Isolation of lytic bacteriophage against Vibrio harveyi


  • C. Crothers-Stomps,

    1.  School of Veterinary and Biomedical Sciences, James Cook University, Townsville, Qld, Australia
    2.  The Australian Institute of Marine Science, PMB No. 3, Townsville, Qld, Australia
    3.  AIMS@JCU, James Cook University, Townsville, Qld, Australia
    Search for more papers by this author
  • L. Høj,

    1.  The Australian Institute of Marine Science, PMB No. 3, Townsville, Qld, Australia
    2.  AIMS@JCU, James Cook University, Townsville, Qld, Australia
    Search for more papers by this author
  • D.G. Bourne,

    1.  The Australian Institute of Marine Science, PMB No. 3, Townsville, Qld, Australia
    2.  AIMS@JCU, James Cook University, Townsville, Qld, Australia
    Search for more papers by this author
  • M.R. Hall,

    1.  The Australian Institute of Marine Science, PMB No. 3, Townsville, Qld, Australia
    2.  AIMS@JCU, James Cook University, Townsville, Qld, Australia
    Search for more papers by this author
  • L. Owens

    1.  School of Veterinary and Biomedical Sciences, James Cook University, Townsville, Qld, Australia
    2.  AIMS@JCU, James Cook University, Townsville, Qld, Australia
    Search for more papers by this author

Leigh Owens, School of Veterinary and Biomedical Sciences and AIMS@JCU, James Cook University, Townsville, Qld 4811, Australia. E-mail:


Aims:  The isolation of lytic bacteriophage of Vibrio harveyi with potential for phage therapy of bacterial pathogens of phyllosoma larvae from the tropical rock lobster Panulirus ornatus.

Methods and Results:  Water samples from discharge channels and grow-out ponds of a prawn farm in northeastern Australia were enriched for 24 h in a broth containing four V. harveyi strains. The bacteriophage-enriched filtrates were spotted onto bacterial lawns demonstrating that the bacteriophage host range for the samples included strains of V. harveyi, Vibrio campbellii, Vibrio rotiferianus, Vibrio parahaemolyticus and Vibrio proteolyticus. Bacteriophage were isolated from eight enriched samples through triple plaque purification. The host range of purified phage included V. harveyi, V. campbellii, V. rotiferianus and V. parahaemolyticus. Transmission electron microscope examination revealed that six purified phage belonged to the family Siphoviridae, whilst two belonged to the family Myoviridae. The Myoviridae appeared to induce bacteriocin production in a limited number of host bacterial strains, suggesting that they were lysogenic rather than lytic. A purified Siphoviridae phage could delay the entry of a broth culture of V. harveyi strain 12 into exponential growth, but could not prevent the overall growth of the bacterial strain.

Conclusions:  Bacteriophage with lytic activity against V. harveyi were isolated from prawn farm samples. Purified phage of the family Siphoviridae had a clear lytic ability and no apparent transducing properties, indicating they are appropriate for phage therapy. Phage resistance is potentially a major constraint to the use of phage therapy in aquaculture as bacteria are not completely eliminated.

Significance and Impact of the Study:  Phage therapy is emerging as a potential antibacterial agent that can be used to control pathogenic bacteria in aquaculture systems. The development of phage therapy for aquaculture requires initial isolation and determination of the bacteriophage host range, with subsequent creation of suitable phage cocktails.


Treatment of bacterial infection by bacteriophage (phage), commonly referred to as phage therapy, has potential for use to control disease in aquaculture systems pending the isolation and identification of phage that specifically kill relevant pathogenic bacteria (Nakai and Park 2002; Park and Nakai 2003; Skurnik et al. 2007). Phage are bacterial viruses that are extremely abundant in nature and are believed to be important in controlling bacterial populations in natural systems (Carlton 1999; Imbeault et al. 2006). Advantageous properties of phage as therapeutic agents include self-replication, which results in increased concentrations as infection persists, and the narrow host range of phage, which prevents harm to beneficial, naturally occurring microflora (Mathur et al. 2003).

Previous studies have successfully demonstrated the application of phage therapy to treat pathogenic bacterial infections in aquaculture systems (Nakai et al. 1999; Nakai and Park 2002; Imbeault et al. 2006; Vinod et al. 2006; Karunasagar et al. 2007). For example, Vinod et al. (2006) isolated phage with lytic activity against Vibrio harveyi and applied phage therapy to luminous vibriosis to improve the survival of Penaeus monodon larvae in India.

Vibrio harveyi has been established as a pathogen of phyllosoma of the tropical rock lobster, Panulirus ornatus (Bourne et al. 2006). Commercial scale, closed life cycle production of tropical rock lobster is not currently possible because of periodic mass mortalities of early-stage hatchery-reared phyllosoma (Bourne et al. 2004, 2006; Payne et al. 2007). Control of the microbial community is therefore considered an essential component to commercial scale culture of P. ornatus phyllosoma (Bourne et al. 2007).

This study aimed to develop and demonstrate methods for the isolation of phage with lytic activity against V. harveyi with the potential for application in phage therapy for aquaculture. The effectiveness of the isolated phage was assessed against a broad collection of Vibrio species and strains, including isolates derived from larval rearing systems and moribund specimens of P. ornatus phyllosoma.

Materials and methods

Bacterial strains

A total of 75 Vibrio species and strains were used, consisting primarily of V. harveyi along with near phylogenetic neighbours Vibrio campbellii, Vibrio rotiferianus, Vibrio proteolyticus and Vibrio parahaemolyticus. Strains were cultured on agar plates of 22 g l−1 nutrient agar with 20 g l−l marine salts (NAMS) at 28°C for 48 h. Bacteria were stored long term in 15% glycerol in a −70°C freezer. Strains were examined for purity on NAMS, thiocitrate bile salts sucrose agar (TCBS) and Gram stains from the NAMS colonies. Working cultures were taken from the TCBS isolates.

Phage isolation

Water samples (70 ml) were obtained by dipping sterile containers into discharge channels and grow-out ponds of a prawn farm in northeastern Queensland, Australia. A total of 21 70 ml environmental samples (E-01–E-21) were collected from the top 30 cm of the water column of discharge channels and the grow-out ponds. Enrichment broths were produced using a combination of 1 ml of broth culture of each of four V. harveyi strains (12, 20, 45, 645) that were cultured in 13 g l−1 nutrient broth with 20 g l−1 marine salt (NBMS) at 28°C for 24 h. These strains were previously identified as not containing endogenous phage (Oakey and Owens 2000). Each water sample (1 ml) was enriched in independent enrichment broths (4 ml) for 16 h at 28°C.

Bacterial-free supernatants were produced by centrifugation (10 000 g for 15 min; Eppendorf centrifuge 5415 D; Eppendorf, Hamburg, Germany) followed by filtration (0·45 μm) of 1·5 ml aliquots of the enriched water samples. The supernatants (10 μl) were spotted onto NAMS plates and grown at 28°C for 24 h to ensure the absence of bacteria. The bacteria-free supernatants were stored at 4°C.

Determination of phage host range

Lawns of the trial bacteria (n = 75) were produced by pipetting (1 ml) bacterial broth cultures in NBMS (24 h at 28°C) onto NAMS plates, swirling the broth to cover the entire plate and pipetting off excess. Plates were dried in a biohazard cabinet at room temperature for 30 min. This method provided an even covering of bacteria producing consistent lawns.

Each bacteria-free supernatant (10 μl) was pipetted onto lawns of each of the 75 trial bacterial strains to form distinct spots. This method allowed up to 21 phage filtrates and a control spot of NBMS broth to be screened on each lawn. Spotted lawns were incubated at 28°C for 24 h and then inspected for zones of clearing.

A modified soft top agar overlay method was used to confirm the presence of phage through plaque formation. For each phage filtrate, a tenfold dilution series (to 10−10) was prepared, and for each dilution, an aliquot (1 ml) was mixed with 9 ml of soft top agar with 2% marine salts (6 g l−1 agar; 13 g l−1 nutrient broth; 20 g l−l marine salts). These mixtures were poured over lawns of V. harveyi strain 12 that had been cultured for 8 h at 28°C. Plates were then incubated for a further 24 h at 28°C before examination for plaques.

Phage purification

Eight bacteria-free supernatants produced from the enriched water samples (E-01, E-02, E-04, E-06, E-19, E-20 and E-21) were selected for further investigation. The selection criteria included having a broad host range, differing infection profiles to other samples and the ability to infect bacterial strains that other selected samples could not. Vibrio harveyi strain 12 was used for the enrichment and purification of samples because it was susceptible to all eight selected enriched water samples; was isolated from sea water rather than from a diseased animal and did not possess endogenous lysogenic phage (Oakey and Owens 2000).

Phage purification of the eight selected bacteria-free supernatants was carried out through threefold plaque purification. The purified phage were named VhCCS-01, VhCCS-02, VhCCS-04, VhCCS-06, VhCCS-17, VhCCS-19, VhCCS-20 and VhCCS-21, with the number corresponding to the sample of origin. The isolated phage filtrates were rescreened to ascertain their new host range using the same protocols as described earlier.

Concentration of phage and transmission electron microscopy (TEM)

Each purified phage filtrate (30 ml) was subjected to ultracentrifugation at 200 000 g for 4 h (Optima L-90K ultracentrifuge, TY 50.2 Ti rotor; Beckman Coulter, Fullerton, CA). Phage pellets were resuspended in 0·01 of the original volume of sterile sodium/magnesium (SM) diluent (NaCl, 5·8 g; MgSO4·7H2O, 2 g; Tris pH 7·5, 50 ml; gelatin, 0·1 g; dH2O, 950 ml).

Aliquots (150 μl) of each of the concentrated filtrates and a control sample of SM diluent were analysed by TEM using negative staining on BUTVAR (polyvinyl butyral resin)-coated gold grids using 2% potassium phosphotungstate stain (pH 6·5).

The effect of phage VhCCS-06 on broth cultures of Vibrio harveyi strain 12

Nine 100 ml cultures of V. harveyi strain 12 were produced by inoculation in NBMS broth and incubation at 28°C. Three treatments (all preformed in triplicate) were tested: addition of phage VhCCS-06 (1 ml) 2 h after inoculation, addition of phage VhCCS-06 (1 ml) 6 h after inoculation and a control with no phage added. Spectrophotometric absorbance measurements (570 nm, Jenway 6300; Barloword Scientific, Dunmow, UK) provided an indication of bacterial growth in each treatment every 2 h over 24 h.

Colony-forming unit (CFU) counts were used to measure viable cells in each treatment. An aliquot (1 ml) was taken from each of the nine broths 4, 6, 12, 18 and 24 h after inoculation and serially diluted into artificial sea water (ASW) (100 μl into 900 μl of ASW, 33 g l−1 marine salt). Each dilution (100 μl) (10−1–10−7) was spread onto replicate TCBS agar plates and incubated at 28°C for 48 h. Vibrio harveyi colonies were counted on dilution plates yielding counts between 30 and 300 cells.

Statistical analysis was performed on CFU data using the program package spss (ver. 14.0; SPSS Inc., Chicago, IL). A nonparametric Kruskal–Wallis test was performed as transformed data remained skewed from a normal distribution. A significant difference was recorded at P < 0·01.


Bacteriophage enrichments

Enrichment of water samples in combination broths of V. harveyi was performed to increase concentrations of phage with lytic activity against this potentially pathogenic species. All enriched samples produced zones of clearing on trial bacterial lawns of some V. harveyi strains, and some enrichments also produced clearing zones for V. campbellii, V. rotiferianus, V. proteolyticus and V. parahaemolyticus strains. Individual enriched samples produced zones of clearing on between 16% and 92% of the 75 host lawns (Fig. 1). Each enriched sample showed a different infection profile, indicating that a diverse array of phage was present in the samples. Most zones of clearing showed distinct edges, however, a small number displayed halos. Control spots showed no zones of clearing.

Figure 1.

 Percentage of the library of bacterial strains cleared by enriched sample, bacteria-free supernatants.

Bacteriophage isolation

Phage were successfully isolated from 8 of 21 enriched samples after threefold plaque purification. The presence of phage was conclusively confirmed by TEM examination of concentrated phage filtrates (Figs 2 and 3). All eight isolated phage displayed binary symmetry and a long tail. Phage VhCCS-01, VhCCS-02, VhCCS-04, VhCCS-06, VhCCS-17 and VhCCS-20 had head diameters of c. 60 nm and tails with no neck/collar region (Fig. 2). In contrast, phage VhCCS-19 and VhCCS-21 had head diameters of c. 100 nm, and their tails displayed a distinct neck/collar region (Fig. 3).

Figure 2.

 Transmission electron microscope of phage VhCCS-01, VhCCS-02, VhCCS-04, VhCCS-06, VhCCS-17 and VhCCS-20. Phage displayed binary symmetry, a head diameter of c. 60 nm with long tails (scale bars = 50 nm).

Figure 3.

 Transmission electron microscope of VhCCS-19 and VhCCS-21. Displaying phage with binary symmetry, heads with a diameter of c. 100 nm and a long tail with a neck/collar region (scale bars = 50 nm).

The plaque-purified phage cleared bacterial strains of V. harveyi, V. campbellii, V. rotiferianus and V. parahaemolyticus, but not V. proteolyticus. All isolated phage produced distinct clearing zones for strains of V. harveyi and V. rotiferianus that were susceptible. For susceptible V. campbellii strains, the two phage morphotypes caused clearing zones of different appearance, with VhCCS-19 and VhCCS-21 producing halos whilst the other purified phage produced distinct clearing zones. Only two phage isolates (VhCCS-02 and VhCCS-20) caused clearing zones for the tested V. parahaemolyticus strain. Phage VhCCS-06 was the overall most efficient isolate, clearing 60% of the trialled bacterial strains (Fig. 4).

Figure 4.

 Percentage of the 75 bacterial strains cleared by the triple, plaque-purified phage.

Effect of phage VhCCS-06 on broth cultures of Vibrio harveyi strain 12

The inoculation of a single phage strain at differing time intervals into a broth culture demonstrated its power to destroy target bacteria (Fig. 5). When phage VhCCS-06 was added early (2 h) to the V. harveyi broth culture, the exponential growth phase was delayed by 4 h and the culture reached a lower maximum density than the control. Addition of phage VhCCS-06 to the V. harveyi culture during the exponential growth phase (6 h) also caused a clear reduction in bacterial numbers. Counts of CFU supported the results obtained through spectrophotometry with significantly lower counts (Kruskal–Wallis test, P < 0·001) obtained after 24 h for cultures where phage VhCCS-06 was added at 2 h (7·3 × 107 CFU) or 6 h (1·2 × 107 CFU) when compared to the control (9·3 × 107 CFU).

Figure 5.

 Mean (± standard deviation) absorbance of broth cultures of Vibrio harveyi strain 12 with three treatments consisting of inoculation with VhCCS-06 at 2, 6 h and the control that was not inoculated with phage. (inline image) Control; (inline image) phage 2 h and (inline image) phage 6 h.


In this study, eight phage strains were isolated from prawn farm effluent after initial enrichment in broths seeded with V. harveyi. The isolated phage had a reduced host range compared to the enriched samples prior to carrying out phage purification. For example, phage VhCCS-06 cleared 60% of tested bacterial strains whilst its parent sample cleared 92% of the strains, indicating that the enriched samples contained multiple phage of different types.

TEM examination of eight presumptive phage isolates confirmed the presence of a single strain of phage in each (Figs 2 and 3). All the isolated phage were assigned to the order Caudovirales because they displayed binary symmetry and were tailed (Ackermann 2006). Phage VhCCS-01, VhCCS-02, VhCCS-04, VhCCS-06, VhCCS-17 and VhCCS-20 were classified as belonging to the family Siphoviridae because of their binary symmetry, head diameter and long tails that lacked a neck/collar region (Fig. 2) (Ackermann 2006). This is consistent with previous studies where phage with lytic activity against V. harveyi were classified into the family Siphoviridae (Pasharawipas et al. 2005; Vinod et al. 2006; Karunasagar et al. 2007; Shivu et al. 2007). Previous studies have indicated that Siphoviridae can control problematic V. harveyi populations causing disease in prawn hatcheries (Vinod et al. 2006), supporting the prospect that phage isolated in this study have potential to control V. harveyi in aquaculture systems. The host range of the purified Siphoviridae was relatively broad, including some but not all tested strains of the closely related species V. harveyi, V. campbellii and V. rotiferianus. For two of the Siphoviridae phage (VhCCS-02 and VhCCS-20), the host range also included V. parahaemolyticus, which is a more distantly related member of the Harveyi clade (Sawabe et al. 2007). In general, most phage have a restricted host range, often including a single species or subspecies, whereas other phage may infect more than one related species or even genus (d’Herelle 1917; Carlton 1999; Fuhrman 1999).

Phage VhCCS-19 and VhCCS-21 were classified into the family Myoviridae based on the morphological characteristics of binary symmetry and long tails with neck/collar regions (Fig. 3) (Ackermann 2006). This is consistent with a previous study by Oakey and Owens (2000) where another phage, termed VHML (Vibrio Harveyi Myo-like Virus), was isolated from V. harveyi and classified into the family Myoviridae. VHML was shown to be lysogenic and infer virulence to its host (Oakey and Owens 2000). In this study, the two Myoviridae phage (VhCCS-19 and VhCCS-21) produced halos rather than distinct clearing zones for a limited range of the tested bacterial isolates, indicating possible bacteriocin production. As the putative bacteriocin was not apparent for these bacterial isolates that had not been infected with VhCCS-19 or VhCCS-21, we hypothesize that these phage integrated as a prophage into the genome of the host bacterium and induced bacteriocin production. Because of this apparent lysogenic activity, we deem these strains to be unsuitable for use in phage therapy in aquaculture without comprehensive investigation into their transducing ability.

Investigation of actively growing broths of V. harveyi strain 12 inoculated with phage VhCCS-06 demonstrated that addition of phage reduced growth, presumably because of lytic activity of the phage. The experiment showed, however, that addition of phage before the onset of exponential bacterial growth (2 h) could delay but not prevent exponential growth. This effect was most likely because of multiplication of phage-resistant cells, which may appear when treating bacteria with phage (Tanji et al. 2005). The presence of phage-resistant cells were also indicated in the spotting assay, where small, nonspreading phage-resistant colonies were observed within clearing zones in a small number of the lawns. Phenotypic alteration in a bacterial strain exposed to phage with lytic activity has previously been reported (Scott et al. 2007). Altered phenotypes include bacterial resistance to infection by the lytic phage and inefficient colonization of the bacteria’s host which may reduce the bacteria’s virulence (Scott et al. 2007). These adaptations may be attributed to a change in selection pressures on the bacterium from virulence to phage resistance. Phage-resistant colonies are not necessarily still pathogenic as selection for resistance could select against virulence (Merril et al. 2006). Skurnik and Strauch (2006) showed that many spontaneous phage-resistant mutants are deficient or have alterations in the receptor required by the phage (Skurnik and Strauch 2006). These authors highlighted the possibility that a mutation that eliminates or alters a bacterial receptor that also functions as a virulence factor can attenuate the virulence of the bacterium. Further investigations are required to determine whether phage-resistant V. harveyi cells emerging during phage therapy have reduced ability to infect the aquaculture-reared animal or have reduced virulence.

The effectiveness of phage therapy decreases dramatically with the emergence of virulent phage-resistant cells. Phage cocktails and multiple doses of phage may assist in inhibiting phage-resistant cells, thus reducing such a problem (Tanji et al. 2005). Finding a large diversity and quantity of phage with lytic action against a pathogenic bacterial species strengthens the likelihood that an effective phage cocktail can be produced that acts against phage-resistant cells. Rational selection of phage for cocktails and the variation of cocktails have the potential to delay or inhibit the emergence of phage resistance (Tanji et al. 2005).

Cocktails of the phage found in this study are being tested for their ability to reduce bacterial-induced mortality in rock lobster phyllosoma, and these results will be reported elsewhere.


As multiple phage were found that cleared all the trialled bacterial strains, it is most likely that if vibriosis occurs in an aquaculture system, a phage with the ability to lyse the causative Vibrio spp. can be isolated using the techniques demonstrated in this study. There is a high probability that phage VhCCS-19 and VhCCS-21 are lysogenic and induce bacteriocin production in V. harveyi strain 12. The inoculation of the isolated phage clearly affects bacterial growth; however, phage resistance is potentially a major obstacle to the use of phage as therapeutic agents. The phage isolated in this study have displayed lytic activity against strains of V. harveyi that have been reported as primary pathogens of phyllosoma of the tropical rock lobster, P. ornatus. Therefore, these phage have potential for use as a biocontrol agent to combat vibriosis in the rearing system of phyllosoma of the tropical rock lobster, P. ornatus. Properly controlled efficacy studies are required to comprehensively demonstrate the effectiveness of phage preparations prior to their application as therapeutic agents in aquaculture.


This work was funded by James Cook University and AIMS@JCU. The authors show their deep appreciation to Howard Prior (Department of Primary Industry Queensland, Animal Research Institute, Yeerongpilly) for his invaluable assistance with TEM.