A Teleost CXCL10 Is Both an Immunoregulator and an Antimicrobial

Chemokines are a group of cytokines that play important roles in cell migration, inflammation, and immune defense. In this study, we identified a CXC chemokine, CXCL10, from Japanese flounder Paralichthys olivaceus (named PoCXCL10) and investigated its immune function. Structurally, PoCXCL10 possesses an N-terminal coil, three β-strands, and a C-terminal α-helix with cationic and amphipathic properties. PoCXCL10 expression occurred in multiple tissues and was upregulated by bacterial pathogens. Recombinant PoCXCL10 (rPoCXCL10) promoted the migration, cytokine expression, and phagocytosis of flounder peripheral blood leukocytes (PBLs). rPoCXCL10 bound to and inhibited the growth of a variety of common Gram-negative and Gram-positive fish pathogens. rPoCXCL10 killed the pathogens by causing bacterial membrane permeabilization and structure destruction. When introduced in vivo, rPoCXCL10 significantly inhibited bacterial dissemination in fish tissues. A peptide derived from the C-terminal α-helix exhibited bactericidal activity and competed with rPoCXCL10 for bacterial binding. Deletion of the α-helix affected the in vitro bactericidal activity but not the chemotaxis or in vivo antimicrobial activity of PoCXCL10. Together, these results indicate that PoCXCL10 exerts the role of both an immunoregulator and a bactericide/bacteriostatic via different structural domains. These findings provide new insights into the immune function and working mechanism of fish CXC chemokines.


INTRODUCTION
Chemokines are a group of small (8 to 12 kDa) proteins with the ability to induce cell migration and immune responses by binding to G protein-coupled receptors on the target cells (1)(2)(3)(4)(5). Chemokines share a common structure, which is stabilized by the disulfide bonds between conserved cysteine residues (6). In mammals, on the basis of the number and distinctive pattern of the N-terminal cysteine residues, chemokines are categorized into four subfamilies: CXC, CC, C (also called XC), and CX 3 C (7). In fish, the CXC, CC, and C subfamilies have been identified, but no CX 3 C has been reported (8,9). The CC and CXC chemokines are the two major subfamilies in both mammals and fish (9).
The functions of many chemokines have been reported. "Homeostatic" chemokines, such as CCL16 and CXCL13, are constitutively expressed and play an important role in regulating the homing and maturation of immune cells in homeostasis (10,11). The CXCL12-CXCR4 axis is known to be vital to the development of the central nervous system in zebrafish and for the development of coronary artery in mice (12)(13)(14). "Inducible" chemokines, such as CCL2 and CXCL8, are induced by inflammatory signals or pathogen infection, and act as key mediators in innate and adaptive immunity (15,16). Some mammalian chemokines have been shown to possess direct bactericidal activities (17). In fish, the functions of a number of chemokines have been studied. For example, turbot Scophthalmus maximus CCL19 was found to exhibit chemotactic activity and immunoregulatory property, which enhanced host resistance against viral and bacterial infection (18); tongue sole Cynoglossus semilaevis CCL21 was able to recruit leukocytes and involved in antibacterial defense (19); rainbow trout Oncorhynchus mykiss CK12a contributed to host antiviral immune response (20). To date, bactericidal activity has been observed only in rainbow trout CK11 and grass carp Ctenopharyngodon Idella CXCL20a/b (21)(22)(23).
Japanese flounder Paralichthys olivaceus is one of the major cultured marine fish species in Asia. In this study, we identified a CXCL10 homologue from flounder and examined its chemotactic activity and antimicrobial activity in relationship to its structure. In addition, we also examined the bactericidal effect of a CXCL10-derived peptide. Our results indicate an important role of CXCL10 in host immune defense against bacterial infection and expand the understanding of the biological function of teleost chemokines.

Animals
Clinically healthy Japanese flounder Paralichthys olivaceus (average weight of 20 g) were purchased from a local commercial fish farm in Shandong Province, China. The fish were acclimatized at 19-20°C in aerated seawater for at least one week before the experiments. Before tissue collection, the fish were euthanized with tricaine methanesulfonate (Sigma, St. Louis, MO, USA) as reported previously (40).

Sequence Analysis of PoCXCL10
The amino acid sequence of PoCXCL10 (accession no. XP_019935349.1) was analyzed with the basic local alignment search tool (BLAST) program at the National Center for Biotechnology Information (NCBI). The signal peptide was predicted with the SignalP5.0 program (http://www.cbs.dtu.dk/ services/SignalP/). The molecular mass (Mw) and theoretical isoelectric point (pI) were obtained using the ExPASy molecular biology server (https://web.expasy.org/compute_pi/). Multiple sequence alignment was created with DNAMAN (version 7.0.2.176, Lynnon Biosoft, San Ramon, CA, USA) software. Phylogenetic analysis was constructed using the Neighborjoining algorithm of MEGA 6.0 software (Tempe, AZ, USA).

Determination of PoCXCL10 Expression by Quantitative Real-Time Reverse Transcription-Polymerase Chain Reaction (qRT-PCR)
qRT-PCR was performed to determine the expression of PoCXCL10 in fish tissues under normal physiological conditions as reported previously (44). Briefly, tissue samples were isolated from the gill, blood, muscle, head kidney, intestine, brain, heart, spleen and liver of healthy Japanese flounder. Total RNA was extracted using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) and used for cDNA synthesis with the First Strand cDNA Synthesis Kit (ToYoBo, Osaka, Japan) according to the manufacturer's protocol. qRT-PCR was carried out with Eppendorf Mastercycler ® ep gradient S (Eppendorf, Hamburg, Germany) using ChamQ Universal SYBR qPCR master mix (Vazyme, Najing, China). Gene expression was calculated using comparative threshold cycle method (2 −DDCT ) and normalized to that of b-actin.
To examine the expression of PoCXCL10 during bacterial infection, Edwardsiella tarda, Vibrio anguillarum and Vibrio harveyi were cultured in Luria-Bertani (LB) medium at 28°C to logarithmic phase, and resuspended in phosphate-buffered saline (PBS) to 1×10 7 CFU/ml. Japanese flounder were infected with 100 µL E. tarda, V. anguillarum, V. harveyi or PBS (control). At 6, 12, 24, and 48 h post infection (hpi), spleen, liver and head kidney were collected. qRT-PCR analysis of PoCXCL10 expression in the collected tissues was performed as described above. 18S ribosomal RNA (18S rRNA), a-tubulin, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were used as internal references in head kidney, spleen, and liver, respectively (45). The sequences of the primers used for qRT-PCR are listed in Table S1. All qRT-PCR assays were performed in triplicate.

Protein Expression and Purification
To obtain recombinant proteins of PoCXCL10 (rPoCXCL10) and a mutant PoCXCL10 lacking the C-terminal a-helix region (rPoCXCL10M), the coding sequence of PoCXCL10 and PoCXCL10M were amplified by PCR with the primer pairs PoCXCL10-F/PoCXCL10-R and PoCXCL10M-F/PoCXCL10M-R, respectively ( Table S1). The PCR products were inserted into pET28a-SUMO plasmid (46) at between the BamH I and Xho I sites. The recombinant plasmids were introduced into E. coli Transetta (DE3) (TransGen Biotech, Beijing, China) via transformation. Recombinant proteins were purified from the transformants as reported previously (44) with slight adjustments. Briefly, the transformants were cultured in LB medium at 37°C to an OD 600 of 0.5. Isopropyl-b-Dthiogalactopyranoside was added to the medium to a final concentration of 0.3 mM. The sample was then transferred to 16°C and cultured for 16 h. Then the cells were collected by centrifugation (6000 rpm) for 10 min, and Sumo-tagged rPoCXCL10/rPoCXCL10M and recombinant Sumo tag (rSumo) (as a control protein) were purified using nickelnitrilotriacetic acid (Ni-NTA) column (QIAGEN, Valencia, USA). The Sumo-tag was cut off with Sumo protease (ubiquitin-like proteases, ULP) as previously reported (46), resulting in tag-free rPoCXCL10 and rPoCXCL10M. The recombinant proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized after staining with Coomassie brilliant blue R-250 ( Figure S1).

Binding of Protein/Peptide to Bacteria and Bacterial Components
Protein-bacteria interaction was determined with enzyme-linked immunosorbent assay (ELISA) as reported previously (44) with slight adjustments. Briefly, bacteria (Bacillus subtilis, Micrococcus luteus, Streptococcus iniae, V. harveyi, V. anguillarum, E. tarda, and Pseudomonas fluorescens) were cultured to logarithmic phase and resuspended in coating buffer (15 mM Na 2 CO 3 , 35 mM NaHCO 3 , pH 9.6) to a final concentration of 1×10 8 CFU/ml. The bacterial suspensions were added to 96-well plates (100 ml/ well), and the plates were incubated at 4°C for overnight. After incubation, the plates were blocked with 5% skim milk and incubated at room temperature for 2 h. The plates were washed three times with PBST (PBS containing 0.1% Tween-20). Then, 100 µL rPoCXCL10 at various concentrations (6.25 nM, 12.5 nM, 25 nM, 50 nM, and 100 nM) or rSumo (100 nM) or PBS (control) was added to the plates. The plates were incubated at room temperature for 2 h, followed by washing three times with PBST. Horseradish peroxidase (HRP)-conjugated mouse anti-Flag antibody (ABclonal, Wuhan, China) (2,000 dilution) was added to the plates, and the plates were incubated at 37°C for 1 h. The plates were washed as above. Color development was performed with the 3, 3′, 5, 5′-tetramethylbenzidine (TMB) kit (Tiangen, Beijing, China), and absorbance at 450 nm was measured using a precision microplate reader (Molecular Devices, Toronto, Canada). The binding index was defined as follows: A 450 of rPoCXCL10/A 450 of PBS. To compare the bacteria-binding abilities of rPoCXCL10 and rPoCXCL10M, V. anguillarum was incubated with 100 nM rPoCXCL10, rPoCXCL10M, rSumo, or PBS for 2 h, and the bacteria-bound protein was determined by ELISA as above.
To examine the binding of rPoCXCL10 to bacteria components, lipopolysaccharide (LPS), peptidoglycan (PGN) and lipoteichoic acid (LTA) (all from Sigma, St Louis, MO, USA) were dissolved in coating buffer to 50 mg/ml and added to 96-well plates (100 ml/well). After incubation at 4°C for overnight, rPoCXCL10 binding was performed with ELISA as above.
To examine the binding of peptide to bacteria, the C-terminal a-helix region of PoCXCL10 (5'-KVVKDRLIRIIKK-3'), named P13, was synthesized by ChinaPeptides Co., Ltd. (Shanghai, China). P13 binding to V. harveyi and V. anguillarum was determined as above with slight adjustments. Briefly, the plates were coated with bacteria and blocked as above, then fluorescein isothiocyanate (FITC) labeled-P13 at various concentrations (1.25 mM, 5 mM, and 20 mM) was added to the plates. The plates were incubated at room temperature for 2h and washed five times with PBST. The fluorescence intensity was detected with a multifunctional microplate reader (TECAN Infinite M200 PRO, Switzerland). The peptide P86P15 was used as a negative control peptide (NC) based on the previous report (47). To examine the effect of P13 on rPoCXCL10 binding to bacteria, V. anguillarum and V. harveyi were incubated with 60 mM P13 or NC or PBS for 2 h and washed as above. The bacteria were then incubated with 25 nM rPoCXCL10 or rSumo or PBS for 2 h. The bacteria-bound protein was detected by ELISA as described above.

Effect of rPoCXCL10 on Bacterial Growth and Survival
To determine the effect of rPoCXCL10 on bacterial growth, V. anguillarum, V. harveyi, E. tarda and S. iniae were cultured to logarithmic phase. V. anguillarum, V. harveyi, and E. tarda were suspended in LB medium to 1×10 5 CFU/ml in 96-well plates (100 ml/well). S. iniae was suspended in tryptic soy broth (TSB) medium (Hopebio, Qingdao, China) to 1×10 5 CFU/ml in 96-well plates (100 ml/well). The bacterial suspension was mixed with or without (control) rPoCXCL10 or rSumo to a final concentration of 4 mM. The plates were incubated at 28°C for 12 h with shaking, and OD 600 was measured at every hour to determine the bacterial growth. To determine the effect of rPoCXCL10 on bacterial survival, V. anguillarum, V. harveyi, E. tarda and S. iniae were suspended in PBS to 1×10 3 CFU/ml and mixed with rPoCXCL10 or rSumo (4 µM) or PBS (control). The samples were then incubated at room temperature for 4 h (at which time point, the bacterial membrane damage could be detected). After incubation, the number of survived bacteria was determined by plate count. The bactericidal activity of rPoCXCL10M was similarly performed. To determine the bactericidal activity of P13, V. anguillarum and V. harveyi were incubated with or without (control) 20 µM P13 or NC for 4 h. The number of survived bacteria was determined as above.
Electron Microscopy and Propidium Iodide Staining Assay V. anguillarum was cultured to logarithmic phase and resuspended in PBS to 1×10 6 CFU/ml. The bacteria were incubated with 4 µM rPoCXCL10 or PBS at room temperature for 4 h. After incubation, the bacteria were washed and resuspended in ultrapure water and observed with a transmission electron microscope (TEM) (Hitachi, HT7700, Japan) as described previously (48). For the PI staining assay, V. anguillarum was pretreated with rPoCXCL10 or PBS as above, and then stained with PI (2 µg/ml) for 10 min. The bacteria were observed with a fluorescence microscope (TiS/L100, Nikon, Tokyo, Japan).

Chemotaxis Assay
Chemotaxis assay was carried out in 24-well Transwell plates (Corning Costar Co., Cambridge, MA, USA) with the pore size of 3 mm. rPoCXCL10 (0.25 mM, 1 mM, or 4 mM) or rSumo (4 mM) or PBS (control) in 600 ml L-15 medium (Sigma, St Louis, MO, USA) was placed at the lower chamber of Transwell. Flounder peripheral blood leukocytes (PBLs) were prepared as reported previously (49) and adjusted to 1×10 7 cells/ml in L-15 medium. The PBLs (100 ml) were loaded onto the upper chamber of the above Transwell plates. The plates were incubated at 22°C for 2 h, and the cells migrated into the lower chamber was observed and counted with a microscope. The chemotactic index was presented as the fold increase in the number of migrated PBLs induced by recombinant protein to that induced by PBS. The chemotactic activity of rPoCXCL10M was similarly determined. To examine the effects of LPS and PGN on the chemotactic activity of rPoCXCL10, 1 mM rPoCXCL10 was pre-incubated with or without 50 mg/ml LPS or PGN for 2 h, and the chemotactic activity of rPoCXCL10 was determined as above.

Binding of rPoCXCL10/M to PBLs
Binding of recombinant proteins to PBLs was detected by immunofluorescence microscopy as reported previously (44) with slight adjustments. Briefly, Japanese flounder PBLs were resuspended in L-15 medium to 1×10 7 cells/ml, and the cells were added to cell culture dish (NEST, USA) for 4 h to allow the cells to settle. The cells were then blocked with 5% skim milk and incubated at 22°C for 1 h. After washing with PBS, 1 mM rPoCXCL10, 1 mM rPoCXCL10M, or PBS were added to the dish and incubated at 22°C for 2 h, followed by washing with PBS for three times. The cells were treated with mouse anti-Flag antibody (ABclonal, Wuhan, China) at 1/1,000 dilution for 1h and then treated with Alexa Fluor ® 488conjugated goat anti-mouse IgG antibody (Abcam, UK) at 1/ 2,000 dilution for 1 h, followed by washing with PBS for three times. The cells were fixed with 4% paraformaldehyde (Biosharp, Shanghai, China) for 30 min and washed as above, then the cells were stained with 1, 1′-dioctadecyl-3, 3, 3, 3′-tetramethylindocarbocyanine perchlorate (DiI) (Invitrogen, Carlsbad, CA, USA) and 4', 6-diamidino-2phenylindole (DAPI) (Beyotime, Shanghai, China). The cells were then observed with a Zeiss LSM 710 confocal microscope (Carl Zeiss, Oberkochen, Germany).

The Effect of rPoCXCL10 on PBLs Gene Expression
Flounder PBLs (1×10 6 cells/ml) were treated with 1 mM rPoCXCL10, 1 mM rSumo, or PBS at 22°C for 3 h, 6 h, and 12 h. Total RNA was extracted at each time point and used for qRT-PCR as described above to analyze the expression of IL-6, TNF-a, CXCL8, IL-1b, and IL-10. The sequences of the primers used for qRT-PCR are listed in Table S1.
The Effects of rPoCXCL10 and rPoCXCL10M on Bacterial Infection of PBLs E. tarda was cultured to logarithmic phase and resuspended in 100 mg/ml FITC (YuanYe Bio-Technology, Shanghai, China), followed by incubation at room temperature for 3 h. The bacteria were collected and washed five times with PBS. The bacteria were resuspended in L-15 medium to 1×10 8 CFU/ml and incubated with or without (control) 1 mM rPoCXCL10, rPoCXCL10M, or rSumo at room temperature for 2 h. Flounder PBLs were mixed with untreated E. tarda or E. tarda treated with rPoCXCL10, rPoCXCL10M, or rSumo at a multiplicity of infection (MOI) of 1:2. After incubation at 22°C for 2 h, the PBLs were centrifuged and washed with PBS. Trypan blue (0.125% in PBS) was added to the cells to quench the extracellular fluorescence. The cells were washed and suspended in PBS. The intracellular fluorescence signal of the bacteria was determined using a FACScan flow cytometer (BD Biosciences, USA).
The Effects of rPoCXCL10 and rPoCXCL10M on Bacterial Infection In Vivo E. tarda and V. anguillarum were prepared as above. Japanese flounder were infected with 100 µL E. tarda (2×10 6 CFU/ml) or 100 µL V. anguillarum (1×10 7 CFU/ml) via intraperitoneal injection in the presence of 5 mM rPoCXCL10, 5 mM rPoCXCL10M, 5 mM rSumo or PBS. Liver, spleen, and head kidney were collected aseptically at 12 and 24 hpi for V. anguillarum infection or 12 and 36 hpi for E. tarda infection. Bacterial recoveries from the tissues were determined by plate count as reported previously (50).

Statistical Analysis
All experiments were performed three times. Statistical analyses were carried out using GraphPad Prism version 6.01 (GraphPad Software Inc., San Diego, CA, USA). Data were analyzed with student's t-test. The results were considered statistically significant when the p < 0.05.

Sequence Characteristics of PoCXCL10
PoCXCL10 is a 94-amino acid protein with an N-terminal signal peptide (residues 1 to 20) and four conserved cysteine residues (Cys-29, 31, 55 and 71) ( Figure 1A). The predicted mature protein (without signal peptide) has a molecular weight of 8.3 kDa and a theoretical pI of 9.64. PoCXCL10 shares 42.11% and 43.24% overall sequence identities with Homo sapiens and Mus musculus CXCL10, respectively. Phylogenetic analysis showed that PoCXCL10 was grouped together with the putative fish CXCL10 homologues, which formed a distinct cluster separated from the clusters formed by the CXCL10 of amphibian, reptile, bird, and mammal ( Figure 1B).

The Expression of PoCXCL10 Is Regulated by Bacterial Pathogens
qRT-PCR showed that PoCXCL10 mRNA was present in nine tissues, i.e., liver, spleen, heart, brain, intestine, head kidney, muscle, blood, and gill, in the order of decreasing expression ( Figure S2A). When flounder were infected with the bacterial pathogen E. tarda, PoCXCL10 expression was significantly upregulated at 12, 24, and 48 hpi in spleen and head kidney, and at 6, 12, 24, and 48 hpi in liver ( Figure S2B). Similarly, infections by the bacterial pathogens V. anguillarum and V. harveyi induced PoCXCL10 expression in flounder tissues in a timedependent manner.  and P. fluorescens) in a dose-dependent manner (Figure 2A). rPoCXCL10 also bound apparently to pathogen-associated molecular patterns (PAMPs), i.e., LPS, PGN, and LTA ( Figure 2B). For V. anguillarum, E. tarda, and S. iniae, their growths were markedly inhibited by rPoCXCL10; for V. harveyi, its growth was inhibited transiently by rPoCXCL10 at the early stage (4-8 h) ( Figure 2C). rPoCXCL10 binding significantly reduced the viability of V. anguillarum and V. harveyi, but had no effect on the survival of E. tarda or S. iniae (Figures 3A, B; Figure S3). Microscopy showed that rPoCXCL10 treatment damaged the cellular structure and membrane of V. anguillarum, which enabled propidium iodide (PI) to pass into the cells (Figures 3C, D).

A PoCXCL10-Derived Peptide Exhibits Bactericidal Activity
Structurally, PoCXCL10 is composed of an extended N-terminal coil, three antiparallel b-strands, and a C-terminal a-helix ( Figure 4A). PoCXCL10 contains abundant positively charged amino acid residues, in particular at the C-terminal a-helix, resulting in an apparent cationic surface ( Figure 4B). Based on this structural feature, the C-terminal a-helix domain (residues 57-69) of PoCXCL10 was synthesized as a peptide named P13,  which has a pI of 11.17. In the helical wheel view of P13, the predominant polar (cationic) amino acids and the hydrophobic amino acids were distributed on two different sides ( Figure 4C). Solvent accessibility representation also showed distinct hydrophobic and hydrophilic areas ( Figure 4C), suggesting an amphipathic characteristic of P13 like that of the classical antimicrobial peptides. In line with this observation, P13 was able to bind V. harveyi and V. anguillarum in a dose-dependent manner and significantly reduced the viability of the bacteria (Figures 4D, E). Furthermore, pre-treatment of the bacteria with P13 significantly reduced the binding of rPoCXCL10 to the bacteria ( Figure 4F). Transwell migration assay showed that rPoCXCL10 could induce the migration of flounder PBLs in a dose-dependent manner ( Figure 5A). However, pre-treatment with LPS or PGN significantly reduced the chemotactic activity of rPoCXCL10 ( Figure 5B). qRT-PCR showed that in rPoCXCL10-treated PBLs, the expressions of IL-6, CXCL8, and IL-1b were significantly up-regulated at 3, 6, and 12 h, and the expressions of TNF-a and IL-10 were significantly up-regulated at 3 and 12 h, respectively ( Figure 5C).
The C-Terminal a-helix of PoCXCL10 Is Required for Direct Bactericidal Activity but not for Chemotaxis or In Vivo Antimicrobial Activity To examine the importance of the C-terminal a-helix to the immune functions of PoCXCL10, a mutant PoCXCL10 lacking this region was prepared as recombinant protein named rPoCXCL10M. Subsequent analysis showed that both rPoCXCL10 and rPoCXCL10M were able to bind to flounder PBLs and exhibited comparable chemotactic activities ( Figures 6A, B). Fluorescence-activated cell sorting (FACS) analysis showed that both rPoCXCL10 and rPoCXCL10M significantly enhanced bacterial phagocytosis by PBLs ( Figure  S4A). However, compared to rPoCXCL10, rPoCXCL10M exhibited significantly reduced abilities to bind and kill bacteria ( Figures 6C, D). When flounder were infected with V. anguillarum in the presence of rPoCXCL10 or rPoCXCL10M, the bacterial loads in liver (12 hpi) and spleen (12 hpi and 24 hpi) were significantly reduced compared to that in the control fish ( Figure 6E). The reduction levels caused by rPoCXCL10M were largely comparable to that caused by rPoCXCL10 ( Figure 6E). Similar results were obtained when the infection was performed with E. tarda ( Figure S4B).

DISCUSSION
The chemokine superfamily has a large number of members belonging to different subfamilies, many of which have been demonstrated to play vital roles in innate and adaptive immunity (3,4,(51)(52)(53). CXCL10 is a member of the CXC chemokine subfamily. To date, several CXCL10 from different fish species have been identified, and their mRNA expression levels before and after pathogen infection were determined (35)(36)(37)(38)(39). However, the biological activity of fish CXCL10 as a protein remains largely unclear. In this study, we characterized the function of a flounder CXC chemokine, PoCXCL10, in association with antimicrobial immunity. Structurally, PoCXCL10 possesses the four cysteine residues conserved in mammalian CXC, which form two intramolecular disulfide bonds that stabilize the molecule (54). Of the few reported fish CXCL10, they were shown to exhibit high levels of mRNA expression in immune relevant organs such as spleen, head kidney, and liver, and their expression levels could change significantly after bacterial challenge (35,37,38). Similarly, we found that PoCXCL10 expression was relatively high in liver and spleen, and was upregulated during the infection of E. tarda, V. anguillarum, and V. harveyi. These results indicate that PoCXCL10 is involved in pathogen-induced immune response.
A previous study showed that in rainbow trout, CXCL10 expression was not induced by LPS (39). It is likely that LPS and live bacteria differ in some aspects of their immunoregulations on the expression of host genes such as CXCL10. Some human chemokines, such as CXCL14 and CXCL9, kill bacteria directly (55,56). In fish, rainbow trout CK11 and grass carp CXCL20a/b were reported to possess bactericidal activities (21)(22)(23). In our study, we found that rPoCXCL10 interacted directly with both Gram-negative and Gram-positive bacteria. These interactions occurred likely via PAMPs on the part of the bacteria, because rPoCXCL10 exhibited apparent binding to LPS, the central components of the outer membrane in Gram-negative bacteria, and PGN, the cell wall of Gram-positive bacteria. rPoCXCL10 binding inhibited the growth of V. anguillarum, V. harveyi, E. tarda, and S. iniae, which are severe pathogens to a wide range of fish species including flounder. rPoCXCL10 displayed apparent bactericidal activity against V. anguillarum and V. harveyi. These observations indicate that rPoCXCL10 can act as a bactericide or a bacteriostat, depending on the target bacteria. Microscopy revealed that the rPoCXCL10 treatment disrupted the membrane and cellular structure of the target bacteria, which was the direct cause of bacterial death.
In mammals, CXCL10 exhibits chemotactic activity towards various types of cells, including T cells, natural killer cells, and macrophages (57)(58)(59). In this study, we found that rPoCXCL10 could induce the migration of PBLs in a dose-dependent manner, indicating that rPoCXCL10 is an active chemoattractant. It is interesting that the presence of LPS or PGN significantly decreased the chemotactic activity of rPoCXCL10. It is possible that binding to LPS or PGN may affect the binding of rPoCXCL10 to its receptors on PBLs, or obstructing the effective signal transduction mediated by rPoCXCL10-receptor interaction. Mammalian CXCL10 is known to induce the production of inflammatory cytokines such as TNF-a, IL-1b, and CCL2 (29). Similarly, we found that rPoCXCL10 enhanced the expression of a variety of inflammatory cytokines in PBLs, indicating an immune regulatory function of rPoCXCL10.
It is well known that most antimicrobial peptides (AMPs) possess cationic charges and amphipathicity, which enable the AMPs to interact with the plasma membrane of the target bacteria (60). Structure modeling revealed that the C-terminal ahelix region of PoCXCL10 is amphipathic and highly positively charged. In agreement with this feature, a peptide, P13, derived from the a-helix was able to bind and kill V. anguillarum and V. harveyi. Hence, P13 probably retained the bactericidal activity of PoCXCL10. In the presence of P13, the rPoCXCL10-bacteria interaction was significantly hindered, implying that P13 competed with rPoCXCL10 for the same or overlapping binding sites in bacteria. These results suggest that the C-terminal a-helix is responsible, at least in part, for the interaction between PoCXCL10 and bacteria. In line with these results, deletion of the C-terminal a- helix region from PoCXCL10 significantly reduced the bacteria binding ability and the bactericidal activity of PoCXCL10, but did not affect the ability of PoCXCL10 to bind PBLs or induce chemotactic signaling, which was consistent with the previous observations that for mammalian chemokines, the N-terminal region was essential for receptor binding (61,62). In the present study, in vivo infection showed that both rPoCXCL10 and rPoCXCL10M significantly reduced V. anguillarum and E. tarda dissemination in flounder tissues. Therefore, although the C-terminal a-helix of PoCXCL10 was essential to in vitro antibacterial activity, it was largely dispensable for in vivo antibacterial immunity under the experimental condition. It is likely that in the in vitro system, which contained no host cells, but only the bacteria and the chemokine, the antibacterial effect was solely dependent on the bactericidal activity of rPoCXCL10 or rPoCXCL10M, while in vivo, the antibacterial effect was mainly dependent on the ability of rPoCXCL10 or rPoCXCL10M to recruit and activate immune cells that eventually led to bacterial clearance. In conclusion, our study demonstrated that PoCXCL10 was an active chemokine that regulated host cell immune response and promoted in vivo bacterial clearance. PoCXCL10 was also a bactericide and bacteriostatic that killed or inhibited the growth of fish pathogens. Importantly, we found that the C-terminal ahelix region of PoCXCL10 was crucial for bactericidal activity but not for chemotactic activity. These results add new insights into the function and mechanism of fish chemokines in antibacterial immunity.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by Ethics Committee of Institute of Oceanology, Chinese Academy of Sciences.