Experimental infection of turbot, Scophthalmus maximus (L.), by Moritella viscosa, vaccination effort and vaccine-induced side-effects

Experimental infection of turbot, Scophthalmus maximus (L.), by Moritella viscosa, vaccination effort and vaccine-induced side-effects B Bjçrnsdóttir, S Gudmundsdóttir, S H Bambir, B Magnadóttir and B K Gudmundsdóttir Institute for Experimental Pathology, University of Iceland, Keldur, Reykjavík, Iceland Abstract Moritella viscosa is the causative agent of winter ulcers in farmed salmonids and Atlantic cod in countries around the North Atlantic. The bacterium has also been isolated from various marine fish species. Bacterial diseases have been a limiting factor in farming of turbot, but M. viscosa has not so far been isolated. In this study, turbot was shown to be sensitive to M. viscosa infection in experimental challenges. Pathological changes in infected turbot were comparable with those previously described for winter ulcers in salmon. A multivalent commercial salmon vaccine, containing M. viscosa as one of five antigens and a mineral oil adjuvant, did not protect turbot against challenge 13 weeks postvaccination. Weight gain of vaccinated turbot compared with controls was not reduced 7 weeks post-vaccination. Vaccination did not induce a specific anti-m. viscosa response, while elevated anti-m. viscosa antibody levels were detected both in vaccinated and unvaccinated fish 5 weeks postchallenge. The vaccine did, however, induce an antibody response against Aeromonas salmonicida, another vaccine component. Minor intra-abdominal adhesions were detected in vaccinated fish and fish injected with a mineral oil adjuvant. The measurement of various innate humoral immune parameters did not reveal significant differences between vaccinated and control groups. Correspondence B K Gudmundsdóttir, Institute for Experimental Pathology, University of Iceland, Keldur v/vesturlandsveg, IS-112 Reykjavík, Iceland (e-mail: bjarngud@hi.is) Keywords: experimental infection, Moritella viscosa, Scophthalmus maximus, turbot, vaccination, winter ulcers. Introduction Winter ulcers, induced by the fish pathogenic bacterium Moritella viscosa (previously Vibrio viscosus) (Benediktsdóttir, Verdonck, Spröer, Helgason & Swings 2000), have caused financial losses in farmed salmonids and cod, Gadus morhua (L.), in several countries around the North Atlantic (Thorarinsson & Lystad 2003; Colquhoun, Hovland, Helleberg, Haug & Nilsen 2004). There are also reports of its isolation from farmed and wild rainbow trout, Oncorhynchus mykiss (Walbaum), and from captive wild plaice, Pleuronectes platessa (L.), lumpsucker, Cyclopterus lumpus (L.), and haddock, Melanogrammus aeglefinus (L.) (Benediktsdóttir et al. 2000; Lunder, Sørum, Holstad, Steigerwalt, Mowinckel & Brenner 2000). The disease most commonly affects salmon reared in sea water, when the temperature is below 10 C (Larsen & Pedersen 1999). Winter ulcers in salmon have been characterized by skin ulcers and degenerative changes in underlying muscle, gill pallor and fin rot. Ecchymotic haemorrhages in liver, pyloric caeca and perivisceral fat, and congestion and necrosis in spleen and kidney are also signs of the disease (Lunder, Evensen, Holstad & Håstein 1995; Benediktsdóttir, Helgason & Sigurjónsdóttir 1998; Bruno, Griffiths, Petrie & Hastings 1998). Moritella viscosa infections have caused considerable damage in salmon, Salmo salar (L.), farming in Iceland. Turbot, Scophthalmus maximus (L.), is an important species in fish farming in Europe and attempts are being made to start turbot farming in 645

several countries, including Iceland. In Iceland, turbot is farmed in land-based tanks with 14 17 C sea water, using heat exchange with geothermal water to elevate the seawater temperature. In intensive farming the amount of geothermal water can be a limiting factor resulting in periodic drops in temperature. Bacterial diseases, mainly caused by Listonella anguillarum (previously Vibrio anguillarum) (Toranzo, Barja & Devesa 1985; Toranzo, Santos, Lemos, Ledo & Bolinches 1987), Vibrio spp. (Austin, Stobie, Robertson, Glass, Stark & Mudarris 1993; Gatesoupe, Lambert & Nicolas 1999; Villamil, Figueras, Toranzo, Planas & Novoa 2003), Aeromonas salmonicida (Pedersen, Kofod, Dalsgaard & Larsen 1994; Lillehaug, Lunestad & Grave 2003), Enterococcus spp. (Toranzo, Cutrin, Nunez, Romalde & Barja 1995) and Strepococcus parauberis (Domenech, Fernandez-Garayzabal, Pascual, Garcia, Cutuli, Moreno, Collins & Dominguez 1996) have been a limiting factor in turbot farming in Europe. A bath challenge experiment of turbot, halibut, Hippoglossus hippoglossus (L.), and cod with M. viscosa has been reported (Gudmundsdóttir, Magnadóttir & Gudmundsdóttir 2001). The results showed that all three fish species were infected by the bacterium and that turbot was found to be as susceptible as cod, but halibut was more resistant. The aims of this work were to study the pathology induced in turbot following an experimental infection by M. viscosa, the effects of a multivalent commercial salmon vaccine in inducing protective immunity in turbot, and to monitor the effects of vaccination on fish growth. Materials and methods Fish Turbot fingerlings with mean weight of about 50 g were used in the experiment. The fish were the offspring of wild turbot caught around Iceland and came from an experimental fish farm (Experimental Laboratory in Mariculture, Marine Research Institute, Stadur, Grindavík) where M. viscosa infection had never been reported. The fish were reared in 8 m 3 tanks at 14 C and fed commercial dry pellets before vaccination and until 11 weeks post-vaccination. For challenge, the fish were transported to another facility (Sandgerdi Marine Centre, which has high quality sea water, but with a constant temperature of 9 C) and acclimatized in 400 L tanks at 9 C. Aerated borehole sea water was used in both facilities. Bacteria Moritella viscosa, strain F288/95, isolated from diseased farmed Atlantic salmon was used in the study. Bacterial preparations were made as follows: a sample from frozen stock was cultured on blood agar with 2% NaCl (BA-S) for 3 days at 15 C and then suspended in phosphate-buffered saline (PBS) (Sigma, Gróco ehf., Reykjavík, Iceland) with 1.5% NaCl and 0.1% peptone (Difco, Gróco ehf.) (PBS-PS). The suspension was spread on BA-S plates and cultured at 15 C for 19 h. The bacterium was then washed off the agar with PBS-PS, diluted in PBS-PS, and kept on ice until injection. All solutions were used ice-cold. Cultures were checked for purity and colonyforming units (CFU) estimated by plate counting. Prechallenge Prechallenge was performed by intraperitoneal (i.p.) and intramuscular (i.m.) injection of 0.1 ml M. viscosa bacterial solution. Four different doses of the bacterium, 10 5,10 6,10 7 and 10 8 CFU per fish were used, and each dose injected into five fish. As a control, five fish were i.p. and five i.m. injected with PBS-PS. Plastic tags were used to identify fish injected with different bacterial doses. Challenge was also performed by bathing two groups of 20 fish each in starting solutions of 10 6 and 10 7 CFU ml )1. Bathing was performed in aerated bacterial solutions, first for 1 h plus another hour in a 10-fold dilution before a continuous flow of sea water was supplied. The fish were observed for 4 weeks, deaths recorded daily and infection established by re-isolation of the bacterium from head kidney. The 50% lethal dose (LD 50 ) was calculated according to the method of Reed & Muench (1938). Experimental vaccination and challenge A commercial polyvalent vaccine, Alphaject 5200 (Alpharma, Pharmaco, Gardabær, Iceland), against M. viscosa, Aeromonas salmonicida subsp. salmonicida, Listonella anguillarum (serotypes O1 and O2) and Vibrio salmonicida was used. The vaccine contained a mineral oil adjuvant and was developed for use in salmon. Two control groups were used in the study. One group was injected with PBS-PS and the other with PBS-PS emulsified 1:1 in Freund s incomplete adjuvant (FIA) (Sigma). A total of 306 turbot fingerlings were included in the vaccination study. The vaccine, as well as 646

control injections of PBS-PS and FIA were administered i.p. in doses of 0.2 ml per fish. Prior to vaccination the fish were lightly anaesthetized with 0.3 ml L )1 phenoxyethanol (Lífsgledi ehf., Gardabær, Iceland) and marked along fin margins with Visible Implant Fluorescent Elastomer Dye (Northwest Marine Technology, Salisbury, UK). The fish were moved to another facility, 11 weeks postvaccination and acclimatized at 9 C. Challenge was performed 13 weeks post-vaccination by i.m. injection of 0.1 ml M. viscosa. Four different injection challenge doses of 4 10 5, 4 10 6, 4 10 7 and 4 10 8 CFU per fish were tested in all three injection groups each containing 12 22 fish. The fish were observed for 5 weeks after challenge and deaths recorded daily. Infection was established by detection of clinical signs and re-isolation of the bacterium from head kidney. Pathology Gross pathology of infected fish was recorded and diseased and moribund turbot were sampled for microscopic examination. Samples of skin and underlying muscle, kidney, liver, spleen, gut, heart and gills were fixed in 10% buffered formalin embedded in paraffin wax, cut into 4 lm sections and stained with Giemsa. Sampling Samples were taken during the vaccination experiment from 20 turbot at vaccination and from 10 turbot from each injection group, 7 and 11 weeks post-vaccination. These fish were reared at 14 C. At the end of the experiment, 5 weeks after challenge and 18 weeks post-vaccination, samples were taken from surviving turbot reared at 9 C (16 from PBS-PS group, 18 from FIA group and 16 from vaccinated group). At sampling, the fish were weighed and evaluated for intra-abdominal adhesions using the Speilberg scale, ranging from 0 to 6 (Midtlyng, Reitan & Speilberg 1996), and blood was drawn from the caudal vessel. The blood was allowed to clot at 4 C overnight, centrifuged and serum isolated and stored in aliquots at )80 C. Serum protein concentration A Coomassie Plus protein assay reagent kit from Pierce (Gróco ehf.), based on the Bradford method (Bradford 1976) was used to measure protein concentration of serum. The sera were diluted 1:100 in PBS and measured in duplicate and bovine serum albumin (Sigma) was used as a standard. Optical density (OD) was read at 590 nm after a 30-min incubation at 22 C. Results were expressed as mg protein ml )1 serum. Spontaneous haemolytic activity (SH 50% ) of serum The SH 50 activity of turbot sera (complement activity by the alternative pathway) was performed as described by Magnadóttir, Jónsdóttir, Helgason, Björnsson, Solem & Pilström (2001), with the modification that the reaction buffer was a gelatine veronal buffer (Sigma). Briefly, serially diluted sera in gelatine veronal buffer were incubated in duplicate for 1 h at 22 C with 0.5% washed red blood cells from sheep. Buffer in place of serum was used as a negative control (0% lysis) and H 2 Oin place of serum as a positive control (100% lysis). The OD of the supernatant was read at 405 nm following centrifugation and a graph of OD values against serum dilutions was plotted. The dilution of serum that gave 50% lysis (SH 50% ) was calculated. Anti-trypsin activity of serum Anti-trypsin activity of turbot serum was measured as described by Ellis (1990), with modifications as described in detail by Magnadóttir, Jónsdóttir, Helgason, Björnsson, Jørgensen & Pilström (1999a). Briefly, 20 ll of turbot serum was incubated with an equal volume of trypsin solution (5 mg ml )1, 1000 2000 BAEE; Sigma T-7409) for 10 min. To detect residual trypsin activity, further incubation was performed with azocasein (Sigma) in 0.1 m phosphate buffer, ph 7, for 1 h and then with trichloroacetic acid for another 30 min. Following centrifugation, the OD of the supernatant diluted with an equal volume of 1 n NaOH was measured at 450 nm. The results were expressed as the percentage of trypsin inhibition of serum compared with a serum free blank. All samples were analysed in duplicate and all steps performed at 22 C. ELISA Antibody activity of turbot serum was measured using an antibody capture ELISA, performed as described by Magnadóttir et al. (2001) with minor modifications. Antibodies against sonicated, whole 647

bacterium (WB) of M. viscosa, isolate F288/95, and A. salmonicida, isolate F19/99, as a control, were measured. The antigens were produced by cultivation of the bacteria at 15 C for 3 days on agar plates covered with cellophane (Bio-Rad, Gróco ehf.). Tryptone soya broth agar (TSB; Difco) with 1.2% bacto-agar (Difco) and 1.5% NaCl was used for M. viscosa cultivation and brain heart infusion agar (Oxoid, Gróco ehf.) for A. salmonicida cultivation. The bacteria were washed off the cellophane with PBS-PS, isolated by centrifugation and sonicated. ELISA trays (Nunc, Gróco ehf.) were coated overnight at 4 C with 100 ll of either antigen (10 lg protein ml )1 ) diluted in ELISA coating buffer (0.05 m carbonate bicarbonate buffer, ph 9.6; Sigma). Turbot sera were diluted 1:100. All samples were tested in duplicate. Bound antibody was detected with a polyvalent antibody against turbot IgM prepared in mouse ascites fluid (prepared at our laboratory), followed by a goat antimouse immunoglobulin labelled with alkaline phosphatase (Dako, Gróco ehf.). P-nitrophenyl phosphate (Sigma) was used for colour development, which was stopped using 3 m NaOH. OD was measured at 405 nm. The same ELISA assay was used for endpoint titration determination (Frey, Di Canzio & Zurakowski 1998). For titration, five sera which were high in the ELISA test were selected from vaccinated fish 11 weeks post-vaccination and from all groups at the end of the experiment. Five sera from PBS-PS-injected turbot, sampled 11 weeks post-vaccination, were used as negative controls. Calculation and statistical analysis Relative percentage survival (RPS) and mean day to death (MDD) were calculated using the following formulae: RPS ¼½1 ðper cent mortality in vaccinated fish= per cent mortality in control fishþš 100 MDD ¼½Rðnumber of mortalities number of days post-challengeþš= total number of mortalities: A chi-square test (Fisher s exact) was used to analyse the significance of differences in mortality between the PBS-PS group and the FIA-injected and vaccinated groups. A non-parametric ANOVA (Kruskal Wallis) was used to analyse the significance of differences in serum measurements and weight between groups. The criterion for significance was set at P < 0.05. Results Prechallenge Mortality of turbot infected experimentally by bathing in a solution containing 10 7 CFU ml )1 was 100%. Deaths started on day 4 and the last fish died on day 7. The bacterium was isolated from the kidney of all fish. No mortality occurred in turbot bathed in 10 6 CFU ml )1. Death of fish injected i.p. and i.m. with M. viscosa started 5 days after injection in those receiving the highest bacterial dose, and subsequently in fish receiving lower doses. The LD 50 dose of i.p.-injected turbot was estimated to be 3.2 10 5 CFU per fish and most of the infected turbot died within the first week. The LD 50 dose of i.m.-injected turbot could not be determined but lies below 10 5 CFU per fish, as all except one fish died within the first 15 days after injection. Table 1 shows the results of M. viscosa re-isolation from head kidney and presence of clinical signs in non-survivors and survivors. Moritella viscosa was re-isolated from 100% of bath, 53% of i.p.- and 74% of i.m.-infected mortalities. Clinical signs of infection (as described below) were detected in all dead fish, except two of 20 from the Table 1 Re-isolation of Moritella viscosa from bath, i.p.- and i.m.-infected turbot and survival and detection of clinical signs in the fish Mv isolation Bath-infected turbot (n ¼ 40) Non-survivors (n ¼ 20) 20/20 18/20 Survivors (n ¼ 20) 0/20 0/20 i.p.-injected turbot (n ¼ 25) Non-survivors (n ¼ 15) Control 0/0 0/0 Mv-injected 8/15 15/15 Survivors (n ¼ 10) Control 0/5 0/5 Mv-injected 0/5 0/5 i.m.-injected turbot (n ¼ 25) Non-survivors (n ¼ 19) Control 0/0 0/0 Mv-injected 14/19 19/19 Survivors (n ¼ 6) Control 0/5 0/5 Mv-injected 0/1 1/1 Clinical signs Mv, Moritella viscosa; i.p., intraperitoneal; i.m., intramuscular; n, number of fish. 648

bath challenge. No mortality occurred and no clinical signs were detected in control fish. Moritella viscosa was not cultivated from any surviving fish. Pathology of M. viscosa-infected turbot Gross examination of M. viscosa-infected turbot revealed pale gills in all fish. In bath-infected fish haemorrhages in skin were detected in various places (Fig. 1a), but most frequently on the head. Bleeding was also seen in internal organs of three fish. Seven days post-bath infection necrotic degeneration was observed in various internal organs and yellowish fluid was detected in the body cavity. In i.m.-injected fish skin ulcers and degenerative changes in underlying tissue at the site of injection were seen in 80% of the fish. Fin rot was also commonly seen. The depth of ulcers extended from the stratum compactum to deep into the skeletal muscle (Fig. 1b). Ulcers were only observed at the injection site. Haemorrhages around the mouth and at fin bases, a swollen abdomen and loss of scales were seen on the dorsal side of the fish. Yellowish fluid and haemorrhages in the body cavity were commonly seen as well as some muscle necrosis. Occasionally, ecchymotic haemorrhages were observed in the liver. Histopathology showed necrosis, inflammatory oedema and haemorrhages in the dermis. Bacteria, liquefactive necrosis and haemorrhages were observed in the skeletal muscle at the site of injection (Fig. 1c). Infiltration of mononuclear cells, regeneration and fibrosis were evident in muscle of fish examined later in the infection. Bacteria were widespread in dermis and muscle and secondary infections were common. Observed pathological changes in gills were hypertrophy and hyperplasia in mucous, chloride and epithelial cells and occasional fusion of the secondary lamellae. A rise in the numbers of melanomacrophages was noted in the Figure 1 Pathological changes in Moritella viscosa-infected turbot. (a) Haemorrhages in skin of bath-infected turbot, 7 days postinfection. (b) Ulcer at the site of injection, 13 days post-i.m. challenge. (c) Bacteria and liquefactive necrosis (circle) in skeletal muscle at the site of i.m. injection (syringe). Diffuse erosion of the epidermis and oedema in stratum spongiosum ( 10). (d) Parenchymal necrosis (circle) in liver ( 40). (e) Tubular (arrows) and parenchymal (circle) necrosis, and disseminated haemorrhages in kidney of i.m.-infected turbot ( 40). 649

primary lamellae. Liquefactive necrosis and disseminated haemorrhages were observed in the liver and parenchyma of kidney (Fig. 1d,e). Tubular necrosis (Fig. 1e) and sometimes tubular dilation were also seen in the kidney. The heart showed modest disseminated hyaline degeneration and mononuclear infiltration in the epicardium (epicarditis). In the spleen, degeneration and hyperaemia were seen but no pathological changes were observed in the intestine or pancreas. Moritella viscosa was not seen in tissues other than skin and muscle, although the bacterium was isolated from the head kidney, indicating that the infection was systemic. Vaccine efficacy No mortalities occurred in turbot challenged with 4 10 5 CFU per fish. Two fish died after challenge with 4 10 6 CFU per fish, one injected with FIA and one turbot vaccinated with Alphaject 5200. The deaths occurred 30 and 31 days after challenge, respectively. When challenged with 4 10 7 CFU per fish, 100% mortality occurred in the vaccinated group while mortalities in the control groups were under 90%. Turbot challenged with 4 10 8 CFU per fish showed 100% mortality in all injection groups within the first 9 days (Table 2). Accumulated mortality and calculated RPS and MDD are shown in Table 2. The results show that neither vaccinated nor FIA-injected turbot survived the M. viscosa infection better than PBS-PS-injected turbot. Furthermore, the MDD values show that the onset of death was not delayed in vaccinated fish. Weight gain and intra-abdominal adhesions The mean weight of each injection group, 7 weeks post-vaccination, is shown in Table 3. The difference in weight gain between vaccinated and control fish was not significant. No signs of intra-abdominal adhesions were detected in turbot injected with PBS-PS, when graded on the Speilberg scale. Turbot injected with adjuvant (FIA) were predominantly graded as 1 (minute adhesions) and turbot vaccinated with Alphaject 5200 as 2 (minor adhesions) as seen in Table 3. No fish was found to have more than moderate adhesions (grades 4 6). Serum analysis Protein concentration, SH 50% and anti-trypsin activity of turbot serum is shown in Table 4. Significant differences between groups were not Table 2 Accumulated mortality (acc. mort.), mean day to death (MDD) and relative percentage survival (RPS) of turbot i.m. challenged with Moritella viscosa, 13 weeks post-vaccination (Alphaject 5200) or control injection (PBS-PS or FIA) Challenge dose (CFU per fish) PBS-PS FIA Alphaject 5200 acc. mort. (%) MDD acc. mort. (%) P-value RPS (%) MDD acc. mort. (%) P-value RPS (%) MDD 4 10 5 0 0 0 4 10 6 0 4.8 0.47 4.6 0.48 4 10 7 89.5 11.2 84.6 1.00 5.4 10.5 100 0.22 )11.8 10.5 4 10 8 100 4.7 100 0 0 100 0 4.8 The P-value is from comparison with acc. mort. in the PBS-PS group. i.m., intramuscular; CFU, colony-forming unit; PBS-PS, phosphate-buffered saline with 0.1% peptone and 1.5% NaCl; FIA, Freund s incomplete adjuvant; Alphaject 5200, a commercial salmon vaccine. Injection group PBS-PS FIA Alphaject 5200 Table 3 Side-effects in post-vaccinated (Alphaject 5200) or control-injected (PBS-PS or FIA) turbot Mean weight 7 weeks p.v. a (g) 84.4 96.4 77.1 Mean weight gain (g) 33.5 45.5 26.2 Mean Speilberg b score 0 1 2 Mean weight at vaccination: 50.9 g. PBS-PS, phosphate-buffered saline with 0.1% peptone and 1.5% NaCl; FIA, Freund s incomplete adjuvant; Alphaject 5200, a commercial salmon vaccine; p.v., post-vaccination. a Comparison of mean weight in vaccinated group and PBS-PS group (P > 0.05). b Midtlyng et al. (1996). 650

Table 4 Measurement of humoral immune parameters in turbot (n ¼ 130) Range Mean SEM Serum protein concentration 11.5 66.4 36.8 0.81 (mg ml )1 ) Spontaneous haemolytic activity 5 125 38.7 2.26 (SH 50% ) Anti-trypsin activity (%) 82.9 98.1 94.5 0.22 n, Number of fish; SEM, standard error of the mean. detected for the three parameters assayed, 7 and 11 weeks post-vaccination, or between groups before and after challenge. A considerable individual variation in SH 50% activity was detected within each group. Antibody response of turbot serum to sonicated WB of M. viscosa and A. salmonicida, 7 and 11 weeks post-vaccination and against M. viscosa 5 weeks after challenge, is shown in Fig. 2. Vaccination did not induce a significant antibody response against M. viscosa antigens in vaccinated turbot, while challenge with M. viscosa raised the antibody response of all injection groups. Antibody levels of FIA-injected turbot were slightly higher than in PBS-PS-injected turbot, but the difference was not significant (Fig. 2a). A significantly increased antibody response to the A. salmonicida antigen was, however, detected in vaccinated turbot both 7 and 11 weeks post-vaccination (Fig. 2b). Endpoint titres against the M. viscosa antigen of five vaccinated turbot sera, sampled 11 weeks postvaccination, were as follows: one of 3200, three of 6400 and one above 12 800. All turbot sera sampled from the three injection groups, 5 weeks after challenge, had anti-m. viscosa endpoint titres of 6400 or higher. Discussion Figure 2 Anti-Moritella viscosa (a) and anti-aeromonas salmonicida (b) antibody activity of turbot serum, sampled 7 and 11 weeks post-vaccination and anti-m. viscosa response (a) after challenge (end) with Moritella viscosa. One group was injected with a commercial salmon vaccine, Alphaject 5200, and two groups injected with PBS-PS or FIA, were used as controls. The results are expressed as box plots. The median is the line within each box, boxes indicate 25 75 percentiles, tails 10 90 percentiles and circles extreme values. PBS-PS, phosphate-buffered saline with 0.1% peptone and 1.5% NaCl; FIA, Freund s incomplete adjuvant; Alphaj., Alphaject 5200; OD 405, optical density read at 405 nm. In this study, turbot was shown to be sensitive to experimental infection by M. viscosa and pathological changes in infected fish were described. Furthermore, it was demonstrated that vaccination with a commercial polyvalent salmon vaccine, Alphaject 5200, containing M. viscosa as one of five antigens, did not induce immune protection in turbot against M. viscosa infection. Challenge of turbot by M. viscosa strain F288/95 showed that fish were sensitive to the infection. The mortality was 100% following a bath in a concentration of 10 7 CFU ml )1, but no mortality occurred in turbot bathed in 10 6 CFU ml )1. The LD 50 of i.p.-infected turbot was estimated at 3.2 10 5 CFU per fish and lower than 10 5 CFU per fish by i.m. infection. Results from a previous study indicate that turbot is as sensitive as cod to M. viscosa infection at 9 C (Gudmundsdóttir et al. 651

2001). The LD 50 of i.m.-injected salmon with M. viscosa strain K58 (isolated Southwest of Iceland in Atlantic salmon) and strain K56 (isolated North of Iceland in Atlantic salmon), has been shown to be less than 3.5 10 3 CFU per fish (Benediktsdóttir et al. 1998, 2000). The bacterium was re-isolated from 100% of bath, 53% of i.p.- and 74% of i.m.-infected mortalities, and 96% of dead fish showed some clinical signs of infection. The low frequency of M. viscosa re-isolation in injected turbot reflects the authors experience with re-isolation of M. viscosa from infected salmon. Other authors have reported that the bacterium is difficult to isolate and often overgrown by fast-growing opportunistic bacteria (Lunder, Evensen, Holstad & Håstein 1995; Benediktsdóttir et al. 1998). No cohabitant infection occurred in control fish and M. viscosa was not re-isolated from any survivors. Pathological changes were not detected in surviving turbot, suggesting that they had recovered from the infection, or that infection had not taken place. In Iceland, turbot is farmed in 14 17 C sea water, where heat exchange with geothermal water is used to elevate water temperature. In intensive farming the amount of geothermal water can be a limiting factor resulting in periodic drops in temperature. In addition, there are farms cultivating turbot, cod and halibut in the same site. Natural M. viscosa infections have only been detected in salmon, rainbow trout and cod reared in sea water below 10 C and thus turbot were infected at 9 C in the present study. Whether M. viscosa can infect turbot at 14 C or higher is not known. The results of this study indicate that M. viscosa infection of farmed turbot may be a risk factor during temporary drops in water temperature. Gross and histological examination revealed that the pathology of M. viscosa infection in turbot was similar to that described in Atlantic salmon (Lunder et al. 1995; Bruno et al. 1998). Diffuse necrosis and disseminated haemorrhages were recorded in the dermis and muscle surrounding the site of injection. Disseminated skin haemorrhages, but no ulcers, were detected in bath-infected fish, which may reflect that all the fish died within a week. The most prominent internal signs were haemorrhages and degeneration in kidney and liver. The results of the vaccination experiment showed that the salmon vaccine, Alphaject 5200, induced neither protection against infection by M. viscosa nor antibody response against the bacterium (14 C). An anti-m. viscosa antibody response was, however, mounted in challenged fish, although the challenge was performed at suboptimal temperature (9 C). Furthermore, an anti-a. salmonicida antibody response was induced by the vaccine (14 C). This demonstrates that the M. viscosa antigen in the vaccine is inadequate for use in turbot. A raised antibody response was detected in FIA-injected turbot, but it was not significantly different from the PBS-PS group. This is in accordance with non-specific immunostimulation by FIA reported previously (Gudmundsdóttir, Jónsdóttir, Steinthórsdóttir, Magnadóttir & Gudmundsdóttir 1997). Specificity of the antibody response of surviving turbot was high, as the endpoint titres of sera sampled after challenge with M. viscosa were all at or above 6400. It would be interesting to test whether sera sampled from challenged turbot could protect against M. viscosa infection, using passive immunization. Challenge with i.m. injection was chosen for the vaccination experiment based on the prechallenge results. In addition, the vaccine was administered i.p. and therefore an i.m. challenge avoids interference with local immune responses at the site of vaccination. The sensitivity of turbot to M. viscosa infection in the two studies, the prechallenge and the vaccination experiment, was very different. A much higher bacterial dose was needed to produce mortality in the vaccination study. The authors have previously observed poor reproducibility between challenges of salmon with M. viscosa. The reason is not known, but environmental factors may be involved. In the present study, the size of turbot and methods used for marking were the only factors known to be different between the two challenge experiments. Turbot used in the prechallenge experiment weighed 50 g and turbot used in the vaccination experiment weighed over 80 g at the time of infection. Plastic tags were injected into muscle for identification in the prechallenge and implants of a fluorescent dye along fin margins were used in the vaccination experiment. Vaccination side-effects were evaluated 7 weeks post-vaccination. The vaccination did not have significant negative effects on the weight of vaccinated fish, although moderate intra-abdominal adhesions were detected in vaccinated fish and in fish injected with the FIA adjuvant. Oil-adjuvant injections have previously been shown to cause intraabdominal adhesions in vaccinated fish, such as halibut (Bowden, Lester, MacLachlan & Bricknell 652

2000; Gudmundsdóttir, Lange, Magnadóttir & Gudmundsdóttir 2003) and salmon (Poppe & Breck 1997; Midtlyng & Lillehaug 1998; Sørum & Damsgård 2004), resulting in growth reduction. Adjuvants are used in most current fish vaccines as antigen carriers and for slow antigen release. An oil emulsion system that resembles FIA is the most widely used (Midtlyng 1997). No significant differences between groups were measured in protein concentration, SH 50%, and anti-trypsin activity of turbot serum 7 and 11 weeks post-vaccination or before and after challenge. The protein concentration measured was comparable with that previously reported in smaller turbot by Hutchinson, Field & Manning (1999). In their study, however, immunization affected serum protein levels negatively. Slightly higher serum protein levels have been measured in halibut, cod and salmon (Magnadóttir & Gudmundsdóttir 1992; Magnadóttir, Gudmundsdóttir & Gudmundsdóttir 1995; Magnadóttir et al. 1999a; Lange, Gudmundsdóttir & Magnadóttir 2001). Various factors, such as environmental temperature (Magnadóttir et al. 1999a), fish size (Magnadóttir, Jónsdóttir, Helgason, Björnsson, Jørgensen & Pilström 1999b) and crowding stress (Yin, Lam & Sin 1995) have been shown to affect serum protein levels in fish. Complement activity by the alternative pathway is an important component of the innate immune system in fish. In this study, the SH 50% of turbot serum was considerably lower than previously reported for sea bass, Dicentrarchus labrax (L.), and halibut by Lange & Magnadóttir (2003). Anti-trypsin activity is an important parameter for the neutralization of extracellular enzymes of pathogenic bacteria. In the present study, anti-trypsin activity was detected in turbot serum. Turbot trypsin-inhibitory activity has previously been compared with halibut, brown trout, Salmo trutta (L.) Arctic char, Salvelinus alpinus (L.), and rainbow trout (Bowden, Butler, Bricknell & Ellis 1997) and was found to be significantly higher in turbot than in Arctic char. The results of this study show that turbot is susceptible to infection by M. viscosa. The fish were infected at a temperature that is lower than the rearing temperature of turbot and further investigation is needed to test if the bacterium can infect turbot at a higher temperature. The results also indicate that the commercial vaccine, Alphaject 5200, which was developed for use in salmon, does not give turbot protection against M. viscosa infections. Some modifications of the vaccine are needed if it is to be used for vaccination against the bacterium in turbot. Acknowledgements This work was funded by the Icelandic Research Council. The authors wish to thank the staff at the Marine Institute s Experimental Station, Stadur, Grindavík and at the Research Centre, Sandgerdi, for providing facilities and looking after the fish. Thanks are also due to Margrét Jónsdóttir for her help with histological samples. References Austin B., Stobie M., Robertson P.A.W., Glass H.G., Stark J.R. & Mudarris M. (1993) Vibrio alginolyticus: the cause of gill disease leading to progressive low-level mortalities among juvenile turbot, Scophthalmus maximus L., in a Scottish aquarium. Journal of Fish Diseases 16, 277 280. Benediktsdóttir E., Helgason S. & Sigurjónsdóttir H. (1998) Vibrio spp. isolated from salmonids with shallow skin lesions and reared at low temperature. Journal of Fish Diseases 21, 19 28. Benediktsdóttir E., Verdonck L., Spröer C., Helgason S. & Swings J. (2000) Characterization of Vibrio viscosus and Vibrio wodanis isolated at different geographical locations: a proposal for reclassification of Vibrio viscosus as Moritella viscosa comb. nov. International Journal of Systematic and Evolutionary Microbiology 50, 479 488. Bowden T.J., Butler R., Bricknell I.R. & Ellis A.E. (1997) Serum trypsin-inhibitory activity in five species of farmed fish. Fish and Shellfish Immunology 7, 377 385. Bowden T.J., Lester K., MacLachlan P. & Bricknell I.R. (2000) Preliminary study into the short term effects of adjuvants on Atlantic halibut (Hippoglossus hippoglossus L.). Bulletin of the European Association of Fish Pathologists 20, 148 152. Bradford M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248 254. Bruno D.W., Griffiths J., Petrie J. & Hastings T.S. (1998) Vibrio viscosus in farmed Atlantic salmon Salmo salar in Scotland: field and experimental observations. Diseases of Aquatic Organisms 34, 161 166. Colquhoun D.J., Hovland H., Helleberg H., Haug T. & Nilsen H. (2004) Moritella viscosa isolated from farmed Atlantic cod (Gadus morhua). Bulletin of the European Association of Fish Pathologists 24, 109 114. Domenech A., Fernandez-Garayzabal J.F., Pascual C., Garcia J.A., Cutuli M.T., Moreno M.A., Collins M.D. & Dominguez L. (1996) Streptococcosis in cultured turbot, Scophthalmus maximus (L.), associated with Streptococcus parauberis. Journal of Fish Diseases 19, 33 38. 653

Ellis A.E. (1990) Serum antiproteases in fish. In: Techniques in Fish Immunology (ed. by J.S. Stolen, T.C. Fletcher, D.P. Anderson, B.S. Roberson & W.B. van Muiswinkel), pp. 95 99. SOS Publications, Fair Haven, NJ. Frey A., Di Canzio J. & Zurakowski D. (1998) A statistically defined endpoint titer determination method for immunoassays. Journal of Immunological Methods 221, 35 41. Gatesoupe F.J., Lambert C. & Nicolas J.L. (1999) Pathogenicity of Vibrio splendidus strains associated with turbot larvae, Scophthalmus maximus. Journal of Applied Microbiology 87, 757 763. Gudmundsdóttir B.K., Jónsdóttir H., Steinthórsdóttir V., Magnadóttir B. & Gudmundsdóttir S. (1997) Survival and humoral antibody response of Atlantic salmon, Salmo salar L., vaccinated against Aeromonas salmonicida ssp. achromogenes. Journal of Fish Diseases 20, 351 360. Gudmundsdóttir B.K., Magnadóttir B. & Gudmundsdóttir S. (2001) Experimental infection of juvenile turbot (Scophthalmus maximus), halibut (Hippoglossus hippoglossus) and cod (Gadus morhua) with the bacterium Moritella viscosa. 10th International Conference, Diseases of Fish and Shellfish. European Association of Fish Pathologists (EAFP), Abstract Book O-88. Gudmundsdóttir S., Lange S., Magnadóttir B. & Gudmundsdóttir B.K. (2003) Protection against atypical furunculosis in Atlantic halibut, Hippoglossus hippoglossus (L.); comparison of a commercial furunculosis vaccine and an autogenous vaccine. Journal of Fish Diseases 26, 331 338. Hutchinson T.H., Field M.D. & Manning M.J. (1999) Evaluation of immune function in juvenile turbot Scophthalmus maximus (L.) exposed to sediments contaminated with polychlorinated biphenyls. Fish and Shellfish Immunology 9, 457 472. Lange S. & Magnadóttir B. (2003) Spontaneous haemolytic activity in Atlantic halibut (Hippoglossus hippoglossus L.) and sea bass (Dicentrarchus labrax) serum. Comparative Biochemistry and Physiology Part B 136, 99 106. Lange S., Gudmundsdóttir B.K. & Magnadóttir B. (2001) Humoral immune parameters of cultured Atlantic halibut (Hippoglossus hippoglossus L.). Fish and Shellfish Immunology 11, 523 535. Larsen J.L. & Pedersen K. (1999) Infeksjoner med Vibriobakterier. In: Fiskehelse og fiskesykdommer (ed. by T. Poppe), pp. 68 83. Universitetsforlaget, Oslo. Lillehaug A., Lunestad B.T. & Grave K. (2003) Epidemiology of bacterial diseases in Norwegian aquaculture a description based on antibiotic prescription data for the ten-year period 1991 to 2000. Diseases of Aquatic Organisms 53, 115 125. Lunder T., Evensen Ø., Holstad G. & Håstein T. (1995) ÔWinter ulcerõ in the Atlantic salmon Salmo salar. Pathological and bacteriological investigations and transmission experiments. Diseases of Aquatic Organisms 23, 39 49. Lunder T., Sørum H., Holstad G., Steigerwalt A.G., Mowinckel P. & Brenner D.J. (2000) Phenotypic and genotypic characterization of Vibrio viscosus sp. nov. and Vibrio wodanis sp. nov. isolated from Atlantic salmon (Salmo salar) with Ôwinter ulcerõ. International Journal of Systematic and Evolutionary Microbiology 50, 427 450. Magnadóttir B. & Gudmundsdóttir B.K. (1992) A comparison of total and specific immunoglobulin levels in healthy Atlantic salmon (Salmo salar L.) and in salmon naturally infected with Aeromonas salmonicida subsp. achromogenes. Veterinary Immunology and Immunopathology 32, 179 189. Magnadóttir B., Gudmundsdóttir S. & Gudmundsdóttir B.K. (1995) Study of the humoral response of Atlantic salmon (Salmo salar L.), naturally infected with Aeromonas salmonicida ssp. achromogenes. Veterinary Immunology and Immunopathology 49, 127 142. Magnadóttir B., Jónsdóttir H., Helgason S., Björnsson B., Jørgensen T.Ø. & Pilström L. (1999a) Humoral immune parameters in Atlantic cod (Gadus morhua L.). I. The effects of environmental temperature. Comparative Biochemistry and Physiology Part B 122, 173 180. Magnadóttir B., Jónsdóttir H., Helgason S., Björnsson B., Jørgensen T.Ø. & Pilström L. (1999b) Humoral immune parameters in Atlantic cod (Gadus morhua L.). II. The effects of size and gender under different environmental conditions. Comparative Biochemistry and Physiology Part B 122, 181 188. Magnadóttir B., Jónsdóttir H., Helgason S., Björnsson B., Solem S.T. & Pilström L. (2001) Immune parameters of immunised cod (Gadus morhua L.). Fish and Shellfish Immunology 11, 75 89. Midtlyng P.J. (1997) Novel vaccines and new vaccination strategies for fish. Bulletin of the European Association of Fish Pathologists 17, 239 244. Midtlyng P.J. & Lillehaug A. (1998) Growth of Atlantic salmon Salmo salar after intraperitoneal administration of vaccines containing adjuvants. Diseases of Aquatic Organisms 32, 91 97. Midtlyng P.J., Reitan L.J. & Speilberg L. (1996) Experimental studies on the efficacy and side-effects of intraperitoneal vaccination of Atlantic salmon (Salmo salar L.) against furunculosis. Fish and Shellfish Immunology 6, 335 350. Pedersen K., Kofod H., Dalsgaard I. & Larsen J.L. (1994) Isolation of oxidase-negative Aeromonas salmonicida from diseased turbot Scophthalmus maximus. Diseases of Aquatic Organisms 18, 149 154. Poppe T.T. & Breck O. (1997) Pathology of Atlantic salmon Salmo salar intraperitoneally immunized with oil-adjuvanted vaccine. A case report. Diseases of Aquatic Organisms 29, 219 226. Reed L.J. & Muench H. (1938) A simple method of estimating fifty percent endpoints. American Journal of Hygiene 27, 493 497. Sørum U. & Damsgård B. (2004) Effects of anaesthetisation and vaccination on feed intake and growth in Atlantic salmon (Salmo salar L.). Aquaculture 232, 333 341. Thorarinsson R. & Lystad Y. (2003) Vintersår: Ny kunnskap skaper optimisme. Norsk Fiskeoppdrett 28, 34 38. Toranzo A.E., Barja J.L. & Devesa S. (1985) First isolation of Vibrio anguillarum biotype 1 causing an epizootic in reared turbot (Scophthalmus maximus L.) in Galicia (NW Spain). Investigacion Pesquera 49, 61 66. 654

Toranzo A.E., Santos Y., Lemos M.L., Ledo A. & Bolinches J. (1987) Homology of Vibrio anguillarum strains causing epizootics in turbot, salmon and trout reared on the Atlantic coast of Spain. Aquaculture 67, 41 52. Toranzo A.E., Cutrin J.M., Nunez S., Romalde J.L. & Barja J.L. (1995) Antigenic characterization of Enterococcus strains pathogenic for turbot and their relationship with other Grampositive bacteria. Diseases of Aquatic Organisms 21, 187 191. Villamil L., Figueras A., Toranzo A.E., Planas M. & Novoa B. (2003) Isolation of a highly pathogenic Vibrio pelagius strain associated with mass mortalities of turbot, Scophthalmus maximus (L.), larvae. Journal of Fish Diseases 26, 293 303. Yin Z., Lam T.J. & Sin Y.M. (1995) The effects of crowding stress on the non-specific immune response in fancy carp (Cyprinus carpio L.). Fish and Shellfish Immunology 5, 519 529. Received: 4 April 2004 Revision received: 25 June 2004 Accepted: 29 July 2004 655