Journal of Fish Diseases 1991, 14, 443-452

Humoral immune response in Atlantic salmon, Salmosalar L., to cellular and extracellular antigens ofAeromonas salmonicida

V. LUND, ' T. J 0 R G E N S E N , ^ K. O. HOLM' & G. EGGSET^ UpothekernesLaboratorium a.s, and ~The Foundation of Applied Research at the University of Troms0, Troms0, Norway

Abstract. The putative virulence factors oi Aeromonas salmonicida, the aetiological agent offurunculosis in salmonids, are candidates for protective antigens in effective vaeeines againstfurunculosis. In this report, the authors have compared the immunogenieily of eell-assoeiatedand extracellular antigens of A. salmonicida in Atlantie salmon, Salmo salar L., to that inrabbit. The animals were immunized with formalin-killed whole cells and formalin-inactivatedextraeellular products (ECP), either separately or in combination. The ability of the antigens toinduce antibody production was studied by EUSA and Western blotting techniques. Theseresults confirm previous reports that far more struetures are immunogenie in rabbit comparedto the antibody responses elicited in salmon. However, in both speeies, some antigens weredominant, including a caseinolytie protease in addition to the A-protein and high andlow MW LPS.

Introduction

The pathogenic mechanisms of Aeromonas salmonicida infection in salmonids are still notclearly understood. A variety of cell-associated and extracellular factors have been investigatedin order to determine their role in virulence and pathogenesis. These include the cell surfaceA-protein, lipopolysaccharides (LPS) and various extracellullar products (ECP).

Several groups have reported correlation between virulence and the A-protein, but avirulentstrains with A-layer (Udey & Fryer 1978) and virulent strains without detectable A-layer(Johnson, Tatner & Horne 1985) have also been reported.

Aeromonas salmonicida produces 15-20 protein components which are released in vitroand probably exert a role in virulence. Because of the complexity of the ECP which includeseveral proteases (Sheeran & Smith 1981; Tajima, Takahashi, Ezura & Kimura 1984; Fyfe,Finley & Coleman 1986; Price, Stevens, Duncan & Snodgrass 1989), haemolysins (Titball &Munn 1981; Fyfe, Coleman & Munro 1987; Nomura, Fujino, Yamakawa & Kawahara 1988;Lee & Ellis 1989) and a leucocidin (Fuller, Pilcher & Fryer 1977), it has been difficult to relatevirulence to one specific component. However, ECP has been shown to reproduce lesionsassociated with the chronic form of furunculosis (Ellis, Hastings & Munro 1981), and bothcrude ECP (Munro, Hastings, Ellis & Liversidge 1980) and purified components like theprotease and haemolysin (Lee & Ellis 1989) have been shown to be lethal in fish.

During the last 50 years, numerous furunculosis vaccines have been tested, such as inactivatedwhole cells, disrupted cells, soluble and particulated extracellular components, attenuated livecells, in addition to passive transfer of immune serum. Also, a variety of bacterial strains andmethods for vaccine production have been used, which may affect the antigenic composition ofthe various vaccines. In addition, different methods of administration, types of adjuvants andspecies of fish were used. In general, the results have been variable and not reproducible.

Correspondence: Vera Lund, Apothekernes Laboratorium a.s., PO Box 2824 Elverlioy, N-9001 Troms0, Norway.

444 V. Lund et al.

While most of the A. salmonicida antigens are immunogenie in rabbit, only a few areimmunogenie in fish, and apparently there are none which might induce specific long-termimnuniity (Hastings & Ellis 1988). On the other hand, passive immunization of trout withimmune serum from rabbit (Ellis, Burrows, Hastings & Stapleton 198Ha) or rainbow trout(Cipriano 19,S2) demonstrated protection against challenge, thus indicating that humoral immuneresponses in fish can be protective if crucial antigens are made more immunogenie.

In the present work, the authors have compared the antigenic properties of both cellularand extraeellular antigens in salmon and rabbit by using the sensitive Western blot technique.

Materials and methods

Bacterial strains

Four different strains of Aeromonas salmonicida subsp. salmonicida were used in this study.The strains F.t.l940 and AL2017 were isolated from Atlantic salmon, Salmo salar L., at twodifferent outbreaks in Norwegian fish farms (1986 and 1988, respectively). Strain MT028, a giftfrom Trevor Hastings, was isolated in the USA from brown trout, Salmo trutta L. These threestrains possess the A-layer (A + ), wherease strain NCMB1102, which was used as a reference,lacks the A-layer (A- ) . All strains produce caseinolytie activity, except MT028.

AJI bacteria were grown in 200 m] brain-heart infusion medium (BHI; Difco, Detroit, MI,USA) with shaking at 20°C for 30h before harvest by centrifugation. The pellet was used forouter membrane (OM) preparations and the supernatant was used for ECP preparations.

Preparation of outer membranes (OM)

Cells (approximately Ig wet weight) were washed once in 20mM Tris HCI (pH 7 4) andresuspended in 20 ml of the same buffer containing 10mM EDTA, 0 1 ng of DNase (Sigma, StLouis, MO, USA) per ml, and Oljig RNase (Sigma) per ml. The cell suspension was broketiby two passes through a French pressure cell. OM-preparations were made from the brokensuspension as described by Munn, Ishiguro, Kay & Trust (1982). Cytoplasmic membranes wereselectively solubilized with sodium lauryl sarcosinate by the method of Filip, Fletcher, Wulff &Earhart (1973). The OM-preparations were stored at -20°C until used in Western blotting.

Extracellular products (FCP)

Culture supernatants were passed trough 0'22-(.tm filters to remove remaining bacterial cells,and then dialysed against 20mM Tris HCI (pH 7 4) before being concentrated (5—lOx) withpolyethylene glyeol (PEG MW 35 000). The concentrated supernatants (ECP preparations)were dialysed once more against the same buffer, and total protein was measured using theBio-Rad Protein Assay. ECP preparations were stored at —20°C until used in Western blotting.

ECP preparations to be used for immunization were inactivated in 5% formalin for 20 h at20°C. Excess formalin was removed by dialysis against PBS.

Antisera

Antisera to cellular antigens. Rabbits and salmon were immunized with whole bacterial cells(WC) of strains Ft.1940 (A+) or NCMB1102 (A- ) inactivated in 0 5% formalin for 4 days.

Immune responses in Atlantic salmon to antigens of A. salmonicida 445

A suspension of killed and washed cells, approximately 210'' cells ml ', was emulsified in anequal volume of Freund's complete adjuvant (WC-FCA). The rabbits received 0-5 ml of themixture in each of the hind legs. After 6 weeks, the animals were boosted with the same doseof bacterial cells in incomplete Freunds adjuvant (WC-IFA) and bled 10 days after. The titresof the sera were determined by ELISA and referred to as R-anti-WC(A-l-) and R-anti-WC(A —).

Salmon (groups of 10—15 individuals, each weighing about lOOg, kept at 10—12''C and fedcommercial fish pellet, were used throughout the experiments), were given 0'2ml i.p. injectionof WC-FCA mixtures (A+ or A - ) . After 12 weeks, the group given the F.t.l940 strain wasbled. The one given NCMB1102 was at the same time boosted with the WC-IFA mixture andbled 7 weeks later. The four sera from each group having the highest antibody titres asscreened by ELISA, were pooled and denoted S-anti-WCCA-t-) and S-anti-WC(A —).

Antisera to extracellular antigens. Two rabbits were immunized with ECP preparations fromstrain AL2017 (400|.ig protein per animal) emulsified in FCA as described above. A boosterinjection (same dose without adjuvant) was given 10 weeks later. Sera were collected 8 dayslater, pooled and referred to as R-anti-ECP.

Salmon received i.p. injections of ECP (80 g protein) in FCA. After 14 weeks, they weregiven the same dose in IFA, and bled another 14 weeks later. The five sera with highestantibody titres were pooled and referred to as S-anti-ECP.

Antisera to combined cellular and extracellular antigens. Salmon were immunized with amixture of whole cells and ECP using aluminium hydroxide gel (Alhydrogel, Superfos Biosectora/s, Vedbaek, Denmark) as adjuvant. The mixture was made as follows: bacterial suspensionsof strains AL2017(A+) and NCMB1102(A—) were mixed (1:1) and added to a mixture ofconcentrated supernatants (1:1 of mg total protein) from the same strains to a final concentrationof 6-10'** cells and 200mg ECP m~' in 1% Alhydrogel. Each fish received an i.p. injection of0 8ml. They were bled 14 weeks later and the sera with highest antibody titres were pooledand referred to as S-anti-WC/ECP.

Monoclonal antibodies. Monoclonal antisera against A. salmonicida A-layer and LPS werekind gifts from Dr R. Nilsen (Apothekernes Laboratorium a/s, Troms0, Norway).

Enzyme-Unked-immunosorbent-assay (ELISA)

Microtitre plates precoated with poly-L-lysine (l\ig PLL in 200|.(l per well) were coated withwashed bacterial cells (100(il of bacterial suspension ODeoo = 0-3 — 0-5), while untreated plateswere used for coating with ECP (20(.ig/ml, 100f.il per well). A carbonate buffer (0-1 M, pH 9-6)was used as coupling buffer. The plates were then blocked with 5% normal goat serum (NGS,Gibco) in PBS (150^1 per well). In the first step, plates were incubated with dilution series ofthe antisera to be assayed (lOOfil per well). In the second step, either a goat anti-rabbit IgGalkaline phosphatase (Sigma, dilution 1:1000 in PBS) or a rabbit anti-salmon IgM conjugatemade in the authors' laboratory (dilution 1:500 in PBS) were used. Normal salmon or rabbitsera (NSS or NRS) were used as negative controls.

p-Nitrophenyl-phosphate (Sigma) in 1-M diethanolamine (Sigma) containing 0 5mM MgCK(Img mP^), pH 9-8, was used as substrate. The binding of antibodies was finally read as ODat 405 nm. Antibody titres were given as the dilution of sera giving the OD405nrn = 0-3 afterbackground subtraction.

446 V. Lund et al.

Western blotting and immunostaining

The OM and ECP preparations were separated by SDS-PAGE (OM- and ECP-PAGE) andblotted onto nitrocellulose filters (OM- and ECP-blot) using a BioRad Miniprotean^'^ unitaccording to the BioRad Instruction Manual. Rainbow marker (Amersham) was used as amolecular weight standard. Unspecific binding sites on filters were blocked with 10% NGS inPBS with ()(IS% Tween before incubation with antisera in an appropriate dilution (see figurelegends). NSS or NRS were used as negative controls. The filters were stained with appropriatealkaline phosphatase conjugates using 5-bromo-4-ch!oro-3-indoIylphosphate and nitrobluetetrazolium chloride (BCIP & NTB, BRL) as substrates. After blotting, gels were proteinsilver stained by the method described by Morrisey (1981).

Results

Antibody titres

Antibody titres in salmon and rabbits to cellular and extracellular antigens of homologousstrains of A. salmonicida are shown in Table 1. After immunization with A( + ) strain, salmondemonstrated a good antibody response to whole cells possessing the A-layer, but a rather lowresponse to A ( - ) cells (data not shown). However, salmon immunized with A(—) straingenerally showed a low antibody titre both against homologous strain and F.t.l940 (notshown). The response to ECP was also poor. Thus, the major antibody production in salmonseems to be directed against the A-layer. However, rabbits responded very well to whole cellswhether they possessed A-Iayer or not. Moreover, the rabbits also responded much better tothe ECP antigens than salmon, as shown by the ELISA titre of 5000 versus 500. When salmonwere immunized with a mixture of whole cells and ECP, they responded to each antigen as ifthey were immunized separately with WC or ECP.

Western blots

To identify the cellular and extracellular antigens able to stimulate antibody production insalmon and rabbit, the antisera were used for immunostaining of Western blots of OM- andECP-preparations (OM- and ECP-blots).

Cellular antigens (OM). Figure lA shows a silver stained OM-PAGE. The A-protein (50kDa,indicated) is demonstrated in all strains except NCMB1102 ( A - ) . The A-protein was alsoidentified by staining of Western blots using a monoelonal antibody specific to the A-protein(data not shown). LPS is located as indicated, and has been identified by silver staining of LPSusing the method described by Hitchcock & Brown (1983).

Staining of OM-blots with S-anti-WC(A-h) or S-anti-WC(A-) are shown in Fig. IB.Except for the A( - ) strain NCMBl 102, the strains MT028, AL2017 and the homologous strainF.t.l940 exhibit the same antibody binding pattern. The humoral immune responses againstthese strains are mainly directed towards the A-protein and in some degree to LPS. However,the response to the A(- ) strain is different, with low MW LPS as the main and only target forthe antibody production. The faint staining of the A-protein is probably due to nonspecificstaining or natural antibodies, since OM-blot of an A(H-) strain stained weakly with normalsalmon serum (NSS) in the A-protein region in addition to another major protein of about 40 kDa.

Immune responses in Atlantic salmoti to antigens of A. salmonicida 447

Table I. Anlibody litres against homologous A. salmonicida antigens in antisera Irom salmon and rabbitdetermined by ELISA

Antigens used forimmunization

F.t.l940(WC)NCMBnO2(WC)AL2O17(ECP)AL2O17/NCMB11O2(WC/ECP)F.t.l940(VVC)NCMBl 102(\VC)AL2017(ECP)

Anliscrum'

S-anti-WC(A + )S-anti-WC(A-)S-anli-ECPS-anti-WC/BCPR-anti-WC(A + )R-anti-WC(A+)R-anti-ECP

Antibody titres

we-

64f>030fJ

50001280(J64fX)

ECP'

500500

5000

' S. salmon; R, rabbit.~ Mierotitre plates were coated with whole cells (WC) of homologous strain.^ Mierotitre plates were coated with AL2017 ECP.

While salmon responded only to a few OM-components, rabbits are good responders tothese antigens as seen in Figure lC. In addition to the A-protein and high and low MW LPS,several eomponents are recognized, especially in the 15-25kDa region. As can also be seenfrom Fig. lC, sera from rabbit immunized with an A( + ) strain seem to recognize more OM-components than rabbit sera raised against an A( - ) strain. This may indicate an adjuvanteffect of the hydrophobic A-Iayer protein, at least in rabbit.

Extracellular antigens (ECP). Aeromonas salmonieida produces a whole range of extracellularproducts as shown on the ECP-PAGE in Fig. 2A. Some variations in the ECP-patterns of thedifferent strains can also be seen, but most of the components are not identified. High and lowMW LPS were revealed by LPS silver staining. The A-protein was not present, as judged bythe lack of staining with an A-protein specific monoclonal antibody (data not shown). StrainMT028 which is protease negative (Ellis, Burrows & Stapleton 1988b), seems to produce fewerECP components than the others. It also lacks the 70 kDa protein which corresponds tothe extracellular caseinolytic protease recently purified in the present authors' laboratory(unpublished data) and which has previously been described by others (Fyfe et al. 1986: Priceetal. 1989).

In order to compare the anti-ECP specificities in rabbit and salmon antisera, ECP-blotswere separately stained with S-anti-ECP and R-anti-ECP (Fig. 2B). While rabbits seemed torespond to most of the ECP-components, salmon responded only to two or four of thecomponents, having MW of approximately 20, 55, 70 and 90kDa. The two or three bandsseen at approximately 55 kDa correspond to high MW LPS and the 70 kDa component ismost probably the protease. Strain MT028 lacks the 20 kDa component in addition to theabove mentioned 70 kDa protease. NSS or NRS gave no staining when incubated with ECPblots (not shown).

Cellular and extracelltdar antigens. Figure 3 shows an OM- and ECP-blot stained with S-anti-WC/ECP. The antibody response to a mixture of cellular and extracellular antigens resemblesthe response obtained after immunizing with WC or ECP separately (Figs IB & 2B). Themajor outer membrane immunogens seem to be the A-protein, the high and low MW LPS,

448 r. Lund cl ul.

3 4

A 30—

-LPS"^A-protein

-LPS

1 2 3* 4 1* 2 3 4

93—69—46—

B 30—

21—

1 4 — mn^-^m^

S-anti--WC(A+)

1 2 3* 4

S-anti-WC(A-) NSS

4

69—

4 6 - *

^ ^ (ju^~* vjiK j | ^ ^ ^^B ^ i ^ ^

21 —

14—

R-anti"WClA+l R-anti-WC(A-) NRS

Figure 1. Humoral immune responses in salmon and rabbit to cellular antigens of A. salmonicida. (A) Outermembrane preparations of four strains of A. salmonicida (lanes: 1, NCMB1102; 2. MT()28; 3, F.t.l940; 4.AL2017) separated by 12% SDS-PAGE and protein silver stained (OM-PAGE). The A-protein and high andlow MW LPS are indicated. (B) Western-blot of OM-PAGE stained with S-anti-WC(A-F), S-anti-WC(A-)and normal salmon serum (NSS), rcspcetively. All sera were diluted 1/250. (C) Western blot of OM-PAGEstained with R-anti-WC(A-), R-anti-WC{A-) and normal rabbit serum (NRS). respectively. All sera werediluted 1/1000. Strains used for immunization are marked with asterix, and molecular weight markers in kDaare indicated.

Immtme responses in Atlantic salmon to antigens of A. salmonicida 449

1 2 3 4

-LPS

21—

1 2 3 4 1 2 3 4

9369

30-

21-

BS-anti-ECP R anti ECP

Figure 2. Humoral immune responses in salmon and rabbit to extracellular antigens of A, salmonicida. (A)Extracellular preparations of four strains of A. salmonicida (lanes: 1, MT028; 2, NCMB1102; 3, F.t. 1940: 4,AL2017) separated by 12% SDS-PAGE and protein silver stained (ECP-PAGE). (B) Western blots of ECP-PAGH stained with S-anti-ECP diluted 1/250 and R-anti-ECP diluted 1/1000. ECP from strain AL2017 was usedfor immunization. Molecular weight markers in kDa are indicated.

and a component of about 35 kDa which is lacking in strain NCMBl 102. The response to ECPseems to be directed mainly against the protease (70kDa), LPS (55kDa) and an unidentifiedcomponent of approximately 20kDa. The response to the 90 kDa component observed withS-anti'ECP (Fig. 2B) seems to be lacking. The strong staining of the high MW LPS ofNCMBl 102 ECP can be explained by the fact that this strain, lacking the A-layer, may releasemore LPS into the culture supernatant.

450 V. Lund et al.

46*-

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

21—

S-anti-WC/ECP

BFigure 3. Humoral immune responses in salmon to a combination of cellular and extracellular antigens. (A)OM- and ECP-preparations from four strains of A. salmonicida (lanes: 1, MT028; 2, NCMB1102: 3, F.t.1940: 4,AL2017) separated by \2% SDS PAGE and protein silver stained. (B) Western blot of the gel from (A) stainedwith S-anti-WC/ECP diluted 1/250. Whole eells and ECP of strains AL20I7 and NCMBU02 were used forimmunization. Moleeular weight markers in kDa are indieated.

Discussion

The present authors' results, obtained by using the Western blot teehnique, confirm previousfindings that nearly all the cellular and extracellular antigens of ^ . salmonicida are immunogenicin rabbit. They have demonstrated that rabbits respond to most OM-components with a majorantibody production against A-layer and high and low MW LPS. Rabbits also respond to about15 of the various ECP-components. This corresponds to the results reported by Ellis, Stapleton& Hastings (1988c) and Hastings & Ellis (1988), who could also identify the 70 kDa proteaseand a haemolysin among the ECP-components immunogenic in rabbits.

On the other hand, only about four OM-components, including A-layer and high and lowMW LPS, were shown to be immunogenic in Atlantic salmon. Salmo salar L. However,Hastings & Ellis (1990) detected antibodies only to A-layer and low MW LPS in rainbow trout.Also, most ECP-components seem to be poorly immunogenic in salmon, since only three tofive components including the caseinolytic protease of 70 kDa, were detected. Ellis et al.(1988c) and Hasting & Ellis (1988) also found three to five immunogenic components inrainbow trout, Oncorhynchus mykiss (Walbaum), but in contrast, they could not detect specificantibodies to the protease. The reason might be that the protease is less immunogenic inrainbow trout, or that crossed immunoelectrophoresis is a less sensitive detection method.Lately, however, Hastings & Ellis (1990) have been able to detect antibodies against theprotease in rainbow trout using the Western blotting technique.

The present authors have shown that most A. salmonicida antigens, including potentialprotective antigens, are not able to induce antibody response in salmon. This might explain thepoor efficiency obtained with experimental as well as commercial furunculosis vaccines. However,several groups have demonstrated protection in salmonids against experimental challenge ofA. salmonicida by passive transfer of immune sera (Cipriano 1982; Ellis et al. 1988a), indicating

Immune responses in Atlantic salmon to antigetis of A. salmonicida 451

that an appropriate antibody response in salmon to A. salmonicida antigens is important forprotection.

In order to make an efficient furunculosis vaccine, the protective antigen(s) has to beidentified, and most probably its immunogenicity has to be enhanced either by chemical orphysical modification. One way of identifying protective antigens is to study purified andmodified antigens' ability to provoke protection against experimental challenge. Alternatively,they can be revealed by passive transfer of rabbit sera to purified antigens followed bychallenge. In addition, more adjuvants should be tested for their ability to increase bothspecific and non-specific defence mechanisms.

References

Cipriano R. C. (1982) Immunization of brook trout (Salvelinus fontinalis) against Aeromonas salmonicida:immunogenicity of virulent and avirulent isolates and proteetive ability of different antigens. CanadianJournal of Fisheries and Aquatic Sciences 39, 218—221.

Ellis A. E., Hastings T. S. & Munro A. L. S. (1981) The role of Aeromonas salmonicida extracellular productsin the pathology of furuneulosis. Journal of Fish Diseases 4, 41-51 .

Elhs A. E.. Burrows A. S., Hastings T. S. & Stapleton K. J. (1988a) Identification of Aeromonas salmonicidaextracellular protease as a protective antigen against furunculosis by passive immunization. Aquaculture70, 207-218.

Ellis A. E., Burrows A. S. & Stapleton K. J. (1988b) Lack of relationship between virulence of Aeromonassalmonicida and the putative virulenee factors: A-iayer, extracellular proteases and extracellular haemolysin.Journal of Fish Diseases 11, 309-323.

EUis A. S., Stapleton K. J. & Hastings T. S. (1988e) The humoral immune response of rainbow trout (Salmogairdneri) immunised by various regimes and preparations of Aeromonas sahnonieida antigens. VeterinaryImmunology and Immunopaihology 19, 153 — 164.

Filip C , Fleteher G., Wulff J. L. & Earhart F. (1973) Solubilization of the cytoplasmic membrane of Escherichiacoli by the ionic detergent sodium-lauryl sareosinate. Journal of Bacteriology 115, 717-722.

Fuller D. W,, Pilcher K. S. & Fr)'er J. L. (1977) A leueocytolytic factor isolated from cultures of Aeromonassalmonicida. Journal of the Fisheries Research Board of Canada 34. 1118—1125.

Fyfe L., Finley A. &. Coleman G. (1986) A study of the pathological effect of isolated Aeromonas salmonicidaextracellular protease on Atlantie salmon. Salmo salar L. Journal of Fish Diseases 9, 403—409.

Fyfe L., Coleman G. & Munro A. L. S. (1987) Identification of major common extraeellular proteins seeretedby Aeromonas salmonicida strains isolated from diseased fish. Applied and Environmental Microbiology53, 722-726.

Hastings T. S. & Elhs A. E. (1988) The humoral immune response of rainbow trout, Salmo gairdneri Richardson,and rabbits to Aeromonas salmonicida extraeellular produets. Journal of Fish Diseases I I , 147—160.

Hastings T. S. & Ellis A. E. (1990) Deteetion of antibodies induced in rainbow trout b)' different Aeromonassalmonicida vaccine preparations. Journal of Aquatic Animal Health 2, 135—140.

Hitchcock P. J. &. Brown T. (1983) Morphological heterogeneity among Salmonella lipopolysaecharide chemotypesin silver-stained polyaerylamide gels. Journal of Bacteriology 154, 269—277.

Johnson C. M., Tatner M. F. & Home M. T. (1985) Comparison of the surface properties of seven strains of afish pathogen, Aeromonas salmonicida. Journal of Fish Biology 27, 445-458.

Lee K.-K. & Ellis A. E. (1989) The quantitative relationship of lethahty between extraeellular proteaseand extraeellular haemolysiu of Aeromonas salmonicida in Atlantic salmon {Salmo salar L.). FEMSMicrobiology Letters 61. 127—132.

Munn C. B.. Ishiguro E. E.. Kay W. W. & Trust T. J. (1982) Role of surface eomponents in serum resistance tovirulent Aeromonas salmonicida. Infection and Immunity 36, 1069—1075.

Munro A. L. S., Hastings T. S., Ellis A. E. & Liversidge J. (1980) Studies on an ichthyotoxic materialprodueed extraecllularly by the furunculosis baeterium Aeromonas salmonicida. hr.Fish Disease, ThirdCOPRAQ-Scssion (ed. by W. Ahne). pp. 98-106. Springer-Verlag. Berlin.

Morrissey J. H. (1981) Silver stain for proteins in polyaerylamide gels: a modified procedure with enhaneeduniform sensitivity. Analytical Biochemistry 117, 307—310.

Nomura S., Fujino M., Yamakawa M. & Kawahara E. (1988) Purifieation and characterization of salmolysin, anextracellular hemolylic toxin from Aeromonas salmonicida. Journal of Bacteriology 170, 3694-3702.

452 \ ; Lund ct al.

Price N, C . Stevens L., Duncan O. « Snodgrass M. (1*789) Proteases secreted by strains of Aeromonassalmonicida. Journal of Fish Diseases 12, 223-232.

Sliccran B. ct Smith P. R. (19<S1) A second extracellular protcolytic activity associated with the fish pathogenAero)nonas sahnonieida. FEMS Microbiology Letters I I , 7 3 - 7 6 .

Tajima K.. Takahashi T.. Ezura Y. & Kimura T. (1984) Enzymatic properties of the purified extraeellularprotease of Aeromonas salmonicida, Ar-4(EFDL). Bulletin of the Japanese Society of Scientific Fisheries50, 145-150.

Titball R. W. & Munn C. B. (1981) Evidence for two haemolytie aetivities from Aeromonas salmonicida. FEMSMicrobiology Letters 12, 27-30.

Udey L. R. & Fryer J. L. (1978) Immunization of fish with bacterins of Aeromonas salmonicida. MarineFisheries Review 40. 12-17.