Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

Bacillus subtilis AB1 controls Aeromonas infection in rainbow trout (Oncorhynchus mykiss, Walbaum)
A. Newaj-Fyzul1,2, A. A. Adesiyun2, A. Mutani2, A. Ramsubhag3, J. Brunt1 and B. Austin1
1 School of Life Sciences, Heriot-Watt University, Edinburgh, Scotland, UK 2 School of Veterinary Medicine, The University of the West Indies, St Augustine, Trinidad, West Indies 3 Department of Life Sciences, The University of the West Indies, St Augustine, Trinidad, West Indies

Keywords Aeromonas, Bacillus subtilis AB1, sh disease, innate immunity, probiotic, specic immunity. Correspondence B. Austin, School of Life Sciences, John Muir Building, Heriot-Watt University, Riccarton, Edinburgh, EH14 4AS, Scotland, UK. E-mail: b.austin@hw.ac.uk

Abstract Aim: To develop a probiotic with effectiveness against Aeromonas sp., which was pathogenic to rainbow trout. Methods and Results: When Bacillus subtilis AB1, which was obtained from sh intestine, was administered for 14 days to rainbow trout in feed at a concentration of 107 cells per gram either as viable, formalized or sonicated cells or as cell-free supernatant, the sh survived challenge with the pathogen. AB1 stimulated immune parameters, specically stimulating respiratory burst, serum and gut lysozyme, peroxidase, phagocytic killing, total and a1-antiprotease and lymphocyte populations. Conclusions: Bacillus subtilis AB1 was effective as a probiotic at controlling infections by a sh-pathogenic Aeromonas sp. in rainbow trout. Signicance and Impact of the Study: Disease control in sh is possible by means of the oral application of live and inactivated cells and their subcellular components with the mode of action reecting stimulation of the innate immune response.

2006 1652: received 27 November 2006, revised 13 March 2007 and accepted 13 March 2007
doi:10.1111/j.1365-2672.2007.03402.x

Introduction Recently, there has been an increasing practice of managing bacterial sh diseases by using naturally antagonistic micro-organisms to control populations of potential pathogens, either by competitive inhibition, enhancement of sh immunity or by the microbial enhancement of the environment. Such organisms have been usually referred to as probiotics (Salminen et al. 1999), and are usually incorporated into the sh feed. Lactic acid bacteria (LAB) are among the most common probiotics used in aquaculture, and have been proposed to function as nonspecic immunostimulants and in environmental enhancement (Vadstein 1997; Ring and Gatesoupe 1998; Skjermo and Vadstein 1999; Robertson et al. 2000). However, a greater variety of micro-organisms has been considered for use as probiotics in aquaculture than in other areas of agriculture (Irianto and Austin 2002). In this study, Bacillus subtilis AB1 was isolated and evaluated as a putative probiotic in preventing disease in rainbow trout caused by a highly virulent strain of Aeromonas sp.

Materials and methods Fish Rainbow trout, Oncorhynchus mykiss (Walbaum) (average weight = 30 g) were obtained from a commercial sh farm in Scotland. The sh were maintained in continuously aerated free-owing dechlorinated fresh water at 17C and fed with a commercial pellet diet (Trouw, Wincham, UK). Representative samples from the sh stock were routinely examined microbiologically and physically to ensure the absence of bacterial diseases and parasites following methods described by Austin and Austin (1989). Bacterial pathogen Aeromonas sp. ABE1 originally isolated from diseased tilapia (Oreochromis sp.) was obtained from the culture collection of the School of Veterinary Medicine, The University of the West Indies, St. Augustine, Trinidad. Pathogenicity of Aeromonas sp. ABE1 against rainbow
1699

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

Control of Aeromonas infection in rainbow trout

A. Newaj-Fyzul et al.

trout was determined by challenging the sh intraperitoneally (i.p.) and intramuscularly (i.m.) with the pathogen at different concentrations (102 to 108 cells per millilitre) and observing for disease development and death over a 7-day period. Putative probiotics Bacterial cultures were obtained from the digestive tract of rainbow trout using the procedures of Brunt and Austin (2005). Briey, the digestive tracts of ve euthanized sh were removed in their entirety, and the intestinal contents including the mucus were emptied into petri dishes before 10 g quantities were added to 9 ml of 09% (w v) saline and vortexed vigorously for 1 min. Then, 10-fold dilutions were prepared in fresh saline to 10)5, and 01 ml volumes were spread over triplicate plates of tryptone soya agar (TSA; Oxoid) with incubation at 22C for up to 7 days. A total of 160 colonies was randomly picked and puried by streaking onto fresh medium, and assessed for inhibition against the sh pathogenic aeromonad using the cross-streak method, spot-on-lawn method and the overlay method as described by Robertson et al. (2000). One inhibitory isolate was evaluated at different concentrations for possible adverse effects in sh. Briey, separate groups of 25 sh were injected i.p. and i.m. with the organism at concentrations ranging from 104 to 109 cells per millilitre as determined by means of a haemocytometer slide (Improved Neubauer type; Merck, Whitehouse Station, NJ, USA) on a Kyowa light microscope at 400 magnication. The sh were observed for disease signs daily for up to 14 days (Brunt and Austin 2005). Additionally, the willingness of sh to accept food with the bacterial culture was tested by feeding separate groups of 25 rainbow trout for 14 days with commercial sh feed supplemented with putative probiotic at concentrations of 103 to 109 cells per gram. Control sh were also fed with commercial sh feed but without any added bacterial culture. The bacterial cultures were examined in the API 20E and API 50CH rapid identication systems (BioMerieux, Basingstoke, UK) and by 16S rRNA sequencing. The cultures were stored in tryptone soya broth (TSB; Oxoid) supplemented with 15% (v v) glycerol at )70C. Challenge experiments to determine the effectiveness of putative probiotics The methods described by Brunt and Austin (2005) were used to determine the potential usefulness of the cultures in preventing disease. Probiotic-supplemented sh feed containing 104, 105, 106, 107, 108 and 109 bacterial cells per gram were prepared spraying 5 ml volumes of appropriate saline suspensions of the organism onto 50 g
1700

batches of feed, with constant mixing. The bacterial counts in the feed and the survival of the putative probiotic on the feed over a 2-month period were determined by means of the total viable counts on TSA. This was achieved by homogenizing (VWR disposable homogenizer) 10 g of feed in 90 ml of saline, preparing 10-fold dilutions and spreading 01 ml amounts over duplicate TSA plates, which were incubated at 22C for 7 days. Additional experiments were conducted to determine the ability of formalin treated cells or cell extracts to confer protection against the pathogen. The putative probiotic as formalized (20 v v formalin for 48 h) cells, sonicated cell suspensions and cell-free extract were added to sh feed to a nal concentration equivalent to107 cells per gram following the methods described by Brunt and Austin (2005). Separate groups of 25 rainbow trout were fed with the modied diets, which were refrigerated until use for 14 days before challenging with the pathogenic aeromonad by i.p. injection with 23 106 cells per millilitre, which was equivalent to 2 LD50 as determined separately (data not shown). Appropriate control groups were included. Each challenge experiment was repeated three times. Determination of the mode of action of probiotics The modes of action on the putative viable probiotics were determined using rainbow trout fed with probiotic dose at 107 cells per gram of sh feed for 14 days. Then, groups of 25 sh were killed by administration of an overdose of anaesthetic (3-aminobenzoic acid ethyl ester; Sigma-Aldrich) before collection of blood, tissue or gut mucus (Brunt and Austin 2005). Each assay was replicated three times. To estimate the number of probiotic cells in the digestive tract of rainbow trout, the following experiment was carried out. For the collection of intestinal mucus, the method of Chabrillon et al. (2005) was followed with modications. Thus, the abdomen of each sh was cut open to expose the gastrointestinal tract. The intestine from the pylorus to the anus was removed, and its outer surface was carefully cleaned of its layers of fat. Pressure was applied to the sides of the intestine so that the mucus exuded out through the open ends. The mucus and gut contents were collected separately in pre-weighed sterile 15-ml Eppendorf tubes and homogenized in sodium phosphate buffer (SPB; 27 mmol l)1 Na2HPO4, 13 mmol l)1 NaH2PO4, 0004 mol l)1; pH 72), and dilutions prepared to 10)4 before 01 ml volumes were spread over duplicate TSA plates with incubation at 30C for up to 3 days. Identication and colony counts were done as described earlier.

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

A. Newaj-Fyzul et al.

Control of Aeromonas infection in rainbow trout

Sample and cell preparations for immunological assays After feeding, the sh were euthanized and exsanguinated by caudal venepuncture using 9 ml capacity Vacuettes containing a Z Serum Sep Clot Activator (Greiner, Stonehouse, UK). Blood was allowed to clot at 4C for 2 h, and the sera were separated by centrifugation (4000 rev min)1 for 25 min at 4C) and stored at )70C for subsequent assays. Gut mucus was collected as described earlier and centrifuged twice at 13 000 rev min)1 for 25 min at 4C to remove particulate and cellular material. The supernatant was removed and stored at )70C for lysozyme analysis. Isolation of head kidney macrophages for evaluation of phagocytic activity, respiratory burst and bacterial killing assay, determination of the number of erythrocytes and leucocytes, the lysozyme and antitrypsin activity, and the total protein quantity of the serum followed the methods of Sakai et al. (1995). Using aseptic techniques, the head kidneys were removed from rainbow trout and forced through a 100 lm nylon mesh with L-15 medium (Sigma-Aldrich) containing 2% (v v) foetal calf serum (FCS; Gibco, Paisley, UK), 100 ll ml )1 of penicillin streptomycin (p s; 10 000 IU ml )1 10 000 UG; Sigma-Aldrich) and 10 ll ml )1 heparin (Sigma-Aldrich) to isolate the leucocytes. Thus, kidney suspensions were layered carefully onto a 3451% (v v) Percoll gradient diluted in Hanks balanced salt solution (HBSS; Sigma-Aldrich). The samples were centrifuged at 400 g for 25 min at 4C before carefully removing the cells lying at the 3451% interface. The cells were adjusted to 106 cells per millilitre in L-15 medium supplemented with 01% (v v) FCS and 100 ll ml )1 p s. Additional assays were carried out as follows: Flow cytometry Analysis of sh blood using ow cytometry was performed following the methods of Takamasa et al. (2002). Thus, stock solutions of 3,3-Dihexyloxacarbocyanine [DiOC6(3); Sigma-Aldrich] were prepared in absolute ethanol to 500 lg ml )1and held in the dark. This stock solution was diluted 10-fold in HBSS immediately prior to use. Fresh rainbow trout blood (10 ll) was added to triplicate test tubes, each containing 1950 ll of HBSS and 40 ll of DiOC6(3) dye solution. This was mixed gently and incubated at room temperature for 10 min. Following staining with DiOC6(3), blood cells were analysed using a ow cytometer (Cyow SL). Forward scatter (FSC), side scatter (SSC) and green uorescence (FL-1) of each cell was measured. All data were analysed using the Flomax (Partec, Munster, Germany) software package.

Bactericidal assay Macrophage killing activity was assessed according to Secombes (1990) with modications. Aeromonas sp. was grown in TSB for 24 h and adjusted to 107 cells per millilitre in saline. Macrophages were adjusted to 106 cells per millilitre in L-15 medium before adding 01 ml volumes to 01 ml of bacterial suspension. Subsequently, 40 ll of pooled fresh rainbow trout serum was added, followed by incubation at 25C for 2 h with shaking every 15 min. Volumes (01 ml) were removed and diluted in 99 ml of sterile (121C 15 min)1) distilled water to release living bacteria from the phagocytes. This was serially diluted to 10)5, and 100 ll volumes were spread onto triplicate TSA plates with incubation overnight at 30C, and the number of colonies was counted (Selvaraj et al. 2005). Control assays were carried out in the absence of macrophages to give 100% survival at all bacterial dilutions. Total and a1-antiprotease activity of serum The antitrypsin activity of sera was measured following the methods described by Magnadottir et al. (2005). Briey, 20 ll of serum was incubated with 20 ll of standard trypsin solution (Sigma-Aldrich, 1000-2000 BAEE, 5 mg ml)1) at room temperature (22C) for 10 min in Eppendorf tubes. Two hundred microlitres of 01 mol l)1 PBS and 250 ll of 2% (w v) azocasein solution (20 mg ml)1 PBS) were added and further incubated for 1 h. Then, 500 ll of 10% trichloro acetic acid (TCA; Fisher) was added and incubated for another 30 min. The tubes were centrifuged at 9000 rev min)1 for 5 min before 100 ll of the supernatant from each tube was placed into the wells of a microtitre at bottom plate (Nalge Nunc, Hereford, UK) containing 100 ll of 1 N sodium hydroxide. The optical density (OD) was read at 450 nm on a Smartspec 3000 spectrophotometer (Bio-Rad). Inhibition of trypsin activity was calculated by comparing with a 100% control sample, which contained the buffer to replace serum, and a negative control where the buffer replaced both serum and trypsin. For a1-antiprotease, the assay was prepared following the method by Ellis (1999) where 10 ll of serum was incubated with 20 lg trypsin dissolved in 100 ll of TrisHCl (50 mmol l)1; pH 82) (Sigma-Aldrich). All tubes were made up to 200 ll with TrisHCl and incubated at room temperature (22C) for 1 h. Then, 2 ml of 01 mmol l)1 Na-benzoyl-DL-arginine-p-nitroanilide HCl (BAPNA; Sigma-Aldrich) was added and incubated for a further 15 min. The reaction was stopped by adding 500 ll of 30% acetic acid and the OD read at 450 nm. The serum blank contained 100 ll of Tris instead of trypsin, and the positive control contained trypsin but no serum.
1701

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

Control of Aeromonas infection in rainbow trout

A. Newaj-Fyzul et al.

a-2 macroglobulin in serum This assay is a modication of a method described by Ellis (1990), and uses an aniline-arginine dye ester as a substrate for trypsin, which hydrolyses the aniline dye resulting in a colour change that can be measured spectrophotometrically. Trypsin, at nal concentrations between 0 and 35 lg (bovine pancreas Type 1, 100 mg ml)1 in 001 mol l)1 TrisHCl; Sigma-Aldrich; pH 82), was incubated with 10 ll of serum at 22C for 5 min. Replicates with a range of trypsin concentrations were used. After the initial incubation, 05 ml of 2 mmol l)1 BAPNA in distilled water was added and the volume made up to 1 ml with 01 mol l)1 TrisHCl, pH 82, followed by incubation at 22C for 25 min. The reaction was stopped by the addition of 150 ll of 30% (v v) acetic acid. Each sample was centrifuged at 462 g and then ltered through a 022 lm lter (Millipore Millex, Edinburgh, UK) into 1 ml cuvettes and read at OD410 in a spectrophotometer against a blank of BAPNA in buffer and acetic acid. Controls consisted of the reaction combination without serum or trypsin. The trypsin hydrolysed BAPNA under standard reaction conditions at a rate that produced a change in OD410 of 0112 for each microgram of trypsin present. Peroxidase content The total peroxidase content present in serum was meas ured according to Daz-Rosales et al. (2006). For this, 15 ll of serum was diluted with 35 ll of Ca+2- and Mg+2-free HBSS (Sigma) in at-bottomed 96-well microtitre plates (Nalge Nunc). Then, 50 ll of 20 mmol l)1 3,3,5,5-tetramethylbenzidine hydrochloride (TMB; Sigma-Aldrich) and 5 mmol l)1 H2O2 (Sigma-Aldrich) were added (both substrates of peroxidase). The serum mixture (150 ll) was transferred from each well to new 96-well microtitre plates. The colour-change reaction was stopped after 2 min by adding 50 ll of 2 mol l)1 sulphuric acid and the OD was read at 540 nm in an ELISA reader (Dynatech, Guernsey, UK). Standard samples without serum were also analysed. Natural haemolytic complement activity The alternative complement pathway (ACH50) activity used sheep red blood cells (SRBC; Sigma-Aldrich) as targets. Equal volumes of SRBC suspension (17 107 cells per millilitre) in phenol red-free Hanks buffer (HBSS; Sigma-Aldrich) containing 01 mmol l)1 Mg+2 and EGTA (Sigma-Aldrich) were mixed with serially diluted serum to give nal serum concentrations ranging from 10% to 0078%. After incubation for 90 min at 22C, the samples
1702

were centrifuged at 400 g for 5 min at 4C. The relative haemoglobin content of the supernatants was assessed by measuring their OD550. The values of maximum (100%) and minimum (spontaneous) haemolysis were obtained by adding 100 ll of distilled water or HBSS to 100 ll samples of SRBC, respectively. The degree of haemolysis (Y) was estimated and a lysis curve for each specimen was obtained by plotting Y (1 ) Y) against the volume of serum added (ml) on a log ) log scaled graph. The volume of serum producing 50% haemolysis (ACH50) was determined and the number of ACH50 units per millilitre obtained for each experimental group. Statistics All sh experiments were repeated three times unless specied. The data were examined by a range of statistical methods including Student t-test for comparing immune responses between probiotic feed and control sh and anova for comparing probiotic treatments in the challenge experiments using Instat 201 statistical software package (GraphPad Software, San Diego, CA, USA). Results The isolate AB1, which was an endospore-forming, Gram-positive bacterium identied as B. subtilis by phenotypic traits and 16S rDNA sequencing (with a sequence homology of 99% when compared with B. subtilis spizizenii strain PDA), was inhibitory to the pathogenic Aeromonas sp. Furthermore, AB1 was harmless to rainbow trout following administration via injection or by feeding. In addition, 14 days after the completion of the feeding regime, the sh appeared healthy, and the organism could not be recovered internally or from around the injection sites. Feeding to sh for 14 days at 107 cells per gram of feed, whole, sonicated or formalized cells as well as cell-free supernatant, led to signicantly (P = 00001) reduced cumulative mortalities after challenge with Aeromonas sp. (Table 1). The survival rates after challenge ranged from 65% to 100% for the probiotic-fed as compared to 5% to 15% in the nonprobiotic control-fed sh. Doses of viable AB1 lower and higher than 107 cells per gram of feed were less successful at controlling the infection by Aeromonas (Table 1). Furthermore, AB1 was present in the intestine at > 104 cells per gram of gut contents and mucus during the feeding regime, but was absent 4 weeks after switching to regular feed (Table 2). Moreover, there was only a slow decline of viability of AB1 in sh feed over 7 weeks at 4C, but at 22C there was a steadier decline in the number of culturable cells from 21 107 g)1 initially to 61 104 g)1 at 56 days (Table 3).

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

A. Newaj-Fyzul et al.

Control of Aeromonas infection in rainbow trout

Table 1 Effect of feeding varying concentration of AB1 on the survival of rainbow trout after challenging with Aeromonas sp.
% survival No. of bacterial cells (CFU g ) 104 105 106 107 108 109 Sonicated cells (107) Formalized cells (107) Cell-free supernatant (107)
)1

Table 4 Immunological response of rainbow trout after feeding AB1 for 14 days
Immunological parameter Probiotic AB1 Control

AB1 65* 72* 86* 100* 68* 65* 100* 100* 100*

Control 15 10 5 15 15 10 5 10 10

means of three replicates of 25 sh used in treatment. *Signicant at < 5% level.

Erythrocytes (10 ml ) Leucocytes (104 ml)1) Phagocytic activity (%) Bactericidal activity (CFU ml)1) Respiratory burst (OD630) Gut mucus lysozyme (U ml)1) Serum lysozyme (U ml)1) Total antiprotease activity (% trypsin inhibition) a1-antiprotease a2-macroglobulin Peroxidase assay (OD 540) Complement activity (ACH50 U ml)1) Total proteins (mg ml)1)

)1

13 05 15 03 28 02* 064 025 69 9* 38 3 21 106 012* 38 104 021 012 002* 006 001 1033 181* 510 45 1269 134* 438 75 86 4* 64 3 93 142 125 68 2* 012 009* 8 84 139 077 64 3 011 006 6

42 3*

33 4

Table 2 Survival of probiotic in the intestines of rainbow trout after feeding for 14 days
Mean number of cells g)1(CFU) in gut Sample Gut contents Gut mucus Treatment Control AB1 Control AB1 Probiotic cells None detected 535 12 104 None detected 81 28 104 Total viable count 46 24 13 36 08 11 06 14 107 104 106 104

*Signicant at < 5% level. Data represent the average standard deviation from a triplicate set of 25 sh.

n = 10.

Table 3 Survival of probiotic (AB1) in feed maintained at different temperatures

Number of cells on feed (CFU g)1) Day 0 1 3 7 14 21 28 35 42 49 56 *Initial count. 4C 21 20 20 69 53 52 47 38 30 24 23 108* 108 108 108 107 107 107 107 107 107 107 22C 21 20 57 74 68 34 12 83 14 52 61 108* 108 107 106 106 106 106 105 105 104 104

Mode of action of the probiotic Generally, there was stimulation of the immune system after administering AB1 to rainbow trout. Specically, the number of leucocytes increased from 064 025 104 ml)1 in the control group to 28 02 104 ml)1 in

AB1-fed sh (P = 00002). The erythrocyte counts for the probiotic treated and control sh were 13 05 109 and 15 03 109, respectively, but the difference was not statistically signicant (Table 4). The phagocytic index of head kidney macrophages of AB1-fed sh (69 9%) was signicantly higher than that of controls (38 3%) (P = 0018). In addition, the bactericidal activity of macrophages from AB1-fed sh (21 106 012 cells per millilitre) was signicantly higher than that of controls (38 104 021 cells per millilitre) (P = 00002) after incubating the pathogen at a dose of 44 107 cells per millilitre with macrophages obtained from head kidney (Table 4). Moreover, there were statistically signicant differences in the respiratory burst activity of blood macrophages from sh which received probiotics (012 002 units of activity) as compared with the controls (006 0005) (P = 00013). The serum lysozyme activity was recorded as 1269 134 and 438 75 U ml)1 after 60 min for AB1 treated and control sh, respectively. The results for gut mucus lysozyme revealed a signicantly higher activity for sh fed with probiotic as compared with controls (1033 181 and 510 45 at 60 min, respectively) (P = 0001). Total antiprotease activity as measured by the mean antitrypsin activity of sera was 86% (4) and 64% (3) for AB1 and controls, respectively. These differences were statistically signicant (P = 00001). In particular, sh fed with AB1 showed a higher activity for a1-antiprotease (93 2) as compared with the controls (84 3), with the differences being signicant (Table 4). Although a2-macroglobulin activity was higher in AB1-fed sh (142 012) than the
1703

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

Control of Aeromonas infection in rainbow trout

A. Newaj-Fyzul et al.

Control blood 4095 4095

AB1 fed blood

SCC

R2 R3 SCC

R1

100 FL1

101

102 -

103

104

100 FL1

101

102 -

103

104

Figure 1 Analysis of sh blood using ow cytometry. Flow cytometry analysis of DiOC6(3) stained blood cells obtained from control sh and AB1 fed sh. The graph illustrates the percentage cell populations within whole blood samples. R1 are erythrocytes and R2 are leucocytes. R3 is the total blood cells.

controls (139 011), these values were not signicantly different (P > 005). Again, the serum peroxidase content from sh fed with probiotic was higher (125 009) than the controls (077 006) (P < 001). Flow cytometry of the whole blood revealed a signicantly higher count of leucocytes in AB1 treated (28 104 020) as compared with control sh (064 104 025) (P = 0001) (Fig. 1). There was no signicant difference (P > 005) in natural haemolytic complement activity levels of serum between AB1 fed (68 8) and control sh (64 6). In contrast, total protein was signicantly higher (P < 001) in AB1fed groups (42 3 g dl)1) as compared with the controls (33 4 g dl)1) (Table 4). Discussion Bacillus subtilis AB1 was able to effectively protect rainbow trout against virulent Aeromonas sp., and thus can be classied as a probiotic agent. The fact that B. subtilis AB1 was isolated from the gut of apparently healthy rainbow trout conrms the potential role of gut micro-organisms in exerting an important role in the wellbeing of the host sh (Cunningham-Rundles and Lin 1998). There is increasing evidence that Bacillus spp. is benecial in protecting against bacterial pathogens. Moreover, B. subtilis has demonstrated antibiosis against pathogenic Vibrio spp., and has also been used to improve pond water quality, leading to increased survival of black tiger prawns (Vaseeharan et al. 2004). The results of the present study indicate that B. subtilis AB1 stimulated both cellular and humoral immune responses, which may have provided the rainbow trout with adequate protection to survive the challenge by the highly virulent Aeromonas sp. The role of probiotics in
1704

inuencing immune responses in sh has been previously reported as having important regulatory effects on the innate and adaptive immune responses of the host (Austin et al. 1995). The immune responses of rainbow trout to B. subtilis AB1 included a signicant increase in the number of leucocytes (Fig. 1) as well as enhanced respiratory burst and phagocytic activity. Similar results have been reported for LAB in turbot (Villamil et al. 2002), Lactobacillus rhamnosus in rainbow trout (Nikoskelainen et al. 2003), Bacillus toyoi in European eel (Chang and Liu 2002), Carnobacterium sp. in rainbow trout (Kim and Austin 2006) and Vibrio sp. in gilthead sea bream (Sparus aurata L.) (Daz-Rosales et al. 2006). The increased respiratory burst activity, which is a measure of the superoxide anion (O2)) and its derivatives, may have contributed to the ability of the probiotic to protect against the pathogen insofar as reactive oxidative species (ROS) compounds are known to contribute extracellular killing of pathogens (Hardie et al. 1996; Itou et al. 1997). The corresponding increase in peroxidase activity was not surprising because these enzymes will be required to remove reactive-free radicals that may be harmful to the sh. However, Daz-Rosales et al. (2006) did not nd any signicant increase in respiratory burst activity or peroxidase activity in gilthead sea bream fed with heat-inactivated probiotic. The role of peroxide activity in protecting sh following exposure to probiotics is, therefore, unclear. However, these cellular responses could provide a mechanism to account for the probiotic properties of select bacteria. Lysozyme is an important humoral innate defence parameter, and is widely distributed in invertebrates and vertebrates (Magnadottir et al. 2005). Lysozyme has an

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

A. Newaj-Fyzul et al.

Control of Aeromonas infection in rainbow trout

antibacterial activity by attacking peptidoglycan in the cell wall of bacteria, predominantly Gram-positive bacteria, thereby causing lysis and stimulation of phagocytosis of bacteria by phagocytic cells (Ellis 1990). An increase in the lysozyme concentration in sh blood can be caused by infections or invasion by foreign material (Siwicki et al. 1998). Certainly, B. subtilis AB1 inuenced the production of higher levels of lysozyme activity in sh serum and gut mucus. This increase may have also contributed to the survival of sh challenged with the pathogen. Panigrahi et al. (2005) and Kim and Austin (2006) have similarly demonstrated increases in gut and serum lysozyme in sh fed with L. rhamnosus and Carnobacterium maltaromaticum B26 and Carnobacterium divergens B33, respectively. It can be suggested, however, that sh fed with AB1 resulted in increased serum and gut lysozyme, which enhanced the immune efciency of sh to withstand challenge with the pathogen. Bacterial pathogens produce proteolytic enzymes to aid in the breakdown of host tissues, but protease inhibitors may be present in sera and other body uids (Bowden et al. 1997). These inhibitors also serve in the homeostasis of body uids, and are involved in acute phase reactions and in defence against pathogens that secrete proteolytic enzymes (Magnadottir 2006). Fish plasma contains a number of protease inhibitors, principally a1-antiprotease, a2-antiplasmin and a2-macroglobulin, which have been demonstrated to have a role in restricting the ability of bacteria to survive in vivo (Ellis 2001). Although several studies have investigated antiprotease levels in sh species, particularly a2-macroglobulin activity (Bowden et al. 1997), there is negligible information concerning modulation in sh other than by infection. It was also signicant that AB1 induced higher levels of total, a1-antiprotease and a2-antiprotease inhibitors in the probiotic-fed sh as compared with the controls. Similarly, Vasudeva Rao and Chakrabarti (2005) reported signicantly higher total and a1-protease levels in carp (Catla catla) administered with feeds supplemented with herbs. These authors noted that the best function of the a2-macroglobulin antiproteases family concerns with the clearance of active proteases from tissue uids. The results of this study indicate that feeding rainbow trout with AB1 enhanced nonspecic factors of the immune system by enhancing the level of natural antiproteases in the serum. Possibly, these may have provided some defence against infection by the pathogen. Previously, Mihal et al. (1990), Nikoskelainen et al. (2003) and Panigrahi et al. (2007) demonstrated that probiotics induced signicant and positive effects on complement levels. In the present study, differences in complement levels were not statistically signicant when compared with the controls.

Overall, the present study reinforces the view that bacterial cultures may well contribute to disease control strategies in aquaculture. Certainly, AB1 was effective in preventing disease caused by highly virulent Aeromonas sp. in rainbow trout. Acknowledgements This study was supported by the Commonwealth Split Site studentship, the Postgraduate Research Fund of The University of the West Indies and Scalar Scientic and Technical Supplies of Trinidad and Tobago. References
Austin, B. and Austin, D.A. (1989) Microbiological Examination of Fish and Shellsh. Chichester: Ellis Horwood. Austin, B., Stuckey, L.F., Robertson, P.A.W., Effendi, J. and Grifth, D.R.W. (1995) A probiotic strain of Vibrio alginolyticus effective in reducing diseases caused by Aeromonas salmonicida, Vibrio anguillarum and Vibrio ordalii. J Fish Dis 18, 9396. Bowden, T.J., Butler, R., Bricknell, I.R. and Ellis, A.E. (1997) Serum trypsin-inhibitory activity in ve species of farmed sh. Fish Shellsh Immunol 7, 377385. Brunt, J.B. and Austin, B. (2005) Use of a probiotic to control lactococcosis and streptococcosis in rainbow trout, Oncorhynchus mykiss (Walbaum). J Fish Dis 28, 693701. Chabrillon, M., Rico, R.M., Arijo, S., Daz-Rosales, P., Bale bona, M.C. and Morinig, M.A. (2005) Interactions of microorganisms isolated from gilthead sea bream, Sparus aurata L., on Vibrio harveyi, a pathogen of farmed Senegalese sole, Solea senegalensis (Kaup). J Fish Dis 28, 531. Chang, C.I. and Liu, W.Y. (2002) An evaluation of two probiotic bacterial strains, Enterococcus faecium SF68 and Bacillus toyoi, for reducing edwardsiellosis in cultured European eel, Anguilla anguilla L. J Fish Dis 25, 311315. Cunningham-Rundles, S. and Lin, D.H. (1998) Nutrition and immune system of the gut. Nutrition 14, 573579. Daz-Rosales, P., Salinas, I., Rodrguez, A., Cuesta, A., Chabrillon, M., Balebona, M.C., Morinigo, M.A., Esteban, M.A. et al. (2006) Gilthead seabream (Sparus aurata L.) innate immune response after dietary administration of heat-inactivated potential probiotics. Fish Shellsh Immunol 20, 482492. Ellis, A.E. (1990) Serum antiproteases in sh. In Techniques in Fish Immunology ed. Stolen, J.S., Fletcher, T.C., Anderson, D.P. and Roberson, B.S. pp. 9599. Fair Haven, NJ: SOS Publications. Ellis, A.E. (1999) Immunity to bacteria in sh. Fish Shellsh Immunol 9, 291308. Ellis, A.E. (2001) Innate host defense mechanisms of sh against viruses and bacteria. Dev Comp Immunol 25, 827 839.

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706

1705

Control of Aeromonas infection in rainbow trout

A. Newaj-Fyzul et al.

Hardie, L.J., Ellis, A.E. and Secombes, C.J. (1996) In vitro activation of rainbow trout macrophages stimulates inhibition of Renibacterium salmoninarum growth concomitant with augmented generation of respiratory products. Dis Aquat Organ 25, 175183. Irianto, A. and Austin, B. (2002) Probiotics in aquaculture. J Fish Dis 25, 110. Itou, T., Iida, T. and Kawatsu, H. (1997) The importance of hydrogen peroxide in phagocytic bactericidal activity of Japanese eel neutrophils. Fish Pathol 32, 121125. Kim, D.H. and Austin, B. (2006) Innate immune responses in rainbow trout (Oncorhynchus mykiss, Walbaum) induced by probiotics. Fish Shellsh Immunol 21, 513524. Magnadottir, B. (2006) Innate immunity of sh (overview). Fish Shellsh Immunol 20, 137151. Magnadottir, B., Lange, S., Gudmundsdottir, S., Bgwald, J. and Dalmo, R.A. (2005) Ontogeny of humoral immune parameters in sh. Fish Shellsh Immunol 19, 429439. Mihal, V., Lackovic, V., Plockova, M. and Brezina, P. (1990) Immunobiologic properties of lactobacilli. Cesk Pediatr 45, 587590. Nikoskelainen, S., Ouwehand, A.C., Bylund, G., Salminen, S. and Lilius, E.M. (2003) Immune enhancement in rainbow trout (Oncorhynchus mykiss) by potential probiotic bacteria (Lactobacillus rhamnosus). Fish Shellsh Immunol 15, 443 452. Panigrahi, A., Kiron, V., Puangkaew, J., Kobayashi, T., Satoh, S. and Sugita, H. (2005) The viability of probiotic bacteria as a factor inuencing the immune response in rainbow trout Oncorhynchus mykiss. Aquaculture 243, 241254. Panigrahi, A., Kiron, V., Satoh, S., Hirono, I., Kobayashi, T., Sugita, H., Puangkaew, J. and Aoki, T. (2007) Immune modulation and expression of cytokine genes in rainbow trout Oncorhynchus mykiss upon probiotic feeding. Dev Comp Immunol 31, 372382. Ring, E. and Gatesoupe, F.J. (1998) Lactic acid bacteria in sh: a review. Aquaculture 160, 177203. Robertson, P.A.W., ODowd, C., Burrells, C., Williams, P. and Austin, B. (2000) Use of Carnobacterium sp. as a probiotic for Atlantic salmon (Salmo salar L.) and rainbow trout (Oncorhynchus mykiss, Walbaum). Aquaculture 185, 235 243.

Sakai, M., Kobayashi, M. and Yoshida, T. (1995) Activation of rainbow trout, Oncorhynchus mykiss, phagocytic cells by administration of bovine lactoferrin. Comp Biochem Physiol 110B, 755759. Salminen, S., Ouwehand, A., Benno, Y. and Lee, Y.K. (1999) Probiotics: how should they be dened? Trends Food Sci Technol 10, 107110. Secombes, C.J. (1990) Isolation of salmonid macrophages and analysis of their killing activity. In Techniques in Fish Immunology, vol. 1 ed. Stolen, J.S., Fletcher, T.C., Anderson, D.P. and Roberson, B.S. pp. 137163. Fair Haven, NJ: SOS Publications. Selvaraj, V., Sampath, K. and Sekar, V. (2005) Administration of yeast glucan enhances survival and some non-specic and specic immune parameters in carp (Cyprinus carpio) infected with Aeromonas hydrophila. Fish Shellsh Immunol 19, 293306. Siwicki, A.K., Studnicka, M., Morand, M., Pozet, F. and Terech-Majewska, E. (1998) Comparative immunotoxicology a new direction. Acta Vet (Brno) 67, 295301. Skjermo, J. and Vadstein, O. (1999) Techniques for microbial control in the intensive rearing of marine larvae. Aquaculture 177, 333343. Takamasa, I., Tadaaki, M., Yumi, T., Sachiko, M., Hiroshi, F. and Teruyuki, N. (2002) A new method for sh leucocyte counting and partial differentiation by ow cytometry. Fish Shellsh Immunol 13, 379390. Vadstein, O. (1997) The use of immunostimulation in marine larviculture: possibilities and challenges. Aquaculture 155, 401417. Vaseeharan, B., Lin, J. and Ramasamy, P. (2004) Effect of probiotics, antibiotic sensitivity, pathogenicity, and plasmid proles of Listonella anguillarum-like bacteria isolated from Penaeus monodon culture systems. Aquaculture 241, 7791. Vasudeva Rao, Y. and Chakrabarti, R. (2005) Stimulation of immunity in Indian major carp Catla catla with herbal feed ingredients. Fish Shellsh Immunol 18, 327334. Villamil, L., Tafalla, C., Figueras, A. and Novoa, B. (2002) Evaluation of immunomodulatory effects of lactic acid bacteria in turbot (Scophthalmus maximus). Clin Diagn Lab Immunol 9, 13181323.

1706

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 103 (2007) 16991706