Agar culture of Piscirickettsia salmonis, a serious pathogen of farmed salmonid and marine fish


  • Editor: Jorge Crosa

Correspondence: Duncan J. Colquhoun, National Veterinary Institute Oslo, Ullevaalsveien 68, 0454, Olso. Tel.: +47 23 21 61 41; fax: +47 23 21 61 01; e-mail:


Piscirickettsia salmonis, a serious bacterial pathogen of farmed marine fish, previously considered culturable only in eukaryotic cell-culture systems, was grown for the first time on agar and broth containing enhanced levels of cysteine, thus greatly increasing the potential for isolation, in vitro culture and study of this organism. Virulence towards Atlantic salmon following passage on agar media was retained in a controlled laboratory trial. Of the studied temperatures, optimal growth on agar was observed at 22 °C.


Piscirickettsia salmonis, a fastidious, Gram-negative bacterium is the causative agent of piscirickettsiosis, or salmonid rickettsial septicaemia, a systemic infection of farmed salmonid and marine fish. Initial reports of salmonids in Chile from 1989 (Bravo & Campos, 1989), were followed by reports in Ireland (Rodger & Drinan, 1993), Norway (Olsen et al., 1997), and Eastern Canada (Cusack et al., 2002), and a P. salmonis-like bacterium has been reported from Atlantic salmon farmed in Tasmania (Corbeil et al., 2005). Reports of P. salmonis infections in marine fish include European seabass, Dicentrarchus labrax (McCarthy et al., 2005), and white sea bass Atractoscion nobilis in California (Arkush et al., 2005).

Diagnosis of piscirickettsiosis is currently based upon macroscopic and histological observation of typical pathological changes including observation of Gram-negative intracellular bacteria and confirmation by immunohistochemistry, PCR on DNA extracted from infected tissues, or by subjecting cell-cultured bacteria to a specific fluorescent antibody test (OIE, 2006). Culture of P. salmonis from infected fish in antibiotic-free chinook salmon-embryo (CHSE) cells is the accepted diagnostic ‘gold standard’, but as this technique is technically challenging, time-consuming and subject to contamination, it is not routinely performed. This means that the relatively few cultured isolates available comprise a very limited resource upon which our understanding of this bacterium is based. Although available culture techniques have recently expanded to include insect-cell lines (Birkbeck et al., 2004), the dependence on culture in eukaryotic cells means that for genetic and antigen related studies, relatively complicated procedures are required to separate eukaryotic DNA and cell debris from the bacterial components (Yuksel et al., 2001). Antibiotic-free cell cultures are also prone to Mycoplasma spp. infection (Thornton et al., 1999).

Culture of P. salmonis on cell-free media constitutes a significant leap forward for research on this bacterium. This paper describes the first agar culture of P. salmonis from an outbreak of piscirickettsiosis in Atlantic salmon in Norway during 2006, together with agar culture of genetically different P. salmonis isolates and the results of an initial phenotypic investigation of a Norwegian and two genetically different Chilean P. salmonis isolates.

Materials and methods

Histological and immunohistochemical investigation

Tissue samples from gills, heart, liver, intestine/pancreas, spleen, skin/muscle and kidney were fixed in buffered formalin and processed for histology using standard procedures. Immunohistochemistry was performed as described by Olsen et al. (1997).

Initial isolation

Isolate NVI 5692 was isolated in December 2006, from a fish with piscirickettsiosis on a farm in the county of Møre og Romsdal, western Norway. The disease had been previously diagnosed in the fish population by histological and immunohistochemical examination of formalin fixed tissue samples. The fish was euthanised at the farm and sent overnight on ice to the laboratory. Kidney and liver samples were processed for cell culture at 15 °C in antibiotic-free CHSE cells as is standard for culture of P. salmonis in our laboratory. Bacteriological samples from anterior kidney tissue were streaked out onto blood agar (BA), BA with 2% NaCl (BAS) and Cysteine Heart Agar (Bacto Heart Infusion broth 25 g L−1; Glucose 10 g L−1; l-cysteine 1 g L−1; agar 15 g L−1; haemoglobin 2 g L−1) supplemented with 5% ovine blood (CHAB). The agar plates were incubated for 14 days (examined every 48 h) at 22 °C (BA and CHAB) and 15 °C (BAS).

Following the initial isolation of isolate NVI 5692, agar cultivation on BA, BAS and CHAB was attempted from tissues archived from diagnostic cases, stored at −70 °C, and cell culture maintained isolates as described above. Isolate identification, geographical origin, and species of initial isolation are listed in Table 1.

Table 1.   CHAB cultivated isolates of Piscirickettsia salmonis
IsolateSpecies isolated fromGroup*CountryAlternative notation
  • *

    Based upon homologies of the ITS region to CHSE cultivated isolates obtained from GenBank.

NVI 5910Rainbow troutLF-89 likeChileP6
NVI 5890Atlantic salmonEM-90-likeChileP7
NVI 5889Atlantic salmonEM-90-likeChileP8
NVI 5786Atlantic salmonEM-90-likeChilePS40 LT5
NVI 5891Atlantic salmonAtlanticNorwayP41
NVI 5895Atlantic salmonAtlanticNorway2001-09-1213
NVI 5892Atlantic salmonAtlanticCanadaNSO811
NVI 5894Atlantic salmonAtlanticNorway2003-09-572
NVI 5692Atlantic salmonAtlanticNorway2006-09-669

Genetic characterization

16S rRNA and 16S–23S internal transcribed sequence (ITS) were amplified and sequenced using PCR and primers described by McCarthy et al. (2005) and Suau et al. (1999). A DYEnamic ET dye terminator cycle sequencing kit and a MEGABACE 1000 capillary sequencer (Amersham Biosciences, NJ) were used. Contiguous sequences were assembled, aligned and compared using Vector NTI (Invitrogen).

Serological characterisation

An indirect fluorescence antibody test (IFAT) utilizing anti-P. salmonis sera (kindly donated by Dr John Fryer, University of Oregon), was performed as described by Lannan et al. (1991), using Francisella philomiragia ssp. noatunensis (Olsen et al., 2006; Mikalsen et al., 2007) and Moritella viscosa as negative control bacteria.

Phenotypical testing

Oxidative/fermentative production of acid from glucose was performed in a basal glucose medium (Difco). Catalase activity was tested by flooding cells on an object glass with 3% H2O2. Production of H2S was detected by taping a lead acetate impregnated paper strip (Sigma) to the inside of the lid of a CHAB plate incubated at 22 °C. Cytochrome oxidase and indole production were tested using Dryslide tests (Becton Dickinson ML,) and enzymatic activity was examined using the APIzym test with 24 h incubation at 22 °C (Biomerieux, Marcy-l'Etoile, France). The results of phenotypical testing for all tested isolates are summarised in Table 3. Antibiotic resistance was tested by the disc-diffusion method on CHAB plates incubated at 22 °C for 3 days. Antibiograms for oxolinic acid (10 μg), oxytetracycline (80 μg), florfenicol (30 μg), streptomycin (100 μg), erythromycin (78 μg), kanamycin (100 μg) and penicillin (5 μg), were established (Table 2). β-Lactamase production was tested using the clover-leaf test (Gots, 1945). CHAB cultures were incubated aerobically at 4, 10, 15, 22 and 30 °C, and under anaerobic and microaerophilic (5% CO2) conditions at 22 °C (7 days).

Table 3.   Summary of phenotypical test results for isolates NVI 5692, NVI 5910 and NVI 5786
TestNVI 5692 (Atlantic)NVI 5910 (LF-89 like)NVI 5786 (EM-90 like)
Oxidative/fermentative production. of acid from glucose−/−−/−−/−
Cytochrome oxidase
β-Lactamase production (Clover-leaf test)++(W)+
Indirect fluorescent antibody test+Not doneNot done
H2S production+++
APIzym testEnzymeNVI 5692P6 (LF-89 like)PS40LT5 (EM-90 like)
  • *

    Very weakly positive on one of four replicate tests

  • W: weakly positive.

1Negative control
2Alkaline phosphatase*W+
3Esterase (C4)+W+
4Esterase Lipase (C8)+W+
5Lipase (C14)WWW
6Leucine arylamidaseWW+
7Valine arylamidase
8Cystine arylamidase
11Acid phosphatase++
13α- Galactosidase
14β- Galactosidase
15β- Glucoronidase
16α- Glucosidase
17β- Glucosidase
19α- mannosidase
20α- fucosidase
Table 2.   Disc diffusion inhibition zones (mm) for Piscirickettsia salmonis
AntibioticNVI 5692
NVI 5910
(LF-89 like)
NVI 5786
(EM 90-like)
Oxolinic acid495460

Infectious challenge

To test the effect of agar plate isolation and culture on virulence, an experiment was performed at VESO Vikan, Namsos, in which Atlantic salmon (Salmo salar) held in seawater at 12 °C, were injected intraperitoneally with bacteria passaged five times on CHAB. Ten fish of c. 200 g in weight were injected with c. 3 × 107 CFU, 10 with 3 × 105 , and 10 with 3 × 103 CFU. A cohabitant group of 10 fish injected with 0.9% saline was also included. The fish were observed for 30 days and dead or moribund fish were removed daily. Macroscopical changes were registered and samples taken for histology and bacteriology (CHAB and BA).

Results and discussion

Macroscopical and histological examination of fish from the primary disease outbreak and from the subsequent infectious challenge trial revealed pathology consistent with P. salmonis infection. Findings included multifocal liver and kidney necrosis. Aggregates/microcolonies of immunohistochemically positive bacteria were seen in the cytoplasm of hepatocytes and in the heart, kidney and spleen.

Bacterial investigation of the primary sampling revealed prolific numbers of a pin-point colony type on CHAB culture following 4 days of incubation, while no growth was observed on BA and only a very sparse mixed bacterial flora was observed on BAS. By Day 6 the colonies on CHAB had reached c. 1 mm in diameter. Primary CHSE cell-cultures were discontinued after several days due to non- P. salmonis bacterial overgrowth. Following subculture of the cysteine dependent bacterium (now awarded our laboratory reference NVI 5692) for 4/5 days incubation at 22 °C on CHAB, the colonies could be described as c. 1 mm in diameter, slightly convex, grey–white, shiny, centrally opaque with translucent, slightly undulating margins. Subculture of isolate 5692 on BA and BAS for 14 days did not produce visible colonies. Similar growth was obtained for the cell culture maintained isolates and tissue from diagnostic cases on CHAB media. Following Gram-staining, the CHAB cultivated cells appeared as Gram-negative irregular cocci, of varying size, which were nonmotile under phase contrast microscopy.

DNA sequencing confirmed CHAB cultivated isolates to be P. salmonis. Alignment of 16S and ITS fragments gave sequences of 420 and 432 bp, respectively, and identified the presence of three genotypes of P. salmonis. Based upon homologies in the ITS region these groups were termed Atlantic (Norwegian and Canadian), EM-90-like and LF-89-like. All sequences displayed over 99% identity with equivalent sequences from P. salmonis in GenBank. Further confirmation of the identity of isolate 5692 was provided by a strongly positive P. salmonis specific IFAT test.

Of the aerobic incubation temperatures tested, no growth was observed at 4 or 30 °C while growth (but not individual colonies) were visible at 10 °C after 5 days incubation. Optimal growth was observed at 22 °C. Microaerophilic culture conditions were found to retard growth. Cysteine heart broth, with and without 5% sheep blood was also found to support growth of isolate 5692, when incubated at 22 °C with gentle shaking. As reported previously (Fryer & Hedrick, 2003), while appearing sensitive to the majority of antibacterial agents tested, a degree of resistance to penicillin was identified. β-Lactamase production was detected using the clover-leaf test in all tested isolates (Table 2). APIzym testing (Table 3.) revealed similar but not identical profiles for Chilean and Norwegian isolates.

In the infectious challenge trial, mortalities were directly correlated to bacterial dose, with the heaviest and earliest mortalities occurring in the highest dose group (Fig. 1). The bacterium was reisolated and the identity confirmed by phenotypical testing, thus fulfilling Koch's criteria, proving that the agar cultivated P. salmonis are pathogenic for Atlantic salmon. Further, successful agar culture of three different genotypes of P. salmonis and reisolation of NVI 5786 (PS40-LT5) following passage through Atlantic salmon confirmed that the ability to grow on agar was not peculiar to isolate NVI 5692.

Figure 1.

 Cumulative mortality following intraperitoneal injection of agar cultivated Piscirickettsia salmonis strain NVI 5692.

In conclusion, following nearly 20 years of culture in eukaryotic cells, agar and broth culture of P. salmonis have for the first time been demonstrated. It has also been shown that agar isolation and culture of P. salmonis does not lead to a sudden drastic loss of virulence. While P. salmonis appears to have a requirement for cysteine, a feature common in several other intracellular bacteria, limited growth was observed on subculture on blood plates. The ability to culture P. salmonis on an agar medium should allow simplified collection of strains, even, as is often the case, in the presence of secondary opportunistic invading bacteria. Isolation in laboratory culture media will also provide a platform for simplified culture and preparation of cells for genetic studies, antisera production for serotyping and diagnostic use, as well as for monitoring antibiotic resistance and production of cells for vaccine development. Diagnosis of disease outbreaks need no longer be confirmed indirectly, but by direct isolation of the bacterium. Although not compared in this study, the rates and overall levels of growth on agar and in broth indicate that growth probably occurs faster and to higher densities on agar and in broth culture than that achieved in cellculture.