Detection and recovery rates achieved using direct plate and ...

Letters in Applied Microbiology 2005, 41, 88–93

doi:10.1111/j.1472-765X.2005.01667.x

Detection and recovery rates achieved using direct plate and enrichment/immunomagnetic separation methods for Escherichia coli O157:H7 in minced beef and on bovine hide S.B. O’Brien1, G. Duffy1, D. Daly1, J.J. Sheridan1, I.S. Blair2 and D.A. McDowell2 1

The National Food Centre, Teagasc, Ashtown, Dublin, Ireland, and 2Food Microbiology Research, NICHE, University of Ulster at Jordanstown, Newtownabbey, Belfast, UK

2004/0221: received 27 February 2004, revised 17 November 2004 and accepted 19 November 2004

ABSTRACT S . B . O ’ B R I E N , G . D U F F Y , D . D A L Y , J . J . S H E R I D A N , I . S . B L A I R A N D D . A . M C D O W E L L . 2005.

Aims: To assess the detection and recovery rates achieved with commonly used cultural methods for the enumeration and recovery of Escherichia coli O157:H7 from minced beef and bovine hide. Methods and Results: Minced beef and bovine hide were inoculated with varying concentrations (log10 1Æ58– 2Æ58 CFU g)1 and log10 2Æ42–4Æ49 CFU 100 cm2 respectively) of E. coli O157:H7 and recovered using a direct plate method or an enrichment/immunomagnetic separation (IMS) method and then plated onto SMAC or SMAC-CT in both cases. The direct plate method detected the pathogen consistently from minced beef samples with an average recovery of 69Æ2–91Æ2%. From faecal material on the bovine hide the recovery of the pathogen ranged from 1Æ80 to 64Æ5% with fresh faeces depending on the inocula while from dried faeces on hide the results ranged from no recovery at all to 25Æ1%. Enrichment/IMS recovered E. coli O157:H7 at all inocula levels tested in minced beef while the pathogen was only detected consistently at an average inocula level of log10 2Æ73 CFU 100 cm2 from fresh faeces and log10 4Æ49 CFU 100 cm2 from dried faeces on bovine hide. Conclusions: The direct count enumeration method for E. coli O157:H7 underestimated the numbers of pathogens present. The enrichment/IMS procedure consistently detected the pathogen from minced beef but did not always detect E. coli O157:H7 from faeces on bovine hide. Significance and Impact of the Study: Overall this study highlights that any microbial data, used in either predictive microbiology or risk assessment, must take account of the sensitivity and associated performance of the methods employed, in order to make an accurate reflection of the true microbiology of the examined sample. Keywords: E. coli O157:H7, enrichment, enumeration, hide, minced beef.

INTRODUCTION Verocytotoxigenic Escherichia coli, in particular serogroup O157:H7 are important food-borne pathogens, which can cause severe infection (Duffy et al. 2000). It is recognized that cattle are a primary reservoir of the pathogen and many outbreaks have been linked to beef (Chapman et al. 1993; Anon 1996; Little and de Louvois 1998; Heuvelink et al. 1999; De Boer and Heuvelink 2001). As such, many controls Correspondence to: G. Duffy, The National Food Centre, Teagasc, Ashtown, Dublin 15, Ireland (e-mail: [email protected]).

and risk management strategies have been targeted at controlling the pathogen in the abattoir environment (Dorsa et al. 1996; Gill and Bryant 1997; Nutsch et al. 1997, 1998; Phebus et al. 1997). The standard cultural method for the recovery of E. coli O157:H7 relies on an initial liquid enrichment stage followed by immunomagnetic separation (IMS) and plating onto selective agar. A wide range of enrichment media have been described for the isolation of E. coli O157:H7 from beef including mTSB (Hara-Kudo et al. 1999), mTSB + vancomycin, cefixime and cefsulodin, E. coli broth (EC) + ª 2005 The Society for Applied Microbiology

RECOVERY RATES FOR E. COLI O157:H7

novobiocin (Ogden et al. 2001); mEC broth + novobiocin (Heuvelink et al. 1997; Hara-Kudo et al. 1999) buffered peptone water (BPW) (Seo et al. 1998), BPW + vancomycin (Ogden et al. 2001), BPW + vancomycin, cefixime and cefsulodin (Ogden et al. 2001). Similarly, a range of media has been used in the isolation of E. coli O157:H7 from bovine faeces and/or hide include BPW (Tutenel et al. 2003) TSB (Barkocy-Gallagher et al. 2002) and sterile brilliant green bile 2% (Elder et al. 2000). The most widely employed agar is sorbitol MacConkey agar (SMAC) with or without the addition of cefixime and tellurite (Baylsis et al. 2001; Chapman 2001). While cultural techniques for E. coli O157:H7 have been subject to considerable investigation over the last number of years, few papers present information on detection limits or the scale of statistical error associated with these techniques. Relatively few studies report the numbers of pathogens present in samples, but some studies have used MPN methods (Fegan et al. 2003) or swab/plate samples directly onto agar (McEvoy et al. 2003) to assess pathogen numbers. The statistical confidence limits of the enumeration method used are generally not reported. With the emergence of quantitative risk assessment and food safety objectives as tools to assess and manage risk associated with food-borne pathogens, there is a growing need for quantitative and epidemiological data on the pathogens in environmental and food samples (Jericho et al. 1997; De Swarte and Donker 2003; Fegan et al. 2003). In addition, for both enrichment and enumeration based methods, it is essential to know the recovery limit and error associated with the microbial technique used so that it can be factored into risk calculations (Lammerding et al. 2001). The study aimed to examine the detection and recovery rates for E. coli O157:H7 achieved using (i) a standard enrichment/IMS procedure. While it has been established that this method performs well for the recovery of the pathogen from minced beef, its performance for problem matrices such as the hide has not yet been established. (ii) The enumeration of E. coli O157:H7 using a newly developed direct plate count method was examined for minced beef and hide to establish the levels of recovery achieved.

MATERIALS AND METHODS Escherichia coli O157:H7 A naturally occurring E. coli O157:H7 isolate, recovered from a bovine faecal sample (F 209) was confirmed as carrying enterohaemolysin A, eaeA, vt1 and vt2 genes by PCR using the method described by Cagney et al. (2004). The strain was stored on ProtectTM beads (Technical Consultant Services Ltd, Lancashire, UK) at )20
C.

89

Preparation of antibiotic-resistant E. coli O157:H7 (ABR) A toxigenic strain of E. coli O157:H7 (F 209), resistant to streptomycin sulfate (1000 lg ml)1) (Sigma Chemical Co., St Louis, MO, USA) and nalidixic acid (50 lg ml)1) (Sigma) was prepared as described by Park (1978). The strain was stored on ProtectTM beads.

BEEF Beef samples Rib steak purchased from a local butcher in the Dublin area was minced through a size 10 plate with 5 mm holes (Crypto Peerless Ltd, London, UK) and dispensed in 25 g amounts in filter stomacher bags (Seward filter bag model 400; Seward, London, UK). Inoculation of beef A ProtectTM bead coated with ABR E. coli O157:H7 was placed into 10 ml of brain–heart infusion broth (BHI) (Oxoid, Basingstoke, UK) and incubated for 24 h at 37
C. A serial dilution of the overnight culture was prepared in maximum recovery diluent (MRD) (Oxoid) (1 : 10, v/v) and 1 ml aliquots were gently dispersed into 25 g samples of minced beef to give final inocula of log10 1Æ58 or 2Æ58 CFU g)1. A duplicate of each 1 ml inocula was plated onto tryptone soya agar (TSA) (Oxoid) containing nalidixic acid (50 lg ml)1) and streptomycin sulfate (1000 lg ml)1) and incubated at 37
C for 24 h; to confirm the numbers and viability of the pathogen added to the beef. Inoculated minced beef samples (n ¼ 12) were examined using a variation of methods described in ISO 16654. In brief, samples were stomached for 1 min at 260 rev min)1 in 225 ml of EC broth (Difco Laboratories, Franklin Lakes, NJ, USA) or EC broth with 25 mg l)1 novobiocin (Sigma) and sampled for the direct count. The samples were then incubated at 37 or 41Æ5
C. Direct count Aliquots (1 ml) of the resultant suspension, and 10)1 and 10)2 dilutions (1 : 10, v/v) in MRD, were plated onto sorbitol MacConkey agar (SMAC) (Oxoid) or SMAC with cefixime (0Æ05 mg l)1) tellurite (2Æ5 mg)1) (Mast Diagnostics, Bootle, UK) (SMAC-CT) containing nalidixic acid (50 lg ml)1) and streptomycin sulfate (1000 lg ml)1). The plates were incubated overnight at 37
C for 22–24 h, and examined for suspect (nonsorbitol fermenting) colonies. Such colonies were enumerated and confirmed as E. coli O157 using the method described below.

ª 2005 The Society for Applied Microbiology, Letters in Applied Microbiology, 41, 88–93, doi:10.1111/j.1472-765X.2005.01667.x

90 S . B . O ’ B R I E N ET AL.

Enrichment After 6 and 24 h, 1Æ0 ml aliquots were removed from the broth cultures and dispensed into sterile 1Æ5 ml Eppendorf tubes (Sartedt Ltd, Co., Wexford, Ireland) containing 0Æ02 ml Dynal E. coli O157 immunomagnetic beads (Dynal A.S., Oslo, Norway). The pathogen/bead complex was extracted according to the manufacturers instructions and (50 ll aliquots) was plated onto SMAC or SMAC-CT containing nalidixic acid (50 lg ml)1) and streptomycin sulfate (1000 lg ml)1) and incubated at 37
C overnight. Suspect (nonsorbitol fermenting) colonies were tested as described below. Confirmatory tests Confirmatory tests were carried out a described by Cagney et al. (2004).

HIDE Hide pieces A fresh bovine hide was obtained from a local abattoir and transported to the laboratory within 1 h. Visibly clean pieces of hide measuring 10 cm · 18 cm were removed, secured in sterilized stainless steel frames, and stored for up to 24 h at 2
C. Escherichia coli O157:H7 inocula A ProtectTM bead coated with the E. coli O157:H7 ABR strain was aseptically placed into 10 ml of BHI and incubated overnight at 37
C for 24 h. A 1Æ0 ml aliquot of the overnight culture was transferred to 100 ml of BHI. The inoculum was then incubated at 37
C for 20 h. The resultant cells were recovered by centrifugation at 5000 g (Eppendorf Centrifuge 5403, Hamberg, Germany) for 10 min, washed three times in MRD, and resuspended in 100 ml of MRD. A serial dilution of the inoculum (1 : 10, v/v) in MRD was prepared to obtain inocula of approximately log10 3Æ0, 4Æ0, 5Æ0, 6Æ0 CFU ml)1. A 1 ml aliquot of each inocula was plated (in duplicate) onto TSA containing nalidixic acid (50 lg ml)1) and streptomycin sulfate (1000 lg ml)1) and incubated at 37
C for 24 h; to confirm the numbers and viability of the pathogen. Preparation of faecal inocula Freshly voided bovine faeces were aseptically recovered from cattle on site at The National Food Centre, transferred to the laboratory within 30 min and used to prepare two forms of faecal inoculum as follows.

i Inoculated faeces inoculum (IFI) prepared by mixing 5 ml of one of the above E. coli O157:H7 inocula levels with 50 g of fresh bovine faeces. ii Diluted inoculated faeces inoculum (DIFI) prepared by mixing 50 ml of one of the above E. coli O157:H7 inocula levels to 16Æ5 g of bovine faeces. iii No faeces inoculum (NFI) of 5Æ0 ml of one of the E. coli O157:H7 inocula was used as control. The entire volume of each of the above inocula was aseptically spread over the entire surface of duplicate pieces of hide giving final inoula of log10 2Æ73, 3Æ53 and 4Æ13 CFU 100 cm2. The first member of each duplicate set of hide pieces was left at room temperature for 15 min, to constitute a fresh sample. The second member of each duplicate set of hide pieces was stored for 48 h at 15
C to simulate dried faeces on the hide. After storage the hide pieces were swabbed (n ¼ 6) in a uniform fashion using a sterile sponge presoaked in 6 ml of Universal Quenching Agent (sodium thiosulfate: 1 g, bacteriological peptone: 1 g, lectin soya extract: 0Æ7 g, tween 20: 5 ml, deionized water: 1 l). Samples were analysed using a variation of the UK national reference method for the detection/isolation and presumptive identification of E. coli O157. The swabs were placed in 90 ml of BPW (Oxoid) and stomached for 1 min at 260 rev min)1. Samples (1 ml), and appropriate dilutions in MRD (1 : 10, v/v), of the stomached suspensions were examined by the direct plate method as described for beef. Stomached suspensions, plus swabs, were incubated at 37
C and examined by the enrichment and IMS procedure as described for beef samples. Statistical analysis Results were analysed using Genstat 5, Release 3.2 (Procedure Library Manual, Lawes Agricultural Trust, Rothamsted Experimental Station, Harpenden, Herts., UK). The data were analysed as a factorial experiment using ANOVA, Friedman’s test (comparing paired groups) and Kruskal–Wallis one-way analysis of variance (comparing unpaired groups) were used as post-tests. For minced beef samples the detection limit of the direct plate method was compared on SMAC and SMAC-CT and the effects of incubation temperature, broth and selective agar were compared within the enrichment/IMS method for E. coli O157:H7. For hide samples the detection limit of the direct plate method was compared on SMAC and SMACCT. In addition for each faecal treatment, the sensitivity of SMAC and SMAC-CT were compared for the recovery of E. coli O157:H7 by enrichment/IMS. Treatments were tested at a 95% confidence level for any significant differences.

ª 2005 The Society for Applied Microbiology, Letters in Applied Microbiology, 41, 88–93, doi:10.1111/j.1472-765X.2005.01667.x

RECOVERY RATES FOR E. COLI O157:H7

RESULTS AND DISCUSSION Direct count The effects of inoculum level and plating media on numbers of E. coli O157:H7 (log10 CFU g)1) recovered from inoculated mince beef using a direct plate method are presented in Table 1. The recovery rates of the pathogen from mince beef ranged from 89Æ1 to 91Æ2% on SMAC and from 69Æ2 to 79Æ4% on SMAC-CT. At a 95% confidence level there was no significant difference between SMAC or SMAC-CT agar. The effects of faecal treatment, inoculum level and plating media on recovery rates of E. coli O157:H7 (log10 CFU 100 cm2) from the bovine hide using a direct plate method, are presented in Table 2. The recovery rates of the pathogen were highly variable ranging from 1Æ80 to 64Æ5% for fresh faeces and from 0Æ02 to 25Æ1% for dried faeces on hide. At a 95% confidence level no significant difference was found

Table 1 Effect of inoculum level and plating media on the number (log10 CFU g)1) and Escherichia coli O157:H7 recovered (%) from mince beef using a direct plate method Inoculum (log10 CFU g)1)

SMAC, n (%)

SMAC-CT, n (%)

1Æ58 2Æ58

1Æ54 (91Æ2) 2Æ53 (89Æ1)

1Æ48 (79Æ4) 2Æ42 (69Æ2)

SMAC, sorbitol MacConkey agar; SMAC-CT, sorbitol MacConkey agar with cefixime and tellurite Inoculum level: averaged value. Each result is an average of seven replications. No significant difference was found between SMAC and SMAC-CT.

91

between either SMAC or SMAC-CT agar with the exception of the NFI treatment at an inoculum of log10 4Æ13 CFU 100 cm2 where, SMAC was found to be significantly better then SMAC-CT for the recovery of the pathogen. This study concluded that the enumeration of E. coli O157:H7 by direct plating on SMAC or SMAC-CT recovered only a small percentage of the cells present in faeces on the bovine hide. At lower levels of inoculation (