Pseudomonas M162 confers protection against rainbow trout fry syndrome by stimulating immunity

Authors


Tiina Korkea-aho, Animal Health and Welfare Unit, Finnish Food Safety Authority Evira, Mustialankatu 3, 00790 Helsinki, Finland. E-mail: tiina.korkea-aho@evira.fi

Abstract

Aims:  To evaluate the antagonistic effect of Pseudomonas M162 against Flavobacterium psychrophilum.

Methods and Results:  The antagonistic activity of M162 was tested in vivo and in vitro, and its mode of action examined by siderophore production and immunological responses of rainbow trout (Oncorhynchus mykiss) fry. Pseudomonas M162 inhibited the growth of Fl. psychrophilum in vitro and increased the resistance of the fish against the pathogen, resulting in a relative per cent survival (RPS) of 39·2%. However, the siderophores produced by M162 did not have an inhibitory effect on Fl. psychrophilum. In fish fed with M162, the probiotic colonized the gastrointestinal tract and stimulated peripheral blood leucocyte counts, serum lysozyme activity and total serum immunoglobulin levels after 3 weeks from the start of feeding.

Conclusions:  This study showed the potential of Pseudomonas M162 as a probiotic by reducing the mortalities that occurred during an experimental Fl. psychrophilum infection, resulting mainly through the immunostimulatory effects of the bacterium.

Significance and Impact of the Study:  Rainbow trout fry syndrome (RTFS) causes high mortalities during the early life stages of the fish’s life cycle, partly because their adaptive immunity has not yet fully developed. Thus, immunomodulation by probiotics could be an effective prophylactic method against RTFS.

Introduction

Flavobacterium psychrophilum causes high levels of mortalities in aquaculture, especially in salmonid culture during the early life stages of rainbow trout (Oncorhynchus mykiss, Walbaum) (Nematollahi et al. 2003). To date, antibiotics have been the only treatment available against Fl. psychrophilum infections in salmonid aquaculture. However, the use of antibiotics is an emotive issue because of the emergence and spread of antibiotic resistance strains of the pathogen (Bruun et al. 2003). Probiotics are defined as beneficial live micro-organisms, when administered to a host at an effective dose (FAO/WHO 2001). The use of probiotics as an alternative strategy for reducing disease outbreaks in aquaculture systems, instead of the application of antibiotics, has received considerable attention in recent years (see reviews Irianto and Austin 2002; Balcázar et al. 2006).

The specific action of a probiotic is often based on its antagonistic activity at the site of colonization on the host’s surface, enhancement of the host’s immune response, competition for nutrients, and production of antimicrobial substances towards the pathogen (Irianto and Austin 2002). In particular, the inhibitory effect of probiotic bacteria on fish pathogens has been demonstrated under iron-depleted conditions, resulting in the production of siderophores by the probiotic in the absence of iron (Gram et al. 2001; Spanggaard et al. 2001; Brunt et al. 2007). Siderophores are high affinity iron acquisition molecules produced by bacteria, giving them a competitive advantage in iron scarce environments (Miethke and Marahiel 2007). Recently, the inhibitory effects of Pseudomonas sp. M174 against Fl. psychrophilum, the causal agent of rainbow trout fry syndrome (RTFS), were demonstrated in vitro and in vivo (Korkea-aho et al. 2011). One of the probable modes of action of the Pseudomonas spp. is thought to be through its ability to produce siderophores (Korkea-aho et al. 2011; Ström-Bestor and Wiklund 2011).

One of the reasons for the high mortality seen in fry during RTFS outbreaks could be the lack of a developed adaptive immune response in these young fish. However, the young fish may be protected by innate immunity (Mulero et al. 2007), and thus, the effect of immunostimulants and probiotics may be beneficial in improving their innate disease resistance. Moreover, natural antibodies are present in the serum of mammals, birds and fish without any apparent antigenic stimulation and are thought to be an important innate humoral defence against invading pathogens (Sinyakov et al. 2002; Kachamakova et al. 2006). A number of immunological parameters have been found to be stimulated in fish fed with probiotics (for review see Nayak 2010), including total serum IgM levels (Nikoskelainen et al. 2003; Balcazar et al. 2007; Salinas et al. 2008).

The aim of the present study was to evaluate the antagonist activity of Pseudomonas M162 against Fl. psychrophilum and to examine the possible modes of action, including siderophore production and its effect both on innate immunity and specific humoral responses in rainbow trout fed with this potential probiotic. The inhibitory activity of serum from these fish against Fl. psychrophilum was also examined in vitro. In addition, a comparison was made between a previously reported probiotic Pseudomonas sp. M174 (Korkea-aho et al. 2011) and Pseudomonas sp. M162 regarding their mode of action, genetic similarity and their ability to inhibit Fl. psychrophilum.

Materials and methods

Isolation and identification of the putative probiotic

Pseudomonas M162 was isolated from the surface of rainbow trout eggs and characterized by 16S rDNA sequencing, using 16S rDNA universal primers (Lane 1991; Korkea-aho et al. 2011). Partial sequence alignment between Pseudomonas M162 and Pseudomonas M174 was performed using Basic Local Alignment Tool (blast, NCBI; Zhang et al. 2000).

Bacterial cultures and growth conditions

For the in vitro experiments, Fl. psychrophilum strains ST2/00 (Madetoja et al. 2002) and 413 (provided by Dr L-R Suomalainen, University of Jyväskylä, Finland) were used. Cultures were maintained on R2A agar (LAB M, Bury, UK) at 4°C and subcultured every 7–10 days. Flavobacterium psychrophilum strains were cultured at 15°C in 10 ml of Shieh medium (Decostere et al. 1997). For the siderophore analysis, the cultures were grown and supernatants collected as described by Korkea-aho et al. (2011).

For the in vivo experiments, cultures of M162 were stored at 4°C and subcultured every 7–10 days in tryptone soya agar (TSA; Oxoid, Basingstoke, UK). A selective medium, Pseudomonas agar supplemented with Pseudomonas C-F-C supplement (Oxoid), was used to culture samples from the gut contents of fish. Flavobacterium psychrophilum JIP02/86 (Duchaud et al. 2007) was used for the experimental infection and was grown and maintained in tryptone yeast extract salt medium (TYES, 0·4% tryptone, 0·04% yeast extract, 0·05% MgSO4·7H2O, 0·05% CaCl2·2H2O, pH 7·2) at 15°C and subcultured every 10–14 days.

Siderophore production

Siderophore production was tested using chrome azurol S (CAS) agar (Schwyn and Neilands 1987), with a slightly modified supplement composition (Korkea-aho et al. 2011). M9 medium with casamino acids (M9C; Atlas 1993), with or without iron, was used as a negative control. Siderophores in culture supernatants were also determined using a CAS solution assay (Schwyn and Neilands 1987; Korkea-aho et al. 2011). The percentage of siderophores in the CAS solution was determined as (Ar − As/Ar) × 100 = siderophore %, where Ar = reference absorbance and As = sample absorbance (Payne 1994).

The presence of siderophores in filter-sterilized supernatants was further analysed using an UV/VIS spectrophotometer (JASCO V-530; Jasco Inc., Easton, MD, USA). The pH of the supernatants was adjusted to 5·5 (±0·2) and the absorbance measured between 500 and 300 nm. Distilled water was used as the blank.

Antagonistic activity against Flavobacterium psychrophilum in vitro

The inhibition of Fl. psychrophilum growth in vitro was assessed using an automated growth chamber and turbidity reader (Bioscreen C; ThermoLabsystems, Helsinki, Finland). In the first trial, inhibition of Fl. psychrophilum was tested using cell-free supernatant filtrates of the probiotic cultured in brain heart infusion broth (BHI; LAB M). A second trial was conducted to elucidate the relationship between the concentration of iron added to the M9C medium and the antagonistic activity of the resulting cell-free filtrate supernatants on Fl. psychrophilum.

For the tests, Fl. psychrophilum was cultured in 10 ml of Shieh medium for 4 days at 15°C. Afterwards, the cells were harvested by centrifugation (3500 g, 5 min) and resuspended in 7 ml of sterile 0·9% (w/v) NaCl. The cell densities used in the first trial were 7 × 107 CFU ml−1 Fl. psychrophilum strain 413 and 1 × 108 CFU ml−1 for Fl. psychrophilum strain ST2/00. In the second trial, the cell densities were 1 × 108 CFU ml−1 for both strains of Fl. psychrophilum, as measured by OD600 values and confirmed with plate counts on R2A agar.

The two bacteria were then added to replicate wells of the supernatant in microplates (Honeycomb; ThermoLabsystems), three replicate wells for the first trial and four wells for the second trial. Each microplate well contained 200 μl Shieh medium, 50 μl of the supernatant and 100 μl of the Fl. psychrophilum culture. Only supernatant and Shieh medium was used as a negative control, whereas Shieh medium (or M9C in the second trial) together with the Fl. psychrophilum culture was used as the positive control. The turbidity of each well was measured every 30 min using a wide-band filter (420–580 nm). The plate was shaken for 10 s prior to each reading.

Safety testing of Pseudomonas M162 in fish

Possible harmful effects of M162 on fish was determined in rainbow trout (average weight = 15 g) as described by Korkea-aho et al. (2011).

Antagonist activity of Pseudomonas M162 in vivo

To assess the antagonistic activity of Pseudomonas M162 against Fl. psychrophilum JIP02/86, rainbow trout fry fed a commercial diet (control) or M162 supplemented feed were experimentally infected with the pathogen. The M162 supplemented diet, containing 2 × 109–5 × 107 CFU g−1, was prepared as described previously by Korkea-aho et al. (2011). The fish (30 per treatment) were divided in duplicate tanks. Initial average weight of the M162 fed fish was 1·9 g and control fish 2·1 g (see challenge experiment two for details in Korkea-aho et al. 2011). For the experimental infection, Fl. psychrophilum JIP02/86 was grown for 5 days at 15°C in TYES broth. Subsequently, the cells were harvested by centrifuging at 3000 g for 10 min and their concentration adjusted to 8 × 105 CFU ml−1 with fresh TYES. After 14 days of feeding, the fish were anaesthetized and injected intramuscularly (i.m.) with Fl. psychrophilum JIP02/86 at 2 × 104 CFU per fish or with TYES (control). Fish were maintained for another 21 days postinjection before ending the trial. To verify specific mortalities, kidney and spleen samples were collected from challenged fish and cultured on TYES agar. Representative yellow colonies were collected from these plates after 7–10 days incubation at 15°C, and bacterial DNA extracted using a NucleoSpin Tissue kit (Macherey-Nagel GmbH, Düren, Germany). The identity of the bacteria was then confirmed using a Fl. psychrophilum-specific PCR as described previously by Korkea-aho et al. (2011). The challenge trial was repeated twice to verify the results.

Assessment of M162 intestinal colonization and immunological analyses

Rainbow trout (initial average weight = 10 ± 0·21 g) were divided into six tanks containing 18 fish per tank. The tanks were supplied with flow-through dechlorinated water at a flow rate of 1·5 l min−1, at an ambient water temperature (average ± SD = 11·8 ± 0·8°C). Three tanks of fish were fed for 2 weeks with Pseudomonas M162 supplemented diet. The other three tanks were fed with unsupplemented feed as controls. The biomass of fish was measured at the start of the experiment and pellets were hand-fed daily at 2% of their biomass over the course of the 30-day trial. After 2 weeks of feeding the probiotic diet, samples of 12 fish per treatment were taken for immunology analyses and nine fish per treatment for microbiology analyses (week 2 sampling). The remaining fish were fed for a further week with the control diet before sampling again (week 3 sampling). Levels of the Pseudomonas M162 were determined from the intestine content of the experimental fish using bacterial counts. The fish were starved for 24 h and nine fish per feeding group were killed by an overdose of anaesthesia (Benzocaine; Sigma-Aldrich Co., St Louis, MO, USA) and exsanguination. Fish were opened aseptically and the whole intestine was removed. The intestinal content was collected in preweighted sterile Eppendorf tubes by slitting the intestine with forceps and removing aseptically the content, which was subsequently diluted in 900 μl sterile phosphate-buffered saline (PBS, 0·02 mol l−1 sodium phosphate, 0·15 mol l−1 NaCl pH 7·2), and homogenized. Serial dilutions of the gut contents made in sterile 0·85% (w/v) NaCl and plated duplicate samples on Pseudomonas agar. After incubating for 3 days at 22°C, bacterial colonies were counted and expressed as the total microbial colony count and M162-specific colony count. The colony identification of Pseudomonas M162 was based on colony morphology, cell morphology, Gram reaction, oxidase reaction (Oxidase strips; Fluka, Buchs, Switzerland) and biochemical identification system API 20E (bioMérieux SA, Marcy-l’Etoile, France).

Innate immunity and haematology

For immunological analyses, blood was withdrawn from the caudal vein of fish from both groups after 2 and 3 weeks from start of the experiment using 1 mL syringes rinsed with heparin (10 IU ml−1; Sigma-Aldrich, Dorset, UK). Samples were used to determine the total erythrocyte and leucocytes counts, where a 10−2 dilution of blood in PBS was used to determine total leucocytes count and a 10−3 dilution was used to count erythrocytes. The cells were counted in four squares of a haemocytometer per sample and expressed as: the number of cells ml−1 = N × dilution factor × 104, where N is the average number of counted blood cells. Plasma was separated from the remaining blood by centrifuging at 3000 g for 10 min and this was stored at −20°C until use.

For the isolation of head kidney macrophages, the head kidney was aseptically removed according to Secombes (1990) and teased through a 100-μm nylon mesh (BD Falcon™; BD Biosciences, Franklin Lakes, NJ, USA) into 2 ml Leibovitz medium (L-15; Sigma-Aldrich) containing 40 μl heparin (10 IU ml−1). The mesh was rinsed with 1 ml of the medium and placed on ice. Two large circles were drawn on a microscope slide with a wax pap pen and 100 μl of the cell suspension was added to the circle. After incubating for 1 h at 15°C, the slides were washed with L-15 to remove adherent cells, and 100 μl yeast suspension (0·025 g in 5 ml of L-15) was place on one circle and L-15 on the second circle (control) for 1 h at 15°C to allow phagocytosis to take place. After washing the slide with L-15, the macrophages were fixed with 100% (v/v) methanol (100 μl) for 5 min and washed five times with 70% (v/v) methanol. The cells were stained with Rapid Romanowsky’s stain (Raymond A lamb, Eastbourne, UK). The slides were examined at ×400 magnification and counted to determine the proportion of cells containing engulfed yeast. Approximately 200 cells were counted in random under a microscope and phagocytic activity (PA) expressed using the following equation: PA = No. of phagocytosing cells/No. of total cells × 100%.

The superoxide anion (inline image) production by head kidney macrophage suspensions was measured by the conversion of nitroblue tetrazolium (NBT; Sigma-Aldrich) to formazan, following the method of Secombes (1990). One hundred microlitres of macrophage suspension was added to the 96-well plate (Iwaki, Tokyo, Japan), incubating at 15°C for 2 h to allow cell attachment. The supernatant was removed and the wells were washed three times with L-15. After washing, 100 μl of L-15 containing 1 mg ml−1 NBT was added to three replicate wells, and this together with phorbol myristic acetate (1 μl ml−1 PMA) was added to another three replicate wells, while 100 μl of lysis buffer (citric acid, 0·1 mol l−1; Tween 20, 1·0% (v/v); crystal violet, 0·05% (w/v); Sigma-Aldrich) was added to two additional replicate wells. The plate was incubated for 60 min at 15°C, the medium was then removed and cells fixed with 100% (v/v) methanol for 2–3 min before washing three times with 70% (v/v) methanol. The plates were air-dried before adding 120 μl of 2 mol l−1 potassium hydroxide (Sigma-Aldrich) and 140 μl of 2 mol l−1 dimethyl sulfoxide (Sigma-Aldrich) to each well to dissolve the resulting formazan. The absorbance was determined at 610 nm using an automated multi-mode microplate reader (Synergy™ HT; BioTek Instruments, Winooski, VT, USA). The number of macrophages attached to the plate was determined by counting the average number of nuclei released by the addition of lysis buffer for two replicate wells. The level of respiratory burst was expressed as an absorbance at 610 nm for 105 cells per sample.

Serum lysozyme activity was based on the lysis of lysozyme sensitive Micrococcus lysodeikticus (Morgan et al. 2008). Briefly, serum (10 μl per well) was placed in triplicate wells of a 96-well plate (Sterilin, Newport, UK), and 190 μl of M. lysodeikticus (Sigma-Aldrich) solution (0·2 mg ml−1) in sodium 0·04 mol l−1 phosphate buffer (SPB, pH 5·8) then added. The absorbance was measured at 540 nm after 1 and 5 min. Two columns of the plate contained 200 μl of SPB used as negative control. A unit of lysozyme activity was defined as the amount of serum causing a decrease in absorbance of 0·001 min−1.

Measurements of total IgM

The level of total IgM in the serum of experimental fish was determined using an enzyme-linked immunosorbent assay (ELISA) with modifications (Nikoskelainen et al. 2003). Flat-bottomed 96-well ELISA plates (Thermo Fisher, Scientific, Waltham, MA) were coated with 100 μl of serum serially diluted from 1 : 50 to 1 : 102 400 in 0·66 mg ml−1 carbonate buffer, pH 9·6 and incubated overnight at 6°C. The wells were washed five times with low-salt washing buffer (LSWB; 0·02 mol l−1 Trizma base, 0·38 mol l−1 NaCl, 0·05% (v/v) Tween 20, pH 7·2) before adding 250 μl of 3% dried skimmed milk (Marvel, Dublin, Ireland) to each well and incubating 1 h at 22°C. After blocking nonspecific binding sites, 100 μl anti-rainbow trout/Atlantic salmon IgM antibody (Aquatic Diagnostics Ltd., Stirling, UK), diluted 1 : 33 in PBS, was added per well and incubated for 1 h at 22°C. Unbound anti-trout/salmon antibody was removed by washing five times with high-salt washing buffer (HSWB; 0·02 mol l−1 Trizma base, 0·5 mol l−1 NaCl, 0·01% (v/v) Tween 20, pH 7·4), before adding 100 μl of anti-mouse IgG conjugated to horseradish peroxidase (HRP; Sigma-Aldrich) at 1 : 2000 dilution in PBS for 1 h at 22°C. The plate was again washed and 100 μl of substrate/chromogen (i.e. 15 ml substrate buffer containing 5 μl hydrogen peroxidase and 150 μl TMB dihydrochloride) added to each well and incubated for 10 min at 22°C. The enzyme reaction was stopped by adding 50 μl of 2 mol l−1 H2SO4 and the absorbance measured at 450 nm. The results were expressed as the average of mean absorbance at a dilution of 1 : 3200, the value which fell within the linear range of the dilution range.

The level of natural antibodies in the serum of experimental fish was determined using an indirect ELISA as described by Kachamakova et al. (2006). Briefly, flat-bottom 96-well ELISA plates were coated with 5 μg ml−1 bovine serum albumin (BSA; Sigma-Aldrich) in coating buffer (carbonate-bicarbonate solution). The plates were left overnight at 4°C and then washed five times with LSWB. After blocking with 3% dried skimmed milk, serum samples were added to the wells, diluted from 1 : 100 to 1 : 25 000 and plates incubated at 22°C for 1 h. After washings with LSWB, 100 μl anti-trout IgM (diluted 1 : 33 in PBS) per well was added and plates incubated for 1 h at 37°C before washing with HSWB. An anti-mouse IgM antibody conjugated with HRP (diluted 1 : 2000 in PBS) was added and plates were incubated for 1 h at 22°C. Substrate for the reaction was added as described above and the absorbance of the reaction measured at 450 nm. The values obtained for a 1 : 200 dilution were chosen as they fell within the linear range of the dilution range.

To measure the specific antibodies against M162, a bacterial protein preparation was used to coat the ELISA plates. The bacterium was cultured as described above, adjusted to an absorbance of 1·383 at 600 nm and washed twice with PBS (centrifugation 1600 g for 15 min). After adding 1 ml PBS, 200 μl protease inhibitor cocktail (Sigma-Aldrich) and 200 μl DNase (1 μg ml−1 in PBS; Sigma-Aldrich), the bacteria were incubated at 4°C for 20 min and sonicated on ice (3 × 10 s) before centrifuging at 10 000 g for 10 min. The supernatant was passed through a 45-μm sterile filter (Ministart; Sartorius Stedim Biotech GmbH, Goettingen, Germany) and filtrate collected in Eppendorf tubes and stored at −20°C. The protein concentration of the filtrate was 476·8 μg ml−1 determined using a Protein Assay Kit (Thermo Scientific, Loughborough, UK). Specific antibodies against Pseudomonas M162 were determined in individual serum samples using the ELISA procedure as described above, with minor modifications: flat-bottom 96-well ELISA plates were coated with 100 μl well−1 of 10 μg M162 bacterial protein ml−1 and incubated overnight at 4°C. The absorbance was measured at 450 nm using the multi-mode microplate reader. The values obtained for a 1 : 200 dilution were again chosen as they fell within the linear range of dilution range.

Bacterial killing activity of serum

The serum bacterial killing activity was measured according to Sharifuzzamann and Austin (2010), by comparing the growth of Fl. psychrophilum strain JIP02/86 in the presence or absence of serum from the experimental fish. Flavobacterium psychrophilum was grown in TYES broth for 3 days at 15°C. The culture was centrifuged at 2500 g for 20 min at 4°C and the concentration adjusted to c. 107 cells ml−1 with fresh TYES broth using OD600. For the assay, 33 μl of serum or TYES (as a control) was placed in wells of a 96-well microtitre plate, before adding 100 μl of the bacterial suspension. This was mixed with a micropipette and incubated for 24 h at 15°C. The plate was then centrifuged at 2500 g for 20 min at 4°C, the supernatant discarded and 100 μl of 0·5 mg ml−1 MTT [3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide; Sigma-Aldrich] added to the wells. After incubating for 15 min in the dark, the plate was read at an OD of 600 nm and the percentage of surviving bacteria calculated as a antibacterial index, that is, (OD600 of each sample/OD600 of control Fl. psychrophilum) × 100.

Statistical analysis

Comparison of the growth curves of Fl. psychrophilum with and without M162 supernatant, siderophore percentage production with and without iron and comparison of cumulative mortalities were analysed using a nonparametric Mann–Whitney test. The relative per cent survival (RPS) was determined according to Amend (1981). Differences in growth curve with or without added iron were tested using a one-way analysis of variances (anova) with least significant difference (LSD) multiple comparison. Comparison between M162 fed fish and control fish in immunological and microbiological analyses, as well as serum killing activity, was tested with a t-test. Significance values of the t-test were interpreted according to Levene’s test results for Equality of Variances either significance results when equal variances or not equal variances were assumed. The statistical analyses were performed using spss 19.0 for Windows software (SPSS Inc., IBM Statistics 19, Chicago, IL, USA).

Results

Identification of M162

Probiotic M162 was characterized as a Gram-negative and oxidase-positive rod. According to API 20 NE results, the isolate had a 99·1% ID match with Pseudomonas fluorescens. Comparison using the nucleotide database blastn confirmed that the 16S rRNA partial gene sequence had a 99% identity with a Pseudomonas sp. This partial sequence has been submitted to EMBL Nucleotide Sequence Database (accession no. FN548143). When 16S rRNA gene sequences of M162 and M174 were aligned, they had a 97% partial sequence identity, suggesting that the organisms represent different species.

Siderophore production

In the CAS agar plate cultures, M162 produced strong orange halos regardless of the iron concentrations used in the medium or number of days grown. The cell-free supernatants produced small orange halos around the wells only after M162 had been grown for 5 and 7 days under iron-deficient conditions. Siderophore production was also detected in the CAS liquid assay, with levels of 82·6% seen on days three, five and seven, and these were statistically significantly different to amount of siderophore produced in cultures supplemented with 10 μmol l−1 added iron, measured at the same time point (Fig. 1; Mann–Whitney P ≤ 0·05). With supernatants from cultures grown under iron depletion, a peak in absorbance was clearly seen at 384 nm, which was absent from supernatants of cultures grown under iron-supplemented conditions (10 μmol l−1 added Fe).

Figure 1.

 Average siderophore percentage (±SD) of Pseudomonas M162 sterile-filtered supernatants, when grown in iron-sufficient (M162 + Fe) or iron-deficient (M162) media. Supernatants collected from cultures grown 1 (inline image), 3 (inline image), 5 (inline image) and 7 (inline image) days. *Statistically significantly different siderophore percentage from same day iron sufficient grown supernatant (Mann–Whitney,  0·05).

Antagonistic activity of Pseudomonas M162 against Flavobacterium psychrophilum

When Pseudomonas M162 was grown in BHI, the supernatant was seen to inhibit the growth of Fl. psychrophilum 413 (Mann–Whitney P = 0·04). Although the growth of Fl. psychrophilum ST2/00 also appeared to be inhibited, it was not statistically different to the growth obtained with the control (Mann–Whitney P = 0·436). No inhibition in growth of either of the Fl. psychrophilum strains was seen when M162 was cultured in M9C medium with or without iron supplementation (anovaP > 0·05).

Safety testing of Pseudomonas M162 in fish

No mortalities or abnormal signs were recorded in fish injected i.p. with Pseudomonas M162.

Antagonistic activity against Flavobacterium psychrophilum in vivo

When experimental fish were injected i.m. with Fl. psychrophilum JIP02/86, mortalities were first seen in both the M162-fed group and control group on day 5 postinjection. Cumulative mortalities at the end of the experiment were 57% in the controls and 35% in M162-fed group (Fig. 2), and these were statistically different between the groups (Mann–Whitney, P = 0·005). The RPS value of the experimental infection was 39·2%.

Figure 2.

 Cumulative percentage mortality of fish fed 2 weeks with Pseudomonas M162 supplemented feed (inline image) or control feed (inline image) then infected with an intramuscular injection of Flavobacterium psychrophilum JIP02/86. Statistically significantly different mortality between groups (Mann–Whitney, = 0·005).

Intestinal M162 colonization and immunology

No differences was observed in the weight gain between the probiotic-fed group (week 2: 13·12 g ± 3·29 and week 3: 14·87 g ± 3·36) and control group (week 2: 13·63 g ± 2·43 and week 3: 16·51 g ± 3·55; P > 0·05, t-test). After 2 weeks of feeding, the average numbers (±SD) of viable Pseudomonas M162 in the intestine of the probiotic-fed group were 1·2 ± 1·5 × 10CFU g−1 of intestine content, which was equivalent to 23·7% of the total colony count (4·9 ± 10 × 104 CFU g−1 of intestine content; Table 1). Following the replacement of Pseudomonas M162 diets with the control diet for 1 week, the level of Pseudomonas M162 in the intestine content comprised 0·8% of the total colony count on Pseudomonas agar (2·6 ± 6·7 × 106 CFU g−1 intestine content). While the numbers of Pseudomonas M162 were statistically higher in the probiotic-fed group compared with the control group (P = 0·039, t-test) after 2 weeks of feeding M162, there was no statistical difference in the amount of M162 present in the intestine content between the M162-fed fish and control fish at week 3 (P = 0·349, t-test; Table 1).

Table 1.   Number of fish, total colony count and Pseudomonas M162 colony count (±SD) on Pseudomonas-specific agar of intestine samples taken on weeks 2 and 3
TreatmentNo. of fish isolated with M162/Total fishTotal colonies (CFU g−1)M162 colonies (CFU g−1)Ratio (M162 colonies/total colonies)
  1. *Statistically significantly different from control.

Week 2
 Probiotic5/94·9 ± 10 × 1041·2 ± 1·5 × 104*23·7%
 Control1/92·6 ± 7·7 × 106222 ± 6670%
Week 3
 Probiotic1/92·6 ± 6·7 × 1062 ± 6 × 1040·8%
 Control1/91·3 ± 2·6 × 107111 ± 3330%

Innate immunity

After 2 weeks of feeding with Pseudomonas M162, statistical differences in leucocyte numbers were recorded compared with the control fish (P = 0·032, t-test), and the week 3 sampling showed a significant increase in the number of leucocytes in probiotic-fed group (7·6 ± 2·6 × 107 ml−1) compared with the control group (5·9 ± 0·9 × 107 ml−1; P = 0·05, t-test; Table 2). However, no significant difference was seen in the number of erythrocytes between M162-fed fish and control group at either week 2 (P = 0·173, t-test) or week 3 (P = 0·111, t-test; Table 2).

Table 2.   Haematological and innate immune analyses (average ± SD) of samples taken on weeks 2 and 3
TreatmentErythrocytes (1 × 108 ml−1)Leucocytes (1 × 107 ml−1)Respiratory activityPhagocytic activity (%)Lysozyme activity (U−1 min−1 ml−1)
  1. *Statistically significantly different from control ( 0·05, t-test).

Week 2
 Probiotic10·6 ± 2·4 (n = 12)4·7 ± 0·97* (n = 12)0·7 ± 0·49 (n = 5)22·4 ± 8·9 (n = 10)119·5 ± 61·2 (n = 12)
 Control9·2 ± 1·99 (n = 10)5·7 ± 0·97 (n = 10)0·4 ± 0·3 (n = 10)20·8 ± 10·9 (n = 11)172·2 ± 171·7 (n = 9)
Week 3
 Probiotic9·5 ± 1·3 (n = 12)7·6 ± 2·6* (n = 12)0·03 ± 0·001 (n = 12)12·6 ± 3·7 (n = 11)893·3 ± 622·7* (n = 12)
 Control10·3 ± 1·2 (n = 12)5·9 ± 0·9 (n = 12)0·03 ± 0·006 (n = 12)17·7 ± 9·4 (n = 11)395·8 ± 332·9 (n = 12)

After 2 weeks of feeding, no significant difference was detected in serum lysozyme activity between the two groups of fish (P = 0·4, t-test; Table 2). However, after 1 week of withdrawing the probiotic feed (week 3 sampling), the serum lysozyme activity of the probiotic-fed group was statistically significantly higher compared with the control group (P = 0·026, t-test).

There was no statistical difference in phagocytic activity of head kidney leucocytes between the probiotic group and the control group at either sampling point (Table 2). Also, no statistical differences were found in the respiratory burst activity of head kidney leucocytes between the probiotic supplemented group and control group at week 2 (P = 0·126, t-test) or week 3 (P = 0·802, t-test; Table 2).

Measurement of antibodies in serum

The total IgM content in the serum of the Pseudomonas M162-fed fish was higher at week 3 sampling when compared with control group (Fig. 3a), with absorbance values (at 450 nm) of 1·34 ± 0·195 and 1·16 ± 0·23 obtained for the probiotic-fed group and control group, respectively (P = 0·045, t-test; Fig. 3a). However, there was no difference in the level of specific antibodies measured against the M162 probiotic by ELISA between the groups at either week 2 (P = 0·093, t-test) or 3 (P = 0·251, t-test; Fig. 3b), or in the level of natural antibodies in the serum between the groups of fish on week 2 or 3 (P = 0·338 and P = 0·227, respectively, t-test; Fig. 3c).

Figure 3.

 Total antibody IgM (a), M162-specific antibody (b) and natural anti-BSA antibody (c) activity of fish fed with M162 suplemented feed (inline image) and control feed (inline image). (average ± SD). Samples taken on week 2 and week 3. *Statistically significantly different with respect to the control.

Bacterial killing activity of serum

The per cent survival of Fl. psychrophilum in the serum killing assay was not statistically significantly different to that of the control group at either week 2 (P = 0·441, t-test) or week 3 (P = 0·437, t-test) of the trial (Fig. 4).

Figure 4.

 Percentage survival of Flavobacterium psychrophilum JIP02/86 after incubation with serum from fish fed with the M162 supplemented feed (inline image) or control feed (inline image). Serum was taken on week 2 and week 3 of the experiment. Data are expressed as mean ± SD.

Discussion

The probiotic effect of the Pseudomonas M162, when tested both in vitro and in vivo against Fl. psychrophilum, is described here. We showed previously that another strain, Pseudomonas M174 was able to produce siderophores in vitro when cultured under iron-depleted conditions, and when grown under these conditions were found it to be antagonistic against Fl. psychrophilum (Korkea-aho et al. 2011). In the present study, Pseudomonas M162 was also found to produce siderophores, but these were not antagonistic against Fl. psychrophilum to the same extent as those of M174. Furthermore, the results of the immunological studies indicated that the main mode of action of M162 appears to be immunomodulation. This is relevant for RTFS infections by promoting increased protection in the young fry through their innate immunity before their adaptive immune response has fully developed (Mulero et al. 2007).

M162 was confirmed as a Pseudomonas sp., which is a bacterial species known to be part of the normal microflora of the gastrointestinal tract of healthy rainbow trout (Spanggaard et al. 2000). Furthermore, Pseudomonas strains have been shown to have antibacterial activity against Aeromonas salmonicida (Smith and Davey 1993; Das et al. 2006), a group of luminous pathogenic vibrios (Spanggaard et al. 2001; Vijayan et al. 2006), and Fl. psychrophilum (Korkea-aho et al. 2011; Ström-Bestor and Wiklund 2011). Pseudomonas M162 and M174 (Korkea-aho et al. 2011) were simultaneously isolated from the eggs of healthy rainbow trout. To confirm that Pseudomonas M174 and M162 are different organisms, the 16S rRNA partial gene sequences between the two organisms were compared and were found to have a 97% identity. Comparison of 16S rRNA gene sequences is not always considered to be the most discriminating method to differentiate between the same bacterial species (Mulet 2009), although the 3% difference found between the two isolates in this study does confirms that Pseudomonas M174 and M162 are indeed in different species, whereas 97% 16S rRNA gene sequence similarity confirms that they are part of the same operational taxonomic unit (Gevers et al. 2005).

In addition to genotypic differences, phenotypic differences were also observed between the two bacteria, such as differences in their siderophore production. Several strains of Pseudomonas sp. are known to produce siderophores, which differ in type and their ability to acquire iron (Miethke and Marahiel 2007). The results of the CAS assays suggest that both M162 and M174 produce siderophores, but with different absorbance spectra for their corresponding filtered culture supernatants (Korkea-aho et al. 2011). Furthermore, siderophores of M174 were shown to inhibit the growth of Fl. psychrophilum, while no inhibition was observed with supernatants from iron-depleted cultures of M162. It is known that competition in iron acquisition exists between the probiotic and the pathogen, and Fl. psychrophilum is known to produce siderophores (Møller et al. 2005). However, this iron acquisition can differ between serotypes of Fl. psychrophilum and is relatively weak compared with other bacterial pathogens affecting fish (Ström-Bestor and Wiklund 2011). Although siderophore production by a probiotic and its depletion of iron resources seems to be an effective method of inhibiting Fl. psychrophilum growth (Korkea-aho et al. 2011; Ström-Bestor and Wiklund 2011), it does not appear to be the mode of the action of M162 against Fl. psychrophilum.

After 2 weeks of feeding the probiotic, viable Pseudomonas M162 could be detected in the intestine of the fish, demonstrating the intestinal colonization capacity of M162. After removing M162 from the diet by re-introducing the unsupplemented feed, the level of Pseudomonas M162 that could be detected in the intestine of these fish dropped considerably. The results of the present study are in agreement with observations made by Balcázar et al. (2007) and Nikoskelainen et al. (2003) who found that lactic acid bacteria (LAB) adhered to and survived in the intestines of the rainbow trout, but the number of the probiotic bacteria decreased after the end of the probiotic supplementation. Together these findings indicate that permanent colonization by externally applied probiotic strains is rarely achieved and M162 intestinal colonization is more likely transitional.

Several studies have suggested that the attachment and colonization of probiotics in the intestine may lead to stimulation of the fish’s innate immune response (Panigrahi et al. 2005; Balcazar et al. 2007), as also shown in the present study for fish fed Pseudomonas M162. In one study feeding Bacillus subtilis to rainbow trout, a higher numbers of leucocytes were found at week 3 in probiotic-fed fish compared with the control group (Newaj-Fyzul et al. 2007). Significantly higher levels of leucocytes were also found in this study at week 3 sampling.

Lysozyme is an important innate defence mechanism, disrupting the peptidoglycan links in bacterial cell walls (Ellis 1999). Several studies have reported that probiotics such as LAB (Panigrahi et al. 2005; Balcázar et al. 2007) and Bacillus spp. (Newaj-Fyzul et al. 2007) can elevate the level of lysozyme activity in salmonids. Also in this study, fish fed with the probiotic M162 had significantly higher levels of serum lysozyme activity compared with the control fish on week 3. Probiotics have also been shown to activate phagocytosis and respiratory burst activity in rainbow trout when used as feed supplements (e.g. Balcázar et al. 2007), but these effects were not observed with M162.

A significant enhancement in the level of total immunoglobulin in the plasma of rainbow trout through the administration of the probiotic LAB has been reported (Nikoskelainen et al. 2003; Panigrahi et al. 2005). These results are in agreement with this study showing that oral administration of Pseudomonas M162 enhanced total IgM production in rainbow trout at the week 3. Elevated levels of immunoglobulin were not detected at week 2, so obviously time was necessary for the immunoglobulin levels to increase in the M162-fed fish’s serum. Indeed, Panigrahi et al. (2005) found elevated immunoglobulin levels by day 20 postfeeding, but not at days 10 or 30, in an experiment feeding rainbow trout with LAB for 30 days. Similarly, Nikoskelainen et al. (2003) found differences in the timing of elevated serum immunoglobulin levels in fish fed for 4 weeks with an LAB probiotic, but seemed to depend on the probiotic concentration used.

Reports on the production of specific antibodies against probiotics in fish are few. In this study no specific antibodies were elicited by the fish against M162 antigens at any of the time points examined. Specific antibody levels produced in response to vaccination or infection are involved with clearing the pathogen from the host. In this respect, presumably specific antibodies produced by the host would be detrimental to the effectiveness of the probiotic bacteria and would limit their use. However, Arijo et al. (2008) found that some probiotics shared common antigens with the pathogen Vibrio harveyi, and one probiotic, Aeromonas sobria A3-51, was able to stimulate production of specific antibodies against V. harveyi in rainbow trout and conferred protection against V. harveyi infection.

Natural antibodies seem to be an important first line of defence against pathogens in fish and have an important role in innate immunity (Sinyakov et al. 2002; Kachamakova et al. 2006). Sinyakov et al. (2002) found that goldfish with high levels of natural antibodies showed better protection against an experimental infection with Aer. salmonicida than goldfish with low levels of natural antibodies. Furthermore, the specific antibody response against Aer. salmonicida was lower in fish with high levels of natural antibodies when compared with fish with low levels of natural antibodies (Sinyakov et al. 2002). This result suggests that there could be a negative correlation between the specific antibody and natural antibody production in fish. In this study, no significant changes were seen in natural antibodies levels in the serum of the Pseudomonas M162-fed group compared with the control fish, assessed by their ability to bind to BSA in ELISA. BSA is most commonly used to measure natural antibodies in fish, although other antigens have also been used (Gonzalez et al. 1988; Kachamakova et al. 2006).

Antibacterial activity of serum can be mediated through an increase in the production of proteolytic enzymes, agglutinins, serum anti-proteases or peroxidase activity (Ellis 1999). In previous studies, significantly higher anti-protease activity against Vibrio anguillarum was reported in the serum of rainbow trout by 2 weeks of feeding with probiotic Kocuria SM1 (Sharifuzzaman and Austin 2010). However, no increased antibacterial activity was observed in the serum of fish fed M162 when the serum was incubated with Fl. psychrophilum.

In our previous study, a significantly higher respiratory burst activity of head kidney macrophages in Pseudomonas M174-fed trout after 2 weeks of feeding with the probiotic was observed (Korkea-aho et al. 2011). Contrary to this, Pseudomonas M162 group did not exhibit any significant change in macrophage respiratory burst activity compared with the control group, but several other immunological responses appear to have been stimulated as discussed earlier. Díaz-Rosales et al. (2009) reported that two probiotic bacteria originating within the same genus had different effects on the respiratory burst activity of Senegalese sole (Solea senegalensis), but both probiotics improved resistance against Photobacterium damselae ssp. piscicida. However, the innate immunology parameters induced by M162 and M174 were not totally comparable as immunological parameters were only analysed after 2 weeks of feeding with M174, but these studies do suggest that although M162 and M174 are from the same origin, and both improve resistance to RTFS, their mode of action seems to differ and further studies combining these probiotics could be useful.

When selecting a potential probiotic both its safety and its efficacy should be considered, and screening their effectiveness in vitro and their applicability in vivo is essential for this end (Irianto and Austin 2002). In this study, Pseudomonas M162 was demonstrated to be a potential probiotic, as it had no unfavourable effects of the health of the fish, it survived in gastrointestinal track of the fish and inhibited the growth of Fl. psychrophilum in vitro and improved resistance to it in vivo. One of the modes of action of this probiotic was its immunostimulatory effect. Immunomodulation of innate immunity is especially beneficial when dealing with infection affecting the early life stages of fish, such as RTFS. This study also showed that closely related micro-organisms can have different modes of action as probiotics.

Acknowledgements

This study was partly funded by the Otto A. Malm Foundation, University of Kuopio Foundation, Academy of Finland (No. 130325) and Marie Curie Intra-European Fellowship within the 7th European Community Framework Programme (PIEF-GA-2009-251821). We would like to thank N. Auchinachie and H. McEwan for technical assistance. All the animal experiments were conducted in accordance with the permission of the Animal Experiment Board, Finland (Permission No. ESLH-2008-00377/Ym-23) and the Animals (Scientific Procedures) Act 1986 Project License, Home Office, UK (Permission No. PPL 60/3850). Authors have no conflict of interests.

Ancillary