An improved enzyme-linked immunosorbent assay (ELISA) for detection of Flavobacterium psychrophilum isolated from salmon and rainbow trout

A double antibody sandwich enzyme-linked immunosorbent assay (ELISA) was developed in order to improve the diagnostic of the diseases caused by Flavobacterium psychrophilum in salmon and trout. With incubation steps of 1 hour at 37oC for coating antibodies and 30 minutes at 37oC for antigens and conjugates, the detection limit was in the range of 1 x 104 cells ml-1. Strains isolated from salmon and rainbow trout were successfully identified in an individual ELISA, which confirms the validity of the method for detection of F. psychrophilum serotypes. This biotin-avidin short protocol ELISA could be the technique of choice for seroepidemiological studies because it is a specific, sensitive, rapid and inexpensive procedure. Introduction Flavobacterium psychrophilum (syn. Cytophaga psychrophila, Flexibacter psychrophilus) remains a major cause of rainbow trout fry syndrome (RTFS) and bacterial cold water disease (BCWD) in wild and hatchery-reared salmonids, particularly in Europe, America and Japan (Austin & Austin, 1999). Additionally, in the past decade, the host range of this bacterium appears to have broadened, with several more non-salmonid fish species being affected (Amita et al., 2000). Currently, a laboratory definitive diagnosis of this infectious disease requires the classic isolation of the bacterium on agar media followed by the taxonomical analysis which may take several days to yield results (Lorenzen et al., 1997). Since there is a need for more rapid diagnosis, methods for detection of antigens could help to obviate the shortcomings of conventional methods. However, the specificity and sensitivity of these serological techniques largely described in the literature is determined by the methodology of the immunoassays and varied depending on the antigen and antiserum used (Crump et al., 2001). Recently, a microtiter agglutination method and an ELISA technique have been developed for detection of F. psychrophilum serotypes (Lorenzen & Olesen, 1997; Izumi & Wakabayashi, 1999). However, these researchers reported the inability of the methods to differentiate between F. psychr ophilum serotypes unless cross-absorbed antisera and “O” antigens were used. Bull. Eur. Ass. Fish Pathol., 21(5) 2001, 196 Thus, here we report the development of a double antibody sandwich short protocol ELISA based on whole-cell antigens and unabsorbed polyclonal rabbit antisera. The method was improved by the use of a biotinavidin system and optimised as simple, routine and inexpensive diagnostic tool for the rapid (four hours or less), specific and sensitive detection of F. psychr ophilum serotypes. Materials and Methods Bacterial strains and growth conditions F. psychrophilum strains included in the study were isolated from salmon (Oncorhynchus kisutch) (NCIMB 1947; NCIMB 2282) and rainbow trout (O. mykiss) (PT 4.1; NCIMB 13384). Other reference bacterial strains from the National Collection of Industrial and Marine Bacteria (NCIMB, Aberdeen, Scotland) and belonging to different species of genus Flavobacterium (F. columnare NCIMB 2281; F. succinicans NCIMB 2277) as well as Flexibacter maritimus (NCIMB 2158) and Aeromonas salmonicida (ATCC 14174) were used to test the specificity of the assay. Bacteria were routinely cultured in modified Anacker and Ordal agar medium (MAOA) (Flavobacterium strains) (Toranzo & Barja, 1993), tryptone soy agar (TSA, Oxoid) (Aeromonas salmonicida isolate) and Flexibacter maritimus medium (FMM) (Pazos et al., 1996). The taxonomical position of the bacterial strains was confirmed using morphological, physiological and biochemical test and API systems as previously described (Pazos et al., 1993; Santos et al., 1992; Santos et al., 1993). Antigens Bacterial cells harvested from a 48-72 hours culture at 18oC on agar plates were pelleted and formalized in a 0.3% (v/v) saline buffer (phosphate buffered saline pH 7.4). Rabbit antibodies Polyclonal antisera for diagnostic purposes were produced against F. psychr ophilum strains NCIMB 1947 and PT 4.1 and named anti-NCIMB 1947 and anti-PT4.1, respectively. Two kilogram New Zealand white rabbits were immunised intravenously at three days intervals with 0.2, 0.4, 0.8 and 1.0 ml of the appropriate formol-fixed bacteria adjusted to 1 x 109 cells ml-1 (McFarland standard no 3). Two weeks later a booster inoculum of four injections (1.0 ml/dose) administered as before were given. Anaesthetised rabbits were bled by cardiac puncture eight days after the last injection. Sera recovered were aliquoted undiluted and frozen at -20oC until use. Preimmune sera were previously collected from the rabbits. Antibody levels were determined in 96-well round-bottomed microtiter plates (Corning, New York) by agglutination as described by Lorenzen & Olesen (1997). The immunoglobulin G fraction (IgG) from antisera was purified through a pre-packed protein A sepharose affinity chromatography column according to the manufacture’s protocol (HiTrapTM Protein A, Amersham Pharmacia Biotech AB, Sweden). Protein concentration of purified antibodies was determined by the method of Bradford (1976) using the commercial kit available from Bio-Rad (Bio-Rad Laboratories). Purified antibodies were kept at -20oC. Conjugates The IgG was biotinylated with a hydroxysuccinimide ester of biotin (Sigma Chemical Co., St. Louis, USA) as described by Bull. Eur. Ass. Fish Pathol., 21(5) 2001, 197 Tijssen (1987). Stock biotin labelled conjugates were stored at -20oC. Enzyme-linked immunosorbent assay The rapid biotin-avidin assay was performed according to the double antibody sandwich method (Rangdale & Way, 1995). Dilutions of the reagents (unlabelled coating antibodies, biotin conjugates and extravidin peroxidase) were determined in previous checkerboard titration with appropriate positive and negative controls. Different volumes, temperatures and incubation times were also evaluated. Working conditions were chosen that showed maximum reactivity of the positive and minimum reactivity of the negative specimen. The wells of flat-bottomed microtiter plates (Nunc-ImmunoTM Plate MaxiSorpTM Surface, Denmark) were sensitised with 100μl of each purified anti-F. psychrophilum serum diluted in 0.05M sodium carbonate-bicarbonate buffer pH 9.6. Plates were incubated for 1 hour at 37oC, washed three times with 0.01M phosphate buffered saline (PBS) pH 7.4 supplemented with 0.05% Tween 20 (PBS-T) and subsequently blocked with 100μl of PBS-T containing 1% bovine serum albumin (PBST-BSA) for 30 min at 37oC to minimise unspecific adsorption. For all the following steps, PBS-T-BSA was used as diluent, each well receiving 100μl and incubations were performed at 37oC for 30 min unless otherwise indicated. Once blocked, plates were washed three times with PBS-T before addition of antigen (formalin killed bacterial cells) and positive and negative controls. After sample incubation, plates were washed as above and diluted biotinylated conjugate of each antiserum (anti-NCIMB 1947 and anti-PT 4.1) were added. Following another incubation and washing step, the wells received ExtrAvidin-Horseradish Peroxidase (Sigma Co.) dilution. Unbound material was removed and plates developed with the substrate 0-Phenylenediamine Dihydrochloride (OPD, Sigma Co.) diluted in 0.05M phosphate-citrate buffer pH 5.0 containing 0.03% sodium perborate (Sigma Co.) and incubated for 15 min in the dark at room temperature. The reaction was stopped by the addition of 1M H2SO4. Absorbance values were immediately read with an ELISA microplate reader (Bio-Rad Model 550) at 492 nm. Results were considered positive if absorbance exceeded at least two standard deviations above the mean of negative controls. Results and Discussion The agglutinating titer of the diagnostic antisera obtained in this study ranged from 1/ 128 to 1/256 against the homologous antigen, similar to that described by other authors (Lorenzen & Olesen, 1997). The protein concentration of the IgG preparations ranged from 6.2 mg ml-1 (anti-NCIMB 1947) to 3.5 mg ml-1 (anti-PT 4.1). In the ELISA, the most consistent results (A492 nm value about 2.0 for the homologous antigen) were obtained using coating antibodies at 2.5 μg/ml, biotinylated IgG at 1/1000 and extravidin peroxidase at 1/5000 dilution. These findings were slightly higher than in another study (Rangdale & Way, 1995), in which an ELISA was designed for detection of F. psychrophilum directly from fish spleen tissue. Bull. Eur. Ass. Fish Pathol., 21(5) 2001, 198 The antisera produced in the present work were specific in ELISA for their homologous bacteria; thus, both salmon strains reacted strongly with anti-NCIMB 1947 but not with anti-PT 4.1, while both rainbow trout strains reacted strongly with anti-PT 4.1 but not with anti-NCIMB 1947. According to these results, the strains were divided into two groups: serotypes 1 and 2. The serotype 1 comprised NCIMB 1947 and NCIMB 2282. The serotype 2 comprised PT 4.1 and NCIMB 13384. In previous works performed with a microtiter agglutination method (Wakabayashi et al., 1994; Izumi & Wakabayashi, 1999), it was indicated that F. psychrophilum isolates from coho salmon and rainbow trout shared common antigen(s) and serotypes were only detected when absorbed antisera and the thermostable “O” antigen were used. Thus, the ELISA results described here compare favourably with the microtiter test, particularly because using whole-cell preparations distinguish 4 isolates into 2 serotypes without the need for absoption. This appears to be a new observation as all other workers (Wakabayashi et al., 1994; Lorenzen & Olesen, 1997; Izumi & Wakabayashi, 1999) have found that antibodies to F. psychrophilum have to be absorbed out to yield serotype-specific antibodies in both agglutination and ELISA methods. Differences in the immunization schedules used to raise rabbit antisera as well as in the concentrations of reagents used in the ELISA methodology could probably explain the results obtained in the present study. Specificity of the ELISA was also investigated by introducing other bacterial antigens into the assay. Thus, cross-reactions were not observed with other closely related Flavobacterium spp. bacteria, neither with Flexibacter maritimus and Aeromonas salmonicida. Sensitivity of this rapid ELISA was analysed with serial dilutions of F. psychrophilum wholecells. The standard curves obtained were very similar for anti-NCIMB 1947 and anti-PT 4.1. The ELISA was able to detect 1 x 104 cells of F. psychrophilum per ml. Although two sandwich ELISA were reported in earlier papers with an overnight incubation step for coating antibodies and 1 or 2 hours for the conjugates (Lorenzen & Olesen, 1997; Rangdale & Way, 1995), the present ELISA consists of one incubation step of 1 hour and three of 30 minutes each with the antigens, biotinylated IgGs and avidin-peroxidase. Moreover, the ELISA can still be adapted if precoated plates are used, in which case the assay takes about three hours. In short, several conditions make this ELISA procedure very attractive as a diagnostic option for F. psychrophilum diseases: 1) can be performed in most laboratories, 2) requires no specialised equipment, 3) visual interpretation of a positive or negative reaction is possible with a naked eye, 4) large numbers of samples can be examined in microtiter plates at the same time, 5) can provide results in a few hours, and 6) allows serotyping and epidemiological studies. In this sense, because the knowledge about serotype distribution of F. psychrophilum is still rare, types should be evaluated in every country in order to ensure an optimal formulation of future vaccines. The introduction of an ELISA system based on a pool of antiF. psychr ophilum sera reacting with serotypes Bull. Eur. Ass. Fish Pathol., 21(5) 2001, 199 most commonly isolated from salmon and rainbow trout might be soon investigated for practical and economical reasons. AcknowledgementsThis work was partially supported by grantsMAR95-1848 from Dirección General deEnseñanza Superior e Investigación y Ciencia(Ministerio de Educación y Ciencia, Spain)and XUGA20008B98 from Xunta de Galicia,Spain. ReferencesAmita, K., Hoshino, M., Honma, T. &Wakabayashi, H. (2000). An investigation onthe distribution of Flavobacteriumpsychrophilum in the Umikawa River. FishPathol. 35: 193-197. Austin B. & Austin A. (1999). Bacterial fishpathogens: disease of farmed and wild fish,3ed rev. ed. (Springer-Praxis series inaquaculture and fisheries). Springer-VerlagBerlin Heidelberg New York. Bradford, M.M. (1976). A rapid and sensitivemethod for the quantitation of microgramquantities of protein utilizing the principle ofprotein-dye binding. Anal. Biochem. 72: 248-254. Crump, E.M., Perry, M.B., Clouthier, S.C. &Kay, W.W. (2001). Antigenic characterizationof the fish pathogen Flavobacteriumpsychrophilum. Appl. Environ. Microbiol. 67:750-759. Izumi, S. & Wakabayashi, H. (1999). Furtherstudy on serotyping of Flavobacteriumpsychrophilum. Fish Pathol. 34: 89-90. Lorenzen, E., Dalsgaard, I. & Bernardet, J-F.(1997). Characterization of isolates of Flavo-bacterium psychrophilum associated withcoldwater disease or rainbow trout fry syn-drome I: phenotypic and genomic studies.Dis. Aquat. Org. 31: 197-208.Lorenzen, E. & Olesen, N.J. (1997). Characteri-zation of isolates of Flavobacteriumpsychrophilum associated with coldwater dis-ease or rainbow trout fry syndrome II: sero-logical studies. Dis. Aquat. Org. 31: 209-220. Pazos, F., Santos, Y., Macías, A.R., Núñez, S.& Toranzo, A.E. (1996). Evaluation of mediafor the successful culture of Flexibactermaritimus. J. Fish Dis. 19: 193-197. Pazos, F., Santos, Y., Núñez, S. & Toranzo, A.E.(1993). Increasing occurrence of Flexibactermaritimus in the marine aquaculture of Spain.FHS News Lett. (Fish Health Section) 21: 1-2. Rangdale, R.E. & Way, K. (1995). Rapid iden-tification of C. psychrophila from infectedspleen tissue using an enzyme-linkedimmunosorbent assay (ELISA). Bull. Eur. Ass.Fish Pathol. 15: 213-216. Santos, Y., Huntly, P.J., Turnbull, A. & Hast-ings, T.S. (1992). Isolation of Cytophagapsychrophila (Flexibacter psychrophilus) in asso-ciation with rainbow trout mortality in theUnited Kingdom. Bull. Eur. Ass. Fish Pathol.12: 209-210. Santos, Y., Romalde, J.L., Bandín, I.,Magariños, B., Núñez, S., Barja, J.L. &Toranzo, A.E. (1993). Usefulness of the API-20E system for the identification of bacterialfish pathogens. Aquaculture 116: 111-120. Tijssen, P. (1987). Practice and theory of en-zyme immunoassays. In: Laboratory Tech-niques in Biochemistry and Molecular Biol-ogy (Burdon, R.H. and van Knippenberg, P.H.,Eds.), Vol. 15. Amsterdam, Elsevier. Toranzo, A.E. & Barja, J.L. (1993). Fry mortal-ity syndrome (FMS) in Spain. Isolation of thecausative bacterium Flexibacter psychrophilus.Bull. Eur. Ass. Fish Pathol. 13: 30-32. Wakabayashi, H., Toyama, T. & Iida, T. (1994).A study on serotyping of Cytophagapsychrophila isolated from fishes in Japan. FishPathol. 29: 101-104.