually work together in innate immunity as complementary partners of the myeloid phagocytic system. The local and global distribution patterns of neutrophils and macrophages are key immune parameter of host, which play critical roles in initiating effective immune defense against invading microbes. Our findings revealed that intraperitoneal injection of PopuCATH effectively elicited neutrophil and monocyte/macrophage influx in mice, and depletion of neutrophils or monocytes/macrophages blocked PopuCATH-mediated protection, indicating that PopuCATH-mediated protection depends on PopuCATH-induced neutrophil and monocyte/macrophage influx. Previous investigations have demonstrated that successful clearance of invading microbes largely depends on efficient migration of these cell types into the infectious sites (Alves-Filho et al., 2010;Li et al., 2013;Nathan, 2006;Scott et al., 2007). These suggest that intraperitoneal injection of PopuCATH enhanced the myeloid phagocytic system of mice, thus providing prophylactic efficacy against bacterial infection.
Macrophages have been shown to phagocytose and directly kill bacteria Scott et al., 2007). In our study, we found that PopuCATH did not promote in vitro phagocytosis of fluorescently labelled bacterial particles by mouse peritoneal macrophages (Figure 8-figure supplement 2), suggesting no direct stimulation of the phagocytic activity of macrophages by PopuCATH. But PopuCATH significantly exhibited immunomodulatory effects on macrophages to induce phagocyte influx. In addition, neutrophils are principal phagocytes in the innate defense system and kill pathogens through mechanisms like oxidative killing activity and release of neutrophil extracellular traps (Neumann et al., 2014;Niyonsaba et al., 2013;Rowe-Magnus et al., 2019), and an influx of neutrophils to the site of infection is pivotal for the clearance of infectious bacteria (Alves-Filho et al., 2010). In our study, PopuCATH was merely demonstrated to promote neutrophil phagocytosis through inducing NET formation, but not significantly elicited oxidative killing activity of neutrophils ( Figure 10-figure supplement 2). These results indicated that macrophages and neutrophils responded to PopuCATH in their own manner.
The expression profiles of chemokines/cytokines in vitro and in vivo are somewhat different, but we observed that the profiles of the major chemokines/cytokines induced by PopuCATH, such as CXCL1, CXCL2, and CXCL3, are similar. We presumed that PopuCATH-induced chemokines/cytokines in vivo are consumed timely. In addition, macrophage is the unique effector cell type of PopuCATH in vitro. While there are many other cell types in vivo, such as monocytes and neutrophils, and we cannot exclude these cells are responsive to PopuCATH and subsequently produce chemokines/cytokines. These may explain the subtle differences of PopuCATH-mediated chemokine/cytokine production in macrophages and mice. In mouse model, the production of chemokines/cytokines in mouse abdominal cavity peaked at 4 hr post injection of PopuCATH. The dynamic of CXCL1, CXCL2, and CXCL3 production is consistent with the dynamic of neutrophil, monocyte/macrophage recruitment in the mouse abdominal cavity and peripheral blood. Although the pretreatment with PopuCATH significantly induced the production of chemokines (CXCL1, CXCL2, and CXCL3) as well as pro-inflammatory cytokines (IL-1β and IL-6), PopuCATH ultimately attenuated the inflammatory response by decrease of TNF-α, IL-1β, and IL-6 levels post Gram-negative and Gram-positive bacterial infection. Cathelicidins are able to block Toll-like receptor (TLR)-mediated inflammatory responses, including those mediated by TLR2 and TLR4 (Coorens et al., 2017;Mookherjee et al., 2006;Wei et al., 2013). In this study, PopuCATH did not affect LTA-and LPS-stimulated inflammatory responses in mouse peritoneal macrophages (Figure 8-figure supplement 3), suggesting that the anti-inflammatory effects of PopuCATH are independent of TLRs, and the attenuation of the inflammation may be secondary to the decrease of bacterial growth.
Some of the properties of PopuCATH are reminiscent of the activities of other cathelicidins like LL-37 (Chen et al., 2000), CRAMP (Kurosaka et al., 2005), and OH-CATH30 (Li et al., 2013), which selectively modulated innate immune responses and have been proposed to mediate protection in animal models. Compared to theses bactericidal cathelicidins with immunomodulatory properties, (i) the usage of PopuCATH is unlikely to induce drug-resistance because the peptide is unable to directly elicit stress on microbes. (ii) PopuCATH showed low side effects unlike LL-37 (Bąbolewska and Brzezińska-Błaszczyk, 2015). (iii) The expression profile of chemokines/cytokines in response to PopuCATH were largely different from those of other cathelicidins. (iv) Key effector cells for Popu-CATH were also largely different from those of LL-37, CRAMP and OH-CATH30. For example, human cathelicidin peptide LL-37 has been shown to directly recruit neutrophils, monocytes, mast cells, and T lymphocytes (Chen et al., 2000;Niyonsaba et al., 2002). While PopuCATH did not directly recruited leukocytes, it just recruited neutrophils and monocytes/macrophages via inducing chemokine/cytokine production in macrophages. In addition, PopuCATH just elicited neutrophil and monocyte/macrophage recruitment, but not T and B lymphocytes. Intraperitoneal injection of PopuCATH significantly drove phagocyte influx in both abdominal cavity and peripheral blood, demonstrating that it effectively regulated both local and global innate immune response. As a result, PopuCATH pretreatment effectively reduced the bacterial load in both abdominal cavity and peripheral blood ( Figure 4-figure supplement 3). (v) Intriguingly, PopuCATH is a glycine-rich cathelicidin containing 21 glycine residues. The amino acid component is different from LL-37, CRAMP, and OH-CATH30, which are not special residue-rich cathelicidins. The substitution of glycine residues of PopuCATH with alanine residues significantly resulted in a reduced efficacy against bacterial infection (Figure 4figure supplement 4). In addition, PopuCATH contains 10 arginine residues and seven serine residues (Supplementary file 1). The substitution of arginine residues or serine residues with alanine residues also significantly led to a decreased efficacy against bacterial infection (Figure 4-figure supplement  4). These data demonstrated that these enriched amino acid residues, including 21 glycine residues, 10 arginine residues, and 7 serine residues, are key structural requirements for PopuCATH-mediated protective efficacy against bacterial infection, and PopuCATH-mediated protection were specifically due to its unique structure. The first frog-derived cathelicidin is also rich in glycine residues. But it has different amino acid sequence with PopuCATH and exhibits direct antibacterial activity unlike PopuCATH (Hao et al., 2012). Cathelicidin antimicrobial peptides display a high structural diversity, and the diverse structures are responsible for their diverse functions. Accordingly, it is not difficult to understand that these two frog cathelicidins have different functions against bacteria.
Recently, many progresses have been achieved in the development of anti-resistance therapy for combatting multidrug resistant bacterial infection. Pre-clinical and clinical data pointed out host-directed therapeutic approaches to enhance 'pauci-inflammatory' microbial killing in myeloid phagocytes merited particular attention (Watson et al., 2020). Host-based therapeutic strategies can maximise microbial clearance and minimise host's harmful consequences induced by inflammatory response, which has great promise. PopuCATH did not show any direct effects on bacteria, but effectively prevented bacterial infection through eliciting phagocyte influx and slightly promoting neutrophil phagocytosis. PopuCATH-mediated protection against bacterial infection can be considered as a classic host-based therapeutic strategy, and the non-bactericidal nature of PopuCATH may reduce the selective pressures that drive bacterial resistance. In an era of emerging and re-emerging infectious diseases, discovery and development of naturally occurring non-bactericidal antimicrobial peptides like PopuCATH may facilitate us to prevent and overcome multidrug-resistant bacterial infection.
In summary, a glycine-rich amphibian cathelicidin, PopuCATH, was identified from tree frog. Popu-CATH didn't show any direct effects on bacteria but provided protection against bacterial infection in vivo. PopuCATH acted as an immune defense regulator against bacterial infection by selective modulation of innate immune response. Our findings provide new insights into the development of non-bactericidal cathelicidins to prevent bacterial infection.
Synthetic peptides were purchased from Synpeptide Co. Ltd (Shanghai, China). The crude peptide was purified by reversed-phase high performance liquid chromatography (RP-HPLC) and analysed by mass spectrometry to confirm the purity higher than 98%.

Experiment animals
Both adult healthy tree frogs of P. puerensis (21-30 g) were captured from Pu'er, Yunnan Province, China (24.786°N, 101.362°E). P. puerensis was not endangered or protected species, and no specific permissions were required for the sampling location/activity. Tree frogs were randomly housed in freshwater tanks in a recirculating system with filtered water, fed with mealworm larvae Tenebrio molitor and refreshed with water once a day. C57BL/6 mice (female, 18-20 g) were purchased from Shanghai Slac Animal Inc, and Rag1 -/mice (female, 18-20 g) were purchased from Model Animal Research Center of Nanjing University. Mice were housed in pathogen-free facility. Animal experiments were performed in accordance with the Institutional Animal Care and Use Committee of Soochow University, and all research protocols were approved by the Animal Ethical Committee of Soochow University. All surgery of animals was performed under pentobarbital sodium anaesthesia with minimum fear, anxiety, and pain.

Mature peptide isolation
Skin secretions were collected according to previous study . Briefly, frogs were stimulated by anhydrous ether, and a total of about 500 mL skin secretions in PBS were quickly collected, centrifuged, and lyophilised. Lyophilised P. puerensis skin secretion was dissolved in phosphatebuffered saline (PBS, 0.1 M, pH 6.0) and separated by molecular sieving fast protein liquid chromatography (FPLC) on GE ÄKTA pure system using a Superdex 75 10/300 GL column (10 × 300 mm, 24 mL volume, GE, USA). Fractions were pooled and further purified by RP-HPLC on a C18 column (25 × 0.46 cm, Waters, USA) for two times. The eluted peaks from RP-HPLC were collected for purity assay using matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) on an UltraFlex I mass spectrometer (Bruker Daltonics, Germany). The amino acid sequence of the purified peptide was obtained by automated Edman degradation analysis on an Applied Biosystemspulsed liquid-phase sequencer (model ABI 491, USA).

cDNA cloning
Skin total RNA extraction, mRNA isolation and cDNA library construction were performed according to previous methods (Wei et al., 2015). About 5.6 × 10 5 independent colonies were produced in the cDNA library. Two primers, an antisense primer, 5'-TTGT CTGC CTCC TCGG CTTCC-3', designed according to the conserved domain of amphibian cathelicidins, and the 5' PCR primer, 5'-AAGC AGTG GTAT CAAC GCAGAGT-3' supplied by cDNA library construction kit, were used to clone the 5' fragment that encoding the precursor of PopuCATH. The full length cDNA encoding the precursor of PopuCATH was obtained by a sense primer, 5'-ATGG CGCT CGCT GCTG CACTC-3' designed according to the 5' fragment of PopuCATH precursor, and 3' PCR primer, 5'-ATTC TAGA GGCC GAGG CGGCCG-3' provided by the kit. PCR procedure for cDNA cloning was 95 ℃ for 5 min, and 30 cycles of 95 ℃ for 30 s, 56 ℃ for 30 s, 72 ℃ for 1 min, followed by an extension step at 72 ℃ for 8 min.

Toxic side effects to mammalian cells and mice
For cytotoxicity assay, mouse peritoneal macrophages or THP-1 cells were seeded into 96-well plates (5 × 10 5 cells/well, 200 µL). PopuCATH (25, 50, 100, and 200 μg/mL) was added to each well. After culture for 24 h, 10 µL of CCK-8 reagent was added to each well. The absorbance at 450 nm was recorded on a microplate reader after incubation for 1 h .
For hemolysis assay, mouse erythrocytes and rabbit erythrocytes were washed with 0.9% saline and incubated with a series of two-fold dilutions of PopuCATH (25, 50, 100, and 200 μg/mL) at 37 ℃. After incubation for 30 min, the erythrocytes were centrifuged at 1,000 g for 5 min and monitored at 540 nm. Triton X-100 (1%) treatment was determined as 100% hemolysis. Hemolytic activity was expressed as the percentage of the Triton X-100-treated group (Wei et al., 2013).

In vitro antimicrobial assay
A standard two-fold broth microdilution method was used to evaluate the MIC of PopuCATH against microbes. Gram-positive bacteria, Gram-negative bacteria, and fungi were diluted with Mueller-Hinton broth, and aquatic pathogenic bacteria were diluted with nutrient broth to 10 5 CFU/mL. Series of twofold PopuCATH dilutions were prepared in 96-well plates (50 μL/well). An equal volume of microbial dilution was added and cultured at 37℃ (for Gram-positive bacteria, Gram-negative bacteria, and fungi) or 25℃ (for aquatic pathogenic bacteria) for 18 hr. Cathelicidin-PY from P. yunnanensis served as positive control. The minimal concentrations at which no visible growth of microbes occurred were defined as MIC values (Wei et al., 2013).
Scanning electron microscope (SEM) assay was used to examine if PopuCATH impairs the bacterial surface morphology. E. coli ATCC25922 and S. aureus ATCC25923 were cultured in Mueller-Hinton broth to exponential phase, washed and diluted using PBS (10 5 CFU/mL). PopuCATH (200 μg/mL), cathelicidin-PY (PY, 1× MIC) or PBS was added into the bacterial dilution and incubated at 37℃. After incubation for 30 min, bacteria were centrifuged (1000 g for 10 min) and fixed for SEM assay according to standard operating protocols. The bacterial surface morphology was observed using a Hitachi SU8010 SEM (Wei et al., 2013).
In order to further investigate its prophylactic efficacy against bacterial infection, PopuCATH (10 mg/kg) was given through intravenous or intramuscular injection at 4 hr prior to E. coli inoculation (2 × 10 7 CFUs/mouse, intraperitoneal injection). At 18 hr post bacterial inoculation, peritoneal lavage was collected for bacterial load assay.
The protective efficacy of PopuCATH was also evaluated in septic mice induced by a lethal bacterial inoculation or CLP. For lethal bacterial challenge, C57BL/6 mice (female, 18-20 g, n = 6) were intraperitoneally injected with PopuCATH (10 mg/kg) 4 hr prior to E. coli (4 × 10 7 CFUs/mouse, intraperitoneal injection) or MRSA (6 × 10 8 CFUs/mouse, intraperitoneal injection) inoculation. The survival rates of mice were monitored for 7 days . To compare the protective efficacy of PopuCATH with other peptides, the protective efficacy of LL-37 and IDR-1 were simultaneously evaluated at the same condition. For CLP-induced sepsis, C57BL/6 mice (female, 18-20 g, n = 6) were intraperitoneally injected with PopuCATH (10 mg/kg) at 8 and 4 hr (two times) prior to CLP. At 4 hr post the last injection of PopuCATH, mice were anaesthesied with ketamine hydrochloride (100 mg/kg), and the abdominal cavity of mice was opened in layers. The cecum was ligated 1.0 cm from the end, a through-and-through puncture was operated using an 18-gauge needle. A small droplet of faeces was extruded for ensuring the patency of the puncture site. Then, the cecum was returned back to the abdominal cavity. A laparotomy but no CLP mice served as control. After CLP, the survival rates of mice were monitored for 7 days .
P. puerensis (21-30 g, n = 5) were intraperitoneally injected with PopuCATH (10 mg/kg, dissolved in 0.2 mL PBS) or PBS. At 4 and 8 hr post injection, total cells in peritoneal lavage were counted using a hemocytometer, and the cells were observed under an optical microscope after Wright-Giemsa staining.

Ethics
Animal experiments were performed in accordance with the Institutional Animal Care and Use Committee of Soochow University, and all research protocols were approved by the Animal Ethical Committee of Soochow University. All surgery of mice was performed under pentobarbital sodium anesthesia with minimum fear, anxiety and pain. • Supplementary file 4. THP-1 cell line authentication report by STR profiling.

Decision letter and Author response
• Supplementary file 5. Primers for qPCR.

Data availability
Sequencing data have been deposited in GenBank (accession number: KY391886). All data generated or analysed during this study are included in the manuscript, supporting files, and source data.
The following dataset was generated: