Isolation and characterization of soilborne virulent bacteriophages infecting the pathogen Rhodococcus equi

Authors


Correspondence

Sophie Foley, School of Life, Sport & Social Sciences, Edinburgh Napier University, Edinburgh EH11 4BN, UK.

E-mail: s.foley@napier.ac.uk

Abstract

Aims

To isolate and characterize a diversity of bacteriophages (phages) that infect the soilborne pathogen Rhodococcus equi.

Methods and Results

Twenty-seven phages were isolated from soil samples from geographically distinct locations using a range of R. equi bacterial strains, including clinical isolates. On the basis of host range, genomic DNA restriction profiles and virion protein profiles, the diversity of these phages was extensive, with phages being divided into 16 groupings.

Conclusions

Based on a range of criteria, these phages could be divided into 16 distinct groupings. The majority of the phages recovered from soil were Siphoviridae, adding to the limited number of Siphoviridae described to date for R. equi. One grouping consisted of phages belonging to the Myoviridae.

Significance and Impact of the Study

This represents the first study looking at the diversity of phages infecting the pathogen R. equi, including the first Myoviridae to be isolated and characterized for the genus Rhodococcus and for the nonmycobacterial actinomycetes. Given their diverse host range, including clinical isolates, this collection of phages offers the potential for the development of phage cocktails for use as a therapeutic or alternatively in the biocontrol of this pathogen in reservoirs of infection relating to animal husbandry.

Introduction

Rhodococcus equi, a member of the mycolic acid–containing actinomycetes, is a soil saprophyte (Barton and Hughes 1984) particularly prevalent in the farm environment. As a major cause of severe bronchopneumonia in foals, with often fatal consequences, it has a major financial impact on the equine industry (Giguère et al. 2011a,b). The main route of infection is through the inhalation of contaminated dust particles, and the titres of airborne R. equi and prevalence of infection are closely related (Muscatello et al. 2007). R. equi is also the causative agent of pyogranulomatous infection in other animals, and opportunistic infections with consequent high mortality are increasingly reported in immunocompromised individuals including HIV-infected patients, patients with cancer and solid-organ transplant recipients (Yamshchikov et al. 2010). Reflecting the intracellular lifestyle of R. equi, treatment is currently based on a lengthy course of antibiotic treatment with limited success (Yamshchikov et al. 2010; Giguère et al. 2011b), and intrinsic resistance to many antimicrobials has been reported (Letek et al. 2010). Treatment is also dependent on effective diagnosis, and incidentally, there are reports of misidentification as mycobacteria (Macías et al. 1997) or diphtheroid (Tuon et al. 2007). In many stud farms worldwide R. equi is endemic and, in the absence of an effective vaccine (Dawson et al. 2010), antibioprophylaxis is used to limit the occurrence of infection in foals (Giguère et al. 2011a,b). The development of alternative approaches to the prevention of infections, and for diagnosis and treatment, is therefore of paramount importance.

As the increase in antibiotic resistance across a broad number of bacterial pathogens continues to be a major cause of concern, a renewed interest in bacteriophages has emerged. For instance, phage-based control strategies are being developed for foodborne pathogens in domesticated animal and bird populations in which inoculations with intact phages result in a reduction in faecal shedding (Goodridge and Bisha 2011). Phages infecting R. equi have until recently been largely ignored; however, due to their lytic properties and host specificity, they offer a potential tool for the treatment of R. equi infections and the control of R. equi populations in particular environments. Studies to date have focussed on the genomic analysis of four R. equi phages isolated from soil by Summer et al. (2011) and of phage RRH1 isolated from activated sludge and capable of infecting five rhodococcal species including R. equi (Petrovski et al. 2011), the genomic and proteomic analysis of R. equi infecting phage E3 also isolated from soil (Salifu et al. 2013) and the demonstration of the antibacterial activity of several phage YF1 proteins against R. equi (Shibayama and Dabbs 2011). The observed reduction in bacterial loads of soil inoculated with R. equi phage DocB7 demonstrates the potential for application of these phages as an antimicrobial (Summer et al. 2011). Furthermore, preliminary studies have demonstrated an inhibitory and/or lytic activity against the bacterial host by specific R. equi phage-encoded proteins (Shibayama and Dabbs 2011; Salifu et al. 2013). When considering phages as therapeutic or biocontrol agents, particular consideration needs to be given to the diversity and host range of phages. Studies on other bacterial genera indicate the development of successful phage-based therapeutic approaches using a cocktail of phages with differing host range (Chan and Abedon 2012).

The objective of this study was therefore to isolate bacteriophages capable of infecting R. equi, including clinical isolates, from a range of soil samples and to undertake a preliminary characterization of the diversity of these phages.

Materials and methods

Phage isolation and propagation

Phages were isolated from soil samples following the method described by Dabbs (1998). Five grams of soil was added to 25 ml glucose yeast extract broth (GYE; 10 g glucose, 10 g yeast extract per litre) supplemented with 10 mmol l−1 CaCl2 and 10 mmol l−1 MgCl2 in a 100-ml conical flask. Flasks were shaken overnight at room temperature at 200 rev min−1 in order to facilitate dissociation of phages from the soil particles and additional decaying organic matter that may be present. The suspension was clarified by centrifugation (10 000 g, 10 min), filtered through a sterile 0·45-μm filter membrane (Nalgene, Thermo Fisher Scientific Inc., Waltham, MA, USA) and stored at 4°C. Soil samples were screened for phages infecting R. equi by the standard spot assay technique using a 3-ml molten GYE (0·7% agar) overlay to which 10 mmol l−1 CaCl2, 10 mmol l−1 MgCl2 and 500 μl of an overnight culture of host bacterium (Table 1) were added prior to pouring over the surface of a GYE agar (1·5%) plate. A volume (10 μl) of filter-sterilized soil extract was spotted on the surface of the solidified agar. Following 2 days of incubation at 27°C, the plates were examined for zones of lysis, which may be indicative of phage presence.

Table 1. Bacterial strains used for phage isolation and screening
Bacterial strainCharacteristicsSource
  1. NCIMB, National Collection of Industrial and Marine Bacteria, Aberdeen, UK; CVS, Prof Alexander & Lindsay, University of Cambridge Veterinary School; EVS, Dr Smith, University of Edinburgh Veterinary School; GVS, Dr D. Taylor, University of Glasgow Veterinary School.

Rhodococcus equi
NCIMB 10027Clinical isolate from equine, type strainNCIMB
103SClinical isolate from equine, type strainLetek et al. (2010)
CV1Clinical isolate from equineCVS
CV2Clinical isolate from equineCVS
CV3Clinical isolate from equineCVS
VI1Clinical isolate from equineEVS
GV1Clinical isolate from equineGVS
GV2Clinical isolate from equineGVS
Rhodococcus erythropolis
SQ1Environmental isolateQuan and Dabbs (1993)
NCIMB 11148Environmental isolate, type strainNCIMB

An enrichment step was also applied to soil samples in which aliquots (100 μl) of the filter-sterilized soil extracts (prepared as described above) were incubated at 27°C for 5 h with 100 μl of an overnight R. equi culture in 10 ml GYE broth supplemented with 10 mmol l−1 MgCl2 and 10 mmol l−1 CaCl2. The samples were then filter-sterilized and screened for phages using the spot assay technique.

Bacterial strains were confirmed as being R. equi on the basis of presence of CAMP-like reaction positive, choE and vapA as defined by PCR (Ladrón et al. 2003; M. Letek, personal communication) and MALDI-TOF analysis (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, personal communication).

Plaque assays and phage purification

A volume (100 μl) of soil extract or phage lysate, appropriately diluted in quarter strength Ringers solution, was combined with 10 mmol l−1 MgCl2 and 10 mmol l−1 CaCl2, and 500 μl of an overnight culture of R. equi. This was incubated for 30 min at room temperature, to allow for phage adsorption, then transferred to a 3-ml GYE overlay and poured over the surface of a GYE agar plate. Following two days of incubation at 27°C, plates were examined for the presence of plaques. Isolated plaques were picked and phages propagated as follows. GYE broth (10 ml) was inoculated with 100 μl of a R. equi overnight culture and incubated at 27°C at 200 rev min−1 for 3 h. A single isolated plaque was then added to the culture, supplemented with 10 mmol l−1 CaCl2 and 10 mmol l−1 MgCl2, and incubation continued until bacterial cell lysis was observed through clearing of the broth or a significant reduction in turbidity. The phage lysate was filter-sterilized. Phages were purified through three rounds of plaque purification.

Phage genomic DNA purification and restriction analysis

Phages were purified and its genomic DNA was isolated following the protocol described by Sambrook et al. (1989) with the following modifications: incubation of the phage lysate with 0·5 mol l−1 NaCl for 1 h at 4°C, prior to centrifugation at 5500 g for 10 min at 4°C, and resuspension of the phage pellet in 1·5 ml phage buffer (40 mmol l−1 Tris–HCl, 100 mmol l−1 NaCl and 10 mmol l−1 MgSO4, pH 7·4) prior to loading on a caesium chloride (CsCl) step gradient. To recover genomic DNA, DNase I and RNase A were added to the purified phage particles at a final concentration of 1 μg ml−1 and incubated at 37°C for 30 min. Subsequently, 20 mmol l−1 EDTA and 50 μg ml−1 proteinase K were added. In the final steps of DNA precipitation, 100% isopropanol was used.

For restriction analysis, 1 μg of phage DNA was digested with 10 U of restriction endonuclease as per the manufacturer's recommendations (Roche). DNA profiles were determined by agarose (0·8%) gel electrophoresis using TAE (40 mmol l−1 Tris–acetate, 1 mmol l−1 EDTA, pH 8·3) buffer and ethidium bromide (0·3 μg ml−1).

SDS-PAGE

Phage protein profiles were determined by SDS-PAGE as described by Laemmli (1970) using 1011 PFU ml−1 of CsCl-purified phage particles, a 12% resolving gel (375 mmol l−1 Tris base, pH 8·8; 0·1%, w/v SDS; 12%, v/v acrylamide mix; 0·1%, w/v ammonium persulfate; 0·0004%, v/v TEMED) and a 4% stacking gel (125 mmol l−1 Tris base, pH 6·8; 0·1% w/v SDS; 4% v/v acrylamide mix; 0·1% w/v ammonium persulfate; 0·001% v/v TEMED). Samples were repeatedly (three times) frozen in liquid nitrogen and thawed at 46°C. Following heating at 75°C for 4 min, samples were mixed with 5× sample buffer (260 mmol l−1 Tris–Cl, pH 6·8; 50% v/v glycerol; 5% SDS; 0·25% w/v bromophenol blue; 0·25 mol l−1 DTT) and reheated at 95°C for 5 min before electrophoresing at 150 volts for 90 min. Gels were stained with Coomassie Blue (25% v/v methanol, 10% v/v acetic acid, 0·2% w/v Coomassie Brilliant Blue).

Negative staining and transmission electron microscopy

CsCl-purified phage particles were dialysed against 1000-fold excess of phage buffer (described above) with three buffer changes. Samples for transmission electron microscopy were examined using both the uranyl acetate, uranyl formate and methylamine vanadate staining techniques. For uranyl acetate staining, 10 μl of phage preparation (105 pfu ml−1) was placed on a formvar-/carbon-coated slot grid for 10 min followed by the addition of 20 μl of 1% uranyl acetate. After 30 s, the excess sample was removed and grids were allowed to dry, before viewing using a Phillips CM120 transmission electron microscope (FEI UK Ltd, Cambridge, UK). For samples stained with 1% uranyl formate, following phage application the grids were placed phage side down onto a droplet of 1% paraformaldehyde/PBS for 15 min and then washed through several droplets of distilled water before staining with 1% uranyl formate which was immediately drawn off and grids left to dry. For methylamine vanadate staining, carbon-/formvar-coated 200-mesh copper grids were glow-discharged and 10 μl of phage preparation was dried down to a thin layer onto the hydrophilic support film. 20 μl of 1% aqueous methylamine vanadate stain (Nanovan; Nanoprobes, Stony Brook, NY, USA) was applied and the mixture dried down immediately using filter paper to remove excess fluid. Air-dried specimens were imaged with a Zeiss 912 energy filtering transmission electron microscope operating at 120 kV. Contrast-enhanced, zero-loss energy-filtered digital images were recorded with a 14-bit/2K Proscan CCD camera.

Phage dimensions were determined, with the average and standard deviation of five individual particles measured for each phage. Head measurements were obtained by measuring the distance between two opposite vertices.

Results

Phage recovery from soil

Soil samples were taken at random from geographically distinct locations within 70 km of Edinburgh (Scotland, UK) over a three-year period. The strategy for collection of samples was based on sampling soil differing in usage (golf courses, gardens, water canals, foot paths, agricultural farmlands, sea soil and woodlands), randomly sampled from different geographical locations, in order to increase the probability of isolating a diverse range of phages. To minimize host-derived selection bias, these samples were screened, with or without enrichment, for the presence of phages infecting R. equi using eight clinical isolates (Table 1), including the type strains NCIMB 10027 and 103S.

It is interesting to note that from some soil samples, phages could be detected directly (i.e. without enrichment) by spotting a soil aqueous extract on a lawn of R. equi bacterial cells. For example, of nine soil samples tested in the initial phase of this study, seven yielded phage titres ranging from 1·2 × 103 to 6·7 × 105 PFUg−1 of soil. Of these soil samples, phages were detected on a limited number of strains, that is, no phages were detected using R. equi bacterial hosts CV2, CV3 and GV1. While high phage titres could be isolated from some soil samples, for others an enrichment step was required in which soil samples were incubated with a combination of R. equi strains prior to preparation of soil extracts (Table 2). This led to the detection of R. equi phages in all soil samples analysed.

Table 2. Host range and plaque morphology of phages isolated from soil capable of infecting Rhodococcus equi
R. equi phageR. equi bacterial strainsR. erythropolis bacterial strainsPlaque morphology on propagating hostGroup by host range
NCIMB 10027GV1GV2CV1CV2CV3VI1103SSQ1NCIMB 11148
  1. Propagating host is shown in boldface.

  2. Phages E1 and CV12 originated from same soil sample, as did phages E3/V2, and phages GV21/GV22.

  3. a

    Represents phages recovered directly from soil, that is, without enrichment.

E1a + +++++++  Faint diffuse, 1 mm diameterE
E2a + +++++++  Faint diffuse, 1 mm diameterE
E3a + +++++++  Faint diffuse, 1 mm diameterE
GV21a + +        Clear, 1 mm diameterA
GV22a + + +++++++Clear 1 mm diameterB
V2a++ ++  + +  Faint diffuse, 1 mm diameterC
V3a++ ++  + +  Clear, 1·5 mm diameterC
V5a++ ++  + +  Clear, 1·5 mm diameterC
CV12a++  + + ++  Clear, 1·5 mm diameterC
V6a + ++  + +  Clear, 3·5 mm diameterD
V7a + ++  + +  Clear, 3·5 mm diameterD
V8a + ++  + +  Clear, 3·5 mm diameterD
V9+ ++++ + +  Clear diffuse, 1 mm diameterF
V10+ ++++ + +  Clear smooth, 5 mm diameterF
V11+ ++++ + +  Clear centre surrounded by turbid zone, 2 mm diameterF
V12+ ++++ + +  Clear, 1 mm diameterF
V13+ ++++ + +  Clear centre turbid edges, irregular shape, 2 mm diameterF
V14+  +++ + +  Smooth edges clear at centre, 2 mm diameterG
V15+  ++  + + +Clear, 1 mm diameterH
V16+ ++++ + +++Irregular shape, 1 mm diameterI
CV13    +   +   Irregular clear, 2 mm diameterJ
CV14    +   +   Clear, 2 mm diameterJ
CV15+ + + ++++++Faint pinpoint, 0·5 mm diameterI
CV16  + + + ++  Clear centre surrounded by turbid zone, 0·5 mm diameterK
GV10  +         Clear, 3 mm diameterL
GV11  + + ++    Clear, 1 mm diameterM
GV12  + + +     Smooth edges, clear centre, 2 mm diameterN

In order to maximize the diversity of phages included for further characterization in the second phase of this study, plaques were selected for phage propagation and plaque purification from a range of propagating host/soil sample combinations. The plaque morphological characteristics of each selected phage on their propagating host are presented in Table 2. When propagated in broth, high titres were obtained for each phage ranging from 3·2 × 108 to 2·8 × 1014 PFU ml−1.

The diversity of R. equi phages recovered from these soil samples was then evaluated by host range analysis, restriction profiles of genomic DNA, virion protein profiles, plaque morphology and phage morphological characteristics as observed in transmission electron microscopy.

Host range

The host range of 27 purified R. equi phage isolates was determined using the spot assay technique on eight clinical isolates of R. equi and two environmental isolates of Rhodococcus erythropolis. Variation in the susceptibility of the Rhodococcus species to an individual phage was observed (Table 2). Phages with a similar host range were grouped together resulting in 14 distinct groups with phages in any given group exhibiting an identical host range. Of note, phages belonging to group E are capable of infecting all clinical isolates tested, including the type strains NCIMB 10027 and R. equi 103S. Although originating from different soil samples, all three phages in group E yield a similar titre and plaque morphology on R. equi NCIMB 10027. The 14 phages isolated with R. equi VI1 as the propagating host could be divided into six groupings (C, D, F, G, H and I) with differing host range profiles. The host range of groups B, H and I was not limited to R. equi, in that infection of R. erythropolis strains SQ1 and NCIMB 11148 was also observed. Most phage groupings are capable of infecting the majority of clinical isolates tested, with the exception of groups A, J, L and N, each infecting less than four of eight strains tested. This is of relevance given the possible application of phages in antimicrobial therapy.

Phage genome restriction and protein profile analysis

To assess the genome diversity of isolated phages, restriction fragment analysis of phage genomic DNA was performed. Genomic DNA restriction profiles are commonly used in evaluating phage diversity and require good quality DNA in sufficient quantity. Preliminary attempts at isolating genomic DNA from phages infecting R. equi using adaptations to commercial kits yielded variable results (data not shown), and therefore, optimization of the protocol described by Sambrook et al. (1989) was undertaken. A series of modifications was tested (Table 3), of which the best quality and quantity of DNA was obtained when nuclease and proteinase treatments were undertaken subsequent to CsCl gradient purification (option 3), in contrast to direct treatment of the crude phage lysate.

Table 3. Optimization of phage genomic DNA isolation
StepModificationBasicaOption 1Option 2Option 3Option 4Option 5
  1. (+) denotes the designated step was carried out, and (−) indicates omitted step.

  2. a

    Basic method represents that described by Sambrook et al. (1989).

1DNase I, RNase A before CsCl purification+++
2DNase I, RNase A after CsCl purification++
320 mmol l−1 EDTA, 50 μg ml−1 proteinase K++++++
4SDS++++
5Phenol–chloroform–isoamyl alcohol++++++
6Repeat step 5++++++
7Chloroform–isoamyl alcohol++++++
8Repeat step 7++++++
9100% isopropanol+++++
10100% ethanol, 0·3 mol l−1 Na-acetate+
1170% ethanol++++++

Of the 24 restriction endonucleases tested, BamHI, PvuII, SalI and SmaI were capable of restricting the genomic DNA from the majority of phages. Phages E1 to E3 showed identical restriction profiles as did phages V6 to V13. Phages GV21, V3, V5 and CV14 could be grouped together on the basis of their restriction profiles, as could V16 and CV16 (see Fig. 1 for a representative comparison). Although GV10, GV11 and GV12 showed similar restriction profiles, they are clearly distinguishable with some distinctive fragments (data not shown). Phage V14 was abandoned at this stage due to difficulties of purification by CsCl gradient centrifugation. Based on the restriction analysis alone, the 26 phages analysed can be divided into 12 groups (Table 4).

Table 4. Phage grouping by phage morphology, host range and DNA restriction profiles
Rhodococcus equi phageTail characteristics by electron microscopyGroup by host rangeGroup by restriction profileFinal phage groupings
E1NDE1I
E2NDE1I
E3Contractile tailE1I
GV21Long noncontractileA2II
GV22Long noncontractileB3III
V2Long noncontractileC4IV
V3Long noncontractileC2V
V5NDC2V
CV12NDC2V
V6Long noncontractileD5VI
V7NDD5VI
V8NDD5VI
V9Long noncontractileF5VII
V10NDF5VII
V11NDF5VII
V12NDF5VII
V13NDF5VII
V14NDGNAInconclusive
V15Long noncontractileH6VIII
V16Long noncontractileI7IX
CV13Long noncontractileJ8X
CV14Long noncontractileJ2XI
CV15Long noncontractileI9XII
CV16Long noncontractileK7XIII
GV10Long noncontractileL10XIV
GV11Long noncontractileM11XV
GV12Long noncontractileN12XVI
Figure 1.

Agarose gel electrophoresis of Rhodococcus equi phages V16 and CV16 genome restriction profiles. The genomic DNA of phages V16 (lanes 2 and 4) and CV16 (lanes 3 and 5) were restricted with PvuII (lanes 2 and 3) and SmaI (lanes 4 and 5). Lane 1 represents DNA hyperladder I (Bioline).

Representatives of each of the 12 phage groups, established on the basis of genome restriction profiles, were selected for SDS-PAGE to analyse the virion-associated (or structural) protein profiles. The results showed individually distinct virion profiles (Fig. 2) supporting the diversity assessed by restriction profiling of genomic DNA. The number of protein bands observed per phage isolate ranged from three to eight with molecular weights of ~ 8 to 60 kDa. While protein bands of a similar molecular weight could be observed in several samples, no single protein band could be observed in all samples.

Figure 2.

SDS PAGE analysis of Rhodococcus equi phages. Protein profiles of intact phage particles from representative phage groups. Lanes 2–10: GV22, V2, V3, V6, V15, CV13, CV15, CV16, GV10. Lanes 12–14: GV12, E3, GV11. Lanes 1 and 11 represent Protein Marker ColorBust™ (Sigma).

Combining the results of host range and genome restriction analysis led to the establishment of 16 phage groupings (I–XVI; Table 4). Phages in the host range groups A, C and J have similar genomes restriction profiles but differ in host range, similarly with phages in groups D and F. In contrast, phages in host range groups C, I and J have distinct restriction profiles.

Phage morphology

Electron microscopy of representative phages from each grouping (I–XVI) revealed that all are of the order Caudovirales, consisting of an isometric head and a tail. The majority are members of the Siphoviridae family, with the exception of phage E3, which is a Myoviridae (Fig. 3). The Siphoviridae phages have head and tail measurements ranging from 48 to 71 nm and 195 to 497 nm, respectively, and are comparable with those previously reported for R. equi phages (66–82 and 236–489 nm for head and tail, respectively; Summer et al. 2011). Phage CV16 is morphologically identical to R. equi phage ReqiPine5 (Summer et al. 2011), with head and tail measurements of 60 and 236 nm, respectively, while phage V2 is morphologically similar to the previously reported R. equi phage ReqiDocB7 (Summer et al. 2011). The head and tail measurements of the myoviridal phage E3 are 93·55 ± 2·53 nm and 94·28 ± 2·07 nm, respectively.

Figure 3.

Transmission electron microscopy of Rhodococcus equi phages. Electron micrographs of representative R. equi phages from phage groupings I–XVI.

Discussion

We report on the first comprehensive study of the diversity of a collection of phages isolated against R. equi. To date, information on phages infecting R. equi is limited to genome analysis for six virulent phages, five of which are Siphoviridae and one Myoviridae (Petrovski et al. 2011; Summer et al. 2011; Salifu et al. 2013); the determination of the antibacterial activity of phage R. equi YF1 encoded proteins (Shibayama and Dabbs 2011); the description of the host range, restriction profile and structural morphotype for a single temperate R. equi phage isolate (Hiddema et al. 1985); and the reporting of phage-like particles associated with chronic infection of virulent R. equi in mice (Nordmann et al. 1994). An objective of our study included the isolation of a diversity of R. equi phages. To this end, a range of soil samples were analysed from distinct geographical locations that included gardens, canal towpaths, woodlands and farmlands. For the majority of samples, an enrichment step was required; however, it is worth noting that the samples for which phages could be isolated at high titres without an enrichment step were all from manured garden soil and horse grazing soil and faeces. It is therefore plausible to hypothesize that these samples contain a high titre of R. equi phages compared to soil samples that are not associated with horses or rich in manure. The difficulty in isolating R. equi phages from some soil samples by direct isolation may also reflect differences in the autochthonous R. equi bacterial strains in the soil compared to those used for phage isolation, although the strains used in this study were all pathogenic isolates of equine origin. The approach taken in this study was to isolate phages directly from soil, rather than by induction from bacterial strains. This therefore does not rule out the possibility that some of the phages recovered are capable of entering into a lysogenic relationship with their host. While lysogeny was not observed in this study, this would need to be investigated further.

The diversity of phages isolated was evaluated on the basis of a number of features including host range, genome restriction and virion-associated protein profiles, together with phage and plaque morphology. Consideration of host range is of importance given the potential for application of these phages in typing R. equi bacterial isolates, in the treatment of infections and in pathogen biocontrol. Noticeably, the group E phages characterized in this study plaqued on all R. equi strains tested. In contrast, phage GV10 has the narrowest host range plaquing only on the original propagating host. The groupings established on the basis of genome restriction profiles were supported by the distinct protein profiles of each phage, the latter corresponding to the virion-associated proteins. For most phages, the groupings established according to restriction profile and host range correlated. For example, phages V6, V7 and V8 have similar host range and restriction profiles even though they were isolated from different geographical locations.

It is of interest to note that while phage DNA could be successfully isolated for phage E3 using commercial λ DNA purification kits, results were variable for other phages. In some cases, failure was due to blocking of the column (despite using the manufacturer's recommended phage titres). Because these kits are designed for the recovery of DNA from phage λ, it is possible that differences in capsid composition between phage λ and R. equi phages may account for the failures. Alternatively, given the generally larger genome size compared to phage λ (data not shown), the greater amount of DNA being released from R. equi phages may result in saturation of the column. However, phage titre reduction did not resolve the issue (data not shown). Furthermore, phage E3 DNA, with a genome size threefold larger than λ (Salifu et al. 2013), was successfully isolated with these kits. In contrast, the genomic DNA purification protocol optimized in this study succeeded in yielding a 4·45-fold increase in DNA compared to the basic method described by Sambrook et al. (1989). In the latter protocol, nuclease treatment is applied prior to phage purification. However, the application of these steps postgradient purification led to a higher yield, while at the same time significantly reducing the costs. The optimized protocol also yielded a higher-quality DNA compared to the basic protocol.

It is interesting to note that R. equi phages were recovered from a diversity of soil samples and that these phages were recovered using bacterial strains that are not indigenous to the soil samples screened. While there has been a renewed interest in recent years in the ecology and impact of phages in aquatic ecosystems, the soil environment remains largely unexplored (Kimura et al. 2008). Attention has focussed on the rhizosphere, on phages infecting significant bacterial plant pathogens and, more recently, on the application of microscopy and metagenomic approaches to evaluating virus abundance and diversity in soil (Williamson et al. 2003; Fierer et al. 2007; Kimura et al. 2008; Frampton et al. 2012). Significantly, the phages isolated from soil in this study are capable of infecting clinical isolates of R. equi. The presence in soil of a diversity of phages infecting the pathogen R. equi, and in some cases at high titre, is of note given that phages act an indicator of bacterial host presence, and also given the role of phages in bacterial adaptation and evolution (Canchaya et al. 2003; Comeau and Krisch 2005; Gomes and Buckling 2011).

In conclusion, the data presented in this study demonstrate the ease with which a diversity of phages infecting R. equi can be isolated. Of the 26 phages characterized, 16 groups were established on the basis of host range and genomic DNA restriction profiles. With the exception of one group, the majority of phages recovered belong to the Siphoviridae, thus adding to those recently described by Summer et al. (2011). Considering the extent of genome sequence diversity amongst the four phages described by Summer et al. (2011), it will be interesting to undertake comparative genomic analysis of representatives of all Siphoviridae groupings defined in this study.

Acknowledgements

We would like to thank Dr Laurence Tetley and Margaret Mullin at the University of Glasgow for the transmission electron microscopy analysis and numerous undergraduate students who have contributed to phage isolation.

Ancillary