Hostname: page-component-586b7cd67f-l7hp2 Total loading time: 0 Render date: 2024-11-23T21:42:47.491Z Has data issue: false hasContentIssue false

Diversity and relatedness of Shiga toxin-producing Escherichia coli and Campylobacter jejuni between farms in a dairy catchment

Published online by Cambridge University Press:  23 November 2015

H. IRSHAD*
Affiliation:
mEpiLab, Infectious Disease Research Centre, Hopkirk Research Institute, Massey University, Palmerston North, New Zealand Animal Health Programme, Animal Sciences Institute, National Agricultural Research Centre, Park Road, Islamabad, Pakistan
A. L. COOKSON
Affiliation:
mEpiLab, Infectious Disease Research Centre, Hopkirk Research Institute, Massey University, Palmerston North, New Zealand AgResearch Ltd, Hopkirk Research Institute, Palmerston North, New Zealand
C. M. ROSS
Affiliation:
AgResearch Ltd, Ruakura Research Centre, Hamilton, New Zealand
P. JAROS
Affiliation:
mEpiLab, Infectious Disease Research Centre, Hopkirk Research Institute, Massey University, Palmerston North, New Zealand
D. J. PRATTLEY
Affiliation:
mEpiLab, Infectious Disease Research Centre, Hopkirk Research Institute, Massey University, Palmerston North, New Zealand
A. DONNISON
Affiliation:
AgResearch Ltd, Ruakura Research Centre, Hamilton, New Zealand
G. McBRIDE
Affiliation:
National Institute of Water and Atmospheric Research (NIWA), Hamilton, New Zealand
J. MARSHALL
Affiliation:
mEpiLab, Infectious Disease Research Centre, Hopkirk Research Institute, Massey University, Palmerston North, New Zealand
N. P. FRENCH
Affiliation:
mEpiLab, Infectious Disease Research Centre, Hopkirk Research Institute, Massey University, Palmerston North, New Zealand Allan Wilson Centre for Molecular Ecology and Evolution, Massey University, Palmerston North, New Zealand
*
*Author for correspondence: Dr H. Irshad, Animal Health Programme, Animal Sciences Institute, National Agricultural Research Centre, Park Road, Islamabad, Pakistan. (Email: [email protected])
Rights & Permissions [Opens in a new window]

Summary

The aim of this study was to examine the population structure, transmission and spatial relationship between genotypes of Shiga toxin-producing Escherichia coli (STEC) and Campylobacter jejuni, on 20 dairy farms in a defined catchment. Pooled faecal samples (n = 72) obtained from 288 calves were analysed by real-time polymerase chain reaction (rtPCR) for E. coli serotypes O26, O103, O111, O145 and O157. The number of samples positive for E. coli O26 (30/72) was high compared to E. coli O103 (7/72), O145 (3/72), O157 (2/72) and O111 (0/72). Eighteen E. coli O26 and 53 C. jejuni isolates were recovered from samples by bacterial culture. E. coli O26 and C. jejuni isolates were genotyped using pulsed-field gel electrophoresis and multilocus sequence typing, respectively. All E. coli O26 isolates could be divided into four clusters and the results indicated that E. coli O26 isolates recovered from calves on the same farm were more similar than isolates recovered from different farms in the catchment. There were 11 different sequence types of C. jejuni isolated from the cattle and 22 from water. An analysis of the population structure of C. jejuni isolated from cattle provided evidence of clustering of genotypes within farms, and among groups of farms separated by road boundaries.

Type
Original Papers
Copyright
Copyright © Cambridge University Press 2015 

INTRODUCTION

Shiga toxin-producing Escherichia coli (STEC) and Campylobacter spp. are important bacterial causes of gastroenteritis in humans worldwide and cattle are considered an important reservoir for both pathogens [Reference Bettelheim1, Reference Mullner2]. STEC have emerged globally as an important foodborne human zoonotic pathogen associated with outbreaks and sporadic cases of diarrhoea, haemorrhagic colitis (HC) and haemolytic uraemic syndrome (HUS) [Reference Bettelheim1]. There are more than 100 STEC serogroups that have been associated with human disease [3], but seven (O26, O45, O103, O111, O121, O145, O157) are considered to be the most important due to their association with large outbreaks and many of the HUS and HC cases [Reference Hsu4]. STEC-associated disease outbreaks have been reported worldwide [Reference Bettelheim1] with the first STEC-associated human case from New Zealand reported in 1980 [Reference Wilson and Bettelheim5]. The rate of reported STEC infections in New Zealand has increased from 1·3 cases/100 000 population in 1998 to 5·1 cases/100 000 population in 2013 [6, p. 53] and is presently higher than that from the UK (2·4 cases/100 000 population) [7] and from the USA (2·2 cases/100 000 population) [Reference D. Gilliss8]. Ruminants are considered to be the main reservoir of STEC [Reference Bettelheim1] and in New Zealand the greatest risk factors for human infection are proximity to cattle, contact with animal manure and contact with recreational water [Reference Jaros9]. Campylobacter jejuni and C. coli are the most important gastroenteritis-causing Campylobacter spp. worldwide [Reference Silva10]. Campylobacteriosis, enteric infection associated with C. jejuni is the most frequently reported notifiable infectious disease in New Zealand. In 2013, 6837 (152·9 cases/100 000 population) human cases of campylobacteriosis were notified [6, pp. 25–26]. Chicken, red meat, milk, water (for drinking and recreational activities) and contact with farm animals are considered to be the most common sources of transmission of Campylobacter to humans [Reference Silva10] and, since the reduction in cases associated with controls in the poultry industry in New Zealand in 2007/2008, ruminant sources (both bovine and ovine) have become relatively more important [Reference Sears11].

Water channels and rural streams can become contaminated with Campylobacter and E. coli by direct deposition of cattle faeces and inputs from surface and subsurface flows following rainfall or irrigation. Elevated Campylobacter concentrations in stream water during flood events appear to arise from local land run-off whereas elevated E. coli concentration can be dominated by sources further upstream [Reference McBride12]. In New Zealand, high levels of faecal bacteria have been reported from rural streams including the catchment where this study was conducted [Reference Wilcock13] in which dairy farming is the major land use. Better understanding of population structure and transmission of zoonotic pathogens would assist with devising appropriate control strategies aimed at reducing environmental contamination, which in turn would reduce transmission to animals and humans. This study was conducted to determine the degree of relatedness of STEC and C. jejuni within farm and also between farms as a function of their separation by a road network and the extent to which they are connected by a network of streams. Understanding the population structure of STEC and C. jejuni within and between communities of farms will allow the development of effective mitigation strategies to reduce both within- and between-farm spread.

MATERIALS AND METHODS

The dairy catchment (~15 km2) where the study was conducted was selected because 70% of catchment land is used for dairy farming and catchment boundaries are well defined. There are 23 farms in this catchment including 18 dairy farms, four dry-stock farms and one horse farm. Characteristics of the catchment and initial observations of the behaviour of generic E. coli and Campylobacter spp. in the waterway draining the dairy farming area have been described previously [Reference Stott14].

Farms that shared a boundary (excluding farms separated by roads, but including those separated by a stream) were grouped together (Fig. 1a ). The region was also divided into farms adjacent to the main stream, and those in the periphery of the catchment (Fig. 1b ).

Fig. 1. River catchment mapping. (a) Five distinct groups marked A–E were evident. The road network is shown in heavy black lines and the stream network in broken grey lines. Presence of E. coli O26 (○), STEC O26 (○*) and C. jejuni sequence types shown at the farm level. W, Water sampling site. (b) An alternative grouping of farms according to whether they were adjacent to the main stream (X, in grey) or away from the main stream (Y, in white).

Animal sampling was carried out over a 3-month period from October to December 2009. Faecal samples were collected from 288 animals, which were either born in July/August 2009 (Y0; average 3 or 4 months) or July/August 2008 (Y1; average 15 or 16 months), from 20 of the farms and each farm was sampled only once. The three farms not sampled were a dry-stock farm that had no calves during the study period and a horse farm. The remaining dairy farm had no Y0 calves but the Y1 animals which were grazed on a neighbouring farm were sampled. Different farms were sampled at each occasion and samples were collected from a freshly voided faecal pat shed by an individual animal.

Sets of four faecal pat samples (grouped on the basis of farm and age group) were thoroughly mixed to form a single composite sample for microbiological analysis. This was done in order to reduce the number of samples required for microbiological analysis, while focusing on the recovery of isolates for genotyping and subsequent phylogenetic comparisons between populations. In total, 72 pooled faecal samples from the 20 farms in the catchment area were transported under cold conditions to the laboratory. These samples were analysed for the presence of STEC (O26, O103, O111, O145, O157) and C. jejuni. In addition 33 water samples were collected from a single site at end of the main stream of the catchment (Fig. 1a ) as part of a separate study that examined changes in total Campylobacter and C. jejuni numbers associated with two separate flood events (11–15 September and 29 September–1 October 2008).

Isolation and characterization of STEC

For STEC analysis, 2 g of each composite sample was enriched in 18 ml buffered peptone water (BPW) for 24 h at 37 °C. The Isolate Faecal DNA kits (Bioline, New Zealand) were used for isolation of DNA from enriched faecal samples according to the manufacturer's instructions. DNA samples were then tested by real-time polymerase chain reaction (rtPCR) for the presence of wzx (O26) [Reference Perelle15], wzx (O103) [Reference Fratamico16], wbdI (O111) [Reference Perelle15], wzx1 (O145) [Reference Fratamico17] and rfbE (O157) [Reference Perelle15] using previously published methods.

Briefly SYTO-9 was used as the fluorescent nucleic acid stain in rtPCR reactions with the thermal melt from 75 °C to 95 °C at a rate of 0·05 °C/s used to confirm the amplification of each serogroup-specific amplicon.

Enriched faecal samples that were E. coli O26, O103, O111, O145 or O157 rtPCR-positive underwent immunomagnetic separation (IMS) for isolation of the respective serogroup using previously published methods [Reference Wright, Chapman and Siddons18]. The bead suspension (100 µl) was inoculated onto sorbitol MacConkey agar supplemented with cefixime (50 µg/ml) and potassium tellurite (2·5 mg/ml) (CT-SMAC; Fort Richard, New Zealand) for isolation of O157, rhamnose MacConkey agar supplemented with cefixime and potassium tellurite (CT-RMAC; Fort Richard) for isolation of E. coli O26 and sorbitol MacConkey agar (SMAC; Fort Richard) for isolation of O103, O111 and O145. The plates were incubated at 37 °C for 24 h and observed for the presence of STEC O157 colonies (grey colour/colourless colonies on CT-SMAC), STEC O26 colonies (grey colour/colourless colonies on CT-RMAC) and O103, O111 and O145 colonies (pink/purple colonies on SMAC). Suspect colonies were subcultured on MacConkey agar. O157 suspect colonies were identified using E. coli O157 latex agglutination kits (Oxoid, New Zealand) and colonies suspected positive for E. coli O26, O103, O111 and O145 were identified by using rtPCR. The allelic profile of positive isolates was then confirmed by multiplex PCR using the method previously described by Paton & Paton [Reference Paton and Paton19] to detect the presence of Shiga toxin 1(stx1), Shiga toxin 2 (stx2), E. coli attaching and effacing (eae) and enterohaemolysin (ehxA) genes.

Molecular typing of each isolate was done by pulsed-field gel electrophoresis (PFGE) following the standard procedure described by PulseNet USA [20]. Bionumerics software (v. 7·2, www.applied-maths.com) was used to analyse and compare PFGE profiles of the E. coli O26 isolates, and to create a dendrogram applying the unweighted pair-group method with arithmetic mean (UPGMA) cluster analysis using the Dice similarity coefficient, with >80% similarity cut-off, and 1% band-matching tolerance.

Isolation and characterization of Campylobacter

Campylobacter isolation was undertaken from 1 g aliquots of each composite sample using the New Zealand Reference Method [21]. Briefly, 1 g of each composite faecal sample was enriched in 9 ml Exeter broth for 48 h at 42 °C. A loopful of enriched broth was plated onto modified charcoal cefoperazone deoxycholate agar (mCCDA) and plates were incubated at 42 °C for 48 h under microaerobic conditions. C. jejuni was confirmed using DNA obtained from individual colonies by PCR as described previously [Reference Inglis and Kalischuk22].

In total 33 water samples were collected from a single site in the main stream of the catchment. C. jejuni were isolated as part of a study enumerating presumptive Campylobacter from the water samples using the most probable number (MPN) method in the format of 100 ml, 10 ml and 1 ml of each water sample in replicates of three (count data not shown here). Water samples of 100 ml and 10 ml were filtered through 0·45 µm filters. Each filter was placed in a tube containing 30 ml and 9 ml Exeter broth, respectively, while each 1 ml water sample was mixed with 9 ml Exeter broth without filtration. These tubes were incubated at 37 °C for 24 h and transferred to 42 °C for a further 24 h. A loopful of inoculum from each tube was inoculated onto mCCDA plates and incubated in a gas jar under microaerobic conditions for 24–48 h. The suspected C. jejuni colonies were identified using PCR as described previously,

Multilocus sequence typing (MLST) of C. jejuni isolates was performed using seven house-keeping genes: aspA (aspartase A), glnA (glutamine synthase), gltA (citrate synthase), glyA (serine hydroxyl methyltransferase), pgm (phosphor glucomutase), tkt (transketolase) and uncA (ATP synthase alpha subunit) based on the method according to Dingle, without modifications [Reference Dingle23]. Sequence data were collated and alleles assigned using the Campylobacter PubMLST database (http://pubmlst.org/Campylobacter/).

Statistical analysis

The relationship between genotypes isolated from cattle and water was examined by constructing minimum spanning trees (MSTs) and by calculating the proportional similarity index (PSI) [Reference Rosef24].

This index provides a value of between 0 and 1, where 1 indicates complete identity between populations from the two sources and 0 indicates no similarity [Reference Rosef24]. Bootstrap confidence intervals (CIs) for PSI were calculated using the method described by Garrett et al. [Reference Garrett25]. The MST was created using Bionumerics v. 7·2 (www.applied-maths.com).

Population structure and differentiation was examined by conducting a permutational multivariate analysis of variance (PERMANOVA) [Reference McArdle and Anderson26]. The dissimilarity measure used was the number of allele differences across all seven loci. The population of C. jejuni isolated on each farm was examined to estimate the components of allelic variation and test the hypotheses that networks of farms that share a boundary (i.e. are not separated by the road network) have isolates that are clustered/more similar compared to those that do not share a boundary (Fig. 1a , groups A–E); and that farms that are located either side of the main catchment stream have isolates that are clustered/more similar compared to farms that are located away from the main catchment stream (Fig. 1b , groups X and Y).

RESULTS

Population structure and characterization of STEC

In total, 72 pooled faecal samples collected from Y0 (n = 46) and Y1 (n = 23) calves were analysed by rtPCR to detect the presence of E. coli O26, O103, O111, O145 and O157. The age of the animals from the three remaining samples was not recorded. Using rtPCR, E. coli O26 was the most commonly detected E. coli serogroup (30/72 samples, 41·6%) compared to O103 (9·7%), O145 (4·1%) and O157 (2·7%), or O111, where no samples were positive (Table 1). The 30 E. coli O26 rtPCR-positive faecal samples were obtained from 13/20 (65%) farms, from which 18 isolates were recovered from 11 farms (Fig. 1a ). Of these 18 isolates, 13 were positive for stx1, eae, and ehxA using multiplex PCR, while the remaining five were positive for eae, and ehxA. Similarly, samples positive for O103 and O145 originated from six and three farms, respectively. Both O157 rtPCR-positive samples were from the same farm, but no O157 isolates were recovered from these enrichments despite the use of serogroup-specific IMS beads and selective media (CT-SMAC). Both of the O103 isolates and a single O145 isolate recovered on SAMC agar were eae, ehxA positive. Most of the composite faecal samples rtPCR positive for E. coli O26, O103, O145 and O157 were from Y0 calves (Table 1). Given the low numbers of PCR-positive samples and isolates of other serotypes, all subsequent analyses focused on the distribution of E. coli O26.

Table 1. Composite faecal samples (n = 72) obtained from cattle in a dairy catchment were analysed for E. coli O157 and non-O157 serogroups using real-time PCR (rtPCR). The samples positive by rtPCR were subjected to isolation. The virulence profile (stx1, stx2, eae, ehxA) of isolates obtained was determined using multiplex PCR

* Calves born in 2009.

† Calves born in 2008.

The 18 E. coli O26 isolates recovered from the 17 composite faecal samples could be broadly grouped into four clusters (>80% similarity) based on their PFGE profile (Fig. 2). All 12 isolates from cluster 4 were stx1-positive and were derived from four regions (B–E). The remaining six isolates, five of which were stx negative, were more heterogeneous based on PFGE profile analysis and were separated into three clusters represented by one isolate (cluster 3), two isolates (cluster 1) and three isolates (cluster 2), respectively. Unlike regions B, C and D (see Fig. 1a ) which produced six, four and five E. coli O26 isolates, respectively, only a single isolate was identified from regions A and E. The isolate from region A was stx negative and present in cluster 1 while the isolate from region E was stx positive and present in cluster 4. Isolates recovered from cattle on the same farm were generally more similar than isolates from different farms although two farms had isolates from different clusters. There was no obvious association between the regions identified in Figure 1(a or b) and genotype of STEC O26, and given the small number of recovered isolates no further analyses were conducted.

Fig. 2. Clustering (unweighted pair group method arithmetic mean, UPGMA dendrogram) of the PFGE profiles of 18 E. coli O26 isolates from 17 composite faecal samples taken from 11 farms in a catchment. Each farm is symbol coded and the catchment region is indicated. Note two isolates from sample EcCa38 are included in the analysis. The last lane is the Salmonella serotype Braenderup reference standard (H9812).

Population structure and characterization of C. jejuni

Fifty (69·4%) of the 72 pooled faecal samples yielded C. jejuni isolates (Fig. 1a). Faecal samples from Y0 calves (n = 46) provided 35 (76·1%) C. jejuni isolates whereas samples from Y1 calves (n = 23) provided 15 (65·2%) isolates. Similarly, the water samples yielded 75 C. jejuni isolates. The distribution of sequence types (STs) from cattle and catchment stream are shown in Table 2. The diversity of genotypes of C. jejuni was higher in the catchment stream compared to the cattle population. Figure 3 shows the rarefaction curves with 95% CIs for the cattle and water samples. For a given sample size (e.g. 50 samples) the number of STs observed was higher in the water samples (22 STs) compared to the cattle samples (11 STs). There was a total of 11 different STs isolated from the cattle, six (54·5%) (STs 42, 50, 53, 61, 474, 2026) of which were also recovered from the catchment stream and these have all been associated with human clinical cases in New Zealand [Reference Mullner2]. Conversely, there was much greater diversity in the water isolates; a total of 22 different C. jejuni STs were isolated from water and of these water C. jejuni STs only 31·8% (7/22) were also recovered from cattle in the catchment. Most of the water isolates (46/75, 61·3%) have previously been associated with wildlife sources, such as ducks, starlings (Sturnus vulgaris) and pukeko (Porphyrio porphyria, a member of the Rallidae family) in New Zealand [Reference French, Marshall and Mohan27] and these are rarely associated with human disease. The C. jejuni population structure displayed as a MST illustrates genotypes, such as ST42 and ST50 that were common to both water and cattle (Fig. 4).

Fig. 3. Genotype (as defined by contrasting C. jejuni sequence types; STs) richness in cattle and water samples. Rarefaction curves indicating the estimated number of unique STs in cattle (black) and water (grey) for varying sample sizes, with 95% confidence intervals (CI).

Fig. 4. Minimum spanning tree showing the multilocus sequence types (STs) of C. jejuni isolated from cattle (black) and water (white) in the catchment stream. Each circle represents a ST; the size of the circle is proportional to the number of isolates and the area of each circle the relative frequency of each source. Isolates are grouped according to their genetic relatedness; short solid lines between STs indicate they vary at just one of the seven MLST loci (i.e. they are single locus variants).

Table 2. The distribution of C. jejuni sequence types (STs) isolated from cattle and water from the catchment stream.

* Indicates STs associated with wildlife in New Zealand.

The area of overlap of the genotype distributions from cattle and water (PSI) was low-moderate, PSI = 0·21 (95% bootstrapped CI = 0·13–0·29).

The PERMANOVA analysis (Tables 3 and 4) indicated that, while most of the allelic variation (72–73%) was residual between-animal variation that could not be accounted for by farm or groups of farms, there was significant clustering of genotypes at the farm level and a component of this may be attributed to the adjacency of farms contained within the road network (22%), but not their adjacency to the main catchment stream (<4%) (Tables 3 and 4). This is consistent with transmission being more likely to occur between farms that share a road boundary, than through the stream network.

Table 3. Permutational multivariate analysis of molecular variance as defined using farm structure grouped according to road network (Fig. 1a)

Table 4. Permutational multivariate analysis of molecular variance as defined using farm structure grouped according to proximity to main stream (Fig. 1b)

DISCUSSION

This study provides information about the genetic diversity of STEC and C. jejuni between farms in a defined catchment environment. It was a small-scale study, limited to a single dairy catchment (~15 km2) and aimed to recover a representative set of isolates for molecular genotyping from as many farms as possible, rather than to conduct a large-scale analysis of the prevalence of carriage of E. coli and C. jejuni in different livestock groups on each farm. Therefore, the strategy of compositing faecal material from four animals for each sample was adopted in an attempt to maximize the recovery of isolates while reducing the number of samples required for culture. The recovery rate for E. coli O26 isolates was higher (60%) compared to O103 (28·5%) and O145 (33·3%) and may reflect the availability of selective media (CT-RMAC) for E. coli O26, whereas no well-established selective media is available for isolation of O103 and O145.

The prevalence of E. coli O26 (41·6%) was higher compared to other serotypes included in this study. This finding has been noted previously in other studies [Reference Bonardi28, Reference Pearce29] although they were not conducted in a single catchment. For example, in a Scottish national survey which focused exclusively on non-O157 STEC, 6086 faecal samples were collected from calves aged >1 year and from 338 farms to determine the prevalence of E. coli O26, O103, O111 and O145 using IMS. The weighted mean prevalence of E. coli O26 (4·6%) was higher than O103 (2·7%), O145 (0·7%) and O111, which was absent. An Italian study which analysed 182 faecal samples collected from cattle from slaughter plants also reported the higher prevalence of E. coli O26 (3%) compared to O157 (0·5%) using IMS; however, no O103, O111 or O145 isolates were obtained [Reference Bonardi28]. The basis for the higher prevalence of E. coli O26 compared to other serogroups is uncertain but E. coli O26 may be better adapted to colonize cattle compared to other clinically important STEC serogroups, such as O103, O111, O145 and O157 [Reference O'Reilly30].

The current New Zealand on-farm prevalence of STEC O26 has not been determined. However an initial cross-sectional study to determine the prevalence of E. coli O26 from recto-anal mucosal swabs (RAMS) taken from 883 cattle and 695 calves (aged <7 days) at slaughter revealed an overall prevalence of 2·0% and 3·5% of stx-positive and stx-negative E. coli O26, respectively [Reference Jaros31]. All twelve stx-positive isolates from cluster 4 were grouped with other stx-positive E. coli O26, clusters 1 and 2 grouped closely with stx-negative E. coli O26, and the single cluster 3 isolate was an outlier in an additional stx-positive group of isolates, distinct from cluster 4.

Infection with STEC O26 is often associated with mild diarrhoeal disease or asymptomatic cases and may not be reported to health practitioners, but may cause complications associated with HUS [Reference Ethelberg32]. Even where cases may be reported, there is a potential underestimation of the rate of infection due to STEC O26 as many laboratories in New Zealand (and internationally) may only test young children with bloody diarrhoea for STEC and that non-bloody diarrhoeal faecal samples may not be assessed for the presence of STEC O26 or other non-O157 STEC [Reference Baker33]. The presence of stx1, eae, and ehxA virulence genes in E. coli O26 isolates from calves has great public health significance as this is the most common virulence profile observed in E. coli O26 isolates from human cases [Reference Schmidt34]. Several other studies have also reported the presence of stx1 and eae virulence genes in E. coli O26 isolates from cattle [Reference Pearce29, Reference Blanco35]. Although, carriage of the stx2 gene in E. coli O26 isolates has also been reported in German cattle [Reference Geue36], no stx2-positive STEC O26 isolates were identified in this study. Five O26, two O103 and one O145 isolates were eae,ehxA-positive which is indicative of the atypical Enteropathogenic E. coli (aEPEC) pathotype. aEPEC has been associated with prolonged diarrhoeal disease and several studies have demonstrated the close genetic relationship and common virulence determinants between aEPEC and STEC strains suggesting a dynamic relationship between aEPEC and STEC through the loss and acquisition of stx-encoding bacteriophage [Reference Bielaszewska37, Reference Whittam38].

PFGE was used in this study to explore the genetic diversity of E. coli O26 isolates and their spatial separation or relatedness from separate farms. Water may represent an important transmission source for a number of zoonotic pathogens including STEC which have previously been isolated from various water sources in New Zealand [Reference Donnison and Ross39]. Although our study did not have the power to explore the population structure of STEC between farms in detail, our data are consistent with a previous molecular epidemiological investigation of 163 STEC O157 isolates that indicated isolates from the same farms had similar molecular profiles compared to isolates from different farms indicating limited transmission of STEC O157 between farms [Reference Faith40].

In this study the prevalence of C. jejuni in cattle faeces was high (70·8%). Other studies have also reported the high prevalence (50–100%) of C. jejuni in cattle faeces [Reference Gilpin41, Reference Stanley42]. Cattle are considered an important reservoir of C. jejuni and potential source of infection to humans [Reference Gilpin41]. A New Zealand study reported cattle (11–18% of human cases) as the second most important source of transmission of C. jejuni to humans after poultry (58–76% of human cases) [Reference Mullner2]. Six of the 11 STs (42, 50, 53, 61, 474, 2026) isolated from cattle in this study are recognized as important human pathogens in New Zealand [Reference Mullner2]. As noted in previous studies [Reference Bae43, Reference Nielsen44] more samples were positive from Y0 calves compared to Y1 calves. Similar to the contrasting STEC prevalence rates associated with Y0 calves and Y1 adult cattle, an increased susceptibility of young calves to colonization with a variety of Campylobacter STs observed in this study and others may be associated with an age-dependent reduction in infection by enteric pathogens [Reference Giacoboni45] due to development of mature intestinal microbiota [Reference Nurmi and Rantala46].

Water is considered an important source of transmission of Campylobacter, and with water-borne campylobacteriosis outbreaks have been reported [Reference Kuusi47]. Campylobacter have been isolated from rivers and waterways [Reference Carter48] and transmission of Campylobacter from water to livestock has been indicated through the isolation of indistinguishable Campylobacter biotypes from water and cattle [Reference Stanley42]. In this study, the genetic diversity of C. jejuni isolates from the catchment stream was higher compared to cattle isolates; observations also noted in other New Zealand studies [Reference Carter48, Reference Devane49]. However, the water samples included in this study were collected during two separate storm events a year earlier than the cattle faecal samples which may contribute to the variation between the C. jejuni STs obtained from water and cattle samples. In the study by Devane et al. [Reference Devane49], 32 heat-stable (HS) serotypes were identified among the 616 isolates examined. Of the 32 HS serotypes 28 were present in isolates from water samples whereas only 12 were present in the isolates from dairy cattle faeces. The contrasting diversity of isolates obtained from cattle and catchment streams, and the association between non-cattle-associated STs and wildlife, suggests that a high proportion of the water isolates may originate from other sources inhabiting the area such as ducks, starlings and pukeko [Reference French, Marshall and Mohan27]. A study conducted in Cheshire, UK to investigate the molecular epidemiology of C. jejuni in dairy cattle, wildlife and the environment also reported that most of the water-related C. jejuni were similar to those present in wild birds [Reference Kwan50]. We observed that the most prevalent ST in water was ST2381 and this ST is also the most prevalent isolate in many other river systems in New Zealand [Reference Carter48]. To date this ST has only been isolated from pukeko [Reference French, Marshall and Mohan27] and takahe (Porphyrio hochstetteri) [Reference French, Sheppard and Meric51], both members of the Rallidae family, and has not been recognised as a human pathogen. However, it is also important to note that some of the STs recovered from water may have originated in cattle, but not been recovered from the faeces sampled. For example, ST45 has been recovered from multiple sources, including cattle, and is strongly associated with human disease in New Zealand, particularly in rural areas and in children [Reference Mullner52, Reference Mullner53]. Kwan et al. [Reference Kwan50] also reported water as a potential source of infection to humans in the UK as six of the eight clonal complexes isolated from environmental water were similar to those isolated from human patients. However, a study conducted in the Taieri river catchment in New Zealand indicated that there was no overlap between genotypes from environmental water and human cases when PFGE profiles and Penner serotyping were used together [Reference Eyles54].

A low-moderate area of overlap of the genotype distributions from cattle and water (PSI = 0·21) was observed. This is greater than the estimated overlap between water isolates and those recovered from human cases (PSI = 0·18) reported in earlier studies, but lower than the PSI for most other source pairwise comparisons (e.g. the PSI for cattle compared to sheep isolates was estimated to be 0·50 and cattle compared to human isolates was estimated to be 0·34 [Reference Mullner2, Reference Garrett25].

PERMANOVA provides evidence of clustering of C. jejuni populations at the farm level and between farms grouped by geographical boundaries. Grove-White et al. [Reference Grove-White55] also reported a similar proportion of genetic variation associated with ruminant C. jejuni populations at the farm level (16–20%, P < 0·001). This finding could have important implications for transmission and control of C. jejuni. Direct contact between cattle in neighbouring farms, sharing of equipment and personnel, and surface water run-off between farms could be important mechanisms for local transmission, underlining the importance of maintaining good between-farm biosecurity.

In summary, STEC and C. jejuni were prevalent in cattle in the catchment, and individual farms were associated with a subset of genotypes circulating among the young stock located on that farm. The pattern of genotypes of E. coli O26 and C. jejuni was suggestive of a high level of within farm transmission, and limited between-farm transmission. Although this study was conducted over a short time-frame on a small number of farms in a single catchment, it has revealed some significant associations that provide insight into the nature and scale of dissemination of potential human pathogens within and between farms. These observations will assist in the development of potential on-farm mitigation strategies, such as restricting animal movement between farms and maintaining individual animal cohorts, to reduce the prevalence of zoonotic pathogens.

ACKNOWLEDGEMENTS

The authors acknowledge the funding from Dairy New Zealand (Schedule BS 808, ‘Transmission of two zoonotic pathogens’), and the sample collectors, James Sukias and Charlotte Yates (NIWA) and Alec McGowan (AgResearch).

DECLARATION OF INTEREST

None.

References

REFERENCES

1. Bettelheim, KA. The non-O157 Shiga-toxigenic (Verocytotoxigenic) Escherichia coli; under-rated pathogens. Critical Reviews in Microbiology 2007; 33: 6787.CrossRefGoogle ScholarPubMed
2. Mullner, P, et al. Assigning the source of human campylobacteriosis in New Zealand: a comparative genetic and epidemiological approach. Infection Genetics and Evolution 2009; 9: 13111319.Google Scholar
3. World Health Organization. Zoonotic non-O157 Shiga toxin-producing Escherichia coli (STEC). Report of a WHO Scientific Working Group meeting. (http://whqlibdoc.who.int/hq/1998/WHO_CSR_APH_98.8.pdf), 1998.Google Scholar
4. Hsu, HY, et al. Effect of high pressure processing on the survival of Shiga toxin-producing Escherichia coli (big six vs. O157:H7) in ground beef. Food Microbiology 2015; 48: 17.Google Scholar
5. Wilson, MW, Bettelheim, KA. Cytotoxic Escherichia coli serotypes. Lancet 1980; 1: 201.Google Scholar
6. Environmental Science and Research. Verocytotoxin-producing E. coli (VTEC/STEC) confirmed by ERL in 2013 (https://surv.esr.cri.nz/PDF_surveillance/ERL/VTEC/VTEC_2013.pdf), 2014.Google Scholar
7. European Food Safety Authority and European Centre for Disease Prevention and Control. The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2011. European Food Safety Authority Journal 2013; 11: 3129.Google Scholar
8. D. Gilliss, , et al. Incidence and trends of infection with pathogens transmitted commonly through food — foodborne diseases active surveillance network, 10 U.S. sites, 1996–2012. Morbidity and Mortality Weekly Report 2013; 62: 283287.Google Scholar
9. Jaros, P, et al. A prospective case-control and molecular epidemiological study of human cases of Shiga toxin-producing Escherichia coli in New Zealand. BMC Infectious Diseases 2013; 13: 450.Google Scholar
10. Silva, J, et al. Campylobacter spp. as a foodborne pathogen: a review. Frontiers in Microbiolgy 2011; 2: 112.Google Scholar
11. Sears, A, et al. Marked campylobacteriosis decline after interventions aimed at poultry, New Zealand. Emerging Infectious Diseases 2011; 17: 10071015.Google Scholar
12. McBride, GB. Explaining differential sources of zoonotic pathogens in intensively farmed-catchment using kinemetic waves. Water Science & Technology 2011; 63: 695703.CrossRefGoogle ScholarPubMed
13. Wilcock, RJ, et al. Land-use impacts and water quality targets in the intensive dairying catchment of the Toenepi Stream, New Zealand. New Zealand Journal of Marine and Freshwater Research 2006; 40: 123140.CrossRefGoogle Scholar
14. Stott, R, et al. Differential behaviour of Escherichia coli and Campylobacter spp. in a stream draining dairy pasture. Journal of Water and Health 2011; 9: 5969.CrossRefGoogle Scholar
15. Perelle, S, et al. Detection by 5′-nuclease PCR of Shiga-toxin producing Escherichia coli O26, O55, O91, O103, O111, O113, O145 and O157:H7, associated with the world's most frequent clinical cases. Molecular and Cellular Probes 2004; 18: 185192.Google Scholar
16. Fratamico, PM, et al. DNA sequence of the Escherichia coli O103 O antigen gene cluster and detection of enterohemorrhagic E. coli O103 by PCR amplification of the wzx and wzy genes. Canadian Journal of Microbiology 2005; 51: 515522.Google Scholar
17. Fratamico, PM, et al. PCR detection of enterohemorrhagic Escherichia coli O145 in food by targeting genes in the E. coli O145 O-antigen gene cluster and the Shiga toxin 1 and Shiga toxin 2 genes. Foodborne Pathogens and Disease 2009; 6: 605611.Google Scholar
18. Wright, DJ, Chapman, PA, Siddons, CA. Immunomagnetic separation as a sensitive method for isolating Escherichia coli O157 from food samples. Epidemiology and Infection 1994; 113: 3139.Google Scholar
19. Paton, AW, Paton, JC. Detection and characterization of Shiga toxigenic Escherichia coli by using multiplex PCR assays for stx1, stx2, eaeA, Enterohaemorrhagic E. colihlyA, rfbO111, and rfbO157. Journal of Clinical Microbiology 1998; 36: 598602.Google Scholar
20. Centre for Disease Control and Prevention. One-day (24–28 h) standardized laboratory protocol for molecular subtyping of Escherichia coli O157:H7, Salmonella serotypes, and Shigella sonnei by pulsed field gel electrophoresis (PFGE). (http://www.pulsenetinternational.org/assets/PulseNet/uploads/pfge/5_201_202_204_2009_PNetStandProtEcolSalShig.pdf), 2009.Google Scholar
21. Ministry of Health. Isolation of thermotolerant Campylobacter – review and methods for New Zealand laboratories (http://www.moh.govt.nz/notebook/nbbooks.nsf/0/73166eb251837f95cc257834000271db/$FILE/IsolationOfThermotolerantCampylobacter.pdf), 2003.Google Scholar
22. Inglis, GD, Kalischuk, LD. Use of PCR for direct detection of Campylobacter species in bovine fecest. Applied and Environmental Microbiology 2003; 69: 34353447.CrossRefGoogle Scholar
23. Dingle, KE, et al. Multilocus sequence typing system for Campylobacter jejuni . Journal of Clinical Microbiology 2001; 39: 1423.Google Scholar
24. Rosef, O, et al. Serotyping of Campylobacter jejuni, Campylobacter coli, and Campylobacter laridis from domestic and wild animals. Applied and Environmental Microbiology 1985; 49: 15071510.Google Scholar
25. Garrett, N, et al. Statistical comparison of Campylobacter jejuni subtypes from human cases and environmental sources. Journal of Applied Microbiology 2007; 103: 21132121.CrossRefGoogle ScholarPubMed
26. McArdle, B, Anderson, M. Fitting multivariate models to community data: a comment on distance-based redundancy analysis. Ecology 2001; 82: 290297.Google Scholar
27. French, NP, Marshall, J, Mohan, V. New and emerging data on typing of Campylobacter spp. strains in animals, environmental matrices and humans. (http://www.foodsafety.govt.nz/elibrary/industry/examining-link-with-public-health/new-and-emerging-data-on-typing-of-campylobacter.pdf), 2011, pp. 1–46.Google Scholar
28. Bonardi, S, et al. Detection of Verocytotoxin-producing Escherichia coli serogroups O157 and O26 in the cecal content and lymphatic tissue of cattle at slaughter in Italy. Journal of Food Protection 2007; 70: 14931497.Google Scholar
29. Pearce, MC, et al. Prevalence and virulence factors of Escherichia coli serogroups O26, O103, O111, and O145 shed by cattle in Scotland. Applied and Environmental Microbiology 2006; 72: 653659.Google Scholar
30. O'Reilly, KM, et al. Associations between the presence of virulence determinants and the epidemiology and ecology of zoonotic Escherichia coli . Applied and Environmental Microbiology 2010; 76: 81108116.Google Scholar
31. Jaros, P, et al. Shedding of Escherichia coli O157:H7 and O26 STEC by slaughter cattle in New Zealand. 56th Annual Meeting of the New Zealand Microbiological Society. Palmerston North, New Zealand, 2011.Google Scholar
32. Ethelberg, S, et al. Outbreak of non-O157 Shiga toxin-producing Escherichia coli infection from consumption of beef sausage. Clinical Infectious Diseases 2009; 48: E78E81.Google Scholar
33. Baker, M, et al. Emergence of Verotoxigenic Escherichia coli (VTEC) in New Zealand. New Zealand Public Health Report 1999; 6: 912.Google Scholar
34. Schmidt, H, et al. Non-O157:H7 pathogenic shiga toxin-producing Escherichia coli: Phenotypic and genetic profiling of virulence traits and evidence for clonality. Journal of Infectious Diseases 1999; 179: 115123.Google Scholar
35. Blanco, M, et al. Distribution and characterization of faecal Verotoxin-producing Escherichia coli (VTEC) isolated from healthy cattle. Veterinary Microbiology 1997; 54: 309319.Google Scholar
36. Geue, L, et al. A long-term study on the prevalence of Shiga toxin-producing Escherichia coli (STEC) on four German cattle farms. Epidemiology and Infection 2002; 129: 173185.Google Scholar
37. Bielaszewska, M, et al. Shiga toxin-negative attaching and effacing Escherichia coli: Distinct clinical associations with bacterial phylogeny and virulence traits and inferred in-host pathogen evolution. Clinical Infectious Diseases 2008; 47: 208217.Google Scholar
38. Whittam, TS, et al. Clonal relationships among Escherichia coli strains that cause hemorrhagic colitis and infantile diarrhea. Infection and Immunity 1993; 61: 16191629.CrossRefGoogle ScholarPubMed
39. Donnison, A, Ross, C. Survival and retention of Escherichia coli O157:H7 and Campylobacter in contrasting soils from the Toenepi catchment. New Zealand Journal of Agricultural Research 2009; 52: 133144.Google Scholar
40. Faith, NG, et al. Prevalence and clonal nature of Escherichia coli O157:H7 on dairy farms in Wisconsin. Applied and Environmental Microbiology 1996; 62: 15191525.Google Scholar
41. Gilpin, BJ, et al. Comparison of Campylobacter jejuni genotypes from dairy cattle and human sources from the Matamata-Piako District of New Zealand. Journal of Applied Microbiology 2008; 105: 13541360.CrossRefGoogle ScholarPubMed
42. Stanley, KN, et al. The seasonal variation of thermophilic Campylobacters in beef cattle, dairy cattle and calves. Journal of Applied Microbiology 1998; 85: 472480.Google Scholar
43. Bae, W, et al. Prevalence and antimicrobial resistance of thermophilic Campylobacter spp. from cattle farms in Washington State. Applied and Environmental Microbiology 2005; 71: 169174.Google Scholar
44. Nielsen, EM. Occurrence and strain diversity of thermophilic campylobacters in cattle of different age groups in dairy herds. Letters in Applied Microbiology 2002; 35: 8589.Google Scholar
45. Giacoboni, GI, et al. Comparison of fecal Campylobacter in calves and cattle of different ages and areas in Japan. Journal of Veterinary Medical Science 1993; 55: 555559.Google Scholar
46. Nurmi, E, Rantala, M. New aspects of Salmonella infection in broiler production. Nature 1973; 241: 210211.Google Scholar
47. Kuusi, M, et al. A large outbreak of campylobacteriosis associated with a municipal water supply in Finland. Epidemiology and Infection 2005; 133: 593601.Google Scholar
48. Carter, PE, et al. Novel clonal complexes with an unknown animal reservoir dominate Campylobacter jejuni isolates from river water in New Zealand. Applied and Environmental Microbiology 2009; 75: 60386046.Google Scholar
49. Devane, ML, et al. The occurrence of Campylobacter subtypes in environmental reservoirs and potential transmission routes. Journal of Applied Microbiology 2005; 98: 980990.Google Scholar
50. Kwan, PSL, et al. Molecular epidemiology of Campylobacter jejuni populations in dairy cattle, wildlife, and the environment in a farmland area. Applied and Environmental Microbiology 2008; 74: 51305138.CrossRefGoogle Scholar
51. French, N, et al. Evolution of Campylobacter species in New Zealand. In: Sheppard, S, Meric, G, eds. Campylobacter Ecology and Evolution. Wymondham, UK: Horizon Scientific Press, 2014.Google Scholar
52. Mullner, P, et al. Source attribution of food-borne zoonoses in New Zealand: a modified Hald model. Risk Analysis 2009; 29: 970984.Google Scholar
53. Mullner, P, et al. Molecular and spatial epidemiology of human campylobacteriosis: source association and genotype-related risk factors. Epidemiology and Infection 2010; 138: 13721383.Google Scholar
54. Eyles, RF, et al. Comparison of Campylobacter jejuni PFGE and Penner subtypes in human infections and in water samples from the Taieri River catchment of New Zealand. Journal of Applied Microbiology 2006; 101: 1825.Google Scholar
55. Grove-White, DH, et al. Molecular epidemiology and genetic diversity of Campylobacter jejuni in ruminants. Epidemiology and Infection 2011; 139: 16611671.Google Scholar
Figure 0

Fig. 1. River catchment mapping. (a) Five distinct groups marked A–E were evident. The road network is shown in heavy black lines and the stream network in broken grey lines. Presence of E. coli O26 (○), STEC O26 (○*) and C. jejuni sequence types shown at the farm level. W, Water sampling site. (b) An alternative grouping of farms according to whether they were adjacent to the main stream (X, in grey) or away from the main stream (Y, in white).

Figure 1

Table 1. Composite faecal samples (n = 72) obtained from cattle in a dairy catchment were analysed for E. coli O157 and non-O157 serogroups using real-time PCR (rtPCR). The samples positive by rtPCR were subjected to isolation. The virulence profile (stx1, stx2, eae, ehxA) of isolates obtained was determined using multiplex PCR

Figure 2

Fig. 2. Clustering (unweighted pair group method arithmetic mean, UPGMA dendrogram) of the PFGE profiles of 18 E. coli O26 isolates from 17 composite faecal samples taken from 11 farms in a catchment. Each farm is symbol coded and the catchment region is indicated. Note two isolates from sample EcCa38 are included in the analysis. The last lane is the Salmonella serotype Braenderup reference standard (H9812).

Figure 3

Fig. 3. Genotype (as defined by contrasting C. jejuni sequence types; STs) richness in cattle and water samples. Rarefaction curves indicating the estimated number of unique STs in cattle (black) and water (grey) for varying sample sizes, with 95% confidence intervals (CI).

Figure 4

Fig. 4. Minimum spanning tree showing the multilocus sequence types (STs) of C. jejuni isolated from cattle (black) and water (white) in the catchment stream. Each circle represents a ST; the size of the circle is proportional to the number of isolates and the area of each circle the relative frequency of each source. Isolates are grouped according to their genetic relatedness; short solid lines between STs indicate they vary at just one of the seven MLST loci (i.e. they are single locus variants).

Figure 5

Table 2. The distribution of C. jejuni sequence types (STs) isolated from cattle and water from the catchment stream.

Figure 6

Table 3. Permutational multivariate analysis of molecular variance as defined using farm structure grouped according to road network (Fig. 1a)

Figure 7

Table 4. Permutational multivariate analysis of molecular variance as defined using farm structure grouped according to proximity to main stream (Fig. 1b)