Molecular Characterization of Escherichia coli O157:H7 Hide Contamination Routes: Feedlot to Harvest

1240 Journal of Food Protection, Vol. 69, No. 6, 2006, Pages 1240 1247 Copyright, International Association for Food Protection Molecular Characterization of Escherichia coli O157:H7 Hide Contamination Routes: Feedlot to Harvest K. D. CHILDS, 1 C. A. SIMPSON, 2 W. WARREN-SERNA, 3 G. BELLENGER, 3 B. CENTRELLA, 3 R. A. BOWLING, 1 J. RUBY, 1 J. STEFANEK, 1 D. J. VOTE, 1 T. CHOAT, 1 J. A. SCANGA, 2 J. N. SOFOS, 2 G. C. SMITH, 2 AND K. E. BELK 2 * 1 Smithfield Beef Group, 2580 University, Green Bay, Wisconsin 54305-3000; 2 Center for Red Meat Safety, Department of Animal Sciences, Colorado State University, Fort Collins, Colorado 80523-1171; and 3 Food Safety Net Services, 221 West Rhapsody, San Antonio, Texas 78216, USA MS 05-347: Received 15 July 2005/Accepted 30 January 2006 ABSTRACT This study was conducted to identify the origin of Escherichia coli O157:H7 contamination on steer hides at the time of harvest. Samples were collected from the feedlot, transport trailers, and packing plant holding pens and from the colons and hides of feedlot steers. A total of 50 hide samples were positive for E. coli O157:H7 in two geographical locations: the Midwest (25 positive hides) and Southwest (25 positive hides). Hide samples were screened, and the presence of E. coli O157: H7 was confirmed. E. coli O157:H7 isolates were fingerprinted by pulsed-field gel electrophoresis and subjected to multiplex PCR procedures for amplification of E. coli O157:H7 genes stx 1, stx 2, eaea, flic, rfbe O157, and hlya. Feedlot water trough, pen floor, feed bunk, loading chute, truck trailer side wall and floor, packing plant holding pen floor and side rail, and packing plant cattle drinking water samples were positive for E. coli O157:H7. Pulsed-field gel electrophoresis banding patterns were analyzed after classifying isolates according to the marker genes present and according to packing plant. In this study, hide samples positive for E. coli O157:H7 were traced to other E. coli O157:H7 positive hide, colon, feedlot pen floor fecal, packing plant holding pen drinking water, and transport trailer side wall samples. Links were found between packing plant side rails, feedlot loading chutes, and feedlot pens and between truck trailer, different feedlots, and colons of multiple cattle. This study is the first in which genotypic matches have been made between E. coli O157:H7 isolates obtained from transport trailer side walls and those from cattle hide samples within the packing plant. Media attention to Escherichia coli O157:H7 related outbreaks have influenced consumers perception of the safety of beef since the early 1990s, primarily because of the human fatalities that can occur in the young and elderly from an infection with E. coli O157:H7. Three children in the Pacific Northwest died after consuming undercooked ground beef contaminated with E. coli O157:H7. Among recorded foodborne outbreaks, up to 25% of infected individuals have been hospitalized, 6% have developed the potentially fatal hemolytic uremic syndrome, and 1% have died from infection with E. coli O157:H7 (15). The U.S. Food Safety and Inspection Service (FSIS) in 1994 declared E. coli O157:H7 an adulterant in ground beef under the Federal Meat Inspection Act. Since 1996, the beef industry has operated under hazard analysis critical control point plans, which emphasize reduction or elimination of hazards including E. coli O157:H7 (8). During the Food Safety Summit in 2003 (sponsored by the National Cattlemen s Beef Association), the Points of Focus for the Producer Sector addressed ways to reduce prevalence of E. coli O157:H7 associated with feedlot cattle. The following best management practices were suggested: (i) good management practices of clean feed, clean water, clean pens, and clean cattle should be implemented and maintained; (ii) interventions or good management practices that have been scientifically validated should be evaluated for adoption; * Author for correspondence. Tel: 970-491-5826; Fax: 970-491-0278; E-mail: keith.belk@colostate.edu. and (iii) packing houses, producers, and the FSIS should maintain open communication and share data within the beef industry regarding preharvest interventions and good management practices (19). Chapman et al. (5), Sofos et al. (24), and Elder et al. (8) have suggested that cattle may be reservoirs or transmission vectors for E. coli O157:H7 that can potentially be transferred from hide to meat during the harvesting and/or dressing processes. Hides of animals are a key source of cross-contamination of carcasses (1, 17, 21). In several studies, links between carcass contamination with E. coli O157 and the presence of E. coli O157:H7 throughout the harvesting process have been examined (4, 11 14). However, these studies did not directly link preharvest sources of E. coli O157 contamination with packing plant hide samples positive for E. coli O157:H7 or identify specific routes of carcass contamination (2). Although cattle hides may become contaminated with E. coli O157:H7 from feedlot pen floors, packing plant holding pens, or restrainer floors (7, 26), there has been no research to determine whether there is a relationship between the E. coli O157:H7 isolates found in feedlots and transport vehicles and the E. coli O157:H7 isolates actually found on animals or carcasses at packing plants. Pulsed-field gel electrophoresis (PFGE) genotyping is a method used by the Centers for Disease Control and Prevention to genetically trace pathogens linked to food safety outbreaks. Barkocy-Gallagher et al. (2) and Tutenel et al.

J. Food Prot., Vol. 69, No. 6 TRANSMISSION OF E. COLI O157:H7 FROM PREHARVEST TO HARVEST 1241 (26) initially used PFGE genotyping to track E. coli O157 on carcasses throughout the harvesting process in various packing plants. Barkocy-Gallagher et al. (2) evaluated the prevalence of E. coli O157 on carcasses and in the slaughter environment and characterized the E. coli O157 strains using PFGE and Shiga toxin gene typing to determine contamination routes and document the wide range of genetic diversity of E. coli O157:H7 isolates recovered during the trial. Tutenel et al. (26), using molecular characterization of E. coli O157 (PFGE profiles and stx types), found that E. coli O157:H7 was widely distributed at packing plants, specifically on the hides of the animals, and that hide-to-hide contamination can occur after departure from the farm during transportation to the packing plant. There are multiple opportunities for cattle hides to become contaminated with E. coli O157:H7 in feedlots, during transport to the packing plants, and while being held in holding pens at packing plants. The present study was designed to verify preharvest origins of E. coli O157:H7 on steer hides via characterization of microbiological isolates using molecular PFGE fingerprinting and multiplex PCR analysis. MATERIALS AND METHODS Cattle, feedlots, and packing plant. Samples were collected from cattle originating from three feedlots in the Midwest and three in the Southwest. Cattle in eight feedlot shipments were examined between July 2004 and February 2005. Cattle were transported 2,411, 1,685, or 55 km from the three midwestern commercial feedlots to two Upper Midwest beef packing plants. In the southwestern region, cattle were transported distances of 81, 86, or 561 km to a single packing plant. For each feedlot shipment, composite samples were collected from eight preharvest environmental locations and two packing plant harvesting processes: feedlot pen floor, feedlot pen water, feed bunk and feed, loading chute, transport truck trailer inner wall and floor, packing plant holding pen, plant pen water, restrainer wall, cattle hides during harvesting, and corresponding colons on the harvest floor. Sample sets were collected until at least 25 hide samples were confirmed positive for E. coli O157:H7 by biochemical analysis from each region (Midwest and Southwest). Feedlot pen sample collection. On the day of or the day before cattle shipment to the packing plant, researchers traveled to selected feedlots and collected samples for microbiological characterization. Each feed bunk was aseptically sampled by swabbing a large area ( 1,000 cm 2 ) of the feed bunk with a hydrated sterile sponge (HydraSponge with peptone water, Biotrace International, Bothwell, Wash.), and random feed samples were collected approximately every 10 ft (3.05 m) for the entire length of the feed bunk in sterile cups (cups held approximately 50 g of feed). Water tanks in each pen were sampled by swabbing the exterior of the trough and the tank ball if applicable (trough base was not sampled) with a hydrated sterile sponge. Water samples also were collected by submersing and gently squeezing (three times) a hydrated sterile sponge into the water tank. All sponges were placed back into individual sterile bags, excess air was removed, and the samples were placed in a foam cooler with ice packs until shipment to a commercial laboratory (Food Safety Net Services, Inc., San Antonio, Tex.) for analysis. Pen floor composite samples of feces and dirt were collected aseptically with a sterile tongue depressor and sterile cup (each cup held approximately 50 g). Each sample was collected by dipping the sterile tongue depressor 1 in. (2.54 cm) deep into the dirt or fecal pile to include the surface and interior in the sample. Each sample represented approximately 10 square yards (9.2 m 2 ) of the pen floor and was collected so that each pen was sampled from side to side and top to bottom. Within the 10 square yards, approximately five subsamples of dirt or feces were taken. Loading chute and truck trailer sample collection. Cattle loading chutes at the feedlots and the associated alley were sampled immediately before loading cattle into the first trailer for shipment. If more than one loading chute was used, both were sampled, and samples were combined. Hydrated sterile sponges were used to take a 500-cm 2 representative sample of the loading chute walls and railing. Samples were taken so that the wall or railing was sampled from the top, middle, and bottom. Sterile tongue depressors and sterile cups were used to aseptically collect composite dirt and fecal samples from the floors of the loading chutes. Before cattle were loaded, the upper deck, lower deck, and back of all trailers used for shipment were sampled. These samples were not combined and were analyzed separately. Trailer walls were sampled by swabbing a 500-cm 2 area with a single hydrated sterile sponge at each location within the trailer (upper deck, lower deck, and back), resulting in a representative wall sample (500 cm 2 ) for each location. If visible trailer floor contamination was present, floor samples were collected with a sterile tongue depressor and sterile cup from the top deck, bottom deck, and back of each trailer (approximately 50 g in each sterile cup). Plant holding pen sample collection. Before cattle were unloaded from trailers at the packing plant, empty pens that were to house the research cattle were sampled with a hydrated sterile sponge. Objects inside the pens that were sampled included side railings, perimeter walls, gates, and pillars. If present, representative fecal material and associated pen bedding were collected into sterile cups from each holding pen floor. Each sample represented approximately 10 square yards of the plant holding pen floor and was collected so that each pen was sampled from side to side and top to bottom. Within the 10 square yards, approximately five subsamples were taken of dirt or feces. Hydrated sterile sponges were used to swab water troughs in each holding pen, including any fecal contamination attached to the inside of the trough. Hydrated sterile sponges also were used to collect water by submersion of the entire sponge in the water tank for approximately 10 s. All sponges were placed into sterile bags after sample collection. Excess air was removed, and the bag was labeled and placed with labeled fecal cups in a foam shipper with ice packs until shipment to the commercial laboratory for analysis. Hide and colon sample collection. After exsanguination but before hide removal, hide samples were collected from randomly selected cattle in each shipped lot of live cattle in each plant. Hide samples were collected with an individual sterile sponge from a 500-cm 2 area of hide at the dorsal midline by swabbing with 20 strokes (one up-and-down or side-to-side movement was counted as one stroke) using enough pressure to remove any dried blood and/or dirt. Each sampled carcass and corresponding colon was tagged to maintain identity. All colons were collected from animals whose hides had been swabbed by excising approximately the last 6 in. (15.2 cm) of the digestive tract at the rectoanalmucosal junction (22), and colon samples were placed with the carcass tag in a labeled Whirl-Pak bag (Nasco, Fort Atkinson, Wis.). Excess air was removed, and the bag was closed and secured firmly so no spilling occurred during transport.

1242 CHILDS ET AL. J. Food Prot., Vol. 69, No. 6 Microbial analysis. Before being shipped, all samples were cooled to 4 C and packed with ice packs in insulated shipping boxes. All samples excluding colons (which were cooled overnight before shipping) were shipped overnight for microbiological analysis at Food Safety Net Services. Upon arrival at the laboratory, the temperature of each shipping container was documented. Samples exceeding 4 C upon arrival were discarded. Intact colons were dissected, and the mucosal membrane of the colon was vigorously swabbed with a hydrated sterile sponge, which was then incubated in 90 ml of tryptic soy broth (TSB; Biotrace International) at 37 C for 24 h. E. coli O157:H7 in all samples was identified following the U.S. Department of Agriculture Meat Animal Research Center method described by Barkocy-Gallagher et al. (3). A 10-ml aliquot of fluid was taken from each sample bag, suspended in 90 ml of TSB, and incubated for 2 h at 25 C and then for 6 h at 42 C. Samples were held at 4 C overnight. A tube was filled with 1 ml of this culture, 20 l of anti E. coli O157: H7 Dynabeads (Dyna Laboratories, Lake Success, N.Y.), and 100 l of 0.05% protamine (Sigma, St. Louis, Mo.), and this mixture was incubated for 30 min on a rocker at room temperature (24 2 C). Tubes were placed in a magnetic separation rack to bind beads and incubated for an additional 5 min at room temperature (24 2 C) on a rocker. A 1-ml portion of supernatant was removed from each tube, and beads were washed three times with 1 ml of phosphate-buffered saline (PBS) plus 0.05% Tween 20 (ph 7.0; Fisher Scientific, Fair Lawn, N.J.) and then resuspended in 100 l of PBS plus 0.05% Tween 20. A 50- l aliquot of the suspended bead solution was spread onto sorbitol MacConkey agar supplemented with 0.05 mg/liter cefeximine and 2.5 mg/liter potassium tellurite (ctsmac; Sigma), and another 50 l was spread on Rainbow-Plus agar (Rainbow agar O157, Bilog Inc., Hayward, Calif.) containing 0.8 g/ml potassium tellurite (Sigma) and 20 g/ml novobiocin (Sigma). Each plate was incubated for 18 hat37 C. Following incubation, up to five typical E. coli O157: H7 colonies identified on ctsmac plates (colorless with or without a dark center) or Rainbow agar (dark slightly blue colonies) were removed and screened with a latex agglutination assay (E. coli O157:H7 Test Kit, Oxoid, Ogdensburg, N.Y.). If agglutination occurred, the colony was tested with a control latex reagent to ensure that the isolate was not an autoagglutinating strain. Initial biochemical analysis of presumptive E. coli O157:H7 isolates included cellobiose, triple sugar iron, and motility tests. Presumptive positive isolates were then screened with an O and H antigen agglutination test (RIM E. coli O157:H7 latex test, Remel, Lenexa, Kans.). Each E. coli O157:H7 isolate from each sample was subjected to additional biochemical testing, including VITEK analysis (biomérieux, Hazelwood, Mo.). Once confirmation testing was complete, isolates were stored in a 20% glycerol- TSB solution and frozen until molecular characterization analysis could be conducted. Multiplex PCR. Each isolate was subjected to each of three multiplex PCR procedures. A combination of different amplification primers was used for (i) Shiga toxin genes 1 and 2 (stx 1 and stx 2 ) as per Wang et al. (27); (ii) genes for enterohemorrhagic E. coli enterohemolysin (hlya), E. coli somatic antigen O157 (rfbe O157 ), and E. coli structural flagella antigen H7 (flic H7 )as per Wang et al. (27); and (iii) the gene for effacing and attachment of E. coli O157:H7 (eaea O157 ) as per Gannon et al. (10). All reactions included a primer set for the 16S rrna gene for E. coli to serve as an internal control as described by Wang et al. (27). DNA was prepared from each E. coli O157:H7 isolate with the Instagene System (Bio-Rad, Hercules, Calif.). All plasticware and pipettes used for the preparation of the PCR master mixes were placed under UV light for 20 min under a laminar flow hood before use to dimerize any residual DNA fragments. The gel was then removed from the destaining solution and placed in Gel Doc EQ (Bio-Rad) for UV photo imaging. Exported gel images were analyzed with the QuantityOne software program (Bio-Rad). PFGE. PFGE was conducted as previously described by Barkocy-Gallagher et al. (2). The gel was then removed from the destaining solution and placed in Gel Doc EQ for UV photo imaging. Exported gel images and band patterns were analyzed with the FingerPrinting II software program (Bio-Rad). Data analysis was conducted with the method described by Tenover et al. (25), as used during epidemiologic studies of foodborne outbreaks. The analysis is based on the use of a single restriction endonuclease. In the current study, one restriction endonuclease was used, and samples were all collected within 1 year. Only isolates with seven or fewer PFGE band differences were used to determine relationships between samples. Tenover et al. (25) suggested the following relationships: isolates with zero or one band difference are indistinguishable and are considered definitely associated with the foodborne outbreak; isolates with two or three band differences are closely related and probably associated with the outbreak; and isolates with four to six band differences are possibly related and possibly associated with the outbreak. RESULTS Distributions of samples positive for E. coli O157:H7 are listed in Table 1. E. coli O157:H7 was detected in samples from feedlot water troughs, pen floors, feed bunks, loading chutes, trailer side walls, trailer floors, and packing plant holding pen floors, side rails, and drinking water. The only sample locations that did not test positive for E. coli O157:H7 were feed and restrainer side walls. The study was not designed to diagnose E. coli O157:H7 prevalence for environmental samples or carcasses but to identify potential cross-contamination points at the feedlot, during transportation, and in holding pens at the packing plant. In both the Midwest and Southwest, all feedlots tested positive for E. coli O157:H7 at both the pen floor and loading chute side panel sample locations. However, E. coli O157:H7 was detected in feed bunk samples and loading chute fecal samples collected at midwestern feedlots but not in loading chute fecal samples from southwestern feedlots (Table 1). E. coli O157:H7 was detected in both midwestern and southwestern transport trailer side walls. Only fecal samples from southwestern trailer floors tested positive for E. coli O157:H7 (Table 1). At the packing plants, E. coli O157:H7 was detected on the receiving pen side rails and in water samples in the receiving pens. The receiving pen floor samples were positive for E. coli O157:H7 in the Midwest but not in the Southwest. Colon samples also tested positive for E. coli O157:H7 in both the Midwest and Southwest. Multiplex PCR analysis. The distribution of E. coli O157:H7 virulence genes is separated by sample location in Table 2. Overall, stx 1, hlya, rfbe O157, flic H7, and eaea O157 genes were detected in at least one isolate from each location. The only locations where E. coli O157:H7 isolates did not include the stx 2 gene were the water at the

J. Food Prot., Vol. 69, No. 6 TRANSMISSION OF E. COLI O157:H7 FROM PREHARVEST TO HARVEST 1243 TABLE 1. Prevalence of E. coli O157:H7 in samples obtained from the feedlot, transport trailers, and packing plants Collection site No. of positive samples Midwest Southwest Total No. of samples collected No. of isolates Feedlot Water 2 0 2 64 3 Pen floor 8 11 19 96 46 Feed 0 0 0 48 0 Feed bunk 1 0 1 32 2 Chute, swab of side panels 1 1 2 20 9 Chute, feces from floor 2 0 2 30 5 Transport trailers Swab of side panels 3 2 5 75 14 Feces from floor 0 3 3 75 5 Processing facility Receiving pen floor 0 2 2 100 3 Receiving pen side rails 1 1 2 50 8 Water 1 0 1 50 5 Restrainer side walls 0 0 0 8 0 Carcasses Hide 25 50 75 442 176 Colon 8 8 16 75 48 feedlot and the receiving pen floor at the packing plant. In more than 50.0% of samples from the feedlot chute side panels, trailer side walls, packing plant receiving pen side rails, and packing plant water, both stx 1 and stx 2 genes were detected (Table 2). The number of hide samples positive for stx 1 and stx 2 genes was greater than that for the companion samples; the hide samples also had the most isolates, 176. There were 48 colon isolates, but only 3 possessed both stx 1 and stx 2 (Table 3). Of the 46 feedlot pen floor isolates, 12 had stx 1 and stx 2 genes. In samples from the carcass (hide and colon), more isolates did not have either stx 1 or stx 2 genes than did express both genes. In environmental samples, with the exception of the feedlot chute floor, more isolates possessed both stx 1 and stx 2 genes than did not express either stx 1 and stx 2 genes. None of the feedlot water and feed bunk isolates ex- TABLE 2. Distribution of E. coli O157:H7 virulence genes among carcass and companion samples Collection site No. of positive samples No. of isolates No. of isolates with: eaea O157 stx 1 stx 2 hlya rfbe O157 flic H7 No. of isolates with all genes Feedlot Water 2 3 2 2 0 3 3 3 0 Pen floor 19 46 43 30 24 44 42 43 9 Feed 0 0 Feed bunk 1 2 1 1 1 1 1 1 0 Chute, swab of side panels 2 9 9 5 8 9 9 9 5 Chute, feces from floor 2 5 4 2 1 4 4 4 1 Transport trailers Swab of side panels 5 14 13 8 12 13 13 13 7 Feces from floor 3 5 3 2 3 2 3 2 1 Processing facility Receiving pen floor 2 3 1 1 0 1 3 1 0 Receiving pen side rails 2 8 7 6 6 6 8 6 6 Water 1 5 5 4 4 5 5 5 4 Restrainer side walls 0 0 Carcasses Hide 75 176 101 59 67 88 136 92 24 Colon 16 48 45 23 17 40 42 40 3 Total 130 322 60

1244 CHILDS ET AL. J. Food Prot., Vol. 69, No. 6 TABLE 3. Distribution of Shiga toxin genes among E. coli O157: H7 isolates from the feedlot, transport trailers, and packing plants Collection site Total no. of isolates No. of isolates without stx 1 or stx 2 a Feedlot Water 3 3 Pen floor 46 34 Feed 0 Feed bunk 2 2 Chute, swab of side panels 9 4 Chute, feces from floor 5 4 Transport trailers Swab of side panels 14 6 Feces from floor 5 4 Processing facility Receiving pen floor 3 3 Receiving pen side rails 8 2 Water 5 1 Restrainer side walls 0 Carcasses Hide 176 142 Colon 48 45 a Isolates that did not contain stx 1, stx 2, or both genes. pressed all genes (Table 2). At the processing facility, none of the receiving pen floor isolates expressed all virulence genes. Of particular interest, more than 50.0% of the isolates from samples collected from the feedlot loading chute side panels, trailer side panels, receiving pen side rails, and receiving pen water samples possessed all virulence genes (Table 2). PFGE analysis. For PFGE analyses, data were first classified into groups of isolates that had the same virulence gene profile and then further separated by packing plant of origin. Plants 1 and 2 are located in the Upper Midwest, and plant 3 is located in the southwestern United States. As illustrated in Tables 4 through 6, isolates from three sample pairs were indistinguishable: those from a feedlot pen floor fecal sample and a packing plant pen floor fecal sample (2,348 km apart) and those from two sets of colon samples. Isolates from 11 samples were determined to be closely related (two or three band differences). Samples comparisons of note were feedlot loading chutes with plant pen side rails, hides from consecutive carcasses, feedlot pens in different locations of the feedlot, feedlot pen feces with hides and colons, and trailer side rails with colons. An isolate from a feedlot pen floor fecal sample was closely related to that of a processing plant pen floor sample, even though the sampling dates were 4 months apart (4 October 2004 and 7 February 2005, respectively). TABLE 4. Comparison of E. coli O157:H7 isolate relationships based on PFGE analysis, plant 1 (Midwest) a Sample pair Genes present in isolates No. of isolate fragment differences Epidemiologic interpretation of isolates Feedlot pen floor fecal plant pen floor fecal eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 1 Indistinguishable Loading chute side panel plant pen side rail eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 2 Closely related Loading chute side panel plant pen side rail eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 4 Possibly related Feedlot pen 1 feedlot pen 2 (same feedlot) eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 6 Possibly related Hide trailer side wall eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 5 Possibly related Feedlot pen floor fecal plant pen side rail eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 6 Possibly related Trailer side wall feed bunk swab eaea O157, stx 2, hlya, rfbe O157, flic 5 Possibly related Hide trailer side wall eaea O157, stx 2, hlya, rfbe O157, flic 6 Possibly related Colon colon eaea O157, stx 1, hlya, rfbe O157, flic 4 Possibly related Hide feedlot pen floor fecal eaea O157, stx 1, hlya, rfbe O157, flic 2 Closely related Hide feedlot pen floor fecal eaea O157, stx 1, hlya, rfbe O157, flic 4 Possibly related Colon colon eaea O157, stx 1, hlya, rfbe O157, flic 6 Possibly related Hide hide eaea O157, hlya, rfbe O157, flic 4 Possibly related Hide hide (consecutive carcasses) eaea O157, hlya, rfbe O157, flic 3 Closely related Hide hide eaea O157, hlya, rfbe O157, flic 4 Possibly related Trailer side wall plant pen water eaea O157, hlya, rfbe O157, flic 4 Possibly related Hide plant pen water eaea O157, hlya, rfbe O157, flic 5 Possibly related Hide colon eaea O157, hlya, rfbe O157, flic 5 Possibly related Hide hide eaea O157, hlya, rfbe O157, flic 4 Possibly related Loading chute floor fecal trailer side wall eaea O157, hlya, rfbe O157, flic 5 Possibly related Hide hide hlya, rfbe O157, flic 4 Possibly related Hide hide hlya, rfbe O157, flic 4 Possibly related Hide hide hlya, rfbe O157, flic 4 Possibly related Hide pen water hlya, rfbe O157, flic 4 Possibly related a PFGE analyses were conducted using the method of Tenover et al. (25). Comparisons were made only for isolates with similar genotypes based on a multiplex PCR analysis.

J. Food Prot., Vol. 69, No. 6 TRANSMISSION OF E. COLI O157:H7 FROM PREHARVEST TO HARVEST 1245 TABLE 5. Comparison of E. coli O157:H7 isolate relationships based on PFGE analysis, plant 2 (Midwest) a Sample pair Genes present in isolates No. of isolate fragment differences Epidemiologic interpretation of isolates Colon colon (consecutive carcasses) eaea O157, stx 1, hlya, rfbe O157, flic 1 Indistinguishable Feedlot pen floor fecal colon eaea O157, stx 1, hlya, rfbe O157, flic 5 Possibly related Feedlot pen 1 floor fecal feedlot pen 2 floor fecal eaea O157, stx 1, hlya, rfbe O157, flic 2 Closely related Feedlot pen 1 floor fecal feedlot pen 2 floor fecal eaea O157, stx 1, hlya, rfbe O157, flic 5 Possibly related Colon feedlot pen floor fecal eaea O157, stx 1, hlya, rfbe O157, flic 5 Possibly related a PFGE analyses were conducted using the method of Tenover et al. (25). Comparisons were made only for isolates with similar genotypes based on a multiplex PCR analysis. Multiplex PCR and PFGE analyses. Davis et al. (6) suggested that the classification system described by Tenover et al. (25) based on one restriction enzyme digest has the potential to misclassify unrelated items as closely related or may not show a close relationship when in fact two PFGE patterns are related. However, Davis et al. (6) stated that when there is a high probability that isolates are epidemiologically linked, PFGE analysis with one restriction enzyme is appropriate. To demonstrate that isolates in the current study had an epidemiological link, PFGE patterns were only compared when isolates had the same molecular characteristics (i.e., presence of virulence gene as determined by multiplex PCR analysis) (Tables 4 through 6). Samples from different packing plants were not analyzed together; only samples from the packing plant and its corresponding feedlot were analyzed for similar PFGE patterns. For plant 1 located in the Midwest (Table 4), the isolate from packing plant pen 1 sample 20 matched those from loading chute side rail sample 6 and feedlot pen sample 23. In turn, the isolate from feedlot pen sample 23 matched that of a fecal sample from feedlot pen 2. For the Midwest, plant 1 companion samples showed relationships that linked the E. coli O157:H7 isolates found in samples from the packing plant holding pen side rail with those from the feedlot loading chute and feedlot pen floor. Analysis of plant 2 samples from the Midwest (Table 5) indicated that isolates from two colons that were collected from two consecutive carcasses were indistinguishable. There was also a relationship between isolates from a feedlot pen floor fecal sample and colon samples. An isolate from colon sample 23 matched that from colon sample 24, and an isolate from colon sample 24 matched that from feedlot pen floor sample 3. These results suggests that multiple cattle within a pen may be shedding a genetically similar E. coli O157:H7 strain, or animals within a pen may become infected with E. coli O157:H7 from other animals within the pen or during transport. TABLE 6. Comparison of E. coli O157:H7 isolate relationships based on PFGE analysis, plant 3 (Southwest) a Sample pair Genes present in isolates No. of isolate fragment differences Epidemiologic interpretation of isolates Trailer side rail colon eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 3 Closely related Feedlot 5 pen floor fecal hide eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 2 Closely related Feedlot 4 pen floor fecal hide eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 4 Possibly related Hide hide eaea O157, stx 1, stx 2, hlya, rfbe O157, flic 3 Closely related Feedlot 4 pen floor fecal feedlot 5 hide eaea O157, stx 2, hlya, rfbe O157, flic 4 Possibly related Feedlot 4 pen floor fecal colon eaea O157, stx 2, hlya, rfbe O157, flic 3 Closely related Feedlot 4 pen floor fecal colon eaea O157, stx 2, hlya, rfbe O157, flic 2 Closely related Feedlot 4 pen floor fecal colon eaea O157, stx 2, hlya, rfbe O157, flic 5 Possibly related Feedlot 4 pen floor fecal colon eaea O157, stx 2, hlya, rfbe O157, flic 4 Possibly related Colon colon eaea O157, stx 2, hlya, rfbe O157, flic 1 Indistinguishable Trailer side rail colon eaea O157, stx 2, hlya, rfbe O157, flic 2 Closely related Trailer side rail feedlot 4 pen floor fecal eaea O157, stx 2, hlya, rfbe O157, flic 5 Possibly related Feedlot pen floor fecal plant pen floor b eaea O157, stx 1, hlya, rfbe O157, flic 3 Closely related Trailer fecal colon c eaea O157, stx 1, hlya, rfbe O157, flic 6 Possibly related Feedlot pen floor fecal colon b eaea O157, stx 1, hlya, rfbe O157, flic 6 Possibly related Feedlot 4 hide Feedlot 5 colon eaea O157, hlya, rfbe O157, flic 6 Possibly related a PFGE analyses were conducted using the method of Tenover et al. (25). Comparisons were made only for isolates with similar genotypes based on a multiplex PCR analysis. b Sample dates were 4 October 2004 and 7 February 2005, respectively. c Sample dates were 13 October 2004 and 7 February 2005, respectively.

1246 CHILDS ET AL. J. Food Prot., Vol. 69, No. 6 Analysis of plant 3 samples from the Southwest (Table 6) revealed numerous isolates that were closely linked. The isolate from feedlot 4 pen floor fecal sample 23 matched that from a feedlot 5 hide sample. An isolate from feedlot 4 pen floor fecal sample 23 also matched those of colon samples 37 and 38. The isolate from colon sample 38 matched those of colon sample 47 and trailer sample 5. Further investigation revealed that the same trailers used to transport cattle from feedlot 5 in the Southwest also transported cattle from feedlot 4. In this case, the trailer was a potential transfer point for E. coli O157:H7 between the two feedlots and multiple cattle. DISCUSSION Results of this study are similar to those described by Tutenal et al. (26). E. coli O157:H7 is detectable in packing plant holding pens, and stx 1 and stx 2 are often but not always present in these isolates. Although Tutenel et al. (26) did not sample any locations outside the packing plant, they concluded that direct hide-to-hide cross-contamination most likely occurred between the time the cattle left the feedlot and the time the cattle arrived at the processing facility. In the current study, potential cross-contamination points were identified from PCR and PFGE analyses and may serve as transmission routes from the companion samples (feedlot pen fecal samples, feedlot loading chute, packing plant holding pen side rails, trailer side rails, and colons) to hides. The current study is the first to identify genotypic matches between isolates from transport trailer side wall samples and those from packing plant cattle hide samples. Genotypic matches for E. coli O157:H7 isolates were also identified for samples from (i) trailer side walls, feedlot feed bunks, and loading chutes and (ii) packing plant receiving pen floors, side rails, and colons. In other studies, contaminated water troughs, feed (not in the current study), feces from feedlot pen floors, and cattle anuses (part of removal process on the harvest floor) are potential cross-contamination points (9, 16, 18, 23). Although Dewell et al. (7) obtained samples from regional locations different from those sampled in the present study, they found regional differences between cattle from eastern Colorado and those from central Nebraska. Although the current study was not designed to determine prevalence, samples that were positive for E. coli O157:H7 in the Midwest were not necessarily the same as those found in the Southwest (Table 1). Among feedlot samples, E. coli O157:H7 was not found in trailer floor fecal samples from the Midwest but was found in trailers from the Southwest. E. coli O157:H7 was not detected in samples of water or feed bunks in southwestern feedlots but was detected in these same samples from midwestern feedlots. For packing plants, E. coli O157:H7 was detected in midwestern water samples but not in those from the Southwest. Southwestern packing plant holding pen floor samples were positive for E. coli O157:H7 but those from the Midwest were not. Regions vary greatly in humidity, sea level, average temperature, wind, and concentration of cattle in feedlots. This study did not account for all environmental factors that could influence prevalence of E. coli O157:H7 or differences in the expression of virulence genes. The findings of this study agree with those of Naylor et al. (20), who determined that E. coli O157:H7 is found in the rectoanal junction of animals. In the current study, E. coli O157:H7 isolates from the colon matched those of numerous environmental samples, as confirmed by PFGE analysis (Tables 4 through 6) and multiplex PCR analysis. Although the mode of transmission for E. coli O157:H7 isolates between the colon and the environmental samples was not investigated, these data suggest that shedding via the colon plays a critical role in distributing E. coli O157: H7 throughout the preharvest environment. In addition to the environmental samples, genotypic matches for E. coli O157:H7 isolates were found between colons and between hides from different animals. These results demonstrate the need for the beef industry to implement standard sanitation practices and/or decontamination interventions that may reduce or eliminate E. coli O157:H7 from cattle colons. In this study, E. coli O157:H7 isolates from preharvest environmental samples genetically matched those from hide samples. Therefore, E. coli O157:H7 hide contamination can be traced to feedlot, transport trailers, and packing plant holding pens. Further research is needed to determine modes of transmission and effective methods for reduction of E. coli O157:H7 contamination in feedlot loading chutes, trailer side panels, and water troughs, both at feedlots and at beef packing plants, and in feed bunks in feedlots. ACKNOWLEDGMENTS This study was funded by beef and veal producers and importers through their $1-per-head checkoff fund administered by the Cattlemen s Beef Board. REFERENCES 1. Barkocy-Gallagher, G. A., T. M. Arthur, M. Riveria-Betancourt, X. Nou, S. D. Shackelford, T. L. Wheeler, and M. Koohmaraie. 2003. 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. 66:1978 1986. 2. Barkocy-Gallagher, G. A., T. M. Arthur, G. R. Siragusa, J. E. Keen, R. O. Elder, W. W. Laegreid, and M. Koohmaraie. 2001. Genotypic analysis 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. 67:3810 3818. 3. Barkocy-Gallagher, G. A., E. D. Berry, M. Riveria-Betancourt, T. M. Arthur, X. Nou, and M. Koohmaraie. 2002. Development of methods for recovery of Escherichia coli O157:H7 and Salmonella from beef carcass sponge samples and bovine fecal and hide samples. J. Food Prot. 65:1527 1534. 4. Bell, R. G. 1997. Distribution and sources of microbial contamination on beef carcasses. J. Appl. Microbiol. 82:292 300. 5. Chapman, P. A., C. A. Siddons, D. J. Wright, P. Norman, J. Fox, and E. Crick. 1993. Cattle as a possible source of verocytotoxin-producing Escherichia coli O157:H7 infections in man. Epidemiol. Infect. 111:439 447. 6. Davis, M. A., D. D. Hancock, T. E. Besser, and D. R. Call. 2003. Evaluation of pulsed field gel electrophoresis as a tool for determining the degree of relatedness between strains of Escherichia coli O157:H7. J. Clin. Microbiol. 41:1843 1849. 7. Dewell, G. D., J. R. Ransom, R. D. Dewell, K. McCurdy, I. A. Gardner, A. E. Hill, J. N. Sofos, K. E. Belk, G. C. Smith, and M. D. Salman. 2005. Prevalence of and risk factors for Escherichia

J. Food Prot., Vol. 69, No. 6 TRANSMISSION OF E. COLI O157:H7 FROM PREHARVEST TO HARVEST 1247 coli O157:H7 in market-ready beef cattle from 12 U.S. feedlots. Foodborne Pathog. Dis. 2:70 76. 8. Elder, R. O., J. E. Keen, G. R. Siragusa, G. A. Barkocy-Gallagher, M. Koohmaraie, and W. W. Lagreid. 2000. Correlation of enterohemorrhagic Escherichia coli prevalence in feces, hides, and carcasses of beef cattle during processing. Proc. Natl. Acad. Sci. USA 97:2999 3003. 9. Fegan, N., G. Higgs, P. Vanderlinde, and P. Desmarchelier. 2005. An investigation of Escherichia coli O157 contamination of cattle during slaughter at an abattoir. J. Food Prot. 68:451 457. 10. Gannon, V. P. J., S. D Souza, T. Graham, R. K. King, K. Rahn, and S. Read. 1997. Use of the flagellar H7 gene as a target in multiplex PCR assays and improved specificity in identification of enterohemorrhagic Escherichia coli strains. J. Clin. Microbiol. 35:656 662. 11. Gill, C. O., and J. C. McGinnis. 1999. Improvement of the hygienic performance of the hindquarters skinning operations at a beef packing plant. Int. J. Food Microbiol. 51:123 132. 12. Gill, C. O., J. C. McGinnis, and M. Badoni. 1996. Assessment of the hygienic characteristics of a beef carcass dressing process. J. Food Prot. 59:136 140. 13. Gill, C. O., J. C. McGinnis, and M. Badoni. 1996. Use of total or Escherichia coli counts to assess the hygienic characteristics of a beef carcass dressing process. Int. J. Food Microbiol. 31:181 196. 14. Grau, F. H. 1987. Prevention of microbial contamination in the export beef abattoir, p. 221 233. In F. J. M. Smulders (ed.), Elimination of pathogenic organisms from meat and poultry. Elsevier Science, Amsterdam. 15. Grauke, L. J., I. T. Kudva, J. W. Yoon, C. W. Hunt, C. J. Williams, and C. J. Hovde. 2002. Gastrointestinal tract location of Escherichia coli O157:H7 in ruminants. Appl. Environ. Microbiol. 68:2269 2277. 16. LeJune, J. T., T. E. Besser, N. L. Merrill, D. H. Rice, and D. D. Hancock. 2001. Livestock drinking water microbiology and the factors influencing the quality of drinking water offered to cattle. J. Dairy Sci. 84:1856 1862. 17. McEvoy, J. M., A. M. Doherty, M. Finnerty, J. J. Sheridan, L. McGuire, I. S. Blair, D. A. McDowell, and D. Harrington. 2000. The relationship between hide cleanliness and bacterial numbers on beef carcasses at a commercial abattoir. Lett. Appl. Microbiol. 30:390 395. 18. McGee, P., D. J. Bolton, J. J. Sheridan, B. Earley, G. Kelly, and N. Leonard. 2002. Survival of Escherichia coli O157:H7 in farm water: its role as a vector in the transmission of the organism within herds. J. Appl. Microbiol. 93:706 713. 19. National Cattlemen s Beef Association. 2003. Production best practices pre-harvest report. Draft (20 October version). National Cattlemen s Beef Association, Centennial, Colo. 20. Naylor, S. W., J. C. Low, T. E. Besser, A. Mahajan, G. J. Gunn, M. C. Pearce, I. J. McKendrick, G. E. Smith, and D. L. Gally. 2003. Lymphoid follicle-dense mucosa at the terminal rectum is the principal site of colonization of enterohemorrhagic Escherichia coli O157:H7 in the bovine host. Infect. Immun. 71:1505 1512. 21. Ransom, J. R., J. N. Sofos, K. E. Belk, G. D. Dewell, K. S. Mc- Curdy, G. C. Smith, and M. D. Salman. 2003. Prevalence of Escherichia coli O157:H7 in feedlot cattle feces and on carcasses from those cattle, p. 116 117. Abstr. 10th Symp. Int. Soc. Vet. Epidemiol. Econ. International Society for Veterinary Epidemiology and Economics, Viña del Mar, Chile. 22. Rice, D. H., H. Q. Sheng, S. A. Wynia, and C. J. Hovde. 2003. Rectoanal mucosal swab culture is more sensitive than fecal culture and distinguishes Escherichia coli O157:H7 colonized cattle and those transiently shedding the same organism. J. Clin. Microbiol. 41: 4924 4929. 23. Smith, D., M. Blackford, S. Younts, R. Moxley, J. Gray, L. Hungerford, T. Milton, and T. Klofenstein. 2001. Ecological relationships between the prevalence of cattle shedding Escherichia coli O157:H7 and characteristics of cattle or conditions of the feedlot pen. J. Food Prot. 64:1899 1903. 24. Sofos, J. N., S. L. Kochevar, G. R. Bellinger, D. R. Buege, D. D. Hancock, S. C. Ingham, J. B. Morgan, J. O. Reagan, and G. C. Smith. 1999. Sources and extent of microbiological contamination of beef carcasses in seven U.S. slaughtering plants. J. Food Prot. 62:140 145. 25. Tenover, F. C., R. D. Arbeit, R. V. Goering, P. A. Mickelsen, B. A. Murray, D. A. Persing, and B. Swaminathan. 1995. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J. Clin. Microbiol. 33: 2233 2239. 26. Tutenel, A. V., D. Pierard, J. Van Hoof, and L. de Zutter. 2003. Molecular characterization of Escherichia coli O157 contamination routes in a cattle slaughterhouse. J. Food Prot. 66:1564 1569. 27. Wang, G., C. G. Clark, and F. G. Rodgers. 2002. Detection in Escherichia coli of the genes encoding the major virulence factors, the genes defining the O157:H7 serotype, and components of the type 2 Shiga toxin family by multiplex PCR. J. Clin. Microbiol. 40:3613 3619.