In this study we isolated novel bacteriophages, infecting the zoonotic bacterium Campylobacter jejuni. These phages may be used in phage therapy of C. jejuni colonized poultry to prevent spreading of the bacteria to meat products causing disease in humans. Many C. jejuni phages have been isolated using NCTC12662 as the indicator strain, which may have biased the selection of phages. A large group of C. jejuni phages rely on the highly diverse capsular polysaccharide (CPS) for infection and recent work identified the O-methyl phosphoramidate modification (MeOPN) of CPS as a phage receptor. We therefore chose seven C. jejuni strains each expressing different CPS structures as indicator strains in a large screening for phages in samples collected from free-range poultry farms. Forty-three phages were isolated using C. jejuni NCTC12658, NCTC12662 and RM1221 as host strains and 20 distinct phages were identified based on host range analysis and genome restriction profiles. Most phages were isolated using C. jejuni strains NCTC12662 and RM1221 and interestingly phage genome size (140 kb vs. 190 kb), host range and morphological appearance correlated with the isolation strain. Thus, according to C. jejuni phage grouping, NCTC12662 and NCTC12658 selected for CP81-type phages, while RM1221 selected for CP220-type phages. Furthermore, using acapsular ∆kpsM mutants we demonstrated that phages isolated on NCTC12658 and NCTC12662 were dependent on the capsule for infection. In contrast, CP220-type phages isolated on RM1221 were unable to infect non-motile ∆motA mutants, hence requiring motility for successful infection. Hence, the primary phage isolation strain determines both phage type (CP81 or CP220) as well as receptors (CPS or flagella) recognised by the isolated phages.
Citation: Sørensen MCH, Gencay YE, Birk T, Baldvinsson SB, Jäckel C, Hammerl JA, et al. (2015) Primary Isolation Strain Determines Both Phage Type and Receptors Recognised by Campylobacter jejuni Bacteriophages. PLoS ONE 10(1): e0116287. https://doi.org/10.1371/journal.pone.0116287
Academic Editor: Mark J. van Raaij, Centro Nacional de Biotecnologia - CSIC, SPAIN
Received: September 10, 2014; Accepted: December 4, 2014; Published: January 13, 2015
Copyright: © 2015 Sørensen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: The farms involved in the study are de-identified to preserve privacy, but otherwise all relevant data are within the paper.
Funding: This work was supported by the Danish Research Council for Technology and Production Sciences by grant 09-070555 (http://ufm.dk/en/research-and-innovation/councils-and-commissions/the-danish-council-for-independent-research/the-council-1/the-danish-council-for-independent-research-technology-and-production-sciences). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Campylobacteriosis is the most commonly reported zoonosis and the leading cause of human bacterial gastroenteritis in EU . Contaminated poultry meat is the major source of Campylobacter jejuni infections and in some countries up to 90% of these domestic birds are colonized and carry this human pathogen as a part of their normal intestinal flora . A number of studies report the application of bacteriophages (phage therapy) to reduce the number of C. jejuni either at the poultry farm setting [2–6] or post-slaughter on the chicken meat [7–9] with significant levels of reduction. Thus, the use of phages is a promising intervention strategy. However, phage treatment against C. jejuni can only be efficient if the phages applied are capable of infecting the diverse population of C. jejuni strains found in the poultry gut or on the contaminated meat .
Attachment of a phage to a specific receptor on the surface of the bacterial host is the first step in phage infection. Several studies have shown that either changes in the capsular polysaccharides (CPS) or loss of motility is involved in phage resistance development in C. jejuni, indicating that CPS and flagella may be the main receptors of phages infecting this bacterium [11–14]. Indeed, mutant analyses and adsorption assays showed that C. jejuni phage F336 binds to CPS, while motility is essential for successful infection by phage F341 [14,15]. Even though the receptor for phage F341 has not yet been identified, transmission electron images demonstrated that phage F341 indeed attaches to the flagella of C. jejuni. Our current hypothesis is that either the widespread flagellar O-linked pseudaminic acid or even the FlaAB proteins may be the receptor of this phage, since mutants lacking legionaminic acid and derivatives of pseudaminic acid attached to the flagella are susceptible to F341 infection .
CPS is the major determinants of the Penner serotyping scheme of C. jejuni, and currently more than 60 different serotypes are known, demonstrating the diversity of these surface components [16,17]. The CPS of C. jejuni are highly variable and complex, due to the many different carbohydrate backbones that also carry highly diverse modifications such as O-methyl, ethanolamine, aminoglycerol and O-methyl phosphoramidate (MeOPN) [18–24]. In addition, phase variable expression of these CPS modifications results in further variations, even among strains encoding the same genes in CPS loci [22,25,26]. The MeOPN modification is common within the Campylobacter genus and has been identified in ∼70% of C. jejuni isolates despite that phosphoramidates are rarely found in nature [27,28]. This commonality suggest an important biological role for this surface structure and loss of MeOPN on the C. jejuni cell surface is associated with a decrease in serum resistance, enhanced invasion of CaCO-2 cells and reduced colonisation in a piglet model . Interestingly, we also identified the MeOPN moiety attached to GalfNAc of CPS in C. jejuni NCTC11168 as a novel phage receptor of several lytic bacteriophages [13,14]. Development of phage resistance caused by changes in the CPS modifications is expected to be rather frequent, due to the phase variable nature of these components. This was confirmed by in vivo chicken colonization experiments where phage resistance by loss of the MeOPN receptor was easily achieved without causing a decline in colonization levels of C. jejuni . Knockout MeOPN transferase mutants in C. jejuni strain 81–176 were also not affected in colonisation levels of chickens, further supporting a minor role of this surface modification in chicken colonisation . Consequently, phage resistant development of C. jejuni may potentially affect the outcome of a phage treatment.
To develop an efficient phage therapy and prevent phage resistance development, it has been suggested to use phage cocktails that target multiple receptors as well as broadening the lytic spectrum of the applied phages [3,6,29]. Noteworthy, a large number of the previously identified and characterized bacteriophages are isolated using C. jejuni NCTC12662 as indicator strain due to its high sensitivity towards phages [4,30–38]. Indeed, almost all Campylobacter phages isolated so far are members of the Myoviridae family and can be categorised into three groups (group I, II and III) based on genome size and morphology . The majority of these phages belong to the Eucampyvirinae subfamily, represented by group II and III phages, that is divided into two genera, the Cp220likevirus (180–190 kb genome, CP220-type phages) and the Cp8unalikevirus (130–140 kb genome, CP81-type phages) . This extensive use of one particular indicator strain may have aided in the recovery of Eucampyvirinae phages that could be biased in respect to recognition of C. jejuni receptors. The current phage collections may thus not represent the phage diversity needed to target numerous receptors in the diverse population of C. jejuni expressing very different surface structures.
Here we isolated novel phages from free-range chicken faecal samples collected throughout Denmark using well characterized Penner serotyped indicator strains expressing different capsular structures and modifications. Our analysis showed that the isolation strain both influence the phage type isolated and the receptors recognised by these C. jejuni phages.
Materials and Methods
Bacterial strains and preparation of bacterial lawns
Strains used in this study are shown in Table 1. Preparations of bacterial lawns were done according to Sørensen et al, 2011 . Briefly, standard growth condition of C. jejuni strains was performed by cultivation on blood agar Base II (Oxoid) supplemented with 5% calf blood (BA) followed by incubation under microaerobic conditions (6% CO2, 6% O2 and 88% N2H2) for 18–24 h at 37°C. For preparation of bacterial lawns, cultivated cells were harvested with cation adjusted (1 mM CaCl2 and 10 mM MgSO4) Brain Hearth Infusion broth (Oxoid) (CBHI). Bacterial suspensions were then adjusted to an optical density at 600 nm (OD600) of 0.35 (∼109 cfu/ml) in CBHI and incubated microaerobically at 37°C for 4 h. Then, 500 μl of cultures were added to 5 ml of molten NZCYM overlay agar (NZCYM broth [Sigma] with 0.6% agar [Sigma]) tempered to 45°C and poured on previously prepared and dried NZCYM basal agar (with 1.2% agar [Sigma] and 10 μg/ml vancomycin [Sigma]) and allowed to dry for 45–60 min in a flow hood.
Isolation, purification and propagation of phages
A total of seventeen free-range chicken farms spread throughout Denmark (13 in Jutland; 1 in Fyn; 3 in Zealand) and one goose farm (Sejerø) were visited in September 2011 (Fig. 1) and fresh faecal samples were collected from five parts of each farm in a plastic bag and taken into analysis in less than 96 hours. As there was no direct contact with the animals and fecal samples were collected randomly from the housing environment without disturbing the daily routines of the birds, formal approval of this study as an animal experiment was not necessary according to Danish law. Sampling was carried out on private land and permission was granted by the farm owners in return of being anonymous. In the laboratory up to five-gram aliquots of the fecal samples were aseptically collected in a 50 ml tube and a 1:10 dilution with sterile SM buffer (100 mM NaCl [Merck], 8 mM MgSO4.7H2O [Merck], 0.01% gelatine [Sigma], 50 mM Tris-HCl [Sigma] in 1 l of distilled water with pH 7.5) was made and left overnight at 4°C on a gyratory shaker (100 rpm) for dissociation of phages. After thorough vortexing, the resulting upper phases were taken in sterile tubes and centrifuged at 3.000 g for 5 min for removal of debris. Supernatants were syringe filtered (0.2 μm filters) and droplets of five times 10 μl were applied on prepared lawns of all indicator strains. Following the settling of droplets, plates were incubated microaerobically at 37°C for 24 h and analysed for the presence of plaques. If plaques were observed the plaque morphologies were recorded.
Black circles represent farms where phages could be isolated.
For purification of phages, single plaques with varying morphologies were transferred to 100–500 μl of SM buffer with sterile pipette tips, vortexed thoroughly and tenfold serial dilutions were prepared and applied on a lawn of the original indicator strain. Following at least three cascades of purification for each plaque, phages were propagated for preparation of phage stocks. Phage propagation was performed on the lawns of the isolating strains by the plate lysis method as described elsewhere  and stocks were stored at 4°C.
Titration, host range and number of plaques
Titration and host range determination of phages by spot assay were conducted as described previously . Briefly, C. jejuni strains were cultivated under standard growth conditions, harvested into CBHI and adjusted to OD600 of 0.35. The bacterial suspensions were then incubated for 4 h at 37°C under microaerobic conditions and bacterial lawns were made as already described. Tenfold serial dilutions of the phage stock suspensions in SM buffer up to 10−7 were made and three times 10 μl of each dilution and the undiluted stock suspension were spotted on the bacterial lawns of the investigated strains. Following incubation under microaerobic conditions for 18–24 h at 37°C, plaque morphologies were observed, plaques counted and mean plaque forming units per ml (pfu/ml) was calculated for each strain. The data represented are the means from between 2–8 independent experiments.
Pulse Field Gel Electrophoresis (PFGE) analysis
PFGE analysis for characterization of phage genomes was done as described by Hansen et al. (2007)  with slight modifications. Briefly, phage stock suspensions (107–109 pfu/ml) were mixed either 1:1, 2:1 or 8:1 with 1%, 2.0% or 4% SeaKem Gold agarose respectively, prepared in TE buffer (10 mM Tris and 1 mM EDTA [pH 7.2]) and liberated overnight in lysis buffer (10 mM Tris, 100 mM EDTA, 1% sarkosyl and 0.1 mg/ml proteinase K) at 55°C with gentle shaking (100 rpm). Following digestion of phage capsids, blocks were washed 4–6 consecutive times (55°C, 100 rpm) for 20 min in washing buffer (20 mM Tris and 50 mM EDTA) and stored in TE buffer at 4°C until the PFGE was performed. Digestion of phage genomes with restriction endonuclease HhaI (New England BioLabs) was carried out, following preparation of 3–5 mm block slices, according to the Manufacturer and completed in 17–19 h at 37°C. For PFGE; intact or digested phage DNA harboring slices were loaded in 1% SeaKem Gold agarose along with low range PFGE Marker (New England BioLabs), run for 14 h at 6 V/cm, included angle 120, and switch time 2–10 s using Biorad CHEF-III apparatus. Finally, gels were stained for 30 min with ethidium bromide and washed in distilled water for 1 hour before being visualized manually.
SmiI restriction profiling of phage genomes
Restriction profiles of the phage genomes were conducted as described by Hammerl et al. (2011) . Phage DNA isolated by proteinase K/SDS treatment of CsCl-purified phages was used for restriction profiling. Restriction endonuclease SmiI was used according to Manufacturer’s recommendations (Fermentas) and analysed by agarose gel electrophoresis using standard procedures.
Construction of ∆kpsM and ∆motA mutants
For the construction of the acapsular NCTC12662∆kpsM mutant strain, the pTA-3 vector (pTZ57R/T::kpsM::aphA, KanR) containing an amplified kpsM::kan fragment  was introduced into the C. jejuni NCTC12662 by electroporation and transformants were selected for kanamycin resistance. The pCV1206-1 vector construct (pCR2.1-TOPO::∆motA::cat, AmpR, KanR, CamR)  used to generate the non-motile NCTC12662∆motA strain, was introduced into C. jejuni NCTC12662 by electroporation and transformants were selected for chloramphenicol resistance. All mutant constructs were verified by PCR amplification and sequencing. Additionally the motA deletion was also confirmed by complete loss of motility in a soft agar motility assay as described below. Attempts to construct ∆kpsM RM1221 and ∆motA RM1221 mutants were unfortunately unsuccessful.
The motility assay was essentially performed as described previously  with minor modifications. Briefly, C. jejuni strains were grown under standard growth conditions on BA, harvested in BHI and OD600 was adjusted to 0.1. One μl of the adjusted cell suspension was placed in the centre of ten 0.25% heart infusion broth (Difco) agar plates pre-dried for 45 min. Plates were incubated under standard growth conditions and growth zones demonstrating the ability to move in the soft agar were measured after 24 and 48 hours. Measurements are given as the mean count standard deviations of two independent experiments.
Transmission Electron Microscopy (TEM) of phages
The basic protocol of Ackermann (2009)  was modified for TEM analysis as follows. Ultra-thin carbon films prepared in a high vacuum coating system (BAL-TEC MED 020; Balzers, Lichtenstein) on a mica sheet were cut (~3×3 mm in size) and floated into a 100 μl drop of phage suspension. After 5–10 min adsorption, films were transferred into a drop of 1% (v/v) of EM-grade glutaraldehyde (20 min) and subsequently into a drop of 2% (w/v) uranyl acetate for negative staining (1–2 min). After two washes for a few seconds in drops of distilled water, samples were picked up with 400-mesh copper grids (Plano, Wetzlar, D). Electron micrographs were taken in a Tecnai 10 transmission electron microscope (FEI Company, Eindhoven, The Netherlands) at an accelerating voltage of 80 kV. Digital micrographs were taken with a Megaview G2 CCD camera (Olympus SIS, Münster, Germany).
Phage isolation from free-range poultry farms
For the isolation of C. jejuni phages from free-range poultry farms, we used seven C. jejuni indicator strains belonging to different Penner serotypes, thus expressing different CPS carbohydrates and modifications of the CPS (Table 1). CPS has structurally been determined for Penner serotypes HS1.44, HS2, HS4, HS6, HS23.36, and HS53 represented by strains NCTC12658, NCTC11168, 1447, 81116, 81–176, and RM1221, respectively. In addition, the Penner serotype HS5j strain NCTC12662 was included, due to its high susceptibility to phage infection. The CPS carbohydrate backbone is not resolved for serotype 5j, but we determined, by high-resolution magic angle spinning NMR spectroscopy, that the CPS carries one or more MeOPN modifications . This has recently been confirmed by the genome sequence of this strain identifying two MeOPN transferase homologues .
In order to isolate as many different C. jejuni phages as possible, all faecal samples were collected from free-range chicken and goose farms where the birds had constant interaction with the environment (Fig. 1). A total of 43 phages showing a variety of plaque morphologies were isolated, purified and propagated from seven of the eighteen free-range poultry farms (Fig. 1, Table 2). Interestingly, phages were only isolated using C. jejuni NCTC12662, NCTC12658 or RM1221 as the indicator strains, while no phages were isolated using the remaining strains (NCTC11168, 1447, 81116, and 81–176). The samples from farm 15 allowed isolation of phages using all three indicator strains, whereas the samples from farm 7 resulted in isolation of phages using strain NCTC12658 and NCTC12662. In contrast, from farms 1, 4, 10, 14, and 17, phages were only isolated using one of these three indicator strains (Table 2). Consequently, phage isolation was not achieved at all farms, and not all C. jejuni strains proved useful for phage isolation.
Host range and genome size of isolated phages
Among the 43 purified phages, we identified 20 distinct phages based on the host ranges and genome restriction analyses (Table 3). Thus, phages originating from the same farm or flock and showing the same host range with similar plaque counts as well as having identical genomic restriction patterns were classified as being replicates of the same phage isolated more than once (Table 3). Although all phages initially were isolated using C. jejuni NCTC12662, NCTC12658 or RM1221 as indicator strains, most phages did infect additional strains such as NCTC11168, 81116, and 1447 that were also used in the isolation procedure (Table 3). Regardless of the initial isolation strain, all phages were able to infect C. jejuni NCTC12662 (Table 3) with high efficiency of plaquing (EOP), confirming the high sensitivity of NCTC12662 to phage predation. In contrast, none of the phages formed plaques on C. jejuni 81–176.
Interestingly, phage genome sizes corresponded to the indicator strain used as phages isolated on C. jejuni NCTC12658 and C. jejuni NCTC12662 all harboured genomes of approximately 140 kb, whilst phages isolated using C. jejuni RM1221 all were close to 190 kb in size (Table 3). Thus, all the isolated phages fall within the two C. jejuni phage groups, group III (130–140 kb genome, CP81-type phages) and group II (180–190 kb genome, CP220-type phages) described previously [39,40]. Restriction profiles of phage genomes revealed that they were highly different and could be categorised into a number of distinct groups that fell within the farm origin (Table 3, Fig. 2). Noteworthy, phages isolated from flock 1 on farm 4 showed one distinct restriction profile, whereas flock 2 from the same farm promoted the isolation of an additional phage with a different profile. Phages isolated from two different flocks on farm 10 showed the same restriction profile regardless of flock origin. Thus, even though the phages were isolated from the same farm they could be very different, which was particularly apparent for farm 15 (Table 3).
A. HhaI profiles of group III (∼140 kb) phage genomes. Lane 1: Marker (New England Biolabs), lane 2: Phage F348 (Type IIIa), lane 3: Phage F360 (Type IIIb), lane 4: F362 (Type IIIc), lane 5: Phage F353 (Type IIId), lane 6: Phage F356 (Type IIIe), lane 7: Phage F371 (Type IIIf), lane 8: Phage F375 (Type IIIg), and lane 9: Marker (New England Biolabs). B. SmiI profiles of group II (∼190 kb) phage genomes. Lane 1: Marker (DNA marker λ Eco130I), lane 2: Phage F376 (Type IIa), lane 3: Phage F379 (Type IIb), lane 4: Phage F381 (Type IId), lane 5: Phage F388 (Type IIe). C. SmiI profile of phage F380 (Type IIc): Lane 1: Marker (DNA marker λ Eco130I), lane 2: phage F380.
Five phages (F348, F362, F375, F376 and F386) isolated from different farms and on different C. jejuni strains were chosen for analysis by transmission electron microscopy as representative of the 20 distinct phages. This analysis showed that all five phages belong to the Myoviridae family having icosahedral heads and contractile tails with morphology resembling other isolated C. jejuni phages [4,30–32,35–38,40,45]. Electron micrographs of a representative group III phage (F362) are shown in Fig. 3 (a and b) illustrating the phage morphology (head diameter: 99 ± 4 nm; n = 22, extended tail length: 107 ± 5 nm; n = 10). Phage particles with contracted tails and empty capsids were frequently sticking in small globular vesicles (most likely bacterial remnants) with their central thin tail tube. An undefined number of flexible tail fibers were detected to be intimately attached to the distal part of the extended tails covering approximately 1/3 of the total tail length in the “folded-up” position. However, in phage particles with contracted tails, these fiber structures are no longer visible. In Fig. 3 (c and d) representative micrographs are shown for the group II phages (F386). Although also belonging to the Myoviridae family, the morphological details differed remarkably from those of the group III phages with respect to the tail size (extended tail length: 140 ± 3 nm; n = 15) and orientation of the tail fibers. The flexible fibers of group II phage F386 were approximately 100 nm in length with small terminal globular structures and were generally expanded uniformly around the tail tip, while tail fibers of phage F362 were sticking on the distal tail regions (Fig. 3). After tail contraction, the tail fibers of F386 were still visible at the distal region of the contracted tail sheaths. Notably phage F386 (genome size 190 kb) has the same capsid dimensions (head diameter: 99 ± 5 nm; n = 15) as phage F362 (genome size: 140 nm) possibly indicating an elevated pressure within the capsid of the group II phages. This correlates with our observations of a significantly higher structural instability of group II phage capsids, as the number of intact F376 and F386 phages were low during microscopy. Thus, these phages fall into the category of the two phage types, the CP81-type (group III phages) and the CP220-type (group II phages), recently described for C. jejuni phages who belong to the proposed genera “CP8unalikevirus” and “Cp220likevirus” within the subfamily Eucampyviriniae .
a. Normal phage particle with extended tail and folded-up fibers. b. Phage particle with contracted tail sticking in small globular vesicles. Note the tail fibers are no longer visible. c. Normal phage particle with extended tail and flexible fibers with attached small terminal globular structures. d. Phage particle with contracted tail. Note tail fibers are still visible. Phages were stained with 2% uranyl acetate and examined on a Tecnai 10 transmission electron microscope.
The role of CPS for phage infection
Looking closer into the host range of the two major groups of phages, we found that the isolation strain used also had a profound effect on the host range of the isolated phages. Many of the phages isolated using C. jejuni NCTC12662 and NCTC12658 could also infect C. jejuni 104–733, while a few phages infect C. jejuni NCTC11168 and 1447. However, none of these phages infect strain RM1221 and 81116 (Table 3). The sugar backbones of the CPS of all strains are very different, but NCTC12662, NCTC12658, 104–733, NCTC11168 and 1447 all carry the MeOPN modification of the CPS (Table 1), which has previously been identified as a receptor of a large group of capsular phages [13,14]
In contrast, phages isolated using C. jejuni RM1221 as indicator strain clearly displayed a different host range than the phages isolated using C. jejuni NCTC12662 and NCTC12658 (Table 3). The most prominent difference is that most of these phages also infect C. jejuni 81116 having a very different CPS structure compared to RM1221 (Table 1 and 3). Comparing the backbone sugars and the modifications of the CPS it is not possible to find any similarities between C. jejuni RM1221 and 81116 (Table 1), suggesting that these phages may be independent of CPS for infection.
To investigate this further, we constructed an acapsular mutant of C. jejuni NCTC12662 by inserting a kanamycin resistance cassette into the kpsM gene and subsequently compared plaque formation to the corresponding wild type. We also included an NCTC12658∆kpsM mutant constructed in a previous study . Here we found that CP81-type phages isolated on the NCTC12658 and NCTC12662 strains neither formed lysis nor plaques on the acapsular mutants, thus showing high dependency on the capsular polysaccharides for successful infection (Table 4). On the contrary, the CP220-type phages isolated on strain RM1221 all infected the acapsular NCTC12662∆kpsM mutant to near wild type levels, often also exhibiting clearer lysis and plaques (Fig. 4). Thus, the isolated CP220-type phages infect C. jejuni independently of the capsule presence.
Clearer lysis and plaque formation can be seen on the acapsular version of the C. jejuni strain.
The role of motility for phage infection
We recently showed that the CP81-type phage F341, which is able to infect ∆kpsM capsule deficient mutants of C. jejuni, is a flagellotropic phage that relies on motility for successful infection . To investigate whether motility is necessary for infection of the capsule independent CP220-type phages isolated in this study, we constructed a non-motile NCTC12662∆motA deletion mutant. This NCTC12662∆motA strain was used in a host range assay with all isolated phages together with an NCTC12658∆motA strain made previously . Interestingly, the host range assay showed that both non-motile mutants (NCTC12662 and NCTC12658 motA) were completely resistant to all the CP220-type phages, as no lysis or plaque formation was observed (Table 4). Consequently, our CP220-type phages depend on motility for successful infection and can be classified as flagellotropic phages. In contrast, all the CP81-type capsular dependent phages infected the non-motile strains with EOPs identical to the wildtype strains (Table 4). Thus, C. jejuni RM1221 promoted isolation of CP220-type flagellotropic phages.
Motility during phage isolation
Noteworthy, all phages isolated in this study were able to infect the NCTC12662 strain with high EOPs, but only phages dependent on the capsule were isolated on this strain. Differences in rotary speed of flagella have previously been shown to have an effect on infection by flagellotropic phages [46,47]. Thus, we therefore speculated if differences in the motility of the three phage isolation strains might explain why no flagellotropic phages were isolated on NCTC12662. To determine this we tested motility of the NCTC12658, NCTC12662 and RM1221 strains by microscopy and by investigating their ability to swarm in soft agar. These results showed a significantly reduced swarming behaviour of NCTC12662 as compared to NCTC12658 and RM1221 (Fig. 5). Growth rates of all three strains in BHI broth were identical, demonstrating that NCTC12662 was indeed impaired in motility compared to NCTC12658 and RM1221. Also, we did observe that lowering the agar concentration from 0.6% to 0.5% or 0.4% in the top agar during plaque assays using NCTC12662 greatly enhanced plaque formation and visibility of CP220-type phages. Plaques formed on 0.6% top agar were small (approx. 0.5 mm in diam.) whereas 0.5% and 0.4% top agar resulted in medium and large size plaques (approx. 1–3 mm in diam.). Thus, the reduced motility of NCTC12662 may explain why no flagellotropic phages were isolated on this strain in spite of its sensitivity towards this group of phages.
Phages infecting C. jejuni have primarily been isolated from various chicken and poultry sites, due to the high colonisation level of the bacterium within the chicken gut [3,4,30–33,35–37,45,48–53]. A large group of these phages are dependent on the CPS structures expressed by C. jejuni in order to establish a successful infection [11,13,14]. Thus, with the aim of isolating novel phage types and investigating the role of the phage isolation strain during this process we used C. jejuni strains expressing different CPS structures for phage isolation from free-range chicken farms.
A total of 43 phages from seven different farms were isolated using three different C. jejuni strains. Characterisation of these phages showed that the phage diversity was restricted to two major groups that fall within the recently described phage types for C. jejuni based on genomic features and morphology, the “Cp220likevirus” and the “Cp8unalikevirus” . These genera are represented by the two type-phages, C. jejuni phage CP220 having a genome of 178 kb and phage CP81 with a genome size of 133 kb, but both being myoviruses . Surprisingly, in this study we found that the phage type correlated with the isolation strain. The two isolation strains NCTC12658 and NCTC12662 promoted the isolation of CP81-type phages, whereas only CP220-type phages where isolated on RM1221. These two phage types have been recovered exclusively over more than 20 years from poultry/chicken settings throughout the world [3,30–33,35–39,41,54–56] indicating that they, and their bacterial hosts, are highly adapted to this environment. Isolating true novel C. jejuni phage types may thus require exploring other niches than poultry settings where this bacterium can be found, although this has proven difficult [49,57]. Low concentration of phages and C. jejuni in other environments have been suggested as limiting factors for phage isolation, but perhaps the choice of isolation strains used in these experiments may also have been a contributing factor. Often the isolation strain has been of a different origin than the environment sampled such as C. jejuni NCTC12662 originating from chicken, due to its high phage sensitivity . We have in this study shown that the applied isolation strain is of great importance for the phage type recovered from a given sample. Hence, using representative C. jejuni strains from the environment of interest or improving the knowledge of the surface composition expressed by such environmental isolates may aid in choosing more appropriate phage isolation strains and thus improve phage recovery from such locations.
Interestingly, we found that the C. jejuni phages isolated in this study were either dependent on CPS or motility for successful infection. Indeed, carbohydrate surface structures such as the capsule and lipopolysaccharides together with flagella are often recognised as receptors of bacteriophages . A recent screening for phages infecting Salmonella enterica serovar Typhimurium also resulted in phages targeting only a limited number of receptor types, including flagella, LPS (lipopolysaccharide) and a membrane transport protein (BtuB) . Noteworthy, we found that the dependency of CPS for successful phage infection correlated with both the isolation strain as well as the phage type, as our CP81-type phages were dependent on CPS for infection, while CP220-type phages relied on motility for successful infection. Using transposon mutants Coward and co-workers also observed that phage-resistance towards group III phages (CP81-type phages) were associated with mutations in the CPS locus, whereas resistant variants to group II phages (CP220-type phages) were non-motile . While this correlation holds true for many C. jejuni phages, we have recently characterised a CP81-type phage F341 isolated on strain NCTC12685 that infects C. jejuni independently of the capsule, but instead requires a motile flagellum for infection . Thus, this exception to the rule shows that the C. jejuni phage type (genome size) is not always linked to receptor type (CPS or flagella) and that the studies of phage receptors provide insight into C. jejuni phage diversity. In addition, more C. jejuni phage genome sequences are required to expand our understanding of the C. jejuni phage types, which currently is based on a limited number of phage genome sequences.
It is both striking and surprising that the C. jejuni NCTC12662 strain only promoted the isolation of one phage type that consistently relied on the capsule for infection, even though both the CPS- and motility-dependent phages isolated in this study proliferate to high numbers on this strain. NCTC12662 (also called PT14) is referred to as the universal Campylobacter bacteriophage host strain and has been extensively used for the isolation of phages infecting C. jejuni due to its general phage sensitivity [4,10,30–38,44]. Nevertheless, several studies also observed that predominantly CP81-type phages were isolated using NCTC12662 as indicator strain [32,33,35–37] supporting our observation that NCTC12662 primarily select for this phage type during the isolation process. However, one might also speculate that the CP81-phage types simply are more predominant in the chicken gut, hence more easily isolated, and we did indeed isolate twice as many phages of this type as compared to CP220-type phages. But interestingly, we discovered that NCTC12662 was impaired in motility as compared to NCTC12658 and RM1221 and we propose that the speed of the flagella rotation is a crucial factor during phage isolation, in particular when phages are embedded in faecal material and are present in relatively low numbers. Indeed, phages using the flagellum as the primary receptor are less frequently isolated from complex material compared to phages binding to the bacterial envelope . Furthermore, it has been reported that the speed of flagella rotation influences the infection efficiency of the flagellotropic phage 7-7-1 infecting Agrobacterium . A similar mechanism may apply for C. jejuni phages, as we recently showed that phage F341 require motile flagella to absorb and infect C. jejuni suggesting that the flagella rotation transports the phage to the bacterial body where the phage DNA is injected . We also found that lowering the top agar concentration greatly enhances plaque sizes of CP220-type phages formed on NCTC12662. Thus, whether lowering the top agar concentration during the phage isolation process or using NCTC12662 mutants with enhanced flagella rotary speed will enhance the recovery of flagellotropic phages using this strain remains to be explored.
The O-methyl phosphoramidate (MeOPN) surface moiety is a common modification of the otherwise very diverse CPS structures expressed by many different C. jejuni strains and is implicated in important biological functions such as serum resistance and colonisation of piglets [27,28]. Interestingly, in the present study the O-methyl phosphoramidate (MeOPN) modification was the only common CPS structure of the strains sensitive to the CP81-type capsular dependent phages. Furthermore, we have previously shown that MeOPN of the CPS of C. jejuni NCTC11168 is a phage receptor recognised by many capsular phages [13,14]. Thus, it is likely that the capsular phages isolated in this study also rely on this structure for successful infection and we speculate that MeOPN comprises a general receptor of this group of C. jejuni phages. Interestingly, even though the C. jejuni 81–176 strain expresses MeOPN it was highly resistant to all phages isolated. This may be explained by its different CPS carbohydrate backbone structure as well as other phase variable modifications that may interfere with phage binding [21,60]. However, as 81–176 is resistant to all phages (i.e. both capsular and flagellotropic), the phage resistant phenotype may not be associated with receptor availability. Indeed, it has been demonstrated that C. jejuni 81–176 encodes a number of different restriction/modification systems that are absent in strains NCTC11168 and RM1221, which may degrade the phage DNA upon entry thereby inhibiting phage proliferation . In C. jejuni NCTC11168, phage resistance is easily achieved by phase variation both in vitro and in vivo by changes in the polyG tract length located in the MeOPN transferase gene cj1421 without affecting growth and colonisation ability of the bacterium [13,14]. Hence, becoming phage resistant by losing the receptor does not have an impact on the ability of C. jejuni NCTC11168 to grow and colonise. Also, the MeOPN capsular modification is phase variable in many other C. jejuni strains [18,19,21,22,24,25,27,60]. However, in NCTC12662 only one of the two MeOPN-transferase homologues shows phase variation at the sequence level . Consequently, the ability not to switch off MeOPN or a lower switching rate could explain why C. jejuni NCTC12662 is consistently sensitive to many capsular phages, thus promoting isolation of this group of phages.
In conclusion, 43 novel phages infecting C. jejuni were isolated and 20 were further characterised to belong to the Eucampyviriniae subfamily either Cp220likevirus or Cp8unalikevirus. Interestingly, we found that the phage type correlated with the isolation strain as well as the receptor dependency, and only C. jejuni phages that were dependent on previously described receptors, namely the CPS and the flagella, were isolated. Thus, even though our attempts to isolate novel C. jejuni phage types in terms of receptor (and morphology) by using strains expressing different CPS structures were unsuccessful, our study emphasises the importance of isolation strain, since the repeatedly used C. jejuni NCTC12662 strain selected for one phage type with a strong bias towards one receptor type. It is well known that the chicken gut is the major ecological niche of C. jejuni providing the perfect environment for growth and colonization of this bacterium. Our study indicates that also C. jejuni myoviruses with limited receptor types and genome sizes are highly predominant and may be well-adapted to this ecological niche of C. jejuni. Identifying true novel Campylobacter phages may thus require exploring other Campylobacter hosts, such as C. fetus and C. upsaliensis and their niche environments in sheep, dog or cat, respectively.
We sincerely thank Christel Galschiøt Buerholt and Gitte Pedersen for excellent technical support. We also thank Dr. Stefan Hertwig (Federal Institute for Risk Assessment) for general support.
Conceived and designed the experiments: MCHS YEG TB SBB CJ JAH LB. Performed the experiments: MCHS YEG TB SBB CJ JAH HN. Analyzed the data: MCHS YEG TB SBB CJ JAH LB. Contributed reagents/materials/analysis tools: LB CSV. Wrote the paper: MCHS YEG LB.
- 1. Anonymous (2014) The European Union Summary Report, Trends and Sources of Zoonoses, Zoonotic Agents and Food-borne Outbreaks in 2012. EFSA Journal 12: 3547.
- 2. Carvalho CM, Gannon BW, Halfhide DE, Santos SB, Hayes CM et al. (2010) The in vivo efficacy of two administration routes of a phage cocktail to reduce numbers of Campylobacter coli and Campylobacter jejuni in chickens. BMC Microbiol 10: 232. pmid:20809975
- 3. El-Shibiny A, Scott A, Timms A, Metawea Y, Connerton P et al. (2009) Application of a group II Campylobacter bacteriophage to reduce strains of Campylobacter jejuni and Campylobacter coli colonizing broiler chickens. J Food Prot 72: 733–740. pmid:19435220
- 4. Loc CC, Atterbury RJ, El-Shibiny A, Connerton PL, Dillon E et al. (2005) Bacteriophage therapy to reduce Campylobacter jejuni colonization of broiler chickens. Appl Environ Microbiol 71: 6554–6563. pmid:16269681
- 5. Wagenaar JA, Van Bergen MA, Mueller MA, Wassenaar TM, Carlton RM (2005) Phage therapy reduces Campylobacter jejuni colonization in broilers. Vet Microbiol 109: 275–283. pmid:16024187
- 6. Fischer S, Kittler S, Klein G, Glunder G (2013) Impact of a Single Phage and a Phage Cocktail Application in Broilers on Reduction of Campylobacter jejuni and Development of Resistance. PLoS One 8: e78543. pmid:24205254
- 7. Atterbury RJ, Connerton PL, Dodd CE, Rees CE, Connerton IF (2003) Application of host-specific bacteriophages to the surface of chicken skin leads to a reduction in recovery of Campylobacter jejuni. Appl Environ Microbiol 69: 6302–6306. pmid:14532096
- 8. Goode D, Allen VM, Barrow PA (2003) Reduction of experimental Salmonella and Campylobacter contamination of chicken skin by application of lytic bacteriophages. Appl Environ Microbiol 69: 5032–5036. pmid:12902308
- 9. Orquera S, Golz G, Hertwig S, Hammerl J, Sparborth D et al. (2012) Control of Campylobacter spp. and Yersinia enterocolitica by virulent bacteriophages. J Mol Genet Med 6: 273–278. pmid:22872802
- 10. El-Shibiny A, Connerton PL, Connerton IF (2005) Enumeration and diversity of Campylobacters and bacteriophages isolated during the rearing cycles of free-range and organic chickens. Appl Environ Microbiol 71: 1259–1266. pmid:15746327
- 11. Coward C, Grant AJ, Swift C, Philp J, Towler R et al. (2006) Phase-variable surface structures are required for infection of Campylobacter jejuni by bacteriophages. Appl Environ Microbiol 72: 4638–4647. pmid:16820455
- 12. Scott AE, Timms AR, Connerton PL, Loc CC, Adzfa RK et al. (2007) Genome dynamics of Campylobacter jejuni in response to bacteriophage predation. PLoS Pathog 3: e119. pmid:17722979
- 13. Holst Sørensen MC, van Alphen LB, Fodor C, Crowley SM, Christensen BB et al. (2012) Phase variable expression of capsular polysaccharide modifications allows Campylobacter jejuni to avoid bacteriophage infection in chickens. Front Cell Infect Microbiol 2: 11. pmid:22919603
- 14. Sørensen MC, van Alphen LB, Harboe A, Li J, Christensen BB et al. (2011) Bacteriophage F336 recognizes the capsular phosphoramidate modification of Campylobacter jejuni NCTC11168. J Bacteriol 193: 6742–6749. pmid:21965558
- 15. Baldvinsson SB, Sørensen MCH, Vegge CS, Clokie MMR, Brøndsted L (2014) Campylobacter jejuni motility is required for infection of the flagellotropic bacteriophage F341. Appl Environ Microbiol: 80: 7096–106. pmid:25261508
- 16. McKay D, Fletcher J, Cooper P, Thomson-Carter FM (2001) Comparison of two methods for serotyping Campylobacter spp. J Clin Microbiol 39: 1917–1921. pmid:11326013
- 17. Penner JL, Aspinall GO (1997) Diversity of lipopolysaccharide structures in Campylobacter jejuni. J Infect Dis 176 Suppl 2: S135–S138. pmid:9396697
- 18. Chen YH, Poly F, Pakulski Z, Guerry P, Monteiro MA (2008) The chemical structure and genetic locus of Campylobacter jejuni CG8486 (serotype HS:4) capsular polysaccharide: the identification of 6-deoxy-D-ido-heptopyranose. Carbohydr Res 343: 1034–1040. pmid:18346720
- 19. Gilbert M, Mandrell RE, Parker CT, Li J, Vinogradov E (2007) Structural analysis of the capsular polysaccharide from Campylobacter jejuni RM1221. Chembiochem 8: 625–631. pmid:17335095
- 20. Kanipes MI, Papp-Szabo E, Guerry P, Monteiro MA (2006) Mutation of waaC, encoding heptosyltransferase I in Campylobacter jejuni 81–176, affects the structure of both lipooligosaccharide and capsular carbohydrate. J Bacteriol 188: 3273–3279. pmid:16621820
- 21. Karlyshev AV, Champion OL, Churcher C, Brisson JR, Jarrell HC et al. (2005) Analysis of Campylobacter jejuni capsular loci reveals multiple mechanisms for the generation of structural diversity and the ability to form complex heptoses. Mol Microbiol 55: 90–103. pmid:15612919
- 22. McNally DJ, Jarrell HC, Li J, Khieu NH, Vinogradov E et al. (2005) The HS:1 serostrain of Campylobacter jejuni has a complex teichoic acid-like capsular polysaccharide with nonstoichiometric fructofuranose branches and O-methyl phosphoramidate groups. FEBS J 272: 4407–4422. pmid:16128810
- 23. Muldoon J, Shashkov AS, Moran AP, Ferris JA, Senchenkova SN et al. (2002) Structures of two polysaccharides of Campylobacter jejuni 81116. Carbohydr Res 337: 2223–2229. pmid:12433486
- 24. St MF, Szymanski CM, Li J, Chan KH, Khieu NH et al. (2002) The structures of the lipooligosaccharide and capsule polysaccharide of Campylobacter jejuni genome sequenced strain NCTC 11168. Eur J Biochem 269: 5119–5136. pmid:12392544
- 25. McNally DJ, Jarrell HC, Khieu NH, Li J, Vinogradov E et al. (2006) The HS:19 serostrain of Campylobacter jejuni has a hyaluronic acid-type capsular polysaccharide with a nonstoichiometric sorbose branch and O-methyl phosphoramidate group. FEBS J 273: 3975–3989. pmid:16879613
- 26. Szymanski CM, Michael FS, Jarrell HC, Li J, Gilbert M et al. (2003) Detection of conserved N-linked glycans and phase-variable lipooligosaccharides and capsules from Campylobacter cells by mass spectrometry and high resolution magic angle spinning NMR spectroscopy. J Biol Chem 278: 24509–24520. pmid:12716884
- 27. McNally DJ, Lamoureux MP, Karlyshev AV, Fiori LM, Li J et al. (2007) Commonality and biosynthesis of the O-methyl phosphoramidate capsule modification in Campylobacter jejuni. J Biol Chem 282: 28566–28576. pmid:17675288
- 28. van Alphen LB, Wenzel CQ, Richards MR, Fodor C, Ashmus RA et al. (2014) Biological roles of the O-methyl phosphoramidate capsule modification in Campylobacter jejuni. PLoS One 9: e87051. pmid:24498018
- 29. Connerton PL, Timms AR, Connerton IF (2011) Campylobacter bacteriophages and bacteriophage therapy. J Appl Microbiol 111: 255–265. pmid:21447013
- 30. Atterbury RJ, Connerton PL, Dodd CE, Rees CE, Connerton IF (2003) Isolation and characterization of Campylobacter bacteriophages from retail poultry. Appl Environ Microbiol 69: 4511–4518. pmid:12902236
- 31. Carvalho C, Susano M, Fernandes E, Santos S, Gannon B et al. (2010) Method for bacteriophage isolation against target Campylobacter strains. Lett Appl Microbiol 50: 192–197. pmid:20002571
- 32. Connerton PL, Loc Carrillo CM, Swift C, Dillon E, Scott A et al. (2004) Longitudinal study of Campylobacter jejuni bacteriophages and their hosts from broiler chickens. Appl Environ Microbiol 70: 3877–3883. pmid:15240258
- 33. El-Shibiny A, Connerton PL, Connerton IF (2005) Enumeration and diversity of Campylobacters and bacteriophages isolated during the rearing cycles of free-range and organic chickens. Appl Environ Microbiol 71: 1259–1266. pmid:15746327
- 34. Frost JA, Kramer JM, Gillanders SA (1999) Phage typing of Campylobacter jejuni and Campylobacter coli and its use as an adjunct to serotyping. Epidemiol Infect 123: 47–55. pmid:10487641
- 35. Hansen VM, Rosenquist H, Baggesen DL, Brown S, Christensen BB (2007) Characterization of Campylobacter phages including analysis of host range by selected Campylobacter Penner serotypes. BMC Microbiol 7: 90. pmid:17945022
- 36. Hwang S, Yun J, Kim KP, Heu S, Lee S et al. (2009) Isolation and characterization of bacteriophages specific for Campylobacter jejuni. Microbiol Immunol 53: 559–566. pmid:19780969
- 37. Loc Carrillo CM, Connerton PL, Pearson T, Connerton IF (2007) Free-range layer chickens as a source of Campylobacter bacteriophage. Antonie Van Leeuwenhoek 92: 275–284. pmid:17387630
- 38. Owens J, Barton MD, Heuzenroeder MW (2013) The isolation and characterization of Campylobacter jejuni bacteriophages from free range and indoor poultry. Vet Microbiol 162: 144–150. pmid:22980913
- 39. Sails AD, Wareing DR, Bolton FJ, Fox AJ, Curry A (1998) Characterisation of 16 Campylobacter jejuni and C. coli typing bacteriophages. J Med Microbiol 47: 123–128. pmid:9879954
- 40. Javed MA, Ackermann HW, Azeredo J, Carvalho CM, Connerton I et al. (2013) A suggested classification for two groups of Campylobacter myoviruses. Arch Virol. 10.1007/s00705-013-1788-2 [doi].
- 41. Hammerl JA, Jackel C, Reetz J, Beck S, Alter T et al. (2011) Campylobacter jejuni group III phage CP81 contains many T4-like genes without belonging to the T4-type phage group: implications for the evolution of T4 phages. J Virol 85: 8597–8605. pmid:21697478
- 42. Ackermann HW (2009) Basic phage electron microscopy. Methods Mol Biol 501: 113–126. pmid:19066816
- 43. Sørensen MCH (2011) Characterization of initial interactions between Campylobacter jejuni and bacteriophages. PhD Thesis, University of Copenhagen.
- 44. Brathwaite KJ, Siringan P, Moreton J, Wilson R, Connerton IF (2013) Complete Genome Sequence of Universal Bacteriophage Host Strain Campylobacter jejuni subsp. jejuni PT14. Genome Announc 1. pmid:24265498
- 45. Bigwood T, Hudson JA, Billington C, Carey-Smith GV, Heinemann JA (2008) Phage inactivation of foodborne pathogens on cooked and raw meat. Food Microbiol 25: 400–406. pmid:18206783
- 46. Choi Y, Shin H, Lee JH, Ryu S (2013) Identification and characterization of a novel flagellum-dependent Salmonella-infecting bacteriophage, iEPS5. Appl Environ Microbiol 79: 4829–4837. pmid:23747700
- 47. Yen JY, Broadway KM, Scharf BE (2012) Minimum Requirements of Flagellation and Motility for Infection of Agrobacterium sp Strain H13-3 by Flagellotropic Bacteriophage 7-7-1. Appl Environ Microbiol 78: 7216–7222. pmid:22865074
- 48. Atterbury RJ, Dillon E, Swift C, Connerton PL, Frost JA et al. (2005) Correlation of Campylobacter bacteriophage with reduced presence of hosts in broiler chicken ceca. Appl Environ Microbiol 71: 4885–4887. pmid:16085889
- 49. Grajewski BA, Kusek JW, Gelfand HM (1985) Development of a bacteriophage typing system for Campylobacter jejuni and Campylobacter coli. J Clin Microbiol 22: 13–18. pmid:4019737
- 50. Khakhria R, Lior H (1992) Extended phage-typing scheme for Campylobacter jejuni and Campylobacter coli. Epidemiol Infect 108: 403–414. pmid:1601075
- 51. Salama S, Bolton FJ, Hutchinson DN (1989) Improved Method for the Isolation of Campylobacter jejuni and Campylobacter coli Bacteriophages. Letters in Applied Microbiology 8: 5–7.
- 52. Scott AE, Timms AR, Connerton PL, El-Shibiny A, Connerton IF (2007) Bacteriophage influence Campylobacter jejuni types populating broiler chickens. Environ Microbiol 9: 2341–2353. pmid:17686030
- 53. Tsuei AC, Carey-Smith GV, Hudson JA, Billington C, Heinemann JA (2007) Prevalence and numbers of coliphages and Campylobacter jejuni bacteriophages in New Zealand foods. Int J Food Microbiol 116: 121–125. pmid:17276534
- 54. Carvalho CM, Kropinski AM, Lingohr EJ, Santos SB, King J et al. (2012) The genome and proteome of a Campylobacter coli bacteriophage vB_CcoM-IBB_35 reveal unusual features. Virol J 9: 35. pmid:22284308
- 55. Hammerl JA, Jackel C, Reetz J, Hertwig S (2012) The complete genome sequence of bacteriophage CP21 reveals modular shuffling in Campylobacter group II phages. J Virol 86: 8896. pmid:22843857
- 56. Kropinski AM, Arutyunov D, Foss M, Cunningham A, Ding W et al. (2011) Genome and proteome of Campylobacter jejuni bacteriophage NCTC 12673. Appl Environ Microbiol 77: 8265–8271. pmid:21965409
- 57. Bigwood T, Hudson JA (2009) Campylobacters and bacteriophages in the surface waters of Canterbury (New Zealand). Lett Appl Microbiol 48: 343–348. pmid:19187501
- 58. Rakhuba DV, Kolomiets EI, Dey ES, Novik GI (2010) Bacteriophage receptors, mechanisms of phage adsorption and penetration into host cell. Pol J Microbiol 59: 145–155. pmid:21033576
- 59. Shin H, Lee JH, Kim H, Choi Y, Heu S et al. (2012) Receptor diversity and host interaction of bacteriophages infecting Salmonella enterica serovar Typhimurium. PLoS One 7: e43392. pmid:22927964
- 60. Kanipes MI, Tan X, Akelaitis A, Li J, Rockabrand D et al. (2008) Genetic analysis of lipooligosaccharide core biosynthesis in Campylobacter jejuni 81–176. J Bacteriol 190: 1568–1574. pmid:18156268
- 61. Hofreuter D, Tsai J, Watson RO, Novik V, Altman B et al. (2006) Unique features of a highly pathogenic Campylobacter jejuni strain. Infect Immun 74: 4694–4707. pmid:16861657
- 62. Fouts DE, Mongodin EF, Mandrell RE, Miller WG, Rasko DA et al. (2005) Major structural differences and novel potential virulence mechanisms from the genomes of multiple campylobacter species. PLoS Biol. 3(1):e15. pmid:15660156