Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Dr H.J. Oakey, Department of Microbiology and Immunology, James Cook University, Townsville, QLD 4811, Australia (e-mail: email@example.com).
Some strains of Vibrio harveyi are known to be pathogenic for fish and many invertebrates including crustaceans. Despite their importance, their modes of virulence have yet to be fully elucidated. Here, we present a previously unreported bacteriophage extracted from a toxin-producing strain of V. harveyi isolated from moribund prawn larvae in tropical Australia. Classification into the family Myoviridae was based upon morphological characteristics (an icosahedral head, a neck/collar region and a sheathed rigid tail) and nucleic acid characteristics (double-stranded linear DNA). We have termed the bacteriophage VHML (Vibrio Harveyi Myovirus Like). VHML is a temperate bacteriophage that has a narrow host range and shows an apparent preference for V. harveyi above other vibrios (63 Vibrio isolates tested) and other genera (10 other genera were tested). The conventional methods for phage concentration and extraction of nucleic acids from phage particles were not efficient and the alternative methods that were used are discussed.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Vibrio harveyi is a Gram-negative bacterium naturally associated with tropical marine environments such as warm marine waters and the intestinal tracts of marine fauna. The majority of isolates appear non-pathogenic, although at least two strains have been shown to cause devastating disease to aquatic invertebrates in northern Australia and there have been other reports from south-east Asia. Reports of disease outbreaks include prawn (Penaeus spp.) hatcheries in the Philippines, Thailand, Indonesia and northern Queensland ( Suranyanto and Miriam 1986; Muir 1987; Lavilla-Pitogo et al. 1990 ; Ruangpan and Kitao 1991; Prayitno and Latchford 1995), spiny lobster (Panuliris homarus) in India ( Jawahar et al. 1996 ) and pearl oyster (Pinctata maximus) in western Australia ( Pass et al. 1987 ). The disease, termed luminous bacteriosis or simply vibriosis, has a high mortality rate and causes dramatic financial loss to marine aquaculturists.
Previous studies to differentiate virulent and apparent avirulent strains have so far showed little success. Pizzutto and Hirst (1995) showed V. harveyi to be a diverse species by genetic profiling and protein profiling. They concluded that there was no defined grouping that differentiated virulent strains and suggested that virulence was probably acquired by association with genetically mobile elements as opposed to an inherent property. Harris (1993) showed that there was no correlation between the presence of plasmids and virulence.
Harris and Owens (1999) isolated two differently sized proteins from the two virulent strains kept in this laboratory. These proteins were demonstrated to cause mortality both in Penaeus monodon and in mice and the two proteins had different LD50s. It was concluded that the two proteins were exotoxins and were the probable virulence factors for the respective strains of V. harveyi. Other researchers have also reported toxic extracellular proteins from strains of this organism known to be virulent to penaeids ( Liu and Lee 1999; Montero and Austin 1999) and virulence-related products such as siderophores have been identified ( Owens et al. 1996 ).
Ruangpan et al. (1999) observed a bacteriophage as well as bacterial cells in P. monodon tissues infected with V. harveyi but, to date, this phage has not been studied further. The current project is designed to investigate the potential role of bacteriophages in the virulence of a toxin-producing strain of V. harveyi isolated from moribund P. monodon larvae on a farm in northern Queensland, Australia. This article will describe the presence of a previously unreported bacteriophage which may have this potential.
Materials and methods
Bacteriophage extraction and concentration
Toxin-producing V. harveyi strain 642 was cultured in broth containing 5 g l−1 peptone, 1 g l−1 yeast extract and 33 g l−1 synthetic sea salts (Aquasonic, Ingleburn, NSW, Australia) (PYSS). The broth was inoculated from frozen bead cultures and incubated for 10–12 h on an orbital shaker at 28 °C. Aliquots of 100 µl were removed for transmission electron microscopy (TEM) and the remaining culture was induced to the lytic cycle by the addition of 30 ng ml−1 mitomycin C (Sigma-Aldrich, Castle Hill, NSW, Australia). Induced cultures were further incubated for 10–12 h on an orbital shaker at 28 °C. Aliquots of 100 µl were removed for TEM. Cultures were centrifuged at 5000 g for 10 min to pellet bacterial cells and the supernatant fluids were filtered through 0·45 µm membranes (Millipore, North Ryde, NSW, Australia).
Two methods of bacteriophage concentration were assessed: polyethylene glycol (PEG) precipitation and ultracentrifugation. The former was modified from Yamamoto et al. (1970) , where 10% (w/v) PEG 6000 was added to the filtered lysate and gently mixed to dissolve. The lysates were incubated at 4 °C for 60 min and the precipitated particles pelleted by spinning at 8000 g for 10 min. Pellets were resuspended in 0·01 original volume sterile SM buffer (5·8 g sodium chloride, 2 g magnesium sulphate, 100 mg gelatin, 50 ml 1 mol l−1 Tris (pH 7·5) and 945 ml distilled water). The ultracentrifugation method was modified from that of Prof. R. Finkelstein (School of Medicine, University of Missouri, USA; personal communication), where filtered extracts were ultracentrifuged at 200 000 g for 4 h. Pellets were resuspended in 0·01 original volume sterile SM buffer. Aliquots of 100 µl were removed for TEM and the remaining concentrates were stored at 4 °C.
Identification of bacterial host strains for purification of the bacteriophage
Bacterial strains were examined for the presence of ‘native’ prophages using a modification of the method described by Jiang and Paul (1998). Vibrio spp. were cultured in 20 ml aliquots of PYSS and Enterobacteriaceae were cultured in nutrient broth (Oxoid, Heidelberg, Victoria, NSW). All cultures were grown at optimal temperatures for a few hours until O.D.600 was approximately 0·2. The cultures were aseptically divided into two equal aliquots and one of each pair was treated with 30 ng ml−1 mitomycin C. Cultures were returned to optimal incubation temperatures and the O.D.600 recorded periodically up to 24 h. Cell lysis, as indicated by a relative decrease in O.D.600 compared with the untreated control, was taken to indicate the presence of ‘native’ prophages.
Those bacterial strains which were seen to have no ‘native’ prophages were challenged with the bacteriophage extracted from V. harveyi 642. Bacteria were cultured in 100-ml aliquots of appropriate growth medium and incubated as above until O.D.600 was approximately 0·2. The cultures were aseptically divided into four equal aliquots. Extracts of the toxin-producing strain (500 µl) were added to two of the aliquots and 500 µl SM buffer added to the other two. All tubes were returned to incubation for 2 h to allow potential infection to occur. All tubes with added extract and one of the ‘blank’ tubes were induced with 30 ng ml−1 mitomycin C and further incubated. Optical density readings were taken periodically through the experiment to a maximum of 30 h post-induction. A decrease in the O.D.600 which was not evident in either the control or the induced control was taken as a presumptive infection from the 642 phage extract. Where infection was suspected, the culture was subjected to extraction as above using the ultracentrifuge method and concentrates analysed using TEM (FX2000; Jeol, MA, USA) with a 1% phosphotungstate stain. All TEM analyses were repeated with ‘blind’ samples, including placebos, by Dr R. Webb (Centre for Microscopy and Microanalysis, University of Queensland) (Jeol 1010) using a 1% ammonium molybdate stain. A sample of extract from V. harveyi 642 was also submitted to Prof. H.W. Ackermann (Department of Medical Biology, Laval University, Quebec, Canada) for further confirmation of the morphology, dimensions and classification.
Nucleic acid extraction from the isolated bacteriophage
Three methods were tested for the extraction of nucleic acid from phage particles. Firstly, a commercial kit for the extraction of DNA from Lambda phage (QIAGEN® Lambda Mini Kit; QIAGEN, Clifton Hill, Victoria, Australia) was applied. The kit was based upon PEG precipitation, sodium dodecyl sulphate (SDS) lysis and low-salt binding of DNA to a resin for washing prior to high-salt elution. The nucleic acid yield using this test could be described at best as ‘trace amounts’ and we attributed this to the salt content of the original growth medium (PYSS). The method was modified to include washing the ultracentrifuged pellet in SM and repelletting.
The second method was modified from that described by Su et al. (1998) . The phage was precipitated from the filtered growth medium with sterile 2 mol l−1 zinc chloride at a ratio of 1 : 50 for 5 min at 37 °C and pelletted at 4000 g for 5 min. The pellet was resuspended in TENS buffer (50 mmol l−1 Tris, pH 8·0, 100 mmol l−1 EDTA, 100 mmol l−1 NaCl and 0·3% SDS) and proteinase K (100 µg ml−1 final concentration) and incubated at 65 °C for 10 min. Proteins were removed by two phenol/chloroform/isoamyl alcohol extractions and DNA was precipitated with isopropanol. After washing in 70% ethanol, the pellet was resuspended in TE (10 mmol l−1 Tris, pH 8·0, 1 mmol l−1 EDTA). This method also appeared to be compromised by the presence of salts in the growth medium and was modified to include ultracentrifuge extraction and washing of the pellet prior to application.
The third method was based upon that described by Sambrook et al. (1989) . Bacteriophage from the concentrated solutions were lysed with the addition of EDTA (final concentration 20 mmol l−1), proteinase K (final concentration 50 µg ml−1) and SDS (final concentration 0·5%) and incubation at 56 °C for 1 h. The nucleic acid was purified using a phenol, a phenol/chloroform and a chloroform extraction. This method was modified whereby the phenol extractant was heated to ∼60 °C, added to the lysate, mixed by inversion, incubated at 60 °C for 30 min, centrifuged at 5000 g for 10 min to separate the phases and the aqueous phase removed for the next extraction. The final aqueous phase was dialysed overnight against TE.
For all three described methods, nucleic acid yield was estimated through agarose gel electrophoresis and comparison of ethidium bromide stain intensity with known DNA standards (Hind III cut lambda phage DNA; Promega, Annandale, NSW).
Nucleic acid characterization
The nucleic acid extracts were diluted to a standard concentration of ∼ 20 ng µl−1. Each extract was subjected to a digestion with Dnase I (Sigma Aldrich), Rnase A (Sigma Aldrich), S1 nuclease (Promega), Bal 31 exonuclease (Promega) and Hinf I (Promega). Each digestion was carried out upon ∼ 250 ng nucleic acid, except the digestions of phage extracted from JM109 (limited DNA yield led to there being ∼200 ng per test). All reactions were terminated using EDTA (10 mmol l−1 final concentration) and analysed using 0·8% agarose gel electrophoresis at 5 V cm−1.
Bacteriophage extraction and concentration from toxin-producing strains
Comparison between the uninduced and induced cultures of V. harveyi 642 using TEM analysis showed that the addition of mitomycin C caused cell lysis. This suggested that prophages were probably present and had switched to a lytic cycle. It was also noted that the bacterial culture visibly cleared over a 12–18-h period after the addition of the mutagen.
Two phage concentration methods were compared with respect to the resulting phage visible by TEM. Although both methods could be used to concentrate the bacteriophage, it was noted that the bacteriophage concentrated using PEG 6000 were more often broken into head and tail components and would obviously not be viable. In contrast, the concentrates obtained though ultracentrifugation contained a larger number of intact bacteriophage particles. The phage-like particles from V. harveyi 642 had an icosahedral-shaped head of 40–50 nm diameter and a sheathed rigid tail of 150–200 nm length. These particles also had a collar/neck zone between the capsid and the tail and, in some instances, an apparent collar structure was visible (see Fig. 1). A bushy complex of tail fibres at the external end of the sheath was observed when the bacteriophages were examined at higher magnifications (> 150 000×) ( Fig. 1).
Purification of single phage types by passage through other host strains
Seventy-six bacterial strains were tested for the presence of native prophages ( Table 1). Of these, 40 (52%) were found to be positive for prophage using optical density as an indicator ( Fig. 2). Of the 63 vibrio isolates tested, 30 (48%) had prophages. Of the 28 V. harveyi isolates tested, 13 (46%) had prophages. Almost all the isolates from prawn gastrointestinal tracts (AO isolates) had resident prophages.
Table 1. Results of analyses showing which strains of test bacteria have ‘native’ prophage, which have no prophage and which strains could be infected with the phage extracts from Vibrio harveyi 642
Strains with native prophage
Strains without native prophage
Strains which can be infected with 642 extract
All bacterial isolates except the AO isolates were obtained from the culture collection of the Department of Microbiology and Immunology, James Cook University. The unspeciated AO isolates were obtained from prawn gastrointestinal flora by Mr A. Oxley, (Department of Microbiology and Immunology, James Cook University).
8; 90-56015; 91-42182 : 1;
45; 20; 645; 12
93-48233 : 2; 664;
90-50605 : 4b; 45; 20;
92-46156 : 1; 90-50405 : 1;
92 : 53864 : 0;
90-55406; 52; PM91; 644;
92-47426 : 102; 650; 643;
656; 645; 648; 12; 301
Vibrio spp. (unidentified)
AO3; AO14; AO17; AO31;
AO9; AO10; AO20; AO27; AOB
AO40; AO41; AO43; AO46;
AO49; AO50; AO54
V. alginolyticus BRO3;
V. alginolyticus ACMM102
V. tubiashi ACMM;
ACMM667; ACMM 669;
V. aesturinus ACMM654;
RC2P666; TGH; ACMM220;
V. vulnificus ACMM76;
V. alginolyticus ACMM102;
V. pelagius ACMM675;
V. furnissi L15; A8;
V. carchariae ACMM117
V. tubiashi ATCC;
V. orientalis ATCC33934;
V. cinncinnati 3592;
V. natriegens ACMM127
Unidentified Aeromonas spp.
A. hydrophila unknown
isolate; A. sobria 10/93;
Plesiomonas spp. AO25;
Photobacterium spp. AO26
Proteus vulgaris ATCC
E. coli JM109
29905; Escherichia coli ATCC 25922;
ATCC 13048; E. coli JM109
Salmonella salford ACM
Pseudomonas nautica ACMM923
Ps. aeruginosa ATCC 27853
Of the 36 isolates observed to contain no native prophages, six could be infected with the phage from the V. harveyi 642 extract. These included four V. harveyi strains (45, 20, 645 and 12), a V. alginolyticus isolate (ACMM 102) and E. coli JM 109 ( Table 1). The numbers of phage particles observed from the V. alginolyuticus and E. coli strains by TEM were markedly lower than those observed from V. harveyi strains. All the phage particles observed from isolates infected with the V. harveyi 642 extract appeared morphologically similar — a tailed phage with an icosahedral head of approx. 40–50 nm diameter. The tail was rigid, partially sheathed and approx. 150 nm in length. A neck was evident between the head and the tail ( Fig. 1).
The mitomycin C-treated control tubes of the cultures which were able to be infected but had no phage extract addition showed no evidence of phage particles. This confirmed the observation that these strains had no resident prophage prior to treatment with V. harveyi 642 extracts.
Nucleic acid extraction
Three methods were examined for the extraction of nucleic acid from the phage concentrates. The QIAGEN® Lambda Mini Kit was not successful for the extraction of nucleic acid from these bacteriophage. This was attributed to salt in the growth medium inhibiting binding of DNA to the resin. The method was modified by washing the concentrated phage pellet to attempt to remove the salt. While the yield of nucleic acid increased slightly (from trace amounts to a maximum of approx. 5 ng µl−1), this method was finally discarded. The method described by Su et al. (1998) was also discarded as a result of low yields, ranging from 0 to approx. 5 ng µl−1, even after the inclusion of an additional washing step. The authors reported that the method was only successful in media containing < 0·1 mol l−1 sodium chloride and that DNA yield decreased at higher salt concentrations. The lack of success with additional washing suggested that the salts in the medium may have bound to or entered the viral particles.
The method described by Sambrook et al. (1989) produced nucleic acid from the phage particles at concentations of up to approx. 40 ng µl−1. However, the nucleic acid was considered unstable and impure as it could not be enzymatically digested in the described laboratory procedures and degraded after 24 h at 4 °C. The method was therefore modified to include the heating of the purification extractants and the resulting nucleic acid showed increased stability (up to 5 d at 4 °C) and was susceptible to experimental enzyme digestion.
Nucleic acid characterization
All phage nucleic acid samples were sensitive to Dnase I, Hinf I and Bal 31 but were resistant to digestion by Rnase A and S1 nuclease. It was concluded that all extracts contained phage with double-stranded linear DNA. In addition, the nucleic acid samples of phage isolated from V. harveyi 642 and V. harveyi strains 12, 20, 45 and 645 and V. alginolyticus 102 infected with phage from strain 642 showed identical restriction digest profiles following digestion with Hinf I, with fragments of approximately 2400, 1750, 1500, 1100, 950 and 700 bp (see Fig. 3). There was also a fragment > 10 kbp for some strains which may be attributed to uncut DNA. The yield from E. coli JM109 was sufficient for only partial characterization as the restriction enzyme digestion yielded a profile too faint to be determined.
Our results, showing 52% of the bacteria tested to be positive for prophage (Vibrio spp. 48%, V. harveyi 46%), concurred with those of Jiang and Paul (1998). They reported frequencies of lysogenic prophages in marine bacteria ranging from 25% to 62% depending upon geographical area (mean frequency of 40·2% among the areas studied).
The results of the current study indicate that V. harveyi 642 contained prophage. Upon induction, V. harveyi 642 cells produced phage particles with small icosahedral heads (approx. 40–50 nm diameter), defined collar/neck regions and sheathed tails of up to 200 nm length. These particles contained linear double-stranded DNA. According to Maniloff and Ackermann (1998) and the International Committee on the Taxonomy of Viruses (ICTV) report on viral taxonomy ( Murphy et al. 1995 ; http://life.anu.edu.au/viruses/ICTVdB, 6 December 1999), such particles would be assigned to the order Caudovirales. Owing to the presence of contractile tail sheaths, the particles would be further placed in the family Myoviridae. However, the heads of these phage were smaller than those previously reported for Myoviridae as 65–80 nm diameter ( Murphy et al. 1995 ) or more recently as 53–160 nm ( Ackermann 1999). The nearest sized myovirus capsid appears to be that of P2, reported to be 57 nm diameter ( Ackermann 1999). Other observations, such as the double-stranded linear DNA genome and the presence of a neck/collar region, support classification into the family Myoviridae.
Initial TEM micrographs did not indicate clearly if the cultures of V. harveyi 642 produced a single phage type. The extraction procedure did not eliminate cellular debris < 0·45 µm and therefore included such structures as bacterial flagella, which initially appeared similar to filamentous phages. Filamentous phages have recently been reported in vibrios, particularly V. cholerae ( Waldor and Mekalanos 1996; Jouravleva et al. 1998 ). Phage extracts were therefore used to treat a large group of potential host bacteria that did not contain native prophages. This procedure allowed for the purification of single phage types. This was based on the assumption that there was a low probability that multiple phage types would all infect the same range of hosts (i.e. that all the different hosts tested would have identical receptor recognition sites to V. harveyi 642). The results showed that treatment of six of the challenged bacteria with the V. harveyi 642 extract resulted in the production of morphologically identical phage. Although probable flagella were often seen in addition to phage particles, no infected culture showed the presence of flagella-like structures alone. We concluded that the V. harveyi 642 extracts contained a single phage type which we have labelled VHML (Vibrio Harveyi Myovirus-Like). This conclusion was supported by the identical restriction digestion patterns obtained from all the phage nucleic acid extracts.
It was concluded from these results that VHML has a narrow host range with a preference for certain strains of V. harveyi. The common factor among strain 642, the four V. harveyi strains which could be infected and the single V. alginolyticus isolate is unknown, but could reasonably be hypothesized to be a common receptor protein. The E. coli strain JM109 was included as a potential host simply because of its common availability and usage in research laboratories. The ability of VHML to infect these cells, and no other non-vibrio tested, could not be explained. While JM109 are commonly used for cloning experiments, such as those using a Lambda phage vector, these cells usually require treatment with maltose in order for the receptor proteins to be expressed. No such treatment was used in the experiments described herein.
The dimensions of VHML prevented further identification of this bacteriophage. It has been reported that tail lengths within bacteriophage species are largely conserved ( Ackermann 1999) although the Myoviridae genus ‘P2-like’ shows a propensity for abnormal tail length (Prof. H.W. Ackermann, personal communication), as does VHML. It has also been reported that in ‘P2-like’ myoviruses the contracted sheath commonly becomes loose and slips down or off the tail tube (Prof. H.W. Ackermann, personal communication). This loosening/loss of tail sheath was also noted for VHML with a fairly high frequency. Evidence suggests therefore that VHML may be a member of the ‘P2-like’ genus. However, further studies with nucleic acid hybridization, sequence similarity and DNA replication mechanisms would be required before this could be fully determined.
There have been other reports of tailed vibriophages. For example, Reidl and Mekalanos (1995) reported K139, a myovirus with a 54-nm head and a 100-nm tail. Chakrabarti et al. (1993) described vibriophage D10, a myovirus with a 60-nm diameter head and a 100-nm tail. Kellogg et al. (1995) reported a myovirus vibriophage with head diameters up to 65 nm and tails up to 100 nm. De Paola et al. (1998) demonstrated the presence of Podoviridae and Myoviridae from V. vulnificus in oysters. These myoviruses had head diameters of approx. 100 nm and tails of approx. 360 nm Matsuzaki et al. (1998) reported a broad host range vibriophage KVP40, a myovirus of the T4-like genus with an elongated head of 140 × 70 nm. The only report of a vibriophage associated with V. harveyi is that of Ruangpan et al. (1999) who described a phage with ‘round-hexagonal’ heads of 60 nm diameter and ‘filamentous’ tails of 100 nm. It is not possible to further classify this virus from the report.
It is evident from this study and comparison with those reports cited above that VHML is a previously unreported bacteriophage. This is concluded from its smaller than usual capsid and a longer tail than previously described for any vibriophages (Prof. H.W. Ackermann, personal communication). Ackermann et al. (1978) recommend that the minimum requirement for reporting a new phage should be novel morphological markers and a description of host range, both of which have been achieved in this study. They also recommend that description of physicochemical properties are an advantage for further comparisons. Factors such as buoyant density have not been determined for VHML. However, the restriction fragment patterns obtained would be useful for such comparative purposes.
This work was funded by Aquaculture CRC Ltd and James Cook University. The authors would like to show their appreciation to Prof. Hans W. Ackermann (Department of Medical Biology, Laval University, Quebec, Canada) for his invaluable assistance.