Association of a bacteriophage with virulence in Vibrio harveyi

The virulence of Vibrio harveyi, which is a serious pathogen of penaeids (Karunasagar, Pai, Malathi & Karunasagar 1994; Pizzuto & Hirst 1995; Alvarez, Austin, Alvarez & Reyes 1998) and finfish (Kraxberger-Beatty, McGarey, Grier & Lim 1990; Ishimaru & Muroga 1997), has been associated with possession of double haemolysin genes (Zhang, Meaden & Austin 2001). The study seeks to investigate a possible relationship between virulence and the previously described bacteriophage of V. harveyi (Oakey & Owens 2000). The bacteriophage, which has been determined to have an icosahedral head and rigid tail and to contain double stranded linear DNA, has been presumptively assigned to the genus Myovirus (Oakey & Owens 2000).

Details of the bacterial cultures used are included in Table 1. Authenticity was verified after Pedersen, Verdonck, Austin, Austin, Blanch, Grimont, Jofre, Koblavi, Larsen, Tiainen, Vigneulle & Swings (1998). Working cultures were maintained on plates of tryptone soya agar (Oxoid, Basingstoke, UK) supplemented with 1% (w/v) NaCl (=TNA) at room temperature with subculturing every 2–3weeks. Stock cultures were maintained in 15% (v/v) glycerol at −70 °C.

A bacteriophage culture, VHML, was acquired from Dr J. Oakey (James Cook University, Townsville, Queensland, Australia) and stored in 15% (v/v) glycerol at −70 °C. The bacteriophage was propagated in V. harveyi VIB 645. Thus, 10-mL volumes of tryptone soya broth (Oxoid) supplemented with 1% (w/v) NaCl (=TNB) were inoculated with 50 μL each of bacterial culture and bacteriophage preparation with incubation on a shaker overnight at 25 °C. Thereafter, 50 ng mL⁻¹ of mitomycin C (Sigma, Basingstoke, UK) was added with further incubation on the shaker at 25 °C, overnight. Then, the cells were removed by centrifugation (3000 g for 15 min at 4 °C), and the supernatant filtered successively through 0.8-, 0.45- and 0.22-μm pore size Millipore (Edinburgh, UK) Millex porosity filters. This preparation, which contained the bacteriophage, was stored at −20 °C until required. For use, the preparation was thawed at room temperature, and mixed with the bacterial cultures in the ratio of one part of bacteriophage and 100 parts of bacterial culture in TNB. This was incubated overnight at 25 °C and used as inoculum for blood agar and for infectivity experiments.

Haemolytic activity was recorded from Columbia agar base (Oxoid) supplemented with 1% (w/v) NaCl and 1% (v/v) of rainbow trout blood (collected freshly by venepuncture from healthy, disease-free rainbow trout) following incubation at 25 °C for 5 days.

Bacterial cultures, with and without bacteriophage, were grown overnight at 25 °C in TNB, centrifuged at 3000 g for 15 min, and the cells resuspended in 10 mL of 0.9% (w/v) saline to ∼10⁷ cells mL⁻¹ using a haemocytometer slide (Improved Neubauer type; Merck, Lutterworth, UK) and a magnification of 400× on a Carl Zeiss (Welwyn Garden City, UK) Axiophot microscope. Confirmation was obtained from total viable counts on TNA after spreading 0.2-mL volumes over the surface of duplicate plates of TNA with incubation at room temperature for 5 days.

Groups of 10 Atlantic salmon, Salmo salar L. (average weight = 22 g), from quarantined stock recognized as disease-free (after Austin & Austin 1989), were infected by intraperitoneal injection with 0.1-mL volumes of the bacterial suspensions to achieve a dose of 10⁶ cells per fish. The infected animals were maintained for up to 14 days in covered polypropylene tanks supplied with dechlorinated aerated static fresh water (50% of the volume was changed daily) at 14 °C. Dead and moribund fish were removed and examined bacteriologically and pathologically (after Austin & Austin 1989). Any survivors after 14 days were sacrificed and examined, as above. Disease signs were recorded and the pathogen identified after Pedersen et al. (1998). Mortalities were considered to be caused by the culture only if it was recovered as dense pure culture growth from the internal organs of the freshly dead or moribund fish.

A second pathogenicity model involved use of the crustacean, Artemia. Disease-free cysts (Waterlife, Longford, UK) were hatched in sea water at 28 °C for 48 h. Then, groups of 20 nauplii were transferred to 50-mL volumes of sea water in conical flasks. These microcosms were seeded with 1.0-mL volumes of the bacterial suspensions with or without the bacteriophage. The conical flasks were covered with tinfoil, incubated at 28 °C, and examined daily for 3 days. Dead nauplii were removed promptly for bacteriological examination. The effects of bacterial cells on the hatching of Artemia cysts were evaluated separately. Thus, groups of 20 cysts were transferred to 50-mL volumes of sea water in conical flasks to which were added 1.0-mL volumes of bacterial suspensions, with or without bacteriophage. The flasks were covered with tinfoil and incubated at 28 °C, with examination at daily intervals for 3 days to determine the level of hatching and the presence of any dead nauplii. These were removed for bacteriological examination, as before.

The presence of the bacteriophage enhanced haemolytic activity of the bacterial cultures (Table 2). Thus, with V. harveyi VIB 571 and VIB 648, haemolysis on trout blood agar was recorded in the presence, but not in the absence of bacteriophage (Table 2). With V. harveyi EO22, VIB 295 and VIB 645, the zones of haemolysis on trout blood agar were certainly larger when bacteriophage was present. In contrast, a control using 50-μL drops of bacteriophage suspension did not produce any evidence of haemolysis.

The presence of bacteriophage led to enhanced mortalities in Atlantic salmon (Table 3). Challenge with V. harveyi EO22, VIB 295 or VIB 391 did not result in any mortalities within 14 days. Moreover, at the end of the experiment, the fish did not display any overt signs of disease. However, the presence of bacteriophage led to substantial mortalities over 14 days (Table 3). Dead fish displayed swollen abdomen, ascites, swollen liquefying kidney and muscle necrosis. With V. harveyi VIB 571, VIB 645 and VIB 648, the presence of bacteriophage led to 100% mortality among the Atlantic salmon within 14 days (Table 3). Also, the addition of bacteriophage influenced hatching of Artemia cysts. Thus, with V. harveyi EO22, VIB 295, VIB 391, VIB 571 and VIB 645, there was a reduction in hatching over 72 h (Table 3). Furthermore, with V. harveyi EO22, VIB 295, VIB 571, VIB 645 and VIB 648, the presence of bacteriophage led to increased mortalities in Artemia nauplii (Table 3).

Previous work revealed that V. harveyi VIB 571 and VIB 645 were the most virulent isolates of V. harveyi to salmonids (Pedersen et al. 1998; Zhang & Austin 2000); most of the other isolates examined being either weakly virulent or non-pathogenic. Subsequently, it was determined that the high virulence was associated with possession of duplicate haemolysin genes (Zhang et al. 2001). However, infection of V. harveyi with the previously described bacteriophage (Oakey & Owens 2000) resulted in enhanced haemolytic activity to rainbow trout blood and increased pathogenicity to Atlantic salmon and Artemia. Thus, the bacteriophage was clearly exerting a synergistic effect on V. harveyi, enhancing the pathogenicity process. In this, there is a parallel with some other pathogens, e.g. V. cholerae and Corynebacterium diphtheriae, for which virulence is associated with the presence of bacteriophage (Rajadhyaksha & Rao 1965; Waldor & Mekalanos 1996). It remains for further work to determine the precise molecular mechanism of the involvement of bacteriophage with virulence.