Campylobacter jejuni - Wiley Online Library

Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

Genetic instability is associated with changes in the colonization potential of Campylobacter jejuni in the avian intestine A.M. Ridley1, M.J. Toszeghy1, S.A. Cawthraw1, T.M. Wassenaar2 and D.G. Newell1 1 Veterinary Laboratories Agency (Weybridge), New Haw, Surrey, United Kingdom 2 Molecular Microbiology and Genomics Consultants, Zotzenheim, Germany

Keywords Campylobacter, genotyping, instability. Correspondence Anne Ridley, Department of Food and Environmental Safety, Veterinary Laboratories Agency (Weybridge), Woodham Lane, New Haw, Addlestone, Surrey, KT15 3NB, UK. E-mail: [email protected]

2007 ⁄ 1474: received 11 September 2007, revised 23 November 2007 and accepted 8 December 2007 doi:10.1111/j.1365-2672.2008.03759.x

Abstract Aims: A panel of pulsed field gel electrophoresis (PFGE) type variants of Campylobacter jejuni, previously identified as of clonal origin, were investigated to determine whether genomic instability could be observed during competitive growth. Methods and Results: Upon recovery from frozen storage, some variants had undergone alterations in PFGE profiles, but subsequent culture produced constant genotypes. Individual variants did not display differences in colonization potential when tested in orally challenged 1-day-old chickens. However, competitive colonization using mixtures of two or three PFGE types generally resulted, by 4 weeks postchallenge, in one predominant PFGE type in all birds. For some variant mixtures, a minor population of novel PFGE types was detected in individual birds. The creation of new variants appeared to be dependent on the extent of competition and of the individual host. Genomic rearrangements most likely explain this increase in genetic diversity, apparently without the involvement of natural transformation or plasmid acquisition. In vitro cultivation of mixed inoculations were again selected for particular variants; but genetic diversity was not generated, suggesting that the selection pressures in vitro differed from those active in vivo. Conclusion: These observations support the hypothesis that by generating genetic diversity, C. jejuni can improve its phenotypic fitness to survive and colonize subsequent hosts. Significance and Impact of the Study: The consequences of such observations for the development of campylobacter control strategies for poultry may be substantial.

Introduction Campylobacter jejuni is a major cause of human acute bacterial enteritis. Although this pathogen is considered to be largely food-borne, C. jejuni is ubiquitous in the environment and accurate source attribution is thus very important. For other organisms, the identification of sources relies on tracking strains through the environment. Such tracking is enabled by molecular epidemiology of strains identified by phenotypic or genotypic properties. The diversity of C. jejuni is well established

and detectable by both genotypic and phenotypic techniques and many such methods are routinely employed (Wassenaar and Newell 2000). However, the value of such molecular approaches in campylobacter epidemiology appears to be restricted by genomic instability (Wassenaar et al. 2000b). Genomic instability generates a weakly clonal bacterial population structure (Dingle et al. 2001) and is the result of genetic events such as intragenomic rearrangements and natural transformation followed by homologous recombination (Wassenaar et al. 2000b). Results of the latter mechanism have been observed in

ª 2008 The Authors Journal compilation ª 2008 The Society for Applied Microbiology, Journal of Applied Microbiology 105 (2008) 95–104

95

Genetic instability in Campylobacter jejuni

wild-type populations at the single locus level, in particular of the flagellin genes (Harrington et al. 1997; Meinersmann and Hiett 2000). Transformation followed by homologous recombination has been modelled in vitro by co-cultivation of isogenic mutants (Wassenaar et al. 1995) and in vivo by co-colonization of the chicken gut between genetically distant, naturally competent strains (de Boer et al. 2002). Localized recombinations (involving the fla locus) have also been observed to occur during chicken colonization (Nuijten et al. 2000), apparently as a result of natural transformation. Genomic instability of C. jejuni at the entire genome level, and detectable by pulsed field gel electrophoresis (PFGE), is also known to occur during the colonization of chickens (Wassenaar et al. 1998; Hanninen et al. 1999; de Boer et al. 2002). The fact that novel PFGE variants may be detected in experimentally colonized chickens or in contaminated carcasses suggests that there are, as yet unknown, selective benefits to such novel subpopulations in such environmental conditions. It can be hypothesized that the creation of genetic diversity increases the survival opportunities for the offspring generation to overcome bottlenecks, such as the environmental stresses experienced during host-to-host transmission. Such bottlenecks are frequently encountered by campylobacters that cannot naturally multiply outside a host. Such stresses can include exposure to atmospheric oxygen, nutrient deprivation, desiccation and temperature extremes. On challenge of a new host, novel environmental stresses (temperature changes, gastric acidity and digestive enzymes) will also be encountered before colonization and clonal expansion can occur. Such stresses require rapid adaptation to changed environments, however there is a paucity of known adaptive mechanisms in the genome of C. jejuni (Parkhill et al. 2000). Moreover, the relatively small genome size of C. jejuni (only 40% of the total number of genes present in Salmonella) excludes the possibility of the evolution of other mechanisms for adaptation based on additional genetic content. Thus, alternative strategies to adaptation may have evolved in C. jejuni. One such strategy is the ability to undergo genetic instability during growth and thereby create genetic diversity. Such a strategy would generate a spectrum of phenotypes, e.g. by relocating, and thus changing the expression levels of, genes or operons. This would generate a range of variants suitable for selection for survival. For this process to be advantageous, every stress response would have to result in the selection of variants, thus increasing the chance of survival of the fittest offspring. To date, the effects of genetic instability on campylobacter phenotypic properties, such as colonization potential, have not been investigated. In a previous study (Wassenaar et al. 1998), a group of clonally related isolates of C. jejuni were derived from a 96

A.M. Ridley et al.

single batch of poultry meat. These isolates were shown to be identical by Penner serotyping, fla-typing and AFLP (amplified fragment length polymorphism), but demonstrated minor differences in PFGE profiles. Based on these observations, it was hypothesized that these strains were variants originating from a single ancestor clone as a result of the stresses associated with poultry meat processing. The investigations indicated that this clone was susceptible to multiple genetic rearrangements resulting in genomic instability. In the current study this strain set was investigated further. First, attempts were made to induce genomic instability as detected by PFGE by mimicking the stresses of poultry processing. Then the strains were used in experimental competitive challenges in the chicken gut to demonstrate that genomic instability occurred and that certain variants were more successful colonizers than others under these conditions. Finally, attempts were made to correlate the in vivo results with an in vitro model of co-challenge. Materials and methods Bacterial strains The sources, genotypes and phenotypes of the panel of C. jejuni isolates investigated have been previously described (Wassenaar et al. 1998). These isolates were derived from multiple swabs (a–d) of multiple sequential packages (119, 121, 123, 124, 125 and 126) of a single batch of poultry meat. Each isolate had been cloned and individual colonies recovered and stored frozen at )80C in 10% glycerol in 1% protease peptone (glycerol broth). Cultures from these individual colonies were analysed for genotypic and phenotypic characteristics. All campylobacter isolates were cultured on blood agar containing selective antibiotics with actidione (100 lg ml)1) and cefoperazone (30 lg ml)1) (BASAC) for 24–48 h at 42C in a microaerobic atmosphere (8% O2, 7% CO2 and 85% N2). Genotyping by PFGE and fla PCR ⁄ RFLP analysis Chromosomal DNA was prepared from C. jejuni isolates cultured as before. The bacteria were harvested into phosphate-buffered saline (PBS) and lysed in 1Æ3% agarose blocks (Gibson et al. 1995). DNA was digested using 20 U SmaI. KpnI was used to confirm isolate identity where appropriate. PFGE was performed on a DRIII (BioRad, Hemel Hempstead, UK) apparatus at 6 V cm)1 for 24 h with pulse times increasing from 10 to 35 s (SmaI) or 2 to 20 s (KpnI).


A.M. Ridley et al.

Restriction fragment length polymorphism analysis of the polymerase chain reaction products (PCR-RFLP) of the fla genes was performed using genomic DNA extracted with a Puregene DNA isolation kit (Gentra Systems, Flowgen, Ashby-de la Zouch, UK) according to the manufacturer’s instructions. The method used to designate fla-type PCR-RFLP profiles has been described previously (Newell et al. 2001; Shreeve et al. 2002). The flaA SVR (short variable region) sequence analysis method of Meinersmann et al. (1997) was used to confirm homogeneity of this known variable region in selected isolates. Phenotyping Selected isolates were serotyped as described by Frost et al. (1998). These analyses were performed by the Campylobacter Reference Unit of the Health Protection Agency (HPA) under the supervision of J.A. Frost. In vitro stress model Individual variants were inoculated into prewarmed biphasic broths (25 ml), in duplicate, and incubated at 42C (the temperature of the avian gut) for c. 21 h. The bacteria were then exposed to temperature stresses to mimic poultry processing. The duplicate broths were mixed and divided into three aliquots of 15 ml. One aliquot was used as a control and was continued in culture at 42C under microaerobic conditions. The second aliquot was incubated at 54C (the approximate temperature of an abattoir dunk tank) for 5 min in a shaking preheated water-bath. The third aliquot was placed at 4C (the chilling temperature for carcasses) for 24 h. Following treatment, the cells were pelleted by centrifugation and re-suspended in 1 ml of PBS before cultivation on BASAC for colony isolation. Ten discrete single colonies, plus a plate sweep, were harvested from each treatment for characterization. Chick model colonization experiments The oral colonization model previously described (Wassenaar et al. 1993) was used with modifications. Briefly, 1-day-old SPF (specific pathogen free) chicks were dosed by gavage and the levels of caecal colonization determined at weeks 2 and 4 postchallenge. Preliminary studies using this model indicated that all isolates tested at doses of c. 106 CFU effectively colonized the chicks to similar levels. A competition colonization model was developed, in which chicks (6–17 per experiment) were dosed orally with an inoculum containing a mixture of 106 CFU of each of two or three variants, distinguishable by PFGE profile. Cloacal swabs were taken after 24–48 h, where feasible, to confirm colonization and per group up to two birds were sacrificed at 48–72 h postchallenge, caecal


contents recovered and cultured for campylobacters to confirm caecal colonization. On days 13 and 27, postchallenge birds were sacrificed and the recovered caecal contents was cultured as before. Up to 20 discrete single colonies, and three plate sweeps were recovered from each bird, and examined using the genotypic and phenotypic methods described before. Seeder bird model An additional model using a colonized seeder bird was used to mimic the conditions during natural transmission of infection. Three SPF chicks were orally dosed with an inoculum of 106 CFU of each of the isolates 123a and 123b. Once colonization was established, as identified by cloacal swabbing, one of these colonized birds was placed in a group of 14 uninfected chicks. The ‘in-contact’ birds were sacrificed at 13 and 27 days postchallenge, as well as one seeder bird per time point. Ten discrete colonies plus a plate sweep were harvested from each bird for characterization. In vitro competition assay The in vitro competitive exclusion model of Barrow and Page (2000) was used with modifications. Briefly, prewarmed biphasic broths (heart infusion; Oxoid, Basingstoke, UK) (25 ml) were inoculated with overnight cultures of mixtures of variants and incubated for 24– 30 h at 42C. The number of bacteria added to the broth for each variant was 105 CFU. Broths were inoculated with: (experiment 1) 1 ml of 123a followed by 1 ml of 123b added after 24 h; (experiment 2) 1 ml of 123b followed by 1 ml of 123a added after 24 h; and (experiment 3) 1 ml of 123a plus 1 ml of 123b simultaneously. Following incubation at 42C for 24 h, 1 ml of cell suspension was aseptically removed from each flask and plated on to BASAC for isolation and characterization of three individual colonies and a sweep sample. The remaining cell suspension was incubated for further 24 h at 42C and another sample was drawn and analysed as before. Results Twenty-one isolates of C. jejuni that were originally derived from a single batch of poultry meat (Wassenaar et al. 1998) were recovered following frozen storage and genomic DNA was analysed to confirm the previously described genotypic differences. All isolates remained a single fla type, 6, 13. However, 4 of the 21 isolates tested (119b, 123a, 125a and 126d) demonstrated a different SmaI PFGE genotype from the previously reported one (Wassenaar et al. 1998), indicating that DNA rearrange-


97


A.M. Ridley et al.

ment had occurred in these strains, presumably as a consequence of frozen storage and recovery. flaA SVR sequence analysis performed on five isolates (119a, 119b, 123a, 123b and 121d) produced a constant genotype, confirming that observed PFGE changes did not involve this known variable region, and once again demonstrated the common origin of these isolates. The genotypic stability of eight selected isolates, two of which had changed PFGE type upon frozen storage (119b and 123a) was tested following repeated culture on agar plates. The results showed that genotypes were stable during culture over 20 subpassages for all isolates. Next, heat and cold stress treatment were applied in vitro. Again, no change in

fla-type was observed (all remained fla 6, 13) and their individual PFGE profiles remained constant (Table 1), as tested on both single colonies and sweep cultures. The colonization potential in the chicken gut was next assessed for each individually selected isolates, applying the chick model as described in the ‘Materials and methods’ section. The fla-types and PFGE profiles remained constant for all isolates following colonization periods of 2 and 4 weeks (Table 1). All isolates colonized to the maximum expected levels (>107 CFU g)1 caecal contents) and no significant differences in colonization potential were observed between isolates. However, minor differences in colonization would be difficult to detect under such conditions. During co-challenge minor differences should be detectable by shifts in bacterial population distribution. Therefore, it was next investigated if mixtures of strains, distinguishable by differences in PFGE genotype, displayed minor differences in colonization efficiency that would result in a population distribution shift. Five combinations of isolates, four duets and a triplet (sets A–E, see Table 2), were selected on the basis of the diversity of their PFGE types, the commonality of certain PFGE types in the originally observed poultry batch (Wassenaar et al. 1998) and on the previously observed profile changes on cryo-storage. A summary of the outcomes of the oral challenges at 2 and 4 weeks postchallenge is presented in Table 2. In set A, a mixture of isolates 121c and 121d (SmaI PFGE types S4 and S1, respectively), was used to orally challenge 11 chicks. Plate sweeps made from caecal isolates recovered at 3 days postchallenge from two birds were of PFGE type S4. At 2 weeks postchallenge, 25 single-colony isolates were analysed (originating from five birds) as well as sweep cultures; all displayed PFGE type

Table 1 Effect of exposure to hot or cold stresses and in vivo passage through the chick model on the pulsed field gel electrophoretic (PFGE) types of selected isolates derived from multiple packages of a single-poultry meat batch

Meat package*

Isolate number

Control genotype

In vitro stress genotype

In vivo passage genotype

119

119a 119b 119c 121c 121d 123a 123b 123c

S1 S2 S3 S4 S1 S13 S12 S4

S1 S2 S3 ND ND S13 ND S4

S1 S2 S3 S4 S1 S13 S12 S4

121 123

*Meat package from which the isolate was originally generated (see Wassenaar et al. 1998). Genotype observed for sweep culture and single colonies after 4 weeks of colonization in chickens. The observed SmaI PFGE type after storage at –80C was different from the PFGE type originally reported (Wassenaar et al. 1998).

Table 2 Pulsed field gel electrophoretic (PFGE) profiles of isolates recovered from in vivo models at 4 weeks post-challenge in in vivo and in vitro models Predominant PFGE type* Strain set

Challenge isolates

Model

Two weeks post-challenge

Four weeks post-challenge

Other types recovered at 4 weeks

Set Set Set Set Set

121c (S4) + 121d (S1) 123a (S13) + 123c (S4) 119a (S1) + 119b (S2) + 119c (S3) 123a (S13) + 123b (S12) 121d (S1) + 123a (S13)

Oral Oral Oral Oral Oral

S4 S13 S3 S12, S13 ND

S4 S13 S3 S12, S13 S13

123a (S13) + 123b (S12) 123a (S13) + 123b (S12)

Seeder bird In vitro competition

S13; S12 S13; S12

S13 S12

S1 S4 (S12, single isolate) S1 S119, S131, S132, S133, S140 S1, S137, S136, S135, S138, S139, S141 S12, S134, S135, S139 S13

A B C D E

Set D Set D

challenge challenge challenge challenge challenge

*PFGE genotype of sweep culture following 2 and 4 weeks of colonization. ND, not determined.

98


A.M. Ridley et al.

S4. At 4 weeks postchallenge, 70 of the 75 recovered isolates from the remaining four birds were of PFGE type S4; the other five isolates, all recovered from a single bird, were of PFGE type S1. These data indicate that for this set of combined strains, S4 was dominantly colonizing over S1. For set B, a mixture of isolates 123a and 123c (PFGE types S13 and S4, respectively) was used to challenge 14 chicks. At 3 days postchallenge, plate sweeps recovered from the caeca of two birds indicated predominant colonization by strains of PFGE type S13. Typing of recovered isolates from six birds sacrificed 2 weeks postchallenge again identified S13 as the dominant type, although S4 was recovered in three cases. Interestingly, a single isolate recovered at week 2 was of PFGE type S12, different from either of the challenge isolates. At 4 weeks postchallenge S13 was the only PFGE type recovered from the six remaining birds. These observations suggest that the S13 variant dominates over S4 during colonization. Twelve birds were orally dosed with an equal mixture of the three isolates 119a, 119b and 119c (set C). Recovered isolates and sweeps from two birds sacrificed at 3 days postchallenge were all of PFGE type S3. The predominance of this PFGE type persisted at 2 weeks (15 ⁄ 15 isolates or plate sweeps tested) and 4 weeks (55 of 60 isolates tested were S3) postchallenge. The residual five isolates at 4 weeks postchallenge, again all recovered from a single bird, were of PFGE type S1. Novel PFGE genotypes were not observed, despite the instability of the original 119a isolate upon cryo-storage. With strains from set D, containing a mixture of 123a and 123b (PFGE types S13 and S12, respectively), isolates of PFGE type S13 predominated (11 ⁄ 15 colonies tested) in three birds at 2 weeks postchallenge. However, sweeps of bacterial growth made at this time comprised a mixture of PFGE types S12 and S13 with a more intense banding pattern representing the S13 PFGE type, suggesting the dominance of this genotype in the caeca (data not shown). By 4 weeks, isolates of PFGE type S13 predominated (10 ⁄ 10 colonies tested) in only one of the six birds investigated. In three birds, PFGE type S12 was recovered from all 40 colonies investigated; from the fourth bird a mixture of isolates of PFGE types S12 and S13 were recovered with the S12 genotype dominating. Thus, dominance of S13 over S12 was not as strong as over S4 observed for set B. The final bird’s caecal content from set D produced a surprising result: apart from the expected PFGE types S12 and S13, five novel PFGE types were recovered from individual colonies. These were designated S119, S131, S132, S133 and S140 and are shown in Fig. 1. DNA preparations were repeated to confirm these findings, with the same result. Moreover, PFGE analysis using KpnI showed corresponding banding


1

2

3

4

5

6

7

8

9

10

11

kb 486 389

291

194 145·5

Figure 1 Representatives of pulsed field gel electrophoresis (PFGE) types used to challenge chicks and recovered following oral gavage with a mixture of strains 123a and 123b (strain set D). Lane 1, dose strain 123a (S13); lane 2, dose strain 123b (S12); lane 6, lambda ladder marker; lanes 3–5 and 7–11, selection of PFGE types recovered at 4 weeks (S12, S140, S131 and S131, S119, S132, S133, S13).

differences in these recovered isolates (not shown). By fla typing, all these isolates were of the same fla-type 6, 13 as all other isolates (not shown). As expected, flaA SVR genotypes of two representative types were indistinguishable to those of the parental isolates. A final set E was challenged with 121d and 123a (PFGE types S1 and S13, respectively); based on the previous results dominance of the S13 type was expected and this was indeed the case. At week 4, 53 ⁄ 60 isolates were S13 and in two birds this was the only genotype detected. However, as was observed for set D, novel PFGE types were identified at 4 weeks postchallenge. Two birds were predominantly colonized with S13 with an S1 minority as well as novel genotypes that differed from those discovered in set D. One bird was predominantly colonized with novel genotype S137 (16 ⁄ 20 colonies) (Fig. 2). Isolates belonging to this new PFGE type remained fla-type 6, 13 and flaA SVR sequence analysis performed on six of the PFGE variant isolates confirmed homogeneity in this region. All observed genotypes and serotypes of parental strains and offspring are summarized in Table 3. The observed changes in serotype are relatively minor, and would have remained undetected by the Penner serotyping procedure (Oza et al. 2002). To investigate whether the same effects were detectable using a model more closely mimicking within-flock transmission, seeder birds were employed. In this model, three seeder birds were orally challenged with the isolates of set D (isolates 123a and 123b), which had resulted in nearly


99


1

2

3

4

5

6

7

8

9 10 11 12 13 14

A.M. Ridley et al.

kb

389

Table 3 Genotypic and phenotypic properties of parent strains and recovered isolates Strain set

Isolate number (colony number)

Source

PFGE type

Serotype

Set A

121d 121c (1) (2) 123a 123c (1) (2) (3) 119a 119b 119c (1) 123a 123b (1) (2) (3) (4) (5) (6) (1) (2) (3) (4) (5) (6) (7) 123a 121d (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Challenge Challenge Caeca Caeca Challenge Challenge Caeca Caeca Caeca Challenge Challenge Challenge Caeca Challenge Challenge Caeca Caeca Caeca Caeca Caeca Caeca Seeder Seeder Seeder Seeder Seeder Seeder Seeder Challenge Challenge Caeca Caeca Caeca Caeca Caeca Caeca Caeca Caeca Caeca Caeca

S1 S4 S4 S1 S13 S4 S13 S13 S4 S1 S2 S3 S3 S13 S12 S131 S132 S12 S13 S133 S12 S13 S134 S135 S139 S135 S12 S12 S13 S1 S137 S13 S138 S119 S1 S136 S139 S12 S12 S1 ⁄ S137

13 18 13 13 UT UT 9 9 UT[9]* 13 18 13 18 5 ⁄ UT 8[13]* 5 ⁄ 13 5 ⁄ 13 13 ⁄ UT 13 ⁄ UT 9 5 5 13 UT 5 13 18 9 UT 13[5]* 9 9 18 UT 9 18 UT UT 18 UT

291

194

Set B

97

Set C

Figure 2 Pulsed field gel electrophoresis (PFGE) showing novel genotype S137, which exclusively colonized one chick from set E. Lanes 1, 3, 9, 14, lambda ladder marker; lane 1, dose strain 121d (S1); lane 3, dose strain 123a (S13); lanes 4–8, recovered isolates representing PFGE type S13; lanes 10–13, recovered isolates representing PFGE type S137.

equal colonization and gave rise to novel genotypes. Once colonization of both birds was established (at 48 h postchallenge), one of the seeder birds was placed among a group of 14 uninfected chicks of the same age. Colonization of all birds was detected within 5 days of the placement of the seeder bird (not shown). At 2 weeks postchallenge with the seeder bird, three of the five birds tested were predominately (23 ⁄ 30 colonies tested) colonized by isolates of PFGE type S13, the remainder were S12; in the other two birds S12 dominated over S13 (13 ⁄ 20 colonies tested). Sweeps of the recovered bacteria confirmed the presence of both genotypes in all birds. At 4 weeks postchallenge, the seeder bird was sacrificed and both a sweep of the recovered bacteria and 10 individual colonies examined were all of PFGE type S13. Similarly, isolates of PFGE type S13 were predominantly recovered from the caecal contents of the remaining nine in-contact birds, with a minority of S12 (Table 2 and Fig. 3). However, 6 of the 11 colonies examined from one of these birds represented three new PFGE types (S134, S135 and S139) (Fig. 4). Distinct KpnI PFGE genotypes were also obtained for these isolates with novel SmaI PFGE types (Fig. 4). All these isolates remained fla 6, 13 (not shown). Two of the newly arisen variants (S135 and S139) had also been observed in the oral challenge model using a mixture of 121d and 123a (set E; Table 2), which suggests that these two variants were derived from 123a (S13), the only variant present in both experiments. The outcome of these colonization experiments clearly demonstrate an order of dominance during competitive colonization for the investigated isolates, with S13 dominating over S4 (as in set B) and over S1 (as in set E), S4 dominating over S1 (as in set A) and S13 showing a 100

Set D

Set E

*Mixed reaction. Tested in duplicate. Genotype comprising a mix of both profiles. UT, untypable.

slightly higher colonization potential than S12 (set D). The outcome of the bacterial population in the latter case seemed to be dependent on the individual bird. A significant finding is that novel PFGE genotypes can be formed and be selected in the chicken gut, but only following challenge with more than one genotype isolate, as all genotypes were found stable following single colonization


A.M. Ridley et al.

1

2

3

4


5

6

7

8

9 10 11 12 13 14 15 kb

291

194

97

Figure 3 Representative pulsed field gel electrophoretic (PFGE) profiles observed after in vivo exposure with strain set D via the seeder bird model. Lanes 1 and 15, lambda ladder marker; lanes 2–9, selected caecal isolates from the birds investigated at 4 weeks postchallenge with the infected seeder bird; lanes 2, 5, 6 and 7, PFGE type S13; lanes 3 and 8, PFGE type S135; lane 4, S139; lane 9, S134; lane 10, sweep recovered from bird colonized with multiple PFGE types; lanes 11, 12 and 14, isolates of PFGE type S13; lane 13, PFGE type S12, all recovered from the second bird.

1

2

3

4

5

6

7

8

9

10 kb

291

194

145·5

97

48·5

Figure 4 KpnI macrorestriction profiles of corresponding to SmaI pulsed field gel electrophoresis (PFGE) types recovered observed after in vivo exposure via the seeder bird model. Lanes 1 and 10, lambda ladder marker; lane 2, isolate from the seeder bird, K12 (S13); lanes 3–9, selected isolates from the first and second birds investigated at 4 weeks postchallenge; lane 3, K22 (S134); lanes 4 and 5, K23 (S135); lane 6, K24 (S139); lanes 7–9, PFGE type K12 (S13).

(Table 1). Novel genotypes were identified in sets D and E (and, in a single colony, in set B), which all shared the presence of 123a in the inoculum. Interestingly, this variant demonstrated changes following frozen storage. The trigger for the formation and ⁄ or selection of novel

genotypes could not be identified from these experiments, but competition seemed to provide a clue. We therefore tried to mimic competition in an in vitro model. Strains 123a and 123b (set D) were first co-cultured on agar plates in vitro for up to 48 h. When 30 colonies were produced from the sweep culture, 23 ⁄ 30 colonies tested were S12 and the other 7 were S13. Novel PFGE types were not detected (not shown). Next, liquid cultures were established for 24 h of either 123a or 123c; the competitor isolate was then added to give an equal ratio and the mixture incubated for another 30 h. A simultaneous inoculum of 123a and 123b was also tested. In all of these experiments, S12 and S13 were recovered but novel genotypes were not observed (results not shown). Apparently, co-cultivation and competition for nutrients was not sufficient, during the period of experimentation, to induce or select for novel genotypes in vitro. Discussion Previous investigations (Wassenaar et al. 1998) described a set of isolates of C. jejuni, derived from a single batch of poultry meat, apparently of the same clonal origin but having undergone genomic rearrangements detectable by differences in PFGE while the fla type, flaA SVR sequence and Penner serotype remained constant. The flaA SVR sequence has been submitted to Genbank (EU292736). The isolates also shared an identical AFLP genotype (T. Wassenaar and B. Duim 2000; unpublished data). Other observations have also described genomic instability in C. jejuni detectable by PFGE (Hanninen et al. 1999; de Boer et al. 2002). Even in outbreaks of campylobacteriosis, PFGE patterns from multiple isolates can demonstrate changes (Barton et al. 2007) suggesting genomic instability. PFGE genotyping has been applied to other food-borne pathogens with such success that it is now regarded as a gold standard for Salmonella typing (Fakhr et al. 2005). It is therefore very important to investigate under which conditions PFGE genotypes of Campylobacter might change. Our data suggest that the genome of C. jejuni, although apparently stable during caecal colonization of individual strains, can undergo changes detectable by differences in PFGE profiles, upon competitive stress in the avian gut. We hypothesize that this is a reflection of subpopulation selection from a genetically diverse parent population by environmental stress and is a mechanism evolved for survival in hostile environments. Campylobacter seem to have evolved a wide repertoire of mechanisms to produce genetic diversity, including natural competence to permit uptake and incorporation of chromosomal DNA, slip-strand mutations in homopolymeric tracts (Wassenaar et al. 2002) and genomic rearrangements as detectable by PFGE. The biological


101


advantage of such genetic diversity at least in terms of colonization potential has become evident from our studies. It seems likely that this diversity enables the generation of multiple phenotypic variants that are capable of displaying a wide range in fitness to survive and overcome environmentally induced evolutionary bottlenecks. So far, phenotypic variation could only be detected by the LEP (HPA Laboratory of Enteric Pathogens) serotyping procedure but the relevance to colonization potential of the detected variation is at present unclear. By extrapolation, where environmental stresses are very hostile to the parent, such variants may out-compete that parent. The key to this hypothetic survival mechanism is, however, to generate sufficient diversity during conditions that allow bacterial growth, so that at least one subpopulation is likely to survive the next environmental bottleneck. However, modelling these competitive stresses in vitro does result in similar changes. Thus, multiple isolates have to compete in the avian gut in order to detect bacterial subpopulations that have undergone genetic changes. Whether the environment of the avian gut triggers such changes, or selects for the produced variants, cannot be determined from these experiments. Changes in PFGE banding patterns can be caused by point mutations affecting the restriction sites of the enzyme used, by recombination within a genome such as reversions, deletions or translocations, or by nonhomologous recombination with the incoming DNA following transformation. Although the precise mechanism by which the observed novel variants were produced cannot be established, several possibilities can be excluded. Point mutations (or changes in methylation patterns affecting restriction enzyme recognition sites) are not responsible for the observed changes in PFGE patterns as similar changes in patterns were obtained with a second restriction enzyme, KpnI. In addition all of the challenge isolates used in our study lacked competence for DNA uptake under laboratory conditions (unpublished observations), which makes it unlikely that natural transformation was involved in the process. Moreover, none of the parent or variant-type isolates possessed plasmids, a finding consistent with the nature of PFGE band variations detected. Temperate bacteriophages that have been shown to affect PFGE genotypes induce relatively minor changes in PFGE banding patterns (Barton et al. 2007), but the substantive changes in banding patterns observed here suggest large and extensive genome rearrangements. Such rearrangements have been described for Salmonella typhi in which the rRNA loci are the apparent loci for recombination (Liu et al. 1996). Campylobacter jejuni contains only three rRNA loci, which limits the possibilities to create recombinant variants. It is unlikely that homologous recombination between the rRNA loci is the mechanism 102

A.M. Ridley et al.

responsible for creation of all variants described here. Possibly, repeat sequences are involved in the process of recombination although global repeats are rare in the C. jejuni genome sequences that are available to date (David W. Ussery, personal communication). Therefore, further investigations are required to resolve the mechanism that causes this observed instability. None of the changes appeared to involve recombination in flagellin genes as indicated by the consistency of the fla type and flaA SVR sequences of selected parent and variant isolates throughout. This is also consistent with the findings of de Boer et al. (2002), who reported intrinsic instability in the C. jejuni PFGE genotype in vivo but which did not affect the flaA gene or any of the house-keeping genes analysed using multilocus sequence typing (MLST). Despite this advanced repertoire to produce genetic diversity, C. jejuni is not a panmictic species (Dingle et al. 2001). Several strains of C. jejuni have been isolated, from a variety of hosts and locations, which remain genetically stable over long periods of time (Wassenaar et al. 2000a; Manning et al. 2001). Most likely, observations on genetic instability made with one strain, or isolate, do not apply to the entire bacterial population. Notably, when C. jejuni 81116 was tested in our competition chick model, its PFGE genotype (by KpnI, as this strain is SmaI-resistant) remained constant (unpublished data). Strain 81116 belongs to a clone of isolates that has been stable for over 14 years (Manning et al. 2001). The dominance in our study of particular variants during competitive colonization is intriguing and suggests that some variants have a greater capacity to survive the environmental pressures and colonize the host gut than others. Even more intriguing is the fact that the variant recovered was dependant on the individual host, suggesting the presence of a host-specific stress. In the noninbred chickens used such stresses could be a particular immune status or gut flora composition. Similar bird-dependant observations were reported when heterologous strains compete during gut colonization in chicks (Konkel et al. 2007) and when transposon mutant libraries of the parent strain C. jejuni 81176 were screened (Grant et al. 2005). One explanation of the generation of novel variants in vivo would be recombination between strains. Certainly, this was reported by de Boer et al. (2002); however, in the study of Konkel et al. (2007) no variant PFGE types were observed, suggesting lack of recombination between the competing strains. Despite the evidence for the generation of novel PFGE types during the in vivo competitive growth, no new types were observed from the in vitro model using the same variant mixture. The static culture method of Barrow and Page (2000) although able to mimic some aspects of


A.M. Ridley et al.

competitive growth of Campylobacter in vivo, clearly cannot reproduce the environmental pressures resulting in the observation of genetic instability and subpopulation selection. This suggests that the selective pressures are unrelated to just nutrient deprivation. The future identification of such selective pressures important in colonization may contribute to the development of targeted interventions for campylobacters in chickens. The observed differences in the capacity of PFGE-type variants of C. jejuni apparently from a single parent clone to colonize chickens in a competitive environment provides supporting evidence for the hypothesis that genetic and phenotypic diversity may be an important mechanism by which some C. jejuni strains can improve fitness to survive and colonize another host. The consequences of such observations for the development of campylobacter control strategies for poultry may be substantial. The successful application of strategies such as competitive exclusion, probiotics, bacteriophage and even vaccination, may all be affected by variation in gene expression consequent to genomic instability, especially if this is dependent on individual hosts. Acknowledgement The work described in this paper was funded by the Department for Environment, Food and Rural Affairs (Defra), United Kingdom (OZ0605). References Barrow, P.A. and Page, K. (2000) Inhibition of colonisation of the alimentary tract in young chickens with Campylobacter jejuni by pre-colonisation with strains of C. jejuni. FEMS Microbiol Lett 182, 87–91. Barton, C., Ng, L.K., Tyler, S.D. and Clark, C.G. (2007) Temperate bacteriophages affect pulsed-field gel electrophoresis patterns of Campylobacter jejuni. J Clin Microbiol 45, 386–391. de Boer, P., Wagenaar, J.A., Achterberg, R.P., van Putten, J.P.M., Schouls, L.M. and Duim, B. (2002) Generation of Campylobacter jejuni genetic diversity in vivo. Mol Microbiol 44, 351–359. Dingle, K.E., Colles, F.M., Wareing, D.R.A., Ure, R., Fox, A.J., Bolton, F.J., Bootsma, H.J., Willems, R.J. et al. (2001) Multilocus sequence typing system for Campylobacter jejuni. J Clin Microbiol 39, 14–23. Fakhr, M.K., Nolan, L.K. and Logue, C.M. (2005) Multilocus sequence typing lacks the discriminatory ability of pulsedfield gel electrophoresis for typing Salmonella enterica serovar Typhimurium. J Clin Microbiol 43, 2215–2219. Frost, J.A., Oza, A.N., Thwaites, R.T. and Rowe, B. (1998) Serotyping scheme for Campylobacter jejuni and Campylo-


bacter coli based on direct agglutination of heat stable antigens. J Clin Microbiol 36, 335–339. Gibson, J.R., Fitzgerald, C. and Owen, R.J. (1995) Comparison of PFGE, ribotyping and phage-typing in the epidemiological analysis of Campylobacter jejuni serotype HS2 infections. Epidemiol Infect 115, 215–225. Grant, A.J., Coward, C., Jones, M.A., Woodall, C.A., Barrow, P.A. and Maskell, D.J. (2005) Signature-tagged transposon mutagenesis studies demonstrate the dynamic nature of cecal colonization of 2-week-old chickens by Campylobacter jejuni. Appl Environ Microbiol 71, 8031–8041. Hanninen, M.-L., Hakkinen, M. and Rautelin, H. (1999) Stability of related human and chicken Campylobacter jejuni genotypes after passage through chick intestine studied by pulsed-field gel electrophoresis. Appl Environ Microbiol 65, 2272–2277. Harrington, C.S., Thomson-Carter, F.M. and Carter, P.E. (1997) Evidence for recombination in the flagellin locus of Campylobacter jejuni: implications for the flagellin gene typing scheme. J Clin Microbiol 35, 2386–2392. Konkel, M.E., Christensen, J.E., Dhillon, A.S., Lane, A.B., Hare-Stanford, R., Schaberg, D.M. and Larson, C.L. (2007) Campylobacter jejuni strains compete for colonization in broiler chicks. Appl Environ Microbiol 73, 2297–2305. Liu, S.L. and Sanderson, K.E. (1996) Highly plastic chromosomal organization in Salmonella typhi. Proc Natl Acad Sci USA 93, 10303–10308. Manning, G., Duim, B., Wassenaar, T., Wagenaar, J.A., Ridley, A. and Newell, D.G. (2001) Evidence for a genetically stable strain of Campylobacter jejuni. Appl Environ Microbiol 67, 1185–1189. Meinersmann, R.J. and Hiett, K.L. (2000) Concerted evolution of duplicate fla genes in Campylobacter. Microbiology 146, 2283–2290. Meinersmann, R.J., Helsel, L.O., Fields, P.I. and Hiett, K.L. (1997) Discrimination of Campylobacter jejuni isolates by fla gene sequencing. J Clin Microbiol 35, 2810–2814. Newell, D.G., Shreeve, J.E., Toszeghy, M., Domingue, G., Bull, S., Humphrey, T. and Mead, G. (2001) Changes in the carriage of Campylobacter strains by poultry carcasses during processing in abattoirs. Appl Environ Microbiol 67, 2636–2640. Nuijten, P.J.M., van den Berg, A.J., Formentini, I., van der Zeijst, B.A.M. and Jacobs, A.A. (2000) DNA rearrangements in the flagellin locus of a flaA mutant of Campylobacter jejuni during colonisation of chicken ceca. Infect Immun 68, 7137–7140. Oza, A.N., Thwaites, R.T., Wareing, D.R., Bolton, F.J. and Frost, J.A. (2002) Detection of heat-stable antigens of Campylobacter jejuni and C. coli by direct agglutination and passive hemagglutination. J Clin Microbiol 40, 996–1000. Parkhill, J., Wren, B.W., Mungall, K., Ketley, J.M., Churcher, C., Basham, D., Chillingworth, T., Davies, R.M. et al. (2000) The genome sequence of the food-borne pathogen


103


Campylobacter jejuni reveals hypervariable sequences. Nature 403, 665–668. Shreeve, J.E., Toszeghy, M., Ridley, A. and Newell, D.G. (2002) The carry-over of Campylobacter isolates between sequential poultry flocks. Avian Dis 46, 378–385. Wassenaar, T.M. and Newell, D.G. (2000) Genotyping of Campylobacter spp. Appl Environ Microbiol 66, 1–9. Wassenaar, T.M., van der Zeijst, B.A.M., Ayling, R. and Newell, D.G. (1993) Colonization of chicks by motility mutants of Campylobacter jejuni demonstrates the importance of flagellin A expression. J Gen Microbiol 139, 1171–1175. Wassenaar, T.M., Fry, B.N. and van der Zeijst, B.A.M. (1995) Variation of the flagellin locus of Campylobacter jejuni by recombination and horizontal gene transfer. Microbiology 141, 95–101. Wassenaar, T.M., Geilhausen, B. and Newell, D.G. (1998) Evidence of genomic instability in Campylobacter jejuni

104

A.M. Ridley et al.

isolated from poultry. Appl Environ Microbiol 64, 1816– 1821. Wassenaar, T.M., Fry, B.N., Lastovica, A.J., Wagenaar, J.A., Coloe, P.J. and Duim, B. (2000a) Genetic characterization of Campylobacter jejuni O:41 isolates in relation with Guillain Barre´ syndrome. J Clin Microbiol 38, 874– 876. Wassenaar, T.M., On, S.L.W. and Meinersmann, R.J. (2000b) Genotyping and the consequences of genetic instability. In Campylobacter, 2nd edn ed. Nachamkin, I. and Blaser, M.J. pp. 369–380 Washington, DC: American Society for Microbiology. Wassenaar, T.M., Wagenaar, J.A., Rigter, A., Fearnley, C., Newell, D.G. and Duim, B. (2002) Homonucleotide stretches in chromosomal DNA of Campylobacter jejuni display high frequency polymorphism as detected by direct PCR analysis. FEMS Microbiol Lett 212, 77–85.
