Virulence traits associated with ... - Wiley Online Library

3 downloads 3906 Views 524KB Size Report
Virulence traits associated with verocytotoxigenic. Escherichia coli O157 recovered from freshwater biofilms. I.R. Cooper1, H.D. Taylor2 and G.W. Hanlon1.
Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

Virulence traits associated with verocytotoxigenic Escherichia coli O157 recovered from freshwater biofilms I.R. Cooper1, H.D. Taylor2 and G.W. Hanlon1 1 School of Pharmacy and Biomolecular Sciences, University of Brighton, Brighton, UK 2 School of the Environment, University of Brighton, Brighton, UK

Keywords biofilm, E. coli O157, faeces, polymerase chain reaction, phenotyping, water. Correspondence I.R. Cooper, School of Pharmacy and Biomolecular Sciences, University of Brighton, Lewes Road, Brighton, BN2 4GJ, UK. E-mail: [email protected]

2006/0065: received 19 January 2006, revised 21 August 2006 and accepted 24 August 2006 doi:10.1111/j.1365-2672.2006.03178.x

Abstract Aim: To investigate whether epilithic biofilms in freshwater streams in a mixed UK agricultural river catchment harbour Escherichia coli O157, and if so, whether they demonstrate an association with those excreted by grazing farm animals. Methods and Results: Flint shingle, native to the study site, was used as a surface for biofilm development within cages of metal lath set into a stream bed at four locations on a chalkland farm. Shingle was collected from all sites once a month, as were pooled faecal samples from five farm animal populations. Subpopulations of E. coli, including E. coli O157 that demonstrated significant phenotypic and genotypic similarity with animal faecal isolates (t-test, P ¼ 0Æ05) were isolated. Of 1002 E. coli isolates from biofilms and animal faeces, 48 were confirmed as the O157 strain by latex agglutination. The presence of five virulence traits associated with incidence of human disease was tested using PCR. Stx2 was the most frequently isolated single gene (30 isolates), while stx1 was the least frequently recovered (four isolates). Conclusion: Escherichia coli O157, expressing up to four virulence factors associated with human disease, reside within freshwater biofilms in this agricultural environment. Significance and Impact of the Study: Aquatic biofilms may potentially act as a reservoir for these pathogens, and the implications of the findings for the protection of drinking water resources should be further investigated.

Introduction Escherichia coli O157 is one of a range of enterohaemorrhagic E. coli (EHEC) strains that can be distinguished by its inability to ferment sorbitol and by its production of the intimin protein and two verocytotoxins, hence the term verocytotoxigenic E. coli (VTEC) (Chaisri et al. 2001; Rogerie et al. 2001). The Shigella dysenteriae type 1 toxin and the two verocytotoxins present high amino acid sequence homology (Paton and Paton 1998). The latter are implicated in two potentially fatal human diseases, namely, haemorrhagic colitis (HC) and haemolytic uraemic syndrome (HUS) (Kudva et al. 1998). The replication of E. coli has been demonstrated in Hawaiian tropical soils (Byappanahalli and Fujioka

1998), and the bacterium has been shown to survive within faecal pellets for up to 9 months (Bailey et al. 2003). Although the majority of E. coli strains are commensal enteric organisms, it has been demonstrated that the species is able to survive for at least 48 h on inorganic plastic and metal substrata (Momba and Kaleni 2002), and in sources of ground water (Moran et al. 1997). Further, the bacterium displays an ability to survive in acidic environments, including food products (apple juice and cider), which might explain its ability to survive passage through the mammalian gastro-intestinal tract (Price et al. 2004). If E. coli O157 were to demonstrate similar resilience in aquatic biofilms, then these biofilms would potentially represent a reservoir of pathogenic organisms, and hence

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

1293

Virulent E. coli O157 in biofilms

I.R. Cooper et al.

a hazard to water consumers and recreational water users. The increased ‘sloughing-off’ of biofilms during storm events may introduce localized high numbers of the pathogen into drinking water supplies. Water sources previously believed to be free of E. coli O157 could in fact harbour the organism in the sediments and biofilm fractions, which are not routinely monitored by water and public health organizations. While a number of longitudinal epidemiological studies have been conducted along the ‘farm-to-fork’ continuum, little is known of the role of aquatic biofilms in the environmental transmission of VTEC E. coli. This paper presents a comparison of E. coli O157 organisms in animal faeces with those found in the biofilms of streams on a chalkland farm in southern England, UK. Materials and methods The mixed farm environment A mixed farm (arable and animal husbandry), situated on the South Downs in East Sussex, UK, was selected for the field studies. The farm is situated approx. 4Æ5 km west of the county town of Lewes, below the northern slopes of the South Downs. A preliminary unpublished study at the farm informed the choice of sites to locate the four biofilm cages. At each of these sites, approx. 1 kg of shingle were placed in metal lath cages (measuring 10 · 25 · 50 cm, with a 7Æ5-cm2 rhomboid grid allowing fairly unobstructed water flow) in order to promote biofilm formation while the cages were submerged in the river. These sites were all located within an area of 1 km2. Site 1 was located next to the main farm buildings and approx. 50-m downstream from the slurry lagoons (where animal faecal matter from the barns and pens was semitreated) and approx. 300 m from the livestock housing. This site was chosen because large quantities of faecal matter were consistently observed in the vicinity, i.e. the presence of the slurry lagoons and animal housing. The stream flow rate ranged from 0Æ06 m3 s)1 in July to 0Æ93 m3 s)1 in November. Site 2 was located in a slight depression in the riverbed of a separate stream to site 1, and was situated between a cattle-grazing and a wheat field during the period of study. A land drainage pipe discharged water following rainfall 50-cm upstream of the cage site. The stream flow rate ranged from 0Æ04 m3 s)1 in July to 0Æ90 m3 s)1 in October. Site 3 was located in a narrow stony section of stream, approx. 500 m downstream of site 1. It was flanked on both banks by fields grazed by horses and cattle, and while cattle grazed intermittently throughout the year (except the winter), horses were only present during 1294

May–July when show-jumping took place. The stream flow rate ranged from 0Æ04 m3 s)1 in April to 1Æ00 m3 s)1 in October. At site 4, the stream was bordered by two wheat fields during the study period. The stream drained much of the run-off from a road running adjacent to the South Downs, and the flow ranged from 0Æ89 m3 s)1 in April to 1Æ43 m3 s)1 in the following January. Sampling procedure Samples of shingle (mean surface area approx. 5 cm2) were collected from all sites at approximately monthly intervals over a 12-month period, and transported to the laboratory in isotonic saline in a light-proof chilled container (approx. 4C) within 2 h. The adherent biofilm was then dislodged by low frequency oscillation (100 oscillations per minute for 5 min). Based on a previous comparison of techniques, it was determined that this procedure recovered the greatest number of viable colony forming units (CFU) per unit area of stone surface area. This approach demonstrated good reproducibility, and proved to be superior to swab and scraping methodologies. On each sampling occasion, eight stones were taken from each cage. On the same day, pooled animal faeces were collected from fresh faeces from each of the pig, cattle, chicken, sheep and goat populations. The eight stones were immediately immersed gently into 200-ml isotonic saline solution, and the animal faecal samples were diluted at a ratio of 1 : 2 (weight : volume, w/v) with 200-ml isotonic saline. Characterization of isolates Ten millilitres of each tenfold serial dilution were filtered through 0Æ45-lm nitrocellulose membranes (Whatman). The membranes were incubated at 30C for a minimum of 18 h on absorbent pads soaked in lauryl sulphate broth (Oxoid). Presumptive E. coli colonies were then enumerated, i.e. all yellow colonies irrespective of size (Anon 2005a). Standard UK IMViC tests were used to characterize the coliform isolates (where ‘I’ stands for indole production from tryptophan, ‘M’ indicates acidic carbohydrate metabolism demonstrated by Methyl Red colour change, ‘Vi’ stands for the Voges-Praskauer test for fermentative metabolism, and ‘C’ for citrate utilisation) (Anon 2002). From the results of these tests, confirmed E. coli isolates were screened for sorbitol fermentation and 4-methylumbelliferyl glucuronidase (MUG) activity as a marker of the O157 strain (Villari et al. 1997; Anon 2005b). These tests were supported by performing a Gram stain on all pre-

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

I.R. Cooper et al.

sumptive E. coli. Latex agglutination was carried out using the DryspotTM O157 test kit (Oxoid, Basingstoke, UK) to detect the somatic O157 antigen, according to the manufacturer’s instructions. In vitro Escherichia coli phenotyping Phenotypic characterization of E. coli isolates was performed using the PhenePlateTM methodology, described in detail by Ku¨hn et al. (2003). In brief, the system uses dehydrated carbon sources (cellobiose, lactose, rhamnose, deoxyribose, sucrose, sorbose, tagatose, arabitol, melbionate, galacturonic acid-lactose and ornithine), stabilized within the 96 wells of standard microtitre plates, to analyse the biochemical kinetics of a bacterial culture added to the wells. Test isolates were harvested from nutrient agar, recultivated to ensure purity and incubated on nutrient agar at 37C for 18 ± 2 h. Three colonies were then picked using a sterile loop, and subsequently suspended in a liquid medium which rehydrates the substrates, thus enabling the reactions to be read. This is termed the PhenePlateTM suspension medium (Ku¨hn et al. 2003). Within the microtitre plates, except for the first column, each column of wells contained a dehydrated substrate which was resuspended with the substrate medium. Using a multipipette (L04281; Labsystems Ltd, Helsinki, Finland), the first row of wells were filled with the suspending medium (325 ll), and the rest with 50 ll. Using sterile loops, the first column of the wells of the microtitre plate were inoculated with three colonies of the test isolate, stirred and allowed to homogenize. The remaining wells were inoculated with 25 ll of inoculated medium from the first row of wells. The plates were then covered with a sterile lid and incubated at 37C. The plates were read in a multiplate reader (ELx800; Bio-Tek Instruments Inc, Winooski, VT, USA) after 7, 24 and 48 h of incubation, so as to yield information on the kinetics of the biochemical reactions using the PhenePlateTM software. A cluster analysis was performed on the data, which grouped isolates based on the kinetics of isolate metabolism. The reaction data were used to calculate phenotypic population similarities using Simpson’s index of diversity (Di). The Di is high in a population consisting of many different phenotypes (PhP-types) and low if certain PhPtypes dominate in the population, with the test result ranging between 1Æ0 and 0Æ0, respectively. A coefficient higher than 0Æ975 (Ku¨hn et al. 1990; Wada˚s et al. 1996) indicates that the isolates compared were of the same clone, in which case they were assigned to the same biochemical PhP-type. A lower correlation coefficient suggests that they were very similar. Isolates with correlation coefficients lower than 0Æ950 were regarded as not related.

Virulent E. coli O157 in biofilms

Simpson’s index of diversity is used to quantify the biodiversity of a habitat by taking into account the number of species present and the abundance of each species (Onaindia et al. 2004). This probability can be calculated by the following equation: D¼1

s X 1 nj ðnj  1Þ NðN  1Þ j ¼ 1

Where N is the total number of strains in the sample population, s is the total number of types described and nj is the number of strains belonging to the jth type (Hunter and Gaston 1988). Polymerase chain reaction Polymerase chain reaction (PCR) was used to determine whether the E. coli O157 isolates obtained from the biofilm and animal faeces possessed the genes necessary to produce virulence factors associated with incidence of human disease. Four virulence factors were tested for: verocytotoxin-1 (stx1), verocytotoxin-2 (stx2), an enterohaemolysin (hly), intimin (eaeA). Alongside this, two enterobacterial repetitive intergenic consensus (ERIC) sequences were tested (no single gene). These sequences have been used to type enterobacteria, but were used here to determine prevalence among the E. coli O157 isolates. A protocol was developed as follows: three colonies of each isolate were inoculated into 100-ll sterile deionized water in an Eppendorf tube, and kept at 95C for 10 min in order to kill all viable cells in the suspension. Each tube was centrifuged at 12 000 g for 10 min. The supernatant liquid was discarded and the pellet was resuspended in 20-ll sterile deionized water. To this, 5 ll of the DNA Taq enzyme mixture was added, followed by 25-ll enzyme premix set L (Cambio Ltd, Cambridge, UK), and 1Æ0-ll primer suspension (diluted to 50 pmol ll)1). The primer sequences used were as follows (Qiagen Ltd., Crawley, UK): STX1: 5¢-GAA GAG TCC GTG GGA TTA CG-3¢ and ‘AGC GAT GCA GCT ATT AAT AA-3¢; STX2 5¢-TTA ACC ACA CCC ACG GCA GT-’3 and 5¢- GCT CTG GAT GCA TCT CTG GT-3¢; eaeA: 5¢-CAG GTC GTC GTG TCT GCT AAA-3¢ and 5¢-TCA GCG TGG TTG GAT CAA CCT-3¢; hly: 5¢-ACG ATG TGG TTT ATT CTG GA-3¢ and 5¢-CTT CAC GTC ACC ATA CAT AT-3¢ (Paton and Paton 1998); and finally ERIC sequences: 5¢-CAC TTA GGG GTC CTC GAA TGT A-3¢ and 5¢-AAG TAA GTG ACT GGG GTG AGC G-3¢ (Leung et al. 2004). Each primer pair was used in combination, and each amplification step was performed according to the following conditions in a Darwin Thermocycler (Labtech International, Ringmer, UK). Initial denaturation was

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

1295

Virulent E. coli O157 in biofilms

I.R. Cooper et al.

performed at 95C for 5 min, 65C annealing temperature for 1 min, followed by separate cycles for each set of primers. For the STX1 primers, this was: 35 cycles of 44Æ3C for 1 min and 58Æ35C for 1 min; for the STX2 primers: 35 cycles of 52Æ3C for 1 min and 62Æ45C for 1 min; for the eaeA primers: 35 cycles of 51Æ2C for 1 min, and 62Æ57C for 1 min; for the hly primers: 35 cycles of 39Æ9C for 1 min, and 56Æ3C for 1 min; and for the ERIC sequences 50Æ15C for 1 min and 63Æ61C for 1 min. All cycles were completed with a final extension at 72C for 10 min and then stored at 4C. The presence of PCR products was visually confirmed by agarose gel (2Æ0% w/v) electrophoresis. Samples were marked using 16Æ6 ll of a 1-in–6 tracking dye suspension (FailSafeTM; Cambio, Cambridge, UK) in order to facilitate gel loading. Gel electrophoresis was run at 200 mV for 30 min. The gel was removed and the DNA stained using 5-ll ethidium bromide (1% w/v ethidium bromide aqueous solution) per 100-ml tris-acetate-EDTA (TAE) buffer in a staining tank for 30 min. Statistical analyses The PhenePlateTM analysis, linkage clustering is applied to the isolates. In brief, the PhenePlateTM software includes mathematical models to calculate the diversity (Di) of the bacterial isolates within samples and the similarities between bacterial populations, expressed as the population similarity coefficient (Sp), using the complete agglomerative and complete linkage techniques to produce a hierarchical clustering of the test isolates. In such methods, the members of inferior-ranking clusters become members of larger, higher-ranking clusters (Legendre and Legendre 2003). Clusters are formed in such a way that the ‘withingroup’ association matrices have a given probability of being homogenous, meaning that the isolates are likely to belong to the same subspecies, or related isolates from the same host organism (Clifford and Goodall 1967). The PhenePlateTM software also uses the Unweighted Pair Group Method with Arithmetic mean (UPGMA) analysis to group clusters, where the highest similarity or smallest difference identifies the next cluster to be formed (Sneath and Sokal 1973). Following initial grouping, the method computes the arithmetic average of the similarities or distances between a candidate object and each of the cluster members, or, in the case of a previously formed cluster, between all members of the two clusters. All objects receive equal weights in the computation. The similarity or distance matrix is updated and reduced in size at each clustering step. Clustering proceeds by agglomeration as the similarity criterion is relaxed, just as it does in single linkage clustering (Legendre and Legendre 2003). 1296

Table 1 Summary of the total Escherichia coli and E. coli O157 isolates recovered from each of the four biofilm sites and from pooled faeces from each of the five animal populations sampled Source of E. coli O157 isolate

Total E. coli (including the O157 strains)

E. coli O157 only

Biofilm site 1 Biofilm site 2 Biofilm site 3 Biofilm site 4 Cattle faeces Goat faeces Pig faeces Chicken faeces Sheep faeces

190 150 159 111 392 50 30 53 14

5 13 4 4 4 14 2 2 0

Results Table 1 shows the number of all E. coli and confirmed E. coli O157 isolates recovered from each source. The total number of E. coli and E. coli O157 isolates varied considerably between both biofilm and animal populations; the highest number of E. coli being recovered from pooled cattle faeces, while the highest number of E. coli O157 was found in pooled goat faeces. E. coli phenotyping Biofilm-derived E. coli O157 isolates Carbon source utilization (CSU) analysis of all the E. coli isolates recovered from the epilithic biofilms indicated that numerous PhP-types (as determined by the PhenePlateTM technique) were isolated from the biofilms. Four E. coli O157 isolates presented a significant similarity to the CSU phenotypes of E. coli O157 from animal faeces (t-test, P ¼ 0Æ05), perhaps suggesting a clonal link between the two populations (Fig. 1). Analysis revealed that while most E. coli isolates recovered in consecutive months showed different biochemical PhP-types, several isolates showed a high phenotypic similarity, suggesting common origin. Animal-derived E. coli O157 isolates Examples of strain persistence were also demonstrated in animal faeces. Cattle faeces isolates recovered through June to October displayed a Di ‡0Æ988, suggesting that E. coli O157 of the same PhP-type were present in the cattle herd during these months. The dendrogram in Fig. 1 shows phenotypic similarity among all biofilm and faeces isolates of E. coli O157 combined with the results from the PCR-based genotyping. This dendrogram reveals two important findings. Firstly, isolates recovered from both biofilm and animal populations cluster closely together, perhaps suggesting a com-

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

I.R. Cooper et al.

Virulent E. coli O157 in biofilms

Similarity index (Di)

0·75

0·80

0·85

0·90

0·95

3(1)MAR stx2, hly, ERIC 2(11)JUN stx2, ERIC 1(11)NOV 4(2)NOV hly 2(46)SEP stx2, eaeA 2(13)SEP stx2, hly, ERIC 1(75)JUN 4(47)OCT stx1, stx2 2(7)SEP GT22 stx2 GT8 stx2 GT5 stx2, eaeA, hly GT7 stx2, eaeA GT9 stx2 CW(62) stx2, ERIC CH25 stx2, eaeA, ERIC PG29 stx1, stx2, eaeA CW2 stx2, ERIC CW57 stx2, hly 1(7)JUL stx2, hly GT18 stx2, eaeA, hly PG20 stx2, eaeA, ERIC 3(55)JUL stx2, hly, ERIC 2(18)OCT stx2 2(67)OCT stx2, ERIC 2(69)OCT stx2, ERIC 2(73)OCT stx2, eaeA 2(52)OCT eaeA 2(22)OCT stx2, ERIC 2(33)OCT stx1, eaeA 3(31)MAY ERIC CW(23) hly, ERIC 2(28)FEB hly, ERIC 1(11)AUG stx1, eaeA, hly 2(13)OCT GT 31 stx2, GT25 stx2, ERIC GT40 hly GT26 stx2 CH18 stx2, eaeA, ERIC GT34 eaeA GT35 stx2, eaeA, hly GT41 stx2, eaeA, hly 4(56)MAY 3(44)JUL ERIC 4(59)MAY ERIC 1(46)MAR

Figure 1 Dendrogram demonstrating phenotypic clustering of all confirmed E. coli O157 isolates using Simpson’s index of diversity (the dotted line at Di ¼ 0Æ975 is minimum species similarity to consider isolates as belonging to the same clonal population), and the virulence genes detected in each isolate using PCR. Key: GT, goat; CH, chicken; CW, cow; PG, pig; 2(11)JUN, biofilm site number (isolate number) month of recovery; stx1, verocytotoxin 1 gene; stx2, verocytotoxin 2 gene; eaeA, intimin gene; hly, haemolysin gene; ERIC, enterobacterial repetitive intergenic consensus sequences.

mon origin, and secondly, isolates expressing the same virulence genes cluster together in some, but not all instances. This may suggest that the phenotypic character of E. coli O157 isolates is loosely associated with the genotype of the bacteria.

1·00

Polymerase chain reaction Biofilm-derived E. coli O157 isolates Among the biofilm isolates, the most common single virulence factor was the stx2 allele and the most common

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

1297

Virulent E. coli O157 in biofilms

I.R. Cooper et al.

+ combination of alleles was stxþ 2 ERIC (Fig. 1). Interestingly, the biofilm isolates included three isolates that possessed the stx1 allele, which was not expressed in the animal-derived isolates.

Animal-derived E. coli O157 isolates No significant similarities were observed between the pattern of virulence factors among the animal and biofilm isolates using standard t-tests (P ¼ 0Æ05) (Fig. 1). Among the animal-derived isolates, the most common single virulence factor was the possession of the stx2 allele, and + + the most common combination was stxþ 2 hly ERIC . No seasonal pattern was associated with the expression of virulence factors by the E. coli O157 isolates recovered from pooled animal faeces (t-test, P ¼ 0Æ05). Discussion In 2005, Landgren et al. reported that of the 2481 E. coli isolates from women diagnosed with lower urinary tract infections from 17 European countries tested by the PhenePlateTM system, 2067 isolates belonged to one of the 74 common CSU PhP-types, and the rest belonged to 414 single types. This suggests that the PhP-type of disease-causing E. coli isolates is geographically independent, and that isolates presenting the same PhP-type in vitro are likely to belong to the same clonal population. Further, Murinda et al. (2004) compared the pathogenic profiles of VTEC O157:H7 isolates from human and animal populations using multiplex PCR to suggest that interspecies transmissibility between humans and cattle is possible, indicating that the bacterium is capable of existing in and moving between different host species, and, in effect, host environments. In the United Kingdom, HUS is the most common cause of acute renal failure in children (Hunter 2003). An understanding of the behaviour of the causative bacterium in the environment is important to the prevention of future cases of disease. The presence of viable and virulent E. coli O157 isolates in aquatic biofilms represent a potential hazard to human health. Although aquatic biofilms may not be a fundamental component of the life cycle of human pathogens, such as E. coli O157, these biofilms are increasingly recognized as an important reservoir of pathogenic bacteria. This paper demonstrates that E. coli O157, related to agricultural animal populations, are able to exist within aquatic epilithic biofilms and retain the potential to cause disease in human populations. References Anon (2002) PHLS standard operating procedure: detection of Escherichia coli. A Health Protection Agency Publication.

1298

Anon (2005a) PHLS standard operating procedure: enumeration of coliforms and Escherichia coli by membrane filtration. A Health Protection Agency Publication. Anon (2005b) PHLS standard operating procedure: identification of Escherichia coli O157. A Health Protection Agency Publication. Bailey, J., Warner, L., Prichard, G., Williamson, S., Carson, T., Willshaw, G. and Cheasty, T. (2003) Wild rabbits – a novel vectors for Vero cytotoxigenic Escherichia coli (VTEC) O157. Commun Dis Public Health 5, 74–75. Byappanahalli, M. and Fujioka, R. (1998) Evidence that tropical soil environment can support the growth of Escherichia coli. Water Sci Technol 38, 171–174. Chaisri, U., Nagata, M., Kurazono, H., Horie, H., Tongtawa, P., Hayashi, H., Watanabe, T., Tapchaisri, P. et al. (2001) Localization of Shiga toxins of enterohaemorrhagic Escherichia coli in kidneys of paediatric and geriatric patients with fatal haemolytic uraemic syndrome. Microb Pathog 31, 59–67. Clifford, H. and Goodall, D. (1967) A numerical contribution to the classification of the Poaceae. Aust J Botanics 15, 499–519. Hunter, P. (2003) Drinking water and diarrhoeal disease due to Escherichia coli. J Water Health 1, 65–72. Hunter, P. and Gaston, M. (1988) Numerical index of the discriminatory ability of typing systems: an application of Simpson’s index of diversity. J Clin Microbiol 26, 2465– 2466. Kudva, I., Blanch, K. and Hovde, C. (1998) Analysis of Escherichia coli O157:H7 survival in ovine or bovine manure and manure slurry. Appl Environ Microbiol 64, 3166–3174. Ku¨hn, I., Brauner, A. and Mo¨llby, R. (1990) Evaluation of numerical typing systems for Escherichia coli using the API 50 CH and the PhP-EC systems as models. Epidemiol Infect 105, 521–531. Ku¨hn, I., Iverson, A., Burman, L., Olsson-Lijequist, B., Franklin, A., Finn, M., Aarestrup, F., Mette Seyfarth, A. et al. (2003) Comparison of enterococcal populations in animals, humans, and the environment – a European study. Int J Food Microbiol 88, 133–145. Landgren, M., Oden, H., Ku¨hn, I., Osterlund, A. and Kahlmeter, G. (2005) Diversity among 2481 Escherichia coli from women with community-acquired lower urinary tract infections in 17 countries. J Antimicrob Chemother 55, 928–937. Legendre, P. and Legendre, L. (2003) Developments in environmental modelling 2.0. Numerical Ecology, 2nd English Ed. Elsevier. Leung, K., Mackereth, R., Yuan-Ching, T. and Topp, E. (2004) A comparison of AFLP and ERIC-PCR analyses for discriminating Escherichia coli from cattle, pig and human sources. FEMS Microbiol Ecol 47, 111–119. Momba, M. and Kaleni, P. (2002) Regrowth and survival of indicator microorganisms on the surfaces of household containers used for the storage of drinking water in rural communities of South Africa. Water Res 36, 3023–3028.

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

I.R. Cooper et al.

Moran, P., Nhandara, C., Hove, I., Charimari, L., Katito, C., Bradley, M. and Williams, M. (1997) Contamination of traditional drinking water sources during a period of extreme drought in the Zvimba communal lands, Zimbabwe. Cent Afr Med 43, 316–321. Murinda, S., Nguyen, L., Landers, T., Draughon, F., Mathew, A., Hogan, S., Smith, K., Hancock, D. et al. (2004) Comparison of Escherichia coli isolates from humans, food, and farm and companion animals for the presence of Shiga toxin-producing E. coli virulence markers. Foodborne Pathogen Dis 1, 178–184. Onaindia, M., Dominguez, I., Albizu, I., Garbisu, C. and Amezaga, I. (2004) Vegetation diversity and vertical structure as indicators of forest disturbance. Forest Ecol Manag 195, 341–354. Paton, A. and Paton, J. (1998) Detection and characterization of Shiga toxigenic Escherichia coli by using multiplex PCR assays for stx1, stx2, eaeA, Enterohemorrhagic E. coli hlyA, rfbO111, and rfbO157:H7:H7. J ClinMicrobiol 36, 598– 602. Price, S., Wright, J., DeGraves, F., Castanie-Cornet, M-P. and Foster, J. (2004) Acid resistance systems required for survi-

Virulent E. coli O157 in biofilms

val of Escherichia coli O157:H7 in the bovine gastrointestinal tract and in apple cider are different. Appl Environ Microbiol 70, 4792–4799. Rogerie, F., Marecat, A., Gambade, S., Dupond, F., Beaubois, P. and Lange, M. (2001) Characterization of Shiga toxin producing E. coli and O157 serotype E. coli isolated in France from healthy domestic cattle. Int J Food Microbiol 63, 217–223. Sneath, P. and Sokal, R. (1973) Numerical Taxomony. pp. 644. San Francisco: W.H. Freeman. Villari, P., Iannuzzo, M. and Torre, I. (1997) An evaluation of the use of 4-methylumbelliferyl-beta-D-glucuronide (MUG) in different solid media for the detection and enumeration of Escherichia coli in foods. Lett Appl Microbiol 24, 286–290. Wada˚s, B., Ku¨hn, I., Lagerstedt, A.-S. and Jonsson, P. (1996) Biochemical phenotypes of Escherichia coli in dogs: comparison of isolates isolated from bitches suffering from pyometra and urinary tract infection with isolates from faeces of healthy dogs. Vet Microbiol 52, 293–300.

ª 2006 University of Brighton Journal compilation ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 102 (2007) 1293–1299

1299