CK11, a Teleost Chemokine with a Potent Antimicrobial Activity

Chemokines are cytokines with chemotactic capacities that regulate the migration of immune cells under both inflammatory and normal physiological conditions. They are small molecules defined by the presence of four conserved cysteine residues. In mammals, they are divided into four subfamilies based on the distinctive pattern of the two N-terminal cysteines: CXC (also referred to as α chemokines), CC (β chemokines), C (γ chemokines), and CX₃C (δ chemokines) classes. Some chemokines are constitutively expressed and regulate the homing, maturation, and microenvironmental segregation of immune cells within lymphoid organs in homeostasis (1, 2). Alternatively, other chemokines are induced in response to pathogenic exposures or inflammatory signals to promote leukocyte mobilization and regulate the immune responses and differentiation of the recruited cells, thus orchestrating the first steps of both innate and acquired immune responses (3, 4). Interestingly, some members of both these “homeostatic” and “inducible” chemokine groups have been shown to exert antimicrobial activity (5). This antibacterial activity, demonstrated in the recent years for a considerable number of mammalian chemokines, provides them with a novel role in immune surveillance and antimicrobial immunity.

In teleost fish, many chemokine genes belonging to different chemokine families have been identified in the past years (6). In all fish species in which the presence of chemokine genes has been explored (mostly aquacultured fish species), the presence of chemokines belonging to both the CXC and CC families has been demonstrated (6). C chemokines, in contrast, have only been reported thus far in zebrafish (Danio rerio) (7). Interestingly, in this species, a further fish-specific chemokine subfamily designated as CX has been identified. These CX chemokines lack one of the two N terminus–conserved cysteine residues but retain the third and the fourth ones in contrast to the C family that only retains the second and fourth of the signature cysteine residues (7). No CX₃C chemokines have ever been reported in fish.

The CC chemokine family is the largest chemokine group both in mammals and teleost fish. Actually, this family has suffered a considerable expansion in some teleost fish species as a result of extensive species-specific intrachromosomal duplications. Thus, although there are 24 CC chemokines in humans (3), 18 CC chemokines have been identified in rainbow trout (Oncorhynchus mykiss) (8), 30 in Atlantic salmon (Salmo salar) (9), 26 in catfish (Ictalurus punctatus) (10), and 81 in zebrafish (7). This expansion, together with the fact that chemokines are proteins that evolve faster than other immune genes in response to the different antigenic experiences (9, 11), has significantly hindered the identification of true homologues of mammalian CC chemokines among fish chemokines. To clarify this issue, in 2007, Peatman and Liu (9) performed an extensive phylogenetic analysis using CC chemokine sequences from rainbow trout, Atlantic salmon, catfish, and zebrafish along with mammalian CC chemokines and established seven major groups of highly related CC chemokines. These included the CCL20 group, the CCL27/28 group, the monocyte chemotactic protein group, the macrophage inflammatory protein (MIP) group, the CCL17/22 group, the CCL19/21/25 group, and a fish-specific group.

CK11 is the only CC chemokine ascribed to the CCL27/28 group established by Peatman and Liu (9) among the rainbow trout CC chemokines known to date. Interestingly, other fish species, such as salmon or zebrafish, express up to four different CCL27/28-like chemokines (9). In mammals, CCL27 and CCL28 signal through the same receptor, CCR10 (12). CCL27 is a chemokine predominantly expressed in the skin, whereas CCL28 is expressed in different mucosal surfaces, such as the intestine, the reproductive tract, or the lungs (12). CCL27 recruits memory T cells that express CCR10, thereby playing an important role in T cell–mediated skin inflammation (12, 13). In contrast, CCL28 has been reported as responsible for the recruitment of IgA-producing CCR10⁺ plasma cells to mucosal sites (4). Remarkably, although human and mouse CCL28 have a potent antimicrobial activity against Candida albicans, Gram-negative, and Gram-positive bacteria, CCL27 showed no significant bactericidal activity and only killed C. albicans at high concentrations (14).

Previous studies performed by our group focused on establishing how a bath infection with viral hemorrhagic septicemia virus (VHSV) affected the levels of transcription of a wide range of chemokines, which established that CK11 was one of the few chemokines significantly upregulated in response to the virus in the gills but not in the fin bases, the main site of entry for this virus (15). CK11 transcription was also significantly upregulated along the digestive tract in response to an infection with infectious pancreatic necrosis virus or to oral vaccination against this virus (16). However, to date, the biological role of rainbow trout CK11 is still unknown, and this is what we have addressed in the current paper. Our results show that although naive leukocyte populations from central immune organs or LPS-stimulated leukocytes from peripheral blood were not significantly attracted by CK11, this chemokine, which is mostly expressed in skin and gills in homeostasis, has a strong antimicrobial activity. We have demonstrated that CK11 strongly inhibits the growth of different rainbow trout Gram-positive and Gram-negative bacterial pathogens, namely Lactococcus garvieae, Aeromonas salmonicida subsp. salmonicida, and Yersinia ruckeri. Our results also showed a significant effect of CK11 on the viability of Ichthyophthirius multifiliis, a parasitic ciliate that infects skin and gills of different fish species (17). To our knowledge, this study constitutes the first report of a fish chemokine with antimicrobial activity, thus establishing a novel role for teleost chemokines in immune surveillance and direct pathogen killing that adds further evidence to the previously suggested evolutionary relationship between chemokines and antimicrobial peptides.

A multiple sequence alignment of rainbow trout CK11 with human and mouse CCL27 and CCL28 was performed using the Clustal Οmega program (https://www.ebi.ac.uk/Tools/msa/clustalo/). The prediction of the secondary structure of rainbow trout CK11 was performed with the Iterative Threading Assembly Refinement server (https://zhanglab.ccmb.med.umich.edu/I-TASSER/) (18, 19) and was visualized using the University of California, San Francisco chimera program (http://www.rbvi.ucsf.edu/chimera) (20). The predictions of secondary structures of human CCL27 (21) and CCL28 (22) were also performed for comparative reasons.

Rainbow trout (O. mykiss) adults of ∼200 g (transcriptional analysis and chemotaxis experiments) and 7–12 g rainbow trout fry (A. salmonicida and L. garvieae infections) were obtained from Centro de Acuicultura El Molino (Madrid, Spain) and maintained at the animal facilities of the Animal Health Research Center in a recirculating water system at 14°C, with a 12:12-h light/dark photoperiod. Fish were fed twice a day with a commercial diet (Skretting). Prior to any experimental procedure, fish were acclimatized to laboratory conditions for at least 2 wk. All of the experiments described comply with the Guidelines of the European Union Council (2010/63/EU) for the use of laboratory animals and have been approved by the Instituto Nacional de Investigación Agraria y Alimentaria Ethics Committee (Code CEEA 2011/044).

The experimental challenges with Y. ruckeri and I. multifiliis were performed at the facilities of the Laboratory of Aquatic Pathobiology at the University of Copenhagen. For this, female rainbow trout hatched from disinfected eggs and reared under pathogen-free conditions at the Bornholm Salmon Hatchery (Nexø, Denmark) were brought to the University of Copenhagen when reaching a mean weight of 7 g (600° days posthatch). These experimental procedures were approved by the Experimental Animal Inspectorate under the Ministry of Food, Agriculture and Fisheries (license: 2015-15-0201-00508/Imm. Stim.).

Fish were killed by benzocaine (Sigma-Aldrich) overdose administered through bath immersion (50 mg/l), following the recommendations from Zahl et al. (23). Before collecting the organs, a transcardial perfusion of each fish was conducted to remove the circulating blood from the tissues using teleost Ringer solution (pH 7.4) containing 0.1% procaine (24). Thereafter, skin, thymus, gills, adipose tissue, brain, heart, foregut, stomach, pyloric caeca, midgut, hindgut, gonad, liver, head, and posterior kidney and spleen were collected and placed in TRI Reagent solution (Invitrogen).

Total RNA was extracted from organs using TRI Reagent solution following the manufacturer’s instructions and then quantified using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific). One microgram of RNA was treated with DNase I to remove any genomic DNA traces using a RapidOut DNA Removal Kit (Thermo Fisher Scientific) and then used to synthesize cDNA using the RevertAid Reverse Transcriptase (Thermo Fisher Scientific), primed with oligo(dT)23VN (1.6 μM), following the manufacturer’s instructions. cDNA was diluted in nuclease-free water and stored at −20°C until use.

To evaluate the levels of transcription of the CK11 gene, a real-time PCR was performed in a LightCycler 96 System instrument (Roche) using FastStart Essential DNA Green Master reagents (Roche) and specific primers previously described (Supplemental Fig. 1) (25). Each sample was subjected, in duplicate, to an initial cycle of denaturation (95°C for 10 min), followed by 40 amplification cycles (95°C for 10 s, 60°C for 10 s, and 72°C for 10 s). A dissociation curve was obtained by reading fluorescence every degree between 60°C and 95°C to ensure only a single product was amplified. Negative controls with no template and minus reverse transcription controls were included in all experiments. The expression of the CK11 gene was normalized to the relative expression of the rainbow trout elongation factor (EF-1α) gene amplified using specific primers previously described (Supplemental Fig. 1) (16). Expression levels were calculated using the 2^−ΔCt method, where ΔCt is determined by subtracting the EF-1α value from the target cycle threshold as described previously (26, 27).

Blood was extracted from the caudal vein of freshly killed rainbow trout using a heparinized (Sigma-Aldrich) syringe and diluted 10 times with Leibovitz (L-15) medium (Life Technologies) containing 100 IU/ml penicillin, 100 μg/ml streptomycin (Life Technologies), and 5% FCS (Life Technologies). PBLs were isolated by placing blood samples onto 51% Percoll (GE Healthcare) density gradients. The gradients were centrifuged at 500 × g for 30 min at 4°C, and the interface cells collected and washed twice at 500 × g for 5 min in L-15 containing 5% FCS. Spleen, head kidney (the main hematopoietic tissue in teleost fish), and gills were aseptically removed from fish and passed through a 100-μm nylon mesh (Corning) using L-15 with 5% FCS. Cell suspensions were placed onto 30/51% discontinuous Percoll density gradients and centrifuged at 500 × g at 4°C for 30 min. The interface cells were collected and washed twice in L-15 containing 5% FCS. In all cases, the viable cell concentration was determined by trypan blue (Sigma-Aldrich) exclusion, and cells were resuspended in L-15 with 5% FCS at a concentration of 1 × 10⁶ cells/ml.

The nucleotide sequence corresponding to the extracellular domain sequence of rainbow trout CK11 (GenBank accession number CDQ77591.1, https://www.ncbi.nlm.nih.gov/protein/CDQ77591.1), including a methionine and a 6xHis-Tag at the N terminus, was synthesized and subcloned into the pET30a expression vector and transformed in Escherichia coli BL21 (DE3) (Abyntek Biopharma). The amino acid sequence corresponding to the cloned CK11 is shown in Fig. 1A. A kanamycin-resistant single-positive colony carrying the recombinant plasmid was grown in Luria–Bertani broth supplemented with kanamycin (50 mg/l) at 37°C. After induction with isopropyl-β-d-thiogalactoside (Sigma-Aldrich), cells carrying the recombinant plasmid were harvested and lysed by sonication (3 s on and 6 s off for a total of 15 min). Inclusion bodies in the cell lysate were then pelleted by low-speed centrifugation (6880 × g for 20 min) and dissolved using urea (8 M). The recombinant CK11 protein was purified by one-step purification using Nickel columns (Sigma-Aldrich) and then pooled. The protein refolding was carried out by dialysis against the storage buffer (50 mM Tris-HCl, 10% glycerol, and 0.5 M l-arginine [pH 8]) using a Slide-A-Lyzer Cassette (Thermo Fisher Scientific) with a refolding ratio 1:100 (v/v) and two buffer changes. After the dialysis, the protein was centrifuged (17,000 × g for 30 min) and filtered through 0.22 μm. The purified protein was analyzed by SDS-PAGE using a 4–20% gel under reducing conditions. The gel was stained with Coomassie Blue to confirm the purity of the expressed protein or transferred onto a polyvinylidene difluoride membrane (Bio-Rad Laboratories) for Western blot analysis. In this case, after blocking the membrane in PBS with 5% skim milk for 1 h, it was incubated with a mouse anti-His mAb (Invitrogen) in blocking solution at 4°C overnight. After three washing steps, the membrane was incubated for 1 h with a sheep anti-mouse IgG-HRP–linked whole Ab (GE Healthcare Life Sciences) as a secondary Ab. The reactive bands were visualized using the ECL system (GE Healthcare Life Sciences).

Protein concentrations were determined in a BCA protein assay (Thermo Fisher Scientific) and stored at −80°C until use. For antimicrobial and some chemotaxis assays, CK11 was dialyzed against 20 mM potassium phosphate buffer (pH 6) given that, in mammals, this buffer has been preferentially used to test the antimicrobial activity of chemokines (14, 28). This dialysis was performed at 4°C overnight using a Slide-A-Lyzer Cassette in a 2-l beaker with three buffer changes. The ToxinSensor Chromogenic LAL Endotoxin Assay Kit (GenScript) was used to confirm the absence of LPS in the recombinant protein.

The chemotactic activity of the recombinant CK11 was performed in chemotaxis chambers introduced in 24-well plates (Costar-Corning Life Sciences). Six hundred microliters of medium (L-15 medium supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin, and 5% FCS) containing different dilutions of recombinant CK11 (100, 500, and 1000 ng/ml) were placed in the bottom of the wells. Negative controls consisted of 600 μl of medium containing the same volume of 20 mM potassium phosphate buffer (pH 6) or refolding buffer as that of CK11. As a positive control, recombinant CK9 (100 ng/ml) was added to the culture medium in some wells. After introducing the chemotaxis chambers in each of the wells, 100 μl of the different leukocyte cell suspensions were loaded on the upper part of the chamber. When needed, blood leukocytes were stimulated with LPS (100 μg/ml) (Sigma-Aldrich) for 24 h at 19°C before conducting the chemotaxis assays. The upper and lower chambers are separated by a 3-μm pore-sized polycarbonate filter. After 2 h of incubation at 19°C, the 600 μl of the bottom chamber were harvested, and the migrated cells were analyzed based on side and forward light scatter parameters on a FACSCalibur flow cytometer equipped with CellQuest software (BD Biosciences) at a constant flow time (1 min).

To elucidate whether CK11 could signal through a potential receptor on rainbow trout leukocytes, we also studied the effect of the recombinant CK11 on the levels of transcription of different immune genes in head kidney leukocytes. For this, head kidney leukocytes obtained as described above were seeded in 24-well plates at a cell density of 2 × 10⁶ cells per well and immediately exposed to recombinant CK11 (1 μg/ml). Wells treated with the same volume of refolding buffer were included as controls. After 24 h of incubation at 19°C, total RNA was extracted from the cells using the Direct-zol RNA Miniprep kit (Zymo Research) that includes a DNase treatment following the manufacturer’s instructions. RNA was then quantified using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific), and 500 ng of RNA were used to synthesize cDNA as described before. The levels of transcription of several immune genes, including proinflammatory cytokines such as IL-1β, IL-8, and TNF-α as well as Ig (IgM, IgT, and IgD), were evaluated by real-time PCR as described before using specific primers (Supplemental Fig. 1) (15, 25, 29, 30). In some experiments, the effect of the native CK11 was compared with that of the heat-denatured CK11 (95°C for 5 min) using a range of concentrations (100, 500, and 1000 ng/ml).

The Gram-positive fish pathogen L. garvieae CF00021 was aerobically grown in de Man, Rogosa, and Sharpe broth (Oxoid, Basingstoke, U.K.) at 30°C. The Gram-negative fish pathogens A. salmonicida subsp. salmonicida CECT4237 and Y. ruckeri LMG3279 were aerobically grown in Tryptone Soya Broth (Oxoid) at 25°C.

The antimicrobial activity of CK11 against three rainbow trout bacterial pathogens was determined using a CFU assay as previously described (28). Briefly, bacteria were grown in broth as described above overnight to exponential phase, and then, cells were harvested, washed, and resuspended in 20 mM potassium phosphate buffer (pH 6). Bacteria (∼1 × 10⁴ CFU/ml) were then incubated with CK11 at different concentrations (0.07–9 μM) or buffer alone at 30°C (L. garvieae CF00021) and 25°C (A. salmonicida CECT4237 and Y. ruckeri LMG3279) for 4 h. After preparing serial dilutions, bacterial cells were spread on agar plates. After incubation for 24–72 h, colonies were counted, and the percentage of viability determined. An irrelevant rainbow trout protein (CD40L) also bearing an N-terminal 6xHis-Tag, was produced in the same conditions, dialyzed against 20 mM potassium phosphate buffer (pH 6), and used as a functional control.

Bacteria were incubated with buffer or rainbow trout CK11 (2.25 and 9 μM) for 4 h as described above and subjected to negative staining using 2% uranyl acetate (Electron Microscopy Sciences). Bacterial samples were examined in a JEOL JEM 1010 transmission electron microscope operated at 80 kV accelerating voltage at the Centro Nacional de Microscopía Electrónica (Facultad de Químicas, Universidad Complutense de Madrid, Madrid, Spain).

Theronts of I. multifiliis were also subjected to transmission electron microscopy (TEM). After 1 h of incubation with rainbow trout CK11 or buffer alone, theronts were collected, fixed in 10% neutral formalin (Merck), and stored at 4°C. Theronts were postfixed in 1% osmium tetroxide (TAAB Laboratories) at room temperature for 1 h, rinsed with distilled water three times, and gradually dehydrated through a series of 30, 50, 70, 80, 90, 95, and 100% acetone for 15 min in each concentration. Samples were then embedded gradually in a series of 1:3 (1 h), 1:1 (1 h), and 3:1 (2 h) v/v Spurr resin-acetone (TAAB Laboratories) mixtures and in pure resin overnight at room temperature. Thereafter, blocks were incubated at 70°C for 48 h until complete polymerization. Ultrathin tissue sections were processed in a Reichert Ultracut S microtome (Leica Biosystems) and mounted on nickel grids, which were stained with 2% aqueous uranyl acetate for 6 min and 1% lead citrate (Electron Microscopy Sciences) for 2 min. Finally, samples were visualized with a JEOL JEM 1010 electron microscope operated at 100 kV at the Centro Nacional de Microscopía Electrónica (Facultad de Químicas, Universidad Complutense de Madrid).

Fixed theronts of I. multifiliis previously treated with CK11 or buffer alone for 1 h were fixed in 10% neutral formalin, washed, and dehydrated in a series of graded ethanol solutions (30, 50, 70, 80, 90, and 100%) for 5 min each. Critical-point drying with CO₂ was performed; samples were attached to stub holders and coated with gold. A scanning electron microscope (JEOL JSM 6400) operated at 20 kV at the Centro Nacional de Microscopía (Facultad de Químicas, Universidad Complutense de Madrid) was used for observation.

For the L. garvieae and A. salmonicida infection experiments, female rainbow trout (7–12 g) were divided into two groups. One group was transferred to 2 l of a bacterial solution containing L. garvieae or A. salmonicida (1 × 10⁷ CFU/ml), whereas the other group (control group) was transferred to 2-l tanks containing an equivalent amount of noninfected culture broth. After 3 h of incubation with strong aeration at 15°C, rainbow trout were transferred to water tanks, maintaining water-quality parameters at optimal levels (31). After 24 h, fish were killed, and skin and gills were sampled from the left side of fish for gene expression analysis as described before.

To carry out the Y. ruckeri infection trial, female rainbow trout (mean weight of 7 g) were bath challenged for 6 h in a concentration of 1.4 × 10⁷ CFU/ml of Y. ruckeri O1 biotype 2 (100415-1/4) and then transferred to their original water tanks (32). After 24 h, fish were killed, and skin and gills were sampled from the left side of the fish for gene expression analysis as described before. A control mock-infected group treated in the same way was also included.

Theronts of I. multifiliis (8–12 per well) were incubated in the presence of CK11 (2.25 and 9 μM) or potassium phosphate buffer (pH 6) at 20°C over time. Because of the slowdown effect on theronts in 20 mM buffer, the final concentration of the buffer used for the assay with or without CK11 was reduced to 10 mM. The survival of theronts was recorded every 15 min for 1 h with an inverse microscope (Olympus CK40-F200). The mortality of theronts was judged according to several criteria that included cessation of ciliary beating, lack of cellular movements, and swelling of the cell (33).

An experiment to determine whether CK11 transcription was regulated in response to I. multifiliis was performed at the facilities of the Laboratory of Aquatic Pathobiology at the University of Copenhagen. For this, a total of 80 female rainbow trout fry (5.5–8.5 g) were acclimated in a 60-l tank containing aerated freshwater, which was changed ∼50% once a day during the experiment. Thereafter, fish were randomly distributed in four tanks. Fish tanks were divided into two groups (infected and control group), each comprising two tanks. In tanks corresponding to the infected group, rainbow trout were exposed to infective theronts (2400 theronts/fish) that had been produced from infected fish using the procedure previously described (34). Five fish from each tank were sampled for gene expression analysis at days 1 and 8 postinfection. Tissue samples (gills and skin) were aseptically dissected, placed into 2-ml tubes containing RNAlater (Sigma-Aldrich), prestored at 4°C for 24 h, and subsequently stored at −80°C until use.

Tissue samples placed in 1.5-ml tubes containing Lysis Solution and 2-ME were homogenized in a TissueLyser II (QIAGEN). Total RNA was then extracted using RTN350 (Sigma-Aldrich) according to the manufacturer’s instructions and subsequently treated with DNase I (Thermo Fisher Scientific). cDNA synthesis was performed with 50 ng of total RNA in a 20-μl setup using the MultiScribe Reverse Transcriptase reagent and random hexamers (Roche) following the manufacturer’s instructions (Thermo Fisher Scientific). Real-time PCR analysis was performed as described above.

Data handling, statistical analyses, and graphic representation were performed using Microsoft Office Excel 2010 (Microsoft) and GraphPad Prism version 7.02 (GraphPad Software). For chemotaxis assays, one-way ANOVA followed by Tukey multiple comparisons test as a post hoc comparison was used to determine statistically significant differences in migrated cells among CK11 treatments and controls. When determining the effect of CK11 on head kidney leukocytes, a paired two-tailed Student t test was used to determine statistically significant differences between transcription levels of immune genes in CK11-treated cells and those obtained in control cells. A repeated measure one-way ANOVA followed by Tukey multiple comparisons was used to determine differences among the levels of IgM mRNA transcription in response to native or heat-denatured CK11. For antimicrobial assays, an unpaired two-tailed Student t test was used after log transformation to determine statistically significant differences in bacterial counts between CK11 treatments and control. Moreover, a dose-response curve was performed to calculate the IC₅₀ of CK11 against bacteria. For in vivo challenge tests and antiparasitic assays, an unpaired two-tailed Student t test was also used to determine statistically significant differences in the expression of CK11 between infected and control groups or in the survival of theronts between CK11 treatment and control, respectively. The statistically significant level was accepted at p < 0.05.

The multiple sequence alignment of the mature proteins of rainbow trout CK11 and human and mouse CCL27 and CCL28 revealed that CK11 has the highest percentage of identity with human CCL27 (30.3%), followed by human CCL28 (22.6%) (Fig. 1A). The predictive model of rainbow trout CK11 shows a secondary structure composed of an extended N-terminal coil followed by a short turn of a 3₁₀ helix, a central three-stranded antiparallel β-sheet, and a C-terminal α helix. When the hydrophobic surface maps are displayed, CK11 shows an amphipathic character, revealing areas of hydrophobic and hydrophilic amino acids (Fig. 1B), similar to human CCL27 (Fig. 1C) and CCL28 (Fig. 1D). Moreover, electrostatic potential maps suggest that the surface of CK11 is cationic (Fig. 1E). However, CK11 shows more uncharged patches on the surface than human CCL28 (Fig. 1G) but less than human CCL27 (Fig. 1F). CK11 has a total of six lysines and nine arginines distributed through the peptide sequence (Fig. 1A), which contribute to this cationic character (isoelectric point [pI] of 10.62). Taking into account that the cationic properties of antimicrobial peptides and chemokines are considered responsible for their antimicrobial properties (35), the similar cationic nature of CK11 and mammalian CCL28 seems to anticipate similar antimicrobial capacities for these chemokines. Interestingly, despite the similar secondary structure of mammalian CCL27 proteins, their lower pI values have been proposed to account for their lack of antimicrobial activity (35).

To elucidate the physiological role of CK11, we first studied its constitutive expression in different tissues obtained from unstimulated perfused rainbow trout. The results showed that CK11 was highly expressed in skin, thymus, and gills and moderately transcribed in adipose tissue and brain (Fig. 2). However, low levels of CK11 constitutive expression were detected in different regions of the digestive tract, gonad, liver, heart, head, and posterior kidney (Fig. 2). Finally, CK11 transcription remained undetected in the spleen (Fig. 2).

To determine its biological activity, rainbow trout CK11 was recombinantly produced in E. coli. The recombinant mature CK11 protein obtained contained 86-aa residues with a calculated pI of 10.62 and a theoretical molecular mass of 9.92 kDa (Supplemental Fig. 2). An endotoxin detection assay confirmed the lack of LPS contamination in this recombinant protein.

As a first step, we evaluated its chemotactic activity on unstimulated leukocyte populations isolated from rainbow trout blood, spleen, head kidney, and gills using a transwell system. No chemotactic activity was observed toward cells belonging to the lymphoid or myeloid gates, defined as previously described (32), using a wide range of CK11 doses (100–1000 ng/ml) (Fig. 3). This chemotaxis assay was undertaken with the CK11 dissolved in 20 mM potassium phosphate buffer such as the one used in the antimicrobial assays, but similar results were obtained when the CK11 dissolved in refolding buffer was used (data not shown). In parallel, CK9, another rainbow trout chemokine produced under the same conditions as CK11, showed a significant chemotactic activity for lymphoid and myeloid leukocyte populations from all tissues tested at 100 ng/ml, as previously reported (36). Given that CCL28 is chemotactic for mammalian plasma cells (4) and taking into account that bacterial LPS differentiates B cells to plasmablasts, we determined whether CK11 chemotaxis increased when blood leukocytes were stimulated with LPS for 24 h prior to conducting the assays. However, the preincubation with LPS had no significant effect on the capacity of leukocytes to migrate toward CK11 (data not shown).

As the lack of chemotactic activity of CK11 could be a consequence of the recombinantly produced CK11 not being able to correctly interact with its corresponding receptor(s) on rainbow trout leukocytes, we determined whether CK11 was capable of modulating immune gene transcription in head kidney leukocytes. Our results show that IL-1β and IL-8 transcription levels were significantly upregulated in cells treated with CK11 when compared with cells treated with the buffer alone (Supplemental Fig. 3A). Additionally, a significant downregulation of IgM mRNA levels was observed in head kidney leukocytes incubated with CK11 (Supplemental Fig. 3A). Alternatively, CK11 had no effect on the levels of transcription of TNF-α, IgT, or IgD (Supplemental Fig. 3A). To confirm that these effects were only exerted by the native form of CK11, the effect of CK11 on IgM transcription levels was compared with that provoked by the heat-denatured protein using a range of concentrations. This experiment confirmed that although CK11 significantly downregulated IgM transcription, the heat-denatured protein showed no effect (Supplemental Fig. 3B). These results demonstrate that the CK11 we have recombinantly produced is capable of interacting with a subpopulation of cells among head kidney leukocytes, despite not being able to induce a significant chemotaxis.

Given that antibacterial activity has been demonstrated in the recent years for a considerable number of mammalian chemokines (14) and taking into account the cationic nature of CK11, we investigated the antimicrobial activity of the recombinant CK11 against three of the most relevant bacterial pathogens for rainbow trout, namely L. garvieae, A. salmonicida, and Y. ruckeri. Our results demonstrate that CK11 exerted a strong antibacterial activity against the three pathogens tested. Interestingly, A. salmonicida and L. garvieae were the most sensitive bacteria to the action of this chemokine (Fig. 4), showing the lowest values of IC₅₀ (Table I). The results showed a significant reduction of bacterial counts of A. salmonicida and L. garvieae in the presence of different doses of CK11 and a complete inhibition when exposed to CK11 concentrations ranging 4.5–9 μM (Fig. 4). CK11 also exerted a significant antimicrobial activity against Y. ruckeri (Fig. 4). In this case, the IC₅₀ value was also higher (Table I). When an irrelevant rainbow trout protein, also bearing an N-terminal 6xHis-Tag, was tested in the same conditions, no significant antibacterial activity was ever observed (data not shown).

For some mammalian chemokines, the mechanism through which the antibacterial activity is achieved is based on the capacity of these molecules to disrupt bacterial membranes (37–39). Thus, we also evaluated the effect of CK11 on the integrity of the bacterial pathogens exposed to CK11 by TEM. Bacteria incubated with CK11 (2.25 μM) clearly showed a breakage of the membrane and leakage of cellular content (Fig. 5B, 5E, 5H) in contrast to untreated control bacteria (exposed to the buffer alone), which conserved the structural integrity of the membrane (Fig. 5A, 5D, 5G). When bacteria were incubated with higher concentrations of CK11 (9 μM), remains of bacterial membranes without structural organization were mostly observed in these preparations (Fig. 5C, 5F, 5I), suggesting a complete destruction of bacteria in the presence of high CK11 doses.

Having established that CK11 exerted a high antimicrobial activity against a range of pathogenic bacteria, we next studied whether CK11 transcription was regulated in mucosal surfaces during the early infection stages in response to an experimental infection with the three pathogens tested. After 24 h of infection with L. garvieae, no significant differences were detected in CK11 transcription in skin and gills between infected and control fish groups (Fig. 6A). In the case of the infection with A. salmonicida, although the levels of CK11 transcription in skin were higher in infected animals than those observed in mock-infected fish, these differences did not reach significance (Fig. 6B). No significant changes in CK11 mRNA levels were observed in response to A. salmonicida in gills either (Fig. 6B). In contrast, a significant upregulation of CK11 transcription was detected in the skin of fish infected with Y. ruckeri in comparison with that of mock-infected controls (Fig. 6C). No effect of the bacteria was observed at this point in CK11 mRNA levels in the gills (Fig. 6C).

Finally, we also addressed whether CK11 was able to inhibit I. multifiliis, a ciliate that colonizes rainbow trout gills and skin. A significant inhibitory effect of CK11 at 9 μM was detected on the viability of I. multifiliis theronts over time (Fig. 7). Theront death seemed evident as theronts lacked swimming movement, ceased their ciliary beating, and swelling of the cell was apparent. This effect was not observed when lower CK11 doses were used (Fig. 7). To verify the degree of affectation of I. multifiliis in the presence of CK11, an analysis of theronts by TEM allowed us to confirm that theronts incubated with CK11 displayed lysed membranes (Fig. 8A, 8B). The cellular structure also showed a great degree of disorganization in comparison with control theronts, with major effects on the macronucleus and the mitochondria (Fig. 8C–F). Interestingly, CK11-treated theronts lost most of the cilia (Fig. 8E, 8F). Similar results were observed by scanning electron microscopy, confirming that cilia had fallen off and the membrane was disrupted in CK11-treated theronts (Fig. 8G, 8H).

Given the antiparasitic activity found for CK11, we also studied the regulation of CK11 transcription in the skin and gills of rainbow trout infected by this ectoparasite. In this case, we decided to test this effect at an early time point (day 1) and at day 8 postinfection as this was the time point when the parasite could be found attached to the host. We observed that the levels of transcription of CK11 were significantly upregulated at day 1 postinfection and then downregulated at day 8 postinfection in skin in response to the parasite when compared with transcription levels observed in noninfected controls (Fig. 9). However, CK11 transcription levels were upregulated at day 8 postinfection in the gills of infected fish (Fig. 9).

Antimicrobial peptides produced by eukaryotes are evolutionary ancient effector molecules that constitute one of the first lines of host defense against bacteria, fungi, enveloped viruses, and protozoa (39–41). Consequently, these molecules are known to play a key role in the innate immune response to these pathogens. Interestingly, recent evidence points to a common origin of antimicrobial peptides and chemokines, given their similar functions. Thus, many antimicrobial peptides have been shown to be chemoattractant for specific leukocyte subsets (42–44). In some cases, these chemoattractant properties are executed through the action of chemokine receptors, such as for example β-defensins, revealed to be potent agonists of the CCR6 chemokine receptor (44). Then again, numerous mammalian chemokines have been found to exert potent antimicrobial activities against Gram-positive and Gram-negative bacteria, fungi, and even protozoa (5, 40). Despite these evidences, very little is known concerning the direct action of chemokines against microorganisms or their relationship with antimicrobial peptides in an ancient animal group such as teleost fish. Hence, this is what we have undertaken in the current study, focusing on rainbow trout CK11, given its phylogenetical relation to mammalian CCL27 and CCL28 and its strong levels of transcription in gills and skin in homeostasis.

Our findings revealed that CK11 has a strong antimicrobial activity against the Gram-positive pathogen L. garvieae and also against the Gram-negative pathogens A. salmonicida and Y. ruckeri. Similar to most antimicrobial peptides (37, 38, 45), CK11 seems to exert its antimicrobial activity by the disruption of the membrane as shown by TEM. The prediction of the secondary structure of CK11 justifies its biological function as an antimicrobial agent because its cationic nature should permit this protein to accumulate and interact with the anionic lipid molecules from bacterial surfaces, such as the lipoteichoic acid at the cell wall of Gram-positive bacteria and the LPS at the outer membrane in Gram-negative bacteria. Moreover, the amphipathic character of CK11 will possibly mediate its insertion into the phospholipids of the cell membrane, leading to its permeabilization and disruption. This mode of action is similar to that reported when Pseudomonas aeruginosa, Streptococcus mutans, and C. albicans were treated with human or mouse CCL28 (14) or when S. pyogenes, S. dysgalactiae subsp. equisimilis, Staphylococcus aureus, P. aeruginosa, and E. coli were exposed to granulocyte chemotactic protein 2 (GCP-2)/CXCL6 (46). In all these cases, similar morphological changes were observed in the cell membrane of the bacteria, which lead to the leakage of bacterial contents as a result of membrane disruption.

Thus, rainbow trout CK11, although phylogenetically related to both CCL27 and CCL28, shares a higher amino acid identity with CCL27 but has maintained the antimicrobial properties of CCL28 that CCL27 does not show, most probably as a consequence of its lower pI value (9.11 in the case of human CCL27) (35). In contrast, CK11 is strongly expressed in homeostasis in the skin, similarly to CCL27 (47), whereas CCL28 is not expressed on the skin but is expressed in other mucosal surfaces, such as intestines, reproductive tracts, lungs, stomach, lactating mammary glands, and salivary glands (14, 48, 49). Likewise, we did observe CK11 transcription in other tissues, including gills and different intestinal segments. Altogether, these results point to CK11 as a common ancestor of both CCL27 and CCL28 and suggest that it was after the emergence of tetrapods that CCL27 lost its antimicrobial activity. Furthermore, these results constitute the first description, to our knowledge, of antimicrobial activities for a teleost fish chemokine. To date, only a couple of studies had previously suggested the antimicrobial role of fish chemokines. One of these studies reported that a synthetic peptide that contained the 12-aa residues (WS12) of the C-terminal helical region of IL-8 from murrel (Channa striatus) showed antimicrobial activity against Bacillus cereus at a concentration of 50 μM (50). Very recently, the α helix domain of rainbow trout and salmon IL-8 was revealed to exert antimicrobial activity against E. coli and P. aeruginosa at 30–40 μM and against S. aureus over 100 μM (51).

Alternatively, our chemotaxis assays revealed that CK11 was unable to significantly recruit leukocytes from the blood, spleen, head kidney, or gills obtained from nonstimulated fish, contrary to what occurs with many other fish chemokines, such as CK9, CK12b, CK1, or IL-8, among others (36, 52–54). These results suggest that CK11 might display chemotactic activity only on a specific leukocyte subpopulation not widely represented in unstimulated leukocyte populations, as shown for mammalian CCL27 and CCL28 (4, 13). In humans, CCL27 selectively attracts circulating cutaneous lymphocyte-associated Ag⁺CD45RO⁺ T lymphocytes expressing CCR10, which have been reported constitute a small percentage of total circulating T cells (13, 55). Likewise, CCL28 specifically attracts IgA-producing CCR10⁺ plasma cells (4). Because LPS is known to be a potent inducer of IgM secretion in rainbow trout (56), we assayed whether we could detect a significant chemotaxis toward CK11 using blood leukocytes previously stimulated with LPS, but no significant increase in the number of migrated cells was observed (data not shown). In fish, a further detection of chemoattractive activities for specific leukocyte subsets is currently a challenge because of the lack of specific markers for the isolation of specific B and T cell subpopulations. Additionally, a true homologue of mammalian CCR10 has not been reported to date in fish, thus the characterization of specific populations on which CK11 signals awaits further study in the future. It is worth noting that despite the lack of chemotactic activity detected by CK11, the chemokine in its native form was able to modulate the levels of transcription of some immune genes in head kidney leukocyte cultures, thus suggesting that at least a small subpopulation of leukocytes within head kidney cultures is able to recognize CK11 and modulate their transcription profile in response to the chemokine. Similarly, the rainbow trout chemokine CK12a also showed no chemotactic activity in vitro, but a capacity to induce the transcription of immune genes (57).

The selective expression of CK11 mRNA in the skin and gills together with its antimicrobial activity suggests that this chemokine is secreted to protect the external mucosal surface against colonizing pathogenic microorganisms. In concordance with this hypothesis, our transcriptional studies showed that CK11 was significantly upregulated in response to Y. ruckeri in the skin but not in the gills, the latter being the initial site of bacterial attachment and early penetration of epithelial cells (58, 59). Similarly, no significant changes on CK11 expression after exposure with A. salmonicida were observed in skin and gills, the two major routes of infection for this microorganism along with the gastrointestinal tract (60). Thus, it seems as if those tissues that induce CK11 are able to control the bacterial infection, whereas in those tissues in which the bacteria establishes an early replication site, CK11 levels are not upregulated. Previous studies performed in response to VHSV also pointed to this lack of CK11 regulation in tissues in which the pathogens are able to establish themselves, given that a significant upregulation of CK11 transcription in response to VHSV was found in the gills where viral replication is almost nondetectable although it was not regulated in the fin bases, the main site of virus entry (15). Likewise, human CXCL14, highly expressed in the epidermis and dermis of healthy human skin and characterized by its broad spectrum of action against cutaneous Gram-positive bacteria as well as E. coli and C. albicans, was downmodulated under conditions of inflammation, that is after treatment with TNF-α and LPS (28). In this sense, several evasive strategies used by pathogens to avoid the action of antimicrobial peptides have been previously described (61, 62). For instance, the pathogen Shigella flexneri suppressed the transcription of genes encoding antimicrobial cationic peptides, principally the human β-defensin hBD-3, which is especially active against this microorganism, and CCL20, leading to a lower recruitment of dendritic cells to the lamina propria of infected tissues (62). In contrast, the lack of response of CK11 to L. garvieae could be due to the fact that this microorganism does not colonize skin or gills, given that in this case the fecal-oral–pathway has been reported as the main route of infection (63).

Finally, as some human chemokines have shown antiparasitic activity against protozoa (35, 61), we investigated whether CK11 was capable of affecting the viability of I. multifiliis, an external protozoan parasite of rainbow trout. Our experiments confirmed that CK11 exerts a direct effect on the cellular membrane of I. multifiliis theronts and, consequently, on their viability. These results have been confirmed by scanning electron microscopy and TEM in which a complete disorganization of the cellular content, loss of cilia, and membrane disruption in response to CK11 have been confirmed. Interestingly, a previous study showed direct parasiticidal effects of several human chemokines on the promastigote form of Leishmania mexicana, CCL28 being the most potent chemokine against this protozoan (35). Furthermore, we demonstrated that through the course of an experimental infection with I. multifiliis, the transcription of CK11 was significantly upregulated in the skin at day 1 postinfection and afterward suppressed at day 8 postinfection. Alternatively, the transcription of CK11 was upregulated in gills at day 8 postinfection but not at day 1 postinfection, revealing a tissue-specific CK11 response during the infectious process. As in the case of Y. ruckeri, it could be possible that the lack of CK11 response in the gills at an earlier stage of infection may facilitate the theronts penetrating the host epithelia. In contrast, it seems that in both cases, the skin is capable of responding at a very early time point. In mammals, it has been shown that the expression of CCL20, a chemokine with direct antiparasitic activity against the enteric protozoan Cryptosporidium parvum, was downregulated in the intestine of neonatal mice during the in vivo infection with this parasite possibly as a result of a parasite escape mechanism (61). Thus, one line of future applications of these studies would be aimed at increasing the production of these antimicrobial chemokines at mucosal surfaces to overcome the effects of the pathogens to control their production.

In conclusion, the phylogenetic studies, together with those related to tissue distribution and biological activity, performed in this work strongly support the hypothesis of rainbow trout CK11 being a common ancestor for both CCL27 and CCL28. Additionally, this study constitutes the first report, to our knowledge, of a chemokine from teleost fish with antimicrobial and antiparasitic activities, thus establishing a novel role for chemokines in antimicrobial immunity in an ancient vertebrate group that supports an evolutionary relationship between chemokines and antimicrobial peptides.

We thank Dr. C. Michel (Institut National de la Recherche Agronomique, Jouy-en-Josas, France) for providing L. garvieae CF00021 and the technical service of the Centro Nacional de Microscopía Electrónica (Facultad de Químicas, Universidad Complutense de Madrid, Madrid, Spain) for assistance in TEM and scanning electron microscopy.

The authors have no financial conflicts of interest.