Mick Bosilevac 1, Brandon Leudtke 1 Rong Wang 1 and Yemi Ogunrinola 2. US Meat Animal Research Center, Clay Center NE

Transcription

1 Development and Application of a Green Fluorescent Protein (GFP) Expressing E. coli O103 Surrogate for Tracking Contamination Through Grinding and Identifying Persistent Points of Contamination Mick Bosilevac 1, Brandon Leudtke 1 Rong Wang 1 and Yemi Ogunrinola 2 1 US Meat Animal Research Center, Clay Center NE Vantage Foods, Chilliwack BC V2R 0E9 USDA is an equal opportunity provider and employer

2 Background Ground beef production lots are based on a unit of time (15, 30, 60 min) where the amount of ground beef produced per minute can vary. Each lot of ground beef gets sampled and tested for E. coli O157:H7. If a lot is found positive, disposition decisions for that lot, the lot before and the lot after are based on the assumption that as more material flows through the equipment, the contaminating bacteria move with the contaminated material. Therefore, it is important to determine the distribution and flow of a contamination event by identifying the length of time required for E. coli O157:H7 to clear the system (no longer detectable).

3 Background Two recent studies described the distribution of E. coli O157:H7 in industrial scale production equipment and provide some scientific basis for decisions. Distribution of Escherichia coli passaged through processing equipment during ground beef production using inoculated trimmings. Koohmaraie et al. J Food Prot Feb;78(2): Use of nonpathogenic, green fluorescent proteinmarked Escherichia coli Biotype I cultures to evaluate the selfcleansing capabilities of a commercial beef grinding system after a contamination event. Wages et al. J Food Prot Nov;77(11):

4 Background Prior information was much earlier and/or not performed at a large commercial scale. Incidence of Escherichia coli O157:H7 in frozen beef patties produced over an 8hour shift. Pruett et al J Food Prot. 65: Distribution patterns of E. coli O157:H7 in ground beef produced by a laboratory scale grinder. Flores and Tamplin J. Food Prot. 65: A statistical method to determine whether microorganisms are randomly distributed in a food matrix, applied to coliforms and Escherichia coli O157 in minced beef. Reinders et al Food Microbiol. 20: Empirical distribution models for Escherichia coli O157:H7 in ground beef produced by a midsize commercial grinder. Flores and Stewart J. Food Sci. 69:M121 M126.

5 Background The described GFP strains have issues Koohmaraie et al. J Food Prot Feb;78(2): Used E. coli Top10 cells. A lab strain for cloning and protein expression; not particularly resistant to interventions and not validated for use as a surrogate. Maintained the GFP, easily observed and detectable at a low level after enrichment (5 CFU/375g). Wages et al. J Food Prot Nov;77(11): Used the accepted FSIS surrogates (P8, P14 and P68) transformed with GFP expression vector. Do not maintain the GFP marker, or express it variably. Not easily observed (must view under microscope). Only direct counts were made, so limit of detection was ~1 CFU/0.5g (750 CFU/375g). To avoid these problems I made my own.

6 Objective Develop and validate an easily trackable E. coli O157:H7/nonO157 STEC surrogate that can be detected to the same level of sensitivity as E. coli O157:H7. Apply the trackable surrogate to model contamination passage through grinding and identify points where contamination may persist.

7 Experimental Design & Analysis A series of E. coli, serogroups O103, O26 and O145 sensitive to ampicillin and lacking virulence genes were transformed with a high copy number GFP expression vector. Transformants that stably expressed the GFP were selected and characterized for: Recovery of ~3 CFU in 375g ground beef Following storage at 4 C for up to 14 days Following storage at 20 C for up to 6 weeks. Detection using direct plating, immunomagnetic separation (IMS) and PCR after enrichment. Resistance to treatments with Lactic Acid, Citrilow and rosemary oil equal to that of E. coli O157:H7. Biofilm formation and surface binding within the range observed for E. coli O157:H7. One strain of E. coli O103 met all the requirements and was selected for use.

8 Experimental Design & Analysis 375g samples Incubated in 1L TSB amp. Directly swabbed and streak plated. 1mL is IMSed for GFPO103 E coli and plated. DNA is extracted from 1mL and PCR ed for GFP gene, visualized on DNA gel. Increasing sensitivity of detection The increased sensitivity is due to concentrating the GFP E. coli and reducing the background bacteria that prevent visualizing the GFP colonies and the fact that GFP gene has 40 copies per CFU

9 Experimental Design & Analysis GFPO103 E. coli was used to inoculate a combo of beef trim. The inoculated combo and four noninoculated combos were used sequentially to produce 4 lb. loaves of ground beef. After each experimental repetition, surface and residual meat samples were collected from all equipment through or across which the inoculated trim and subsequently produced ground beef had passed. The presence of the GFPO103 was monitored through the ground beef loaves in three replicate experiments using a high inoculation dose (6 log CFU/combo) and one replicate using a low inoculation dose (4 log CFU/combo) of GFPO103. Every 20 th loaf was tested to identify approximate end points of GFPO103 passage, then every 10 th and 5 th loaf around the approximate end point were tested in turn to identify the last GFP O103 containing loaf.

10 Experimental Design & Analysis Repetition 4, low inoculation x4 The seeded combo run through ground beef production followed by 4 additional combos to identify end point of contaminating GFP. Approximately 17,000 CFU GFP E. coli inoculated onto trim, that was then seeded into center of combo. 2,000 lbs. followed by 8,000 lbs. 4 ¼ lb loaves were packaged and every 20 th loaf was analyzed for the GFP E. coli. This is equal to 7 cells of E coli in each 375g sample from that combo. For additional resolution Batch 4 was retested at every 10 th loaf, and Batch 5 was retested at every 5 th loaf.

11 Process

12 Results of testing every 20 th 4 ¼ lb loaf for the GFP E. coli by direct plating, IMS and PCR Batch 1 (seeded) Batch 2 Batch 3 Batch 4 Batch 5 Plate IMS PCR Plate IMS PCR Plate IMS PCR Plate IMS PCR Plate IMS PCR % Pos

13 Results of testing every 5 th 4 ¼ lb loaf of Batch 5 for the GFP E. coli by direct plating, IMS and PCR Loaf # Plate IMS PCR Loaf # Plate IMS PCR Loaf # Plate IMS PCR % Pos

14 Percentage positive tests (plating, IMS and PCR) for GFP E. coli over a 10 sample window in total Plate IMS PCR 0 0 Batch 1 Batch 2 Batch 503 Batch Batch 5150 (seeded) Every 20 th loaf tested Every 10 th loaf tested Every 5 th loaf tested

15 Analysis Second Order Polynomial Regression Analysis to predict the point (in total number of pounds) where each test method may reach zero Percentage Positive of Samples In A 10 Sample WIndow Plate: y = 8E07x x R² = IMS: y = 1E06x2 5E05x R² = PCR: y = 1E10x3 3E06x x R² = Total Pounds of Grond Beef Processed Plate IMS PCR Poly. (Plate) Poly. (IMS) Poly. (PCR) Plating predicted to reach zero at 9,013 ± 985 total pounds processed IMS predicted to reach zero at 10,242 ± 919 total pounds processed

16 General Conclusions GFP E. coli at 17,000 CFU/2,000 lb combo (about 7CFU/375g sample) persisted through the next four combos of noninoculated trim. Best fit curves suggest 0% detectable GFP E. coli by any test method would be reached by ~11,000 pounds (6 combos). Therefore, for this processor, if a positive event is detected, it would be advisable to divert 12,000 pounds preceding and following the positive batch. Why didn t the GFP E. coli exit the system as in other studies? Everything we think about E. coli traveling with the meat is wrong. The various strains of GFP E. coli may behave differently. Top10 E. coli may be more sensitive to grinding forces The GFPO103 E. coli may not adhere to meat surfaces as tightly. The GFP plasmid is in such high copy number that it basically contaminates everything. The experimental nature of this production (multiple starts, stops, allowing each combo to clear before adding the following combo) may permit contaminated material to linger in the system, compared to normal daily production. This plant is different than those previously reported, and its system may contain points where contaminated meat may persist and recontaminate later batches.

17 After Production, surface sponge swabs and residual meat was collected for testing from all belts and equipment Type Location Plate IMS PCR 1 Sponge trim belt 2 Meat auger to 1st blender 3 Meat bars top of blender Sponge belt out of 1st blender 4 Meat belt out of 1st blender 5 Meat auger to 2nd blender Increasing sensitivity of detection sponge meat

18 After Production, surface sponge swabs and residual meat was collected for testing from all belts and equipment Type Location Plate IMS PCR 6 Meat auger to grinder 7 Meat grinder belt 8 Meat Vmag Horn 9 Meat Sponge Loaf belt Loaf belt meat sponge Meat from the grinder belt and sponge swab of the loaf belt were very low contamination, such that only IMS or PCR were positive. The IMS had just one GFP colony observed, while PCR had a low intensity band present.

19 Key Results A nonpathogenic strain of E. coli serogroup O103 was developed for use as an E. coli O157:H7 surrogate. The GFP surrogate had a similar growth curve, resistance antimicrobial treatments, and surface adherence (measured by biofilm forming ability) as strains of E. coli O157:H7. The GFPO103 could be easily identified in ground beef to a limit of detection of 15 CFU/325g (the same level as E. coli O157:H7 typically is), and could be recovered from ground beef stored at 4 C for two weeks and at 20 C for six weeks. The grinding study showed that neither the high nor the low dose inoculum was cleared completely through production after the fourth noninoculated combo had been processed. Regression analysis estimated that under the conditions of this investigation (processing equipment types, line configuration and complexity) an additional 8,542 ± 919 lbs of beef trim is required to clear a contaminant to a nondetectable level. Post experiment surface samples and residual meat still present on belts, augers and VMAG horn were identified as points where the GFPO103 persisted through the production of ground beef. Therefore in a complex operations where larger mixing and blending reservoirs & equipment are in use the microbes could linger for extended periods while various pockets of contaminated material clear out.

20