Genetic Relatedness of Escherichia coli O157 Isolates from Cattle Feces and Preintervention Beef Carcasses

Abstract

Our objective was to define and compare pulsed-field gel electrophoresis (PFGE) profiles of Escherichia coli O157 isolated from cattle feces and carcass samples to evaluate relationships between beef carcass contamination and fecal shedding of E. coli O157 at harvest. We used PFGE separation of Xba1-digested DNA to characterize E. coli O157 isolates (n = 174) from preevisceration carcasses (n = 39) and feces (n = 135) that were recovered from 37 E. coli O157–positive truckloads sampled at a commercial abattoir. Semiquantitative fecal culture techniques differentiated high-shedding, low-shedding, and negative cattle. Among all isolates, there were 17 PFGE types (95% homology) and 37 subtypes (100% homology). Specific subtypes were detected on multiple occasions and from different sample types within loads, among loads, and among days. Seventeen subtypes were recovered from carcasses; most were also recovered from feces of high-shedding cattle (13) and low-shedding cattle (14). Within truckload, the percentages of carcass isolates that were identical to high-shedder or low-shedder fecal isolates, as determined by PFGE, were 69.2% and 46.0%, respectively, whereas among different truckloads within the same study day, the percentages of carcass isolates that were the same subtype as high-shedder or low-shedder fecal isolates were 35.3% and 58.8%, respectively. Our results suggest that cattle feces from both low- and high-shedders pose a potential risk for E. coli O157 contamination of carcasses. Truckload may be an important factor in the potential transmission of E. coli O157, but isolates from carcasses also may be similar to those from feces of cattle on different truckloads and harvest days.

Introduction

E scherichia coli O157 is a significant cause of foodborne illness in the United States. From 1982 to 2002, there were 350 reported outbreaks of E. coli O157 including 8598 human illnesses, 1493 hospitalizations, and 40 deaths (Rangel et al., 2005). Although there are more cases of foodborne illness due to other pathogens each year, E. coli O157 outbreaks and related product recalls have generated substantial consumer risk awareness, impacting consumer demand for beef (Marsh et al., 2004; CDC, 2008). Properly cooking beef will kill the organisms, yet pre- and postharvest interventions are still needed to lower E. coli O157 risk since cross-contamination and under-cooking still occur (Caprioli et al., 2006; Verbeke et al., 2007).

A primary pathway of E. coli O157 contamination of beef involves colonized cattle that shed organisms in feces, which subsequently contaminate hides (Loneragan and Brashears, 2005; Baker et al., 2007). Most carcass contamination is believed to occur during the hide-removal and evisceration process (Barkocy-Gallagher et al., 2003); E. coli O157 in cattle feces or on hides is associated with detectable levels of E. coli O157 on carcasses (Elder et al., 2000; McEvoy et al., 2003; Woerner et al., 2006). Some cattle with E. coli O157 shed greater levels, more than 10³ or 10⁴ colony forming units (CFU) per g of feces, than other cattle within a given population (Omisakin et al., 2003). High-shedders may pose a higher risk of carcass or hide contamination than cattle shedding lower concentrations, and may lead to situations where thresholds for effective intervention strategies are exceeded (Matthews et al., 2006; Chase-Topping et al., 2008; Fox et al., 2008).

Assessing relatedness among E. coli O157 isolates recovered from points throughout beef production systems helps to better define the ecology and epidemiology of this pathogen during preharvest and harvest processes. Previously, Fox et al. (2008) described associations at harvest between fecal shedding (both low- and high-shedding levels) and preevisceration carcass contamination within truckloads of finished cattle. Inference was limited by factors inherent to the study design; however, genetic analysis of the recovered isolates could further define the potential transmission routes for carcass contamination. Our objective was to define and compare, within and among cattle cohorts, the pulsed-field gel electrophoresis (PFGE) profiles of E. coli O157 previously isolated from fecal and carcass samples to further define relationships between beef carcass contamination and fecal shedding of E. coli O157 at harvest.

Materials and Methods

Source of isolates

As previously described (Fox et al., 2008), fecal and carcass samples were obtained from 1503 cattle that arrived in 50 truckloads to a commercial abattoir in the Midwest United States during a 5-week period. Data on cattle origin were not available. Up to 32 cattle per truckload were sampled. Preevisceration carcass swab samples, collected posthide removal, were matched within animal to postevisceration fecal samples collected from intact rectums. Gloves were changed between samples and measures were taken to prevent cross-contamination among samples (Fox et al., 2008). All samples were transported in coolers with ice packs to the Preharvest Food Safety Laboratory, College of Veterinary Medicine, Kansas State University, and stored under refrigeration (4°C) for processing within 48 h of collection.

Isolation of E. coli O157

Isolation and identification of E. coli O157 from carcass and fecal samples were previously described (Fox et al., 2008). Briefly, carcass sponge samples were enriched in 2% brilliant green bile broth (Difco–Becton Dickinson, Sparks, MD), and fecal samples were enriched in Gram-negative (GN) broth (BD, Franklin Lakes, NJ) containing cefixime (Sigma-Aldrich, St. Louis, MO; 50 ng/mL), cefsulodin (Sigma-Aldrich; 10 μg/mL), and vancomycin (Sigma-Aldrich; 8 μg/mL, GNccv). To detect high-shedding cattle (Sanderson et al., 2007), a preenrichment fecal suspension was directly streaked onto sorbitol MacConkey agar (BD) containing cefixime (50 ng/mL) and potassium tellurite (2.5 μg/mL; CT-SMAC). Enriched samples were subjected to immunomagnetic separation (IMS) and plated onto CT-SMAC. After 16 to 18 h of incubation at 37°C, up to six nonsorbitol-fermenting colonies were transferred onto blood agar (Remel, Lenexa, KS) and incubated overnight at 37°C. Colonies were tested for indole production and latex agglutination of the O157 antigen (Oxoid Limited, Basingstoke, Hampshire, England); positives were tested by polymerase chain reaction for eae, stx ₁, stx ₂, and hylA genes (Fagan et al., 1999). Isolates were considered E. coli O157 if they had eae and hylA genes and at least one stx gene. Fecal samples were also classified based on the relative concentration of E. coli O157: cattle with fecal samples positive only on the IMS procedure were classified as low-shedder, while those with feces positive by the direct plate technique were classified as high-shedder (Fox et al., 2007, 2008; Sanderson et al., 2007). Isolates were stored on protect beads at −80°C for future characterization.

PFGE

Isolates were analyzed by PFGE separation of Xba1-digested genomic DNA in accordance with the PulseNet United States Standardized Laboratory Protocol (Ribot et al., 2006). Stored isolates were transferred onto blood agar and incubated at 37°C for 18 h. Colonies were added to cell suspension buffer at room temperature, and concentration was adjusted to an optical density of 1.3 to 1.4 at 610 nm. Cell suspensions were incorporated into gel plugs consisting of 1% SeaKem Gold^®, 1% sodium dodecyl sulfate (SDS) agarose (BioWhittaker Molecular Applications, Rockland, ME), and Proteinase K (Fisher Scientific, Fair Lawn, NJ). Cast gel plugs were then subjected to cell lysis in TE buffer (50 mM Tris:50 mM ethylenediaminetetraacetic acid, pH 8.0 + 1% Sarcosyl). After washing in reagent-grade type 1 water, extracted DNA samples underwent restriction digestion with the Xba1 enzyme (Promega Corporation, Madison, WI). Along with Salmonella ser. Braenderup H9812 standards, each sample was cast into agarose plugs and loaded into wells. Restricted plug slices were subjected to electrophoresis, stained with an ethidium bromide solution, and viewed under UV light. Gel imaging software (Quantity One^® [Bio-Rad, Hercules, CA] and BioNumerics^® [Applied Maths, Austin, TX]) was used to digitize, normalize, and assign bands for each isolate image. Band-sharing similarity coefficients were generated from the DNA fragments in the 10- to 550-kb range. Dendograms were constructed to provide a visual representation of the relationship among E. coli O157 isolates. Isolates were grouped based on banding pattern similarities where types and subtypes were defined as isolates having PFGE patterns of >95% or 100% Dice similarity, respectively (Sargeant et al., 2006).

Data analysis

Truckload, animal, sample type, and all laboratory results were recorded and managed in spreadsheet format (Microsoft^® Office Excel 2007; Microsoft, Redmond, WA). The PFGE results for the isolates were assessed in conjunction with epidemiologic data on the source of the isolates, including sample type, truckload, and day. Descriptive analyses of the frequency distribution of types and subtypes within each sample type, truckload, and sampling day were evaluated, and exact 95% binomial confidence intervals (CI) were calculated for proportions using the BETAINV function of Microsoft Office Excel. Exact CI are reported in parentheses for frequency measures. Comparisons of the 95% CI were used to assess the potential for proportions to differ (p < 0.05) while recognizing the strengths and weaknesses of this approach (Schenker and Gentleman, 2001). In addition, we assessed and described the presence or absence of genetically similar E. coli O157 isolates longitudinally, including whether strains were persistently recovered from carcasses sampled on different days within the studied abattoir.

Results

Prevalence and sources of isolates

E. coli O157 was isolated from fecal or carcass samples from 157 of 1503 (10.4%) cattle originating from 37 of 50 (73.5%) truckloads (Fox et al., 2008). Of the 174 recovered E. coli O157 isolates, 135 (77.6%) were from fecal samples and 39 (22.4%) were from carcass samples. Thirty-nine of 1503 carcass samples (2.6%) representing 15 truckloads and 127 of 1495 fecal samples (8.5%) representing 37 truckloads were positive. There were 55 fecal-positive cattle that were positive on direct plating (high-shedders) and 80 cattle that were only positive based on enrichment/IMS of fecal samples (low-shedders) (Table 1). Of 39 carcass-positive cattle, 9 also had positive fecal samples (5 high-shedders and 4 low-shedders). There were 120 fecal-positive cattle that were carcass-negative. The percentages of truckloads that had at least one high-shedder or low-shedder were 52.0% (37.4% to 66.3%) and 62% (47.1% to 75.3%), respectively (Fox et al., 2008). All isolates had the stx ₂ gene, and 90 (51.7%) isolates had the stx ₁ gene.

Table 1.

Counts of Escherichia coli O157 Isolates ^a and Corresponding Pulsed-Field Gel Electrophoresis (PFGE) Subtypes ^b by Truckload, Sampling Date, and Sample Type

		Number of isolates (PFGE subtypes)			Number of fecal isolates with a carcass-matching PFGE subtype
Truck load	Sampling date	Carcass	Fecal high-shedder ^c	Fecal low-shedder ^d	High-shedder	Low-shedder
A1	5/14/07	0	0	0	–	–
A2		0	0	0	–	–
A3		3(1)	2(1)	5(2)	2	4
B1	5/15/07	0	1(1)	1(1)	–	–
B2		0	0	0	–	–
B3		0	0	1(1)	–	–
B4		1(1)	1(1)	3(1)	0	0
B5		0	0	3(2)	–	–
B6		0	1(1)	3(1)	–	–
B7		0	0	1(1)	–	–
C1	5/21/07	0	1(1)	0	–	–
C2		1(1)	1(1)	0	1	–
C3		0	0	1(1)	–	–
C4		0	0	0	–	–
C5		1(1)	3(3)	3(2)	1	2
C6		0	0	1(1)	–	–
D1	5/22/07	0	0	0	–	–
D2		0	1(1)	0	–	–
E1	6/4/07	1(1)	6(2)	14(5)	4	10
E2		6(4)	4(4)	3(2)	3	3
E3		3(2)	7(2)	1(1)	6	0
F1	6/6/07	0	5(1)	1(1)	–	–
F2		0	1(1)	4(2)	–	–
F3		0	0	1(1)	–	–
F4		0	0	0	–	–
F5		0	1(1)	2(2)	–	–
G1	6/11/07	1(1)	1(1)	1(1)	1	0
G2		2(2)	1(1)	0	1	–
G3		0	0	0	–	–
G4		0	0	1(1)	–	–
G5		2(2)	2(1)	5(2)	2	3
G6		0	1(1)	1(1)	–	–
G7		0	3(2)	1(1)	–	–
H1	6/13/07	0	1(1)	0	–	–
H2		0	0	1(1)	–	–
H3		0	0	1(1)	–	–
H4		0	0	0	–	–
H5		3(3)	2(1)	5(2)	2	5
H6		8(2)	3(1)	3(2)	3	2
I1	6/18/07	4(1)	2(2)	1(1)	1	0
I2		0	0	2(2)	–	–
I3		2(1)	2(1)	5(1)	2	5
J1	6/19/07	1(1)	1(1)	3(1)	1	3
J2		0	0	0	–	–
J3		0	0	2(2)	–	–
K1	6/20/07	0	0	0	–	–
K2		0	0	0	–	–
K3		0	0	0	–	–
K4		0	1(1)	0	–	–
Total		39(17)	55(24)	80(28)	30	37

There was only one isolate characterized per sample.

Subtypes represent unique PFGE banding patterns following Xba1 digestion.

Cattle with a fecal sample that was culture positive by a direct plate technique were classified as high-shedders.

Cattle with a fecal sample that was culture positive only on an immunomagnetic separation (IMS) procedure were classified as low-shedders.

PFGE

Among 174 isolates, there were a total of 17 distinguishable PFGE types (95% Dice similarity) and 37 subtypes (100% Dice similarity; no band differences). Among all isolates, two PFGE types were detected at least 30 times and four subtypes were detected at least 10 times (Table 2). Two of 17 types (11.8%), and 12 of 37 subtypes (32.4%) were represented only once. Fifty percentage of all isolates were one of five subtypes, and 48.9% of all isolates were one of three types.

Table 2.

Counts of Escherichia coli O157 Isolates by Pulsed-Field Gel Electrophoresis (PFGE) Subtype ^a and Date of Sampling for Each Truckload Where Escherichia coli O157 Was Recovered from Carcass or Fecal Samples

		PFGE subtypes from all sample types
Load	Date	a	b	c	d	e	f	g	h	i	j	k	l	m	n	o	p	q	r	s	t	u	v	w	x	y	z	aa	bb	cc	dd	ee	ff	gg	hh	ii	jj	kk
A3	5/14/07	–	9	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–
B1	5/15/07	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	2	–	–	–	–	–	–	–	–	–	–	–	–
B3		–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
B4		–	–	–	–	–	–	–	–	–	–	–	4	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–
B5		–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	2	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
B6		–	4	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
B7		–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
C1	5/21/07	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
C2		–	–	–	–	–	–	–	–	–	–	–	–	2	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
C3		–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
C5		–	2	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	4	–	–	–	–	–	–	–	–	–	–	–	–	–	–
C6		–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
D2	5/22/07	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
E1	6/4/07	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	1	15	–	–	1	2	–
E2		–	–	–	–	–	–	–	–	1	–	2	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	3	4	3	–
E3		–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	8	–	–	–	–	1	1	–	1	–	–
F1	6/6/07	–	–	–	6	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
F2		–	–	–	4	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1
F3		–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
F5		–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	2	–	–	–	–	–	–	–	–	–
G1	6/11/07	–	–	–	–	–	–	2	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
G2		–	–	–	–	–	–	–	–	–	–	–	–	2	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
G4		–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
G5		–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	6	–	–	1	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–
G6		–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–
G7		2	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	2	–	–	–	–	–	–	–	–	–	–	–
H1	6/13/07	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
H2		–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
H3		–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–
H5		–	5	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	4	–	–	–	–	–	–	–	–	–	–	–	–	–	–
H6		–	3	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	10	1	–	–	–	–	–	–	–	–	–	–	–	–	–
I1	6/18/07	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	5	–	–	–	–	–	–	1	–	–	–	–
I2		1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–
I3		–	–	–	–	–	–	–	–	–	9	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
J1	6/19/07	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	5	–	–	–	–	–
J3		–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–
K4	6/20/07	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	1	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–

Subtypes represent unique PFGE banding patterns following Xba1 digestion (designations correspond to those of Table 3).

Overall, there were 17 E. coli O157 subtypes recovered from carcasses and 36 subtypes recovered from fecal samples. Twenty-four of the fecal isolate subtypes were recovered from high-shedding cattle, and 28 were recovered from low-shedding cattle (Table 1). Sixteen subtypes were found in the feces of both high-shedding and low-shedding cattle. Thirty high-shedder (54.6%; 40.6% to 68.0%) and 37 low-shedder (46.3%; 35.0% to 57.8%) fecal isolates had identical PFGE subtypes as carcass isolates on the same load (Table 1).

Eleven of 37 (29.7%; 15.9% to 47.0%) subtypes were recovered from fecal or carcass samples from multiple truckloads within the same day (Table 2). Ten out of 37 (27.1%; 13.8% to 44.1%) subtypes were found on multiple truckloads on different days (Table 2), while 4 out of 17 (25.5%; 6.8% to 49.9%) subtypes from carcasses were found on carcasses from multiple loads on different days (Table 1). Only three (8.1%; 1.7% to 21.9%) subtypes from any sample type were found on multiple loads on consecutive days.

Thirteen of the 37 total subtypes (35.1%; 20.2% to 52.5%) were recovered from both carcasses and high-shedding cattle feces, whereas 14 (37.8%; 22.5% to 55.2%) were recovered from both carcass and low-shedding cattle feces. Eleven subtypes (29.7%) were recovered from all three sample types. All but one of the subtypes recovered from a carcass were also found in a fecal sample, and the subtype not found in feces was recovered from only one carcass.

Several E. coli O157 subtypes were recovered from multiple carcass or fecal samples within truckloads (Tables 1 and 2). Eight of 37 subtypes (21.6%; 9.8% to 38.2%) were found on more than one carcass within a load, and 18 (48.7%; 31.9% to 65.6%) were found in more than one fecal sample within truckload. One subtype was detected 10 times within a truckload and another was detected 15 times within a truckload (Table 2). There was a mean of 1.9 (range: 1–6) subtypes per positive truckload. Eleven subtypes (29.7%) were recovered from both carcass and fecal samples within a truckload (Table 3). The percentages of subtypes from carcasses that also were detected from a high-shedder, low-shedder, or any feces from within the same truckload were 70.6% (44.0% to 89.7%), 47.12% (23.0% to 72.2%), and 70.6% (44.0% to 89.7%), respectively. Within truckload, the percentages of carcass isolates that were identical to high-shedder, low-shedder, or any fecal isolate, as determined by PFGE, were 69.2% (52.4% to 83.0%), 46.0% (30.1% to 62.8%), and 69.2% (52.4% to 83.0%), respectively (Fig. 1). Nine of 15 truckloads with a positive carcass had carcass subtypes that matched at least 1 high-shedder fecal isolate, 11 of 15 carcass-positive truckloads had a carcass subtype that matched a low-shedder subtype, while 8 out of 15 carcass-positive truckloads had a carcass subtype that matched both low- and high-shedder subtypes in the same truckload (Table 3).

FIG. 1.

Percentages of Escherichia coli O157 carcass isolates that had the same pulsed-field gel electrophoresis banding pattern (following Xba1 digestion) as fecal isolates recovered from high- or low-shedding cattle within the same truckload or different truckloads on the same sampling day. Error bars represent 95% exact binomial confidence intervals.

Table 3.

Distribution of Escherichia coli O157 Pulsed-Field Gel Electrophoresis (PFGE) Subtypes ^a Recovered by Carcass Sampled, as well as Negative (−) Preevisceration Carcasses Within Cattle Truckload by Order of Slaughter

		Carcass order within load
Truck load	Sample date	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26	27	28	29	30	31	32
A3	5/14/07	−	−	−	b^b –d	−	−	−	−	−	−	−	−	−	−	−	−	−	b^b –d	−	−	−	−	−	−	−	−	−	−	b^b –d	−	−	−
B4	5/15/07	−	−	−	−	−	−	−	−	−	−	y	−	−	−	−	−	−	−	−	−	−	−	−	−	−	.	.	.	.	.	.	.
C2	5/21/07	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	m^b	−	−	−	−	−
C5		−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	w^b,c	−	−	−	−	−	−	−	−
E1	6/4/07	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	ff^b,c	−	−	−	−	−	−	−	−	−
E2		−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	ii^b,c	−	−	k^d	k^d	jj^b,c	−	hh^b,d	−	hh^b,d	−	−	−
E3		−	−	−	−	−	−	−	−	−	−	−	−	ff	aa^b,d	aa^b,d	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	.
G1	6/11/07	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	g^b	−	−	−
G2		−	−	−	−	−	t	−	−	−	−	m^b	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−
G5		−	−	−	−	−	−	−	−	−	s^b,c	−	−	−	−	−	−	−	−	−	v	−	−	−	−	−	−	−	−	−	−	−	−
H5	6/13/07	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	w^c	e	−	−	−	−	−	−	b^b,c	−	−
H6		−	b^d	−	−	−	−	−	w^b –d	−	−	w^b –d	−	−	−	−	−	w^b –d	b^d	−	w^b –d	−	w^b –d	b^d	−	−	−	−	−	−	−	−	−
I1	6/18/07	−	−	−	−	−	−	−	−	−	−	−	z^b,d	−	−	−	−	−	−	z^b,d	−	z^b,d	−	−	−	z^b,d	−	−	−	−	−	−	.
I3		−	−	−	−	−	−	−	−	−	−	−	−	−	−	j^b –d	−	−	−	−	j^b –d	−	−	−	−	−	−	−	−	−	−	−	−
J1	6/19/07	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	ff^b,c	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−

Subtypes represent unique PFGE banding patterns following Xba1 digestion (designations correspond to those of Table 2). Superscripts (b–d) indicate where carcass isolates were the same subtype as other isolates recovered within the load.

This subtype was also present in a fecal isolate from a high-shedder within the same load.

This subtype was also present in a fecal isolate from a low-shedder within the same load.

This subtype was also present in another carcass isolate within the same load.

(·) represents missing values as three carcass-positive truckloads had less than 32 cattle.

Four of 17 carcass subtypes (23.5%; 6.8% to 49.9%) were detected on carcasses from multiple truckloads, and 3 of these were on the same day. There were 11 fecal subtypes recovered from multiple fecal samples from different truckloads. Seven fecal subtypes were recovered from high-shedders from different truckloads, and four were found on truckloads harvested on the same day. Ten fecal subtypes were detected in feces from low-shedders from different truckloads with six of these on the same day. The percentages of carcass subtypes that also were detected from a high-shedder, low-shedder, or any feces among different truckloads within the same day were 35.3% (14.2% to 61.7%), 58.8% (32.9% to 81.6%), and 58.8% (32.9% to 81.6%), respectively. Among different truckloads within the same day, the percentages of carcass isolates that were identical to high-shedder, low-shedder, or any fecal isolate, as determined by PFGE, were 20.5% (9.3% to 36.5%), 48.7% (32.4% to 65.2%), and 51.3% (34.8% to 67.6%), respectively (Fig. 1).

Discussion

We found that most of the E. coli O157 isolates recovered from preevisceration carcasses were the same PFGE subtype as isolates recovered from postevisceration fecal samples from cattle within the same truckload. The percentages of carcass isolates that had identical PFGE patterns to that of fecal isolates from high-shedding (69.2%) and low-shedding (46.0%) cattle within a truckload were fairly similar. However, comparing those numbers to corresponding percentages (35.3% and 58.8% respectively) among different truckloads within the same day revealed that the importance of high- versus low-shedding cattle may depend on whether the consideration of E. coli O157 dynamics is being made within or between truckloads (Fig. 1). The transport cohort appears to be an important factor in the transmission of E. coli O157 at harvest, yet 32% of the subtypes were recovered from multiple truckloads and sampling days. Approximately half of the fecal isolates from high-shedders and low-shedders were the same PFGE subtype as carcass isolates on the same truckload. Although some have suggested that detecting high-shedders may be more efficient than detecting low-shedders and prioritizing the detection and mitigation of high-shedders within cattle cohorts might reduce risk of carcass contamination (Matthews et al., 2006; Fox et al., 2007; Sanderson et al., 2007; Chase-Topping et al., 2008), our findings suggest that preharvest intervention strategies need to mitigate the effects of both high- and low-shedding cattle within and among transport cohorts.

Defining associations between fecal shedding and carcass contamination may enhance the development of monitoring and/or intervention strategies to reduce E. coli O157 contamination of beef products. Molecular subtyping by PFGE has been used to study the genetic relatedness of E. coli O157 strains in foodborne disease outbreaks as well as in epidemiologic research (Faith et al., 1996; Swaminathan et al., 2001; Rangel et al., 2005). Although PFGE protocols with different restriction enzyme combinations have been employed, the comparison of banding patterns produced by the same restriction enzyme(s) appears to provide useful estimates of relatedness among E. coli O157 isolates (Davis et al., 2003). In this study, the total number of PFGE subtypes among isolates from carcasses and feces reflect the recognized genetic diversity of E. coli O157 within a population at an abattoir. The results of our PFGE analysis are consistent with observed frequencies among live cattle and beef carcasses (Rice, 1999; Barkocy-Gallagher et al., 2001; Renter et al., 2003). These data may not completely reflect genetic relationships among isolates, yet within the context of accompanying epidemiologic data, they allow us to infer reasonable estimates of relatedness (Davis et al., 2003; Singer et al., 2004). We chose a standard PFGE method using one restriction enzyme, which limited our observation of restriction fragment patterns to approximately 20 bands (Singer et al., 2004). These classification limitations may have affected our ability to differentiate some E. coli O157 isolates that we considered indistinguishable based on PFGE.

Some E. coli O157 PFGE subtypes were represented much more frequently than others, which is similar to other Xba1-PFGE comparison studies in cattle and beef carcasses (Rice, 1999; Barkocy-Gallagher et al., 2001; Renter et al., 2003). In this study, only one isolate from each sample underwent PFGE analysis, so the existence of other subtypes within a positive sample was unknown. Other studies have shown considerable heterogeneity of subtypes among individual samples (Faith et al., 1996; Renter et al., 2003), so it is possible that our study underestimated the existence of related E. coli O157 isolates among cattle, as well as the diversity of isolates within this population. Despite this limitation of the study, the genetic relatedness of detected isolates supports associations of E. coli O157 among cattle and the potential transmission of specific subtypes from feces to carcass.

Our E. coli O157 isolates came from a study that used two culture methods to differentiate relatively low- versus high-shedding levels of E. coli O157 in fecal-positive cattle (Fox et al., 2008). They found that the presence of a high-shedder within a truckload of cattle was the strongest predictor of carcass contamination. High E. coli O157 concentrations in cattle feces usually can be detected with the rapid, less expensive, direct-plating technique, which has an estimated diagnostic sensitivity and specificity of 82.6% and 92.3%, respectively, for a breakpoint of 5 × 10⁴ CFU/g feces (Sanderson et al., 2007). Concentrations as low as 10² CFU/g can be detected with approximately 90% sensitivity by culture methods incorporating IMS (Fox et al., 2007). A previous study demonstrated that a majority of cattle with low-level E. coli O157 carriage were fecal-positive only by using an IMS protocol and suggested a breakpoint fecal concentration of E. coli O157 of 10³ CFU/g for differentiating high-shedders (Low et al., 2005). Although some previous studies have used experimentally inoculated cattle instead of naturally colonized cattle to estimate concentrations and test parameters, any misclassification in our study likely would have been nondifferential with regard to classifying fecal-positive cattle into two groups based on relative concentration of fecal shedding. Although we did not make a direct attempt to enumerate E. coli O157 organisms within positive samples in this study, this classification scheme was still useful in examining associations while considering the heterogeneity of E. coli O157 fecal shedding in cattle populations.

High-shedders can contribute to the hide contamination of cohorts during transport to harvest and may be directly associated with carcass contamination (Bach et al., 2004; Fox et al., 2008). Overall prevalence of fecal shedding within cattle populations is also associated with hide and carcass contamination (Elder et al., 2000), yet overall fecal prevalence may be affected by the presence or frequency of high-shedders (Cobbold et al., 2007). Therefore, both fecal shedding concentration and overall prevalence of E. coli O157 within a pen or truckload may contribute to hide and eventual carcass contamination. In this study, 1 in 3 cattle testing positive for E. coli O157, and 1 in 27 (3.7%) of the cattle sampled were identified as high-shedders (Fox et al., 2008). Another study, which enumerated fecal isolates in harvested cattle, determined that only 1 in 11 (9.0%) cattle testing positive for E. coli O157 at harvest was a high-shedder, but estimated that over 96% of the shed organisms originated from these high-shedders (Omisakin et al., 2003). Some suggest that testing methods that primarily detect high-shedders, such as direct plating, may be a valid and feasible approach to strategically monitoring the risk of E. coli O157 contamination at harvest (Fox et al., 2007; Sanderson et al., 2007).

Preevisceration carcass swab samples and postevisceration fecal samples were chosen to examine potential pre- and postharvest associations of E. coli O157. Associations among subtypes from fecal and carcass samples may indicate that E. coli O157 isolates were transmitted among cattle, particularly since only 20.5% of carcass-positive cattle were also positive in their feces. Approximately the same number of E. coli O157 subtypes from high-shedder and low-shedder cattle feces was also recovered from carcasses. This suggests that the risk of subsequent carcass contamination may be similar for both high- and low-shedding cattle; however, our data suggest that the cohort effect may modify this relationship. Only one subtype from a carcass was not found in any fecal sample, which suggests that a majority of E. coli O157 on carcasses at harvest may be detected in feces. However, previous literature suggests that E. coli O157 isolates detected on beef carcasses also can be detected on cattle hides (Barkocy-Gallagher et al., 2001; Tutenel et al., 2003).

Within the truckload, several subtypes were found on more than one carcass, in more than one fecal sample, or in both carcass and fecal samples. Some subtypes appeared to cluster among loads and on consecutive sampling days (Table 2). Although these data do not reveal extensive temporal associations, they support the hypothesis of transmission of E. coli O157 in feces between cattle at the feedlot or during transportation, lairage, or harvest (Akiba et al., 2000; Minihan et al., 2003). Data from pretransport, posttransport, and lairage fecal and hide samples could have revealed further insight into the movement of E. coli O157 isolates among finished cattle within our study population.

The frequency of high-shedder subtypes matching carcass subtypes within truckload was not higher than the frequency of low-shedder subtypes matching carcass subtypes. Similarities among proportions may have been due to low statistical power in this study, as there were only 17 subtypes isolated from carcasses. However, our determination of genetic relatedness suggests that more than half of E. coli O157 subtypes (58.8%) from carcasses were also present in cattle feces within the same load. The plausibility that the number of high-shedding cattle may pose a greater risk of carcass contamination than the overall number of cattle shedding E. coli O157 has been suggested in other studies (Omisakin et al., 2003; Fegan et al., 2004; Matthews et al., 2006). Fox et al. studied the same population as we did, and found that high-shedder cattle, low-shedder cattle, and combined fecal prevalence were all significantly associated with carcass-positive cattle within a truckload. However, the odds of carcass contamination were highest when a high-shedder was present within the truckload (Fox et al., 2008).

Percentages of carcass isolates that matched high-shedder or low-shedder fecal isolates were similar both within truckloads and among truckloads within the same day (Fig. 1). However, comparing the distributional trends of these data, within versus among truckloads, may indicate that high-shedders could play a more important role in transmission of E. coli O157 within truckloads than among truckloads. These data suggest that the truckload remains an important transmission factor in the ecology of E. coli O157 immediately before harvest, and perhaps the relative contribution of high- versus low-shedding cattle on carcass contamination risk depends on whether the assessment is made within or among truckloads.

Several identical subtypes were recovered from fecal samples taken from different truckloads, and several carcass subtypes also were detected from fecal samples among loads within the same day. This apparent transmission or persistence of E. coli O157 isolates among cattle in different truckloads, or their environments, may be related to the pen or feedlot origin of cattle, the comingling of cattle during lairage, or dispersion of individual E. coli O157 isolates during the harvest process.

We described the relatedness of E. coli O157 isolates from feces and preevisceration carcasses within and among truckloads of finished beef cattle at harvest. We demonstrated that specific PFGE subtypes may be detected on multiple occasions and from different sample types within loads, among loads, and among days. Our results support previously suggested associations between overall E. coli O157 prevalence, fecal shedding, and carcass contamination within truckload. Hence, this study provides additional evidence that preevisceration carcass contamination might be decreased by mitigating the effect of high- and low-shedders within a truckload. However, our results also suggest that the frequency distribution of carcass isolates that have identical PFGE patterns to that of fecal isolates from high- and low-shedding cattle within a truckload may differ from the corresponding distribution among different truckloads on the same day. Further investigations are needed to assess the relative importance of mitigating high-shedders, low-shedders, or any fecal shedder within and among transport cohorts to decrease the risk of carcass contamination.

Footnotes

Acknowledgments

Financial support for this research was provided in part by U.S. Department of Agriculture grant (2005-34359-13008) and by USDA-CREES NC-1007 Multistate Research, Enteric Diseases of Swine and Cattle: Prevention, Control, and Food Safety. Contribution no. 10–015-J, from the Kansas Agricultural Experiment Station (Manhattan, KS). We extend special thanks to Dr. Abbey Nutsch, Neil Wallace, Nathan Hoffman, Miguel Barrios, Clay Adair, and the rest of the staff at the Preharvest Food Safety Labs, College of Veterinary Medicine, Kansas State University, as well as the workers and administration of the commercial slaughter facility where sampling occurred.

Disclosure Statement

No competing financial interests exist.

References

Akiba

, Rice

, Davis

et al. A comparison of Escherichia coli O157 isolates from cattle in Japan and the USA by molecular biological methods. Epidemiol Infect, 2000; 125:221–224.

Bach

, McAllister

, Mears

et al. Long-haul transport and lack of pre-conditioning increases fecal shedding of Escherichia coli and Escherichia coli O157:H7 by calves. J Food Prot, 2004; 67:672–678.

Baker

, Moxley

, Steele

et al. Differences in virulence among Escherichia coli O157:H7 strains isolated from humans during disease outbreaks and from healthy cattle. Appl Environ Microbiol, 2007; 73:7338–7346.

Barkocy-Gallagher

, Arthur

, Rivera-Betancourt

et al. Seasonal prevalence of shiga toxin-producing Escherichia coli, including O157:H7 and non-O157 serotypes, and Salmonella in commercial beef processing plants. J Food Prot, 2003; 66:1978–1986.

Barkocy-Gallagher

, Arthur

, Siragusa

et al. Genotypic analyses of Escherichia coli O157:H7 and O157 nonmotile isolates recovered from beef cattle and carcasses at processing plants in the Midwestern states of the United States. Appl Environ Microbiol, 2001; 67:3810–3818.

Caprioli

, Morabito

, Scheutz

et al. Pathogenesis of verocytotoxin/Shiga toxin–producing Escherichia coli infection. Emerg Infect Dis, 2006. www.cdc.gov/ncidod/EID/vol12no08/06–0170.htm(online)

CDC. Preliminary FoodNet data on the incidence of infection with pathogens transmitted commonly through food—10 states, 2007. Morb Mortal Wkly Rep, 2008; 57:366–370.

Chase-Topping

, Gally

, Low

et al. Super-shedding and the link between human infection and livestock carriage of Escherichia coli O157. Nat Rev Microbiol, 2008; 6:904–912.

Cobbold

, Hancock

, Rice

et al. Rectoanal junction colonization of feedlot cattle by Escherichia coli O157:H7 and its association with supershedders and excretion dynamics. Appl Environ Microbiol, 2007; 73:1563–1568.

10.

Davis

, Hancock

, Besser

et al. Evaluation of pulsed-field gel electrophoresis as a tool for determining the degree of genetic relatedness between strains of Escherichia coli O157:H7. J Clin Microbiol, 2003; 41:1843–1849.

11.

Elder

, Keen

, Siragusa

et al. Correlation of enterohemorrhagic Escherichia coli O157 prevalence in feces, hides, and carcasses of beef cattle during processing. Proc Natl Acad Sci, 2000; 97:2999–3003.

12.

Fagan

, Hornitzky

, Bettelheim

et al. Detection of Shiga-like toxin (stx1 and stx2), intimin (eaeA), and enterohemorrhagic Escherichia coli (EHEC) hemolysin (EHEC hlyA) genes in animal feces by multiplex PCR. Appl Environ Microbiol, 1999; 65:868–872.

13.

Faith

, Shere

, Brosch

et al. Prevalence and clonal nature of Escherichia coli O157:H7 on dairy farms in Wisconsin. Appl Environ Microbiol, 1996; 62:1519–1525.

14.

Fegan

, Vanderlinde

, Higgs

et al. The prevalence and concentration of Escherichia coli O157 in faeces of cattle from different production systems at slaughter. J Appl Microbiol, 2004; 97:362–370.

15.

Fox

, Renter

, Sanderson

et al. Associations between the presence and magnitude of Escherichia coli O157 in feces at harvest and contamination of pre-intervention beef carcasses. J Food Prot, 2008; 71:1761–1767.

16.

Fox

, Renter

, Sanderson

et al. Evaluation of culture methods to identify bovine feces with high concentrations of Escherichia coli O157. Appl Environ Microbiol, 2007; 73:5253–5260.

17.

Loneragan

, Brashears

. Pre-harvest interventions to reduce carriage of E. coli O157 by harvest-ready feedlot cattle. Meat Sci, 2005; 71:72–78.

18.

Low

, McKendrick

, McKechnie

et al. Rectal carriage of enterohemorrhagic Escherichia coli O157 in slaughtered cattle. Appl Environ Microbiol, 2005; 71:93–97.

19.

Marsh

, Schroeder

, Mintert

. Impacts of meat product recalls on consumer demand in the USA. Appl Econ, 2004; 36:897–909.

20.

Matthews

, Low

, Gally

et al. Heterogeneous shedding of Escherichia coli O157 in cattle and its implications for control. Proc Natl Acad Sci, 2006; 103:547–552.

21.

Matthews

, McKendrick

, Ternent

et al. Super-shedding cattle and the transmission dynamics of Escherichia coli O157. Epidemiol Infect, 2006; 134:131–142.

22.

McEvoy

, Doherty

, Sheridan

et al. The prevalence and spread of Escherichia coli O157:H7 at a commercial beef abattoir. J Appl Microbiol, 2003; 95:256–266.

23.

Minihan

, O'Mahony

, Whyte

et al. An investigation on the effect of transport and lairage on the faecal shedding prevalence of Escherichia coli O157 in cattle. J Vet Med Ser B, 2003; 50:378–382.

24.

Omisakin

, MacRae

, Ogden

et al. Concentration and prevalence of Escherichia coli O157 in cattle feces at slaughter. Appl Environ Microbiol, 2003; 69:2444–2447.

25.

Rangel

, Sparling

, Crowe

et al. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982–2002. Emerg Infect Dis, 2005; 11:603–609.

26.

Renter

, Sargeant

, Oberst

et al. Diversity, frequency, and persistence of Escherichia coli O157 strains from range cattle environments. Appl Environ Microbiol, 2003; 69:542–547.

27.

Ribot

, Fair

, Gautom

et al. Standardization of pulsed-field gel electrophoresis protocols for the subtyping of Escherichia coli O157:H7, Salmonella, and Shigella for PulseNet. Foodborne Pathog Dis, 2006; 3:59–67.

28.

Rice

. Genetic subtyping of Escherichia coli O157 isolates from 41 Pacific Northwest USA cattle farms. Epidemiol Infect, 1999; 122:479–484.

29.

Sanderson

, Sreerama

, Nagaraja

. Sensitivity of direct plating for detection of high levels of E. coli O157:H7 in bovine fecal samples. Curr Microbiol, 2007; 55:158–161.

30.

Sargeant

, Shi

, Sanderson

et al. Pulsed-field gel electrophoresis patterns of Escherichia coli O157 isolates from Kansas feedlots. Foodborne Pathog Dis, 2006; 3:251–258.

31.

Schenker

, Gentleman

. On judging the significance of differences by examining the overlap between confidence intervals. Am Stat, 2001; 55:182–186.

32.

Singer

, Sischo

, Carpenter

. Exploration of biases that affect the interpretation of restriction fragment patterns produced by pulsed-field gel electrophoresis. J Clin Microbiol, 2004; 42:5502–5511.

33.

Swaminathan

, Barrett

, Hunter

et al. PulseNet: the molecular subtyping network for foodborne bacterial disease surveillance, United States. Emerg Infect Dis, 2001; 7:382–389.

34.

Tutenel

, Pierard

, Van Hoof

et al. Molecular characterization of Escherichia coli O157 contamination routes in a cattle slaughterhouse. J Food Prot, 2003; 66:1564–1569.

35.

Verbeke

, Frewer

, Scholderer

et al. Why consumers behave as they do with respect to food safety and risk information. Anal Chim Acta, 2007; 586:2–587.

36.

Woerner

, Ransom

, Sofos

et al. Determining the prevalence of Escherichia coli O157 in cattle and beef from the feedlot to the cooler. J Food Prot, 2006; 69:2824–2827.