Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 36

NOTE columnare associated with mortality of salmonids farmed in Chile: a case report of two outbreaks

R. Avendaño-Herrera1*, V. Gherardelli2, P. Olmos2, M. G. Godoy2, A. Heisinger3 and J. Fernández4

1 Universidad Andrés Bello, Facultad de Ciencias Básicas, Departamento de Ciencias Básica. Viña del Mar, Chile;2 Laboratorio Aquagestion, Puerto Mon, Chile;3Multiexport Foods, Puerto Mon, Chile;4 Instituto de Salud Pública, Santiago

Abstract Since 2007 when infectious salmon anaemia virus was reported in Atlantic salmon (Salmo salar), the Chilean aquaculture industry has been restructured and intensification of Pacific salmon (Oncorhynchus kisutch) and rainbow trout (Oncorhynchus mykiss) farming has resulted in growing problems of emergent freshwater pathogens. This work reports the isolation, identification and characterization of a group of Chilean Flavobacterium columnare isolates from recent episodes of outbreaks in Pacific salmon and rainbow trout farms in Rupanco Lake in Chile.

Flavobacterium columnare is the etiological agent at least 36 species of fish in several geographical of disease, one of the most predomi- areas have been identified as susceptible to F. nant bacterial diseases of freshwater fish. This columnare (Shoemaker et al., 2003). Columnaris Gram-negative bacterium infects wild, farmed disease generally causes characteristic white to and ornamental fish species worldwide and yellow erosion in the tegument, severe necrosis oVen causes remarkable financial and mate- of the gill and skin epithelium, and in some rial losses yearly for the fish farming industry instances it also can be isolated from moribund (Bader and Starliper, 2002; Wagner et al., 2002). fish, exhibiting lile or no gross clinical signs of In fact, infections are the second leading cause infection (Decostere et al., 1998; Olivares-Fuster of mortality in pond raised channel catfish (Ic- et al., 2007). talurus punctatus) in the south-eastern United States (Wagner et al., 2002; Durborrow et al., AVer Norway, Chile was the second major 1988). salmon producer during several years, espe- cially Atlantic salmon (Salmo salar), with an The species was first described by Davis (1922) in estimated production of approximately 400000 numerous freshwater fishes, and subsequently ton during 2007 (hp://www.salmonchile.cl),

* Corresponding author’s email: [email protected] or [email protected] Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 37 surprisingly no cases of columnaris disease were The mortality rate in Farm A was lower than noted in Chilean aquaculture. However, aVer the Farm B, but continuous during the course of the outbreaks of infectious salmon anaemia virus in outbreak (one month). The distance between the marine farmed Atlantic salmon in June of 2007 two farms is approximately 4 km. Both farms (Godoy et al., 2008), the Chilean aquaculture were using water from the Rupanco Lake in industry has restructured itself and currently is their fish cultures. The origin of all fish was showing an increased interest in the cultivation from eggs spawned in Chile. The affected fish of Pacific salmon (Oncorhynchus kisutch) and were farmed at a water temperature of 16 ± 1ºC rainbow trout (Oncorhynchus mykiss). When during the different outbreaks. compared to February 2009, current exports of rainbow trout have increased 36% (www. A total of 5 moribund fish from each disease ifop.cl). outbreak was transported to the laboratory facilities in plastic bags for pathological and The intensification of fish farming in freshwater bacteriology examination. Scrapings obtained environments has resulted in the emergence of from gills and skin lesions were used for micro- freshwater bacterial pathogens such as atypical scopic observation and microbiological analysis. (Godoy et al., 2010) and Ten slides (equivalent of 2 different samples more recently F. columnare. In this study, we per fish) were also stained using Gram stain. report recent outbreaks of columnaris disease Moribund fish were euthanized by an overdose in Pacific salmon and rainbow trout farms in a of benzocaine and subjected to post-mortem ex- freshwater lake in Chile and describe, for the amination. Internal analysis showed pale liver, first time, the characterization of F. columnare splenomegalia, swollen kidney that contained associated with mortality. white nodules, ascitis and enteritis. Samples from gills and kidney were streaked onto TYES In February 2010 mortalities occurred in young agar plates (tryptone yeast extract salts medium: rainbow trout (25–35 g) at a rearing site (Farm 0.4% tryptone, 0.05% yeast extract, 0.02% an- A) in Rupanco Lake located 112 km at north- hydrous calcium chloride, 0.05% magnesium west Osorno (X Región) of Chile. The external sulphate heptahydrate, pH 7.2) and incubated gross finding noted in affected fish included aerobically at 15ºC for up to 10 days. eroded fins and tail rot and pale gills with ac- cumulations of mucus of yellow color on the Microscopic examinations of wet mount and filaments. Gram stained smears from gills lesions of both fish species revealed the presence of many long Twenty five days later, another outbreak oc- rods, thin with the ability to flex. Yellow pig- curred in smolt of Pacific salmon in the same mented colonies inoculated from gill lesions area, but in another farm (Farm B). The size of of all rainbow trout samples grew on all the the affected fish ranged from 60 to 70 g and the TYES plates and were observed aVer 4 to 6 pathology observed was similar to the described days post-incubation. Pure cultures of yellow for rainbow trout. colonies were also obtained from both rainbow trout and Pacific salmon kidney samples. Three Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 38 yellow-pigmented bacterial isolates were taken tochrome oxidase and catalase tests. Colonies for identification using biochemical and genetic were flat with rhizoid edges and strongly adher- studies. The isolates LM-01-Fc and LM-03-Fc ent to the medium and color shiVed from yellow were recovered from rainbow trout gills and to pink in presence of 3% sodium hydroxide. All kidney, respectively, while the isolate LM-02- isolates contained a cell-wall-associated flexiru- Fc was obtained from Pacific salmon kidney. bin-type pigment (KOH method according to Then, total genomic DNA was extracted from Reichenbach, 1989), absorbed Congo red and each isolate in pure colonies using InstaGeneTM produced diffusible gelatin degrading enzyme. Matrix (Bio-Rad) according to the manufacturer Chondroitinase activity from the culture su- instructions. All DNAs were maintained at – pernatant of the three isolates (Stringer-Roth 20ºC until they were used for PCR reactions. et al., 2002) was detected only in the Pacific Stock cultures were maintained frozen at –80ºC salmon isolate. Interestingly, the production of in Criobilles tubes (AES Laboratories). chondroitinase has been related to the virulence of F. columnare due to it causes necrotic lesions Initially, the recovered isolates were presump- by degrading chondroitin in the extracellular tively diagnosed as members of the F. psy- matrix of the fish tissue (Stringer-Roth et al., chrophilum species. However, the three strains 2002; Suomalainen et al., 2006). were not recognized by the Urdaci et al., (1998) and Izumi et al., (2003) pair of primers, indicat- Other phenotypic features of these isolates were ing that the pigmented isolates were distinct performed using API ZYM (bioMérieux) strips, from F. psychrophilum. These and all other PCR according to the manufacturer’s instruction with assays described in this work were performed the exception of the incubation temperature, following the protocol described by each author which was fixed at 25°C. The type strain F. co- with the exception that PCR reactions were lumnare ATCC 23462T from the American Type carried out using the commercial kit PuReTaqTM Culture Collection was included for compara- Ready-To-GoTM PCR beads (GE Healthcare), tive purpose. The presence and activity of 19 which included all the reagents needed for enzymes in the API ZYM gallery showed that the PCR reactions (buffer, nucleotides and Taq all isolates were similar in number of detected DNA polymerase), with the exception of the enzymes and level of enzymatic activity pro- specific primers and DNA template. The reac- duced, i.e. the typical profile of the F. columnare tion mixtures were amplified in a Mastercycler reference strain. personal (Eppendorf) apparatus. As for the PCR assays, similar finding was observed when In particular, none of the enzymes involved bacterial colonies were tentatively identified in the metabolism of carbohydrates could be by an indirect fluorescent antibody tests (IFAT, detected, similar to the reports of Michel et al., BiosChile), confirming that all Chilean isolates (2002) and Austin and Austin (2007). were not F. psychrophilum. Antimicrobial tests were applied by disc diffu- Biochemically, the isolated were Gram- sion method on dilute versions of Mueller-Hin- negative, non-motile, and positive for the cy- ton medium as recommended by the Clinical Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 39

Laboratory Standards Institute (CLSI, 2006) for (CLSI, 2006). Despite the fact that the majority use with F. columnare and F. psychrophilum. The of the drugs tested were effective in vitro, fish reference strain A. salmonicida subsp. salmonicida mortality was reduced following oral treatments ATCC 33658 was grown on Mueller–Hinton with florfenicol (15 mg kg-1 fish for 15 days). agar and used as control. All Chilean strains presented an identical antimicrobial suscep- Based on the phenotypic and biochemical char- tibility paern. They were highly susceptible acteristics, the bacteria recovered from necrop- to amoxicillin (AMX, 25 μg), enrofloxacine sied fish were consistent with characteristics of (ENR, 5 μg), florfenicol (FFC, 30 μg) and ox- F. columnare. To confirm this observation, the ytetracycline (OT, 30 μg) with mean inhibition specific PCR-analysis described by Welker et al., zones sizes ranging from 50 to 60 mm, while (2005) to identify the intergenic spacer region the oxolinic acid (AO, 2 μg) and trimethoprim- (ISR) between 16S–23S rRNA of F. columnare was sulphamethoxazole (SXT [1.25 μg/23.75 μg]; 25 employed, the bacterial strains were capable of μg) produced significantly smaller inhibition amplifying a single fragment of approximately zones (34 and 14 mm, respectively). These drugs 550 bp (Figure 1). PCR with the same primers were selected because they are authorized for yielded a nearly identical product paern for use in Chilean aquaculture. The susceptibility the reference strain ATCC 23462T, allowing us paern of the control strain ATCC 33658 was to confirm the assignment of all strains to the within the acceptable limits given in M42-A F. columnare species. These results are similar

MW 1 2 3 4 5 MW

800 550 400

Figure 1. Specific PCR products amplified with FCISRFL and FCISRR1 primer set. Lanes: MW: 100 bp DNA ladder (100-1000 bp, Bioron); 1, negative control (no DNA); 2, Flavobacterium columnare ATCC 23462T; 3, LM- 01-Fc; 4, LM-02-Fc and 5, LM-03-Fc. Numbers on the leV indicate the position of molecular size marker in bp. Numbers on the right indicate the size of the specific amplicons in bp. Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 40 to those previously described by Welker et al. phic DNA (RAPD) assays and the restriction (2005). fragment length polymorphic (RFLP) analysis. Both methods have been successfully used to On the other hand, the rainbow trout LM-01- fingerprint the F. columnare isolates (Triyanto Fc and Pacific salmon LM-02-Fc isolates were and Wakabayashi 1999; Michel et al., 2002; analyzed by 16S rRNA gene sequences using Arias et al., 2004; Thomas-Jinu and Goodwin the universal primer pair pA (5’–AGAGTTT- 2004; Darwish and Ismaiel 2005). The RAPD GATCCTGGCTCAG–3’) and pH (5’–AAG- reactions were performed using Ready-To-Go GAGGTGATCCAGCCGCA–3’) (Edwards et RAPD analysis beads (GE Healthcare) as pre- al., 1989), and PCR products were purified viously described by Avendaño-Herrera et al., using QIAquick PCR purification kits (Qiagen). (2004) and Thomas-Jinu and Goodwin (2004). The 16S rRNA gene sequence was determined These commercial beads have been optimized directly using the PCR-amplified DNA as a for RAPD-PCR reactions and, as with the PCR sequencing template on an ABI PRISM 3130 se- beads above, contained buffer, nucleotides and quencer (Applied Biosystems) according to the Taq DNA polymerase; the only reagents added manufacturer’s recommendations. The resulting to the reaction were template DNA (1 μl), 100 sequences were compared to those available in pmol of respective RAPD primers (supplied in Genbank, EMBL (hp://www.ebi.ac.uk) and the kit) and water to make the reaction up to a the Ribosomal Database Project using the Blast volume of 25 μl. Before the RAPD technique was program. Sequencing analysis of the nearly applied to the three isolates of F. columnare, the complete 16S rDNA gene (1436 bp) revealed six distinct random 10-mer primers (included that all the Chilean F. columnare isolates studied in the kit) were screened with purified genomic were identical, and that the sequence obtained DNA of the strain ATCC 23462T. Similarly to showed 99.9% similarity with F. columnare ATCC that reported previously by Thomas-Jinu and 49512. The next nearest isolate was F. columnare Goodwin 2004), all six primers amplified seg- ATCC 49513 with a similarity of 98.8% (data ments, but only P1 (5’-d[GGTGCGGGAA]-3’) not shown). Two nucleotide sequences for 16S and P5 (5’-d[AACGCGCAAC]- 3’) generated rRNA generated in this study were deposited in an appropriate paern of amplified products the GenBank database with accession numbers suitable for accurate analysis. Then, they were HM235501 and HM235502. selected for further studies with all strains. The paerns obtained with primer 1 showed a The process of typing is essential for recogniz- similar profile among all the strains with seven ing disease outbreaks, investigating the cross- major amplification bands ranging in size from transmission of pathogens, studying geographi- 100 to 1000 bp and, therefore, a unique group cal and host distributions of possible variant could be established (Figure 2A). When the of a specific pathogen, detecting particularly analysis was repeated with primer 5 a differ- virulent strains and monitoring vaccination ent RAPD paern was obtained between the programmes (Olive and Bean, 1999). Genetic Chilean isolates and the F. columnare type strain characterization of the Chilean isolates was (Figure 2B). performed using randomly amplified polymor- Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 41

MW A B C D E F MW 3000

1000

500

200

Figure 2. RAPD fingerprinting obtained for Flavobacterium columnare strains with P1 (A) and P5 (B). Lanes: MW, GeneRulerTM 100bp DNA Ladder Plus (100-3000 bp; Fermentas); A: negative control (no DNA); B, Positive control (included in RAPD kit); C, ATCC 23462T; D, LM-01-Fc; E, LM-03-Fc and F, LM-02-Fc. Numbers on the leV indicate the position of molecular size marker in bp.

In the case of RFLP, the amplicons of the 16S necessary to demonstrate that the Chilean iso- rDNA were separately digested with two re- lates do cause the observed disease and to fulfill striction endonucleases, HaeIII and RsaI (New Koch’s postulates. England Biolabs), according to the instruction of the manufacturer. Independent cleavage using Typically, columnaris disease is not spontaneous these enzymes gave one restriction paern (data (Shos and Starliper, 1999) and the outbreaks not shown). Thus, in accordance with the results are generally associated with poor environ- obtained the Chilean F. columnare isolates and mental conditions, such as organic load, high the type strain from Chinook salmon corre- temperature, high fish densities and handling sponded to gemovar I, as previously reported stress (Decostere et al., 1999; Plumb, 1999). Until by Arias et al., (2004) for the strain ATCC 23463T. very recently, farms in Chile cultured Atlantic In fact, the RFLP paern was identical to the salmon; however, increasing Pacific salmon and profile reported by Triyanto and Wakabayashi rainbow trout culture, especially in Rupanco (1999). Lake, may have spurred the F. columnare out- breaks. Although the three F. columnare strains were associated with yellow necrotic gill lesions at To date, the origin of the columnaris disease the tips of the lamellae and with skin disease is unknown and its focus has been only in the in freshwater farmed rainbow trout and Pacific Rupanco Lake. Given that F. columnare is an salmon, challenge experiments with fish are environmental bacterium, it may gain access Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 42

Genetic fingerprinting of Flavobacterium to the host by horizontal transmission in water. columnare isolates from cultured fish. Journal Our RAPD and RFLP findings show that the of Applied Microbiology 97, 421-428. isolates studied are genetically homogeneous Austin B and Austin DA (2007). “Bacterial which seem to be strongly related to the water Fish Pathogens: Diseases of Farmed and source. Therefore, one question is whether these Wild Fish”. Springer Praxis Publishing, F. columnare strains are endemic of the Rupanco Chichester, UK. ISBN: 9781402060687. water or the result of introduction of this patho- Avendaño-Herrera R, Rodrıíguez J, Magariños gen from other geographic areas. Moreover, the B, Romalde JL and Toranzo AE (2004). Intraspecific diversity of the marine possibility that the increased technology avail- fish pathogen Tenacibaculum maritimum able to detect this microorganism from cases as determined by randomly amplified categorized routinely as “normal mortality” polymorphic DNA-PCR. Journal of Applied together with the active programs of sanitary Microbiology 96, 871-877. monitoring cannot be ruled out. Bader JA and Starliper CE (2002). The genera Flavobacterium and Flexibacter. In “Molecular diagnosis of salmonid diseases” (C.O. To our knowledge, this is the first report and Cunningham, Ed.), pp 99-139. Kluwer characterization of F. columnare in Chilean Academic Publisher, Boston. ISBN: salmonids. All Chilean isolates studied were 9781402005060. biochemically and genetically similar to the F. Clinical and Laboratory Standards Institute columnare type strain. Due to the damages that (2006). “Methods for Antimicrobial it can cause, it can be considered as a disease Disk Susceptibility Testing of Bacteria isolated from Aquatic Animal, Approved with potential risk for the local culture of these Guidelines”. Clinical and Laboratory fish species, capable of causing great negative Standard Institute document M42-A. . ISBN: economic impact. Therefore, future studies are 1-56238-611-5. necessary to establish the clinical significance Darwish AM and Ismaiel AA (2005). Genetic of this bacterial species for the salmonid aqua- diversity of Flavobacterium columnare culture in Chile and, subsequently, to initiate examined by restriction fragment length polymorphism and sequencing of the 16S vaccine development. ribosomal RNA gene and the 16S-23S rDNA spacer. Molecular and Cellular Probe 19, 267- Acknowledements 274. Funding for this study was provided in part by Davis HS (1922). A new bacterial disease of Grant FONDECYT 1090054 from the Comisión fresh-water fishes. Bulletin of US Bureau of Nacional de Investigación Científica y Tecnológ- Fisheries 38, 261-280. ica–CONICYT (Chile) and also by Grant DI- Decostere A, Haesebrouck F and Devriese 01-10/R from the Universidad Andres Bello. LA (1998). Characterization of four Flavobacterium columnare (Flexibacter Thanks are expressed to G. Arancibia for his columnaris) strains isolated from tropical technical help. fish. Veterinary Microbiology 62, 35-45. Decostere A, Haesebrouck F, Turnbull JF and References Charlier G (1999). Influence of water quality Arias CR, Welker TL, Shoemaker CA, and temperature on adhesion of high and Abernathy JW and Klesius PH (2004). low virulence Flavobacterium columnare Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 43

strains to isolated gill arches. Journal of Fish Olivares-Fuster O, Baker JL, Terhune JS, Diseases 22, 1-22. Shoemaker CA, Klesius PH and Arias CR (2007). Host-specific association between Durborrow RM, Thune RL, Hawke JP Flavobacterium columnare genomovars and Camus AC (1988). “Columnaris and fish species. Systematic and Applied Disease - A Bacterial Infection Caused Microbiology 30, 624-633. by Flavobacterium columnare”. SRAC Publication, No. 479. hps://srac.tamu. Plumb JA (1999). “Health maintenance and edu/index.cfm/event/getFactSheet/ principal microbial diseases of cultured whichfactsheet/128/ fish”. Iowa State University Press, Ames, IA. ISBN: 978-0-8138-2298-3. Edwards U, Rogall T, Bloecker H, Emde M and Boeger EC (1989). Isolation and direct Reichenbach H. (1989). Order I. Cytophagales complete nucleotide determination of entire Leadbeer 1974. In “Bergey’s Manual genes. Characterization of a gene coding for of Systematic Bacteriology” (J.T. Staley, 16S ribosomal RNA. Nucleic Acids Research Bryant M.P., Pfennig N. and J.G. Holt, Eds.), 28, 1942–1946. vol. 3, pp. 2011–2013. Williams & Wilkins, Baltimore. ISBN: 978-0683079081. Godoy MG, Aedo A, Kibenge MJ, Groman DB, Yason CV, Grothusen H, Lisperguer A, Shoemaker CA, Klesius PH, Lim C and Yildirim Calbucura M, Avendaño F, Imilán M, Jarpa M (2003). Feed deprivation of channel M and Kibenge FS (2008). First detection, catfish, Ictalurus punctatus (Rafinesque), isolation and molecular characterization of influences organosomatic indices, infectious salmon anaemia virus associated chemical composition and susceptibility with clinical disease in farmed Atlantic to Flavobacterium columnare. Journal of Fish salmon (Salmo salar) in Chile. BMC Veterinary Diseases 26, 553-561. Research 4, 28. Shotts EB Jr and Starliper CE (1999). Godoy M, Gherardelli V, Heisinger A, Fernández Flavobacterial diseases: Columnaris disease, J, Olmos P, Ovalle L, Ilardi P and Avendaño- cold-water disease and bacterial gill disease. Herrera R (2010). First description of atypical In “Fish diseases and disorders, volume 3: furunculosis in freshwater farmed Atlantic viral, bacterial and fungal infection” (P.T.K. salmon, Salmo salar L., in Chile. Journal of Woo and D.W. Bruno, Eds.), pp. 560-576. Fish Diseases 33, 441-449. Cab. International. ISBN: 0851991947. Izumi S, Aranishi F and Wakabayashi H (2003). Stringer-Roth KM, Yunghans W and Caslake LF Genotyping of Flavobacterium psychrophilum (2002). Differences in chondroitin AC lyase using PCR-RFLP analysis. Diseases of Aquatic activity of Flavobacterium columnare isolates. Organisms 56, 207-214. Journal of Fish Diseases 25, 687-691. Michel C, Messiaen S and Bernardet J-F Suomalainen L-R, Tiirola M and Valtonen (2002). Muscle infections in imported ET (2006). Chondroitin AC lyase activity neon tetra, Paracheirodon innesi Myers: is related to virulence of fish pathogenic limited occurrence of microsporidia and Flavobacterium columnare. Journal of Fish predominance of severe forms of columnaris Diseases 29, 757-763. disease caused by an Asian gemovar of Thomas-Jinu S and Goodwin AE (2004). Flavobacterium columnare. Journal of Fish Morphological and genetic characteristics Diseases 25, 253-263. of Flavobacterium columnare isolates: Olive DM and Bean P (1999). Principles and correlations with virulence in fish. Journal applications of methods for DNA-based of Fish Diseases 27, 29-35. typing of microbial organisms. Journal of Triyanto A and Wakabayashi H (1999). Clinical Microbiology 37, 1661-1669. Genotypic diversity of strains of Bull. Eur. Ass. Fish Pathol., 31(1) 2011, 44

Flavobacterium columnare from diseased fishes. Fish Pathology 34, 65-71. Urdaci MC, Chakroun C, Faure D and Bernardet J-F (1998). Development of a polymerase chain reaction assay for identification and detection of the fish pathogen Flavobacterium psychrophilum. Research in Microbiology 149, 519-530. Wagner BA, Wise DJ, Khoo LH and Terhune JS (2002). The epidemiology of bacterial diseases in food-size channel catfish. Journal of Aquatic Animal Health 14, 263-72. Welker TL, Shoemaker CA, Arias CR and Klesius PH (2005). Transmission and detection of Flavobacterium columnare in channel catfish Ictalurus punctatus. Diseases of Aquatic Organisms 63, 129-138.