Research Areas

Animal Health - Food Safety 2005-2007

Researcher: Jose Renaldi Feitosa Brito


To prospect the opportunities of collaboration, and establish initial cooperation linkages between the ARS/USDA and Embrapa to conduct research to better recover, characterize and control pathogens at various points in the food supply, and develop strategies and technologies to prevent bacterial pathogens from entering the food chain or to cause their destruction if present, for the mutual benefit of both Brazil and the United States.

Projects developed

Identification and characterization of bacterial pathogens in the dairy and pork food chains. (Embrapa project MP2 – 02-04-3-08)
Participating Institutions: CNPAT, CNPC, CNPGL, CNPSA, CTAA, and ERRC.

Objectives: Identify and characterize bacterial pathogens in selected animal products; identify the source of contamination, to understand the epidemiology; identify interventions to better manage food borne pathogens in Brazilian products; and expand interactions and collaborations among a network of Embrapa laboratories on food safety and ERRC.
Approach: Bacterial food borne pathogens were recovered from pasteurized milk (Staphylococcus aureus), Minas Frescal Cheese (Listeria monocytogenes, S. aureus) and pork (Salmonella spp.) Molecular subtyping was used to characterize these pathogens and to link them to the production environment, to confirm the source of contamination and to test the result of interventions.
Finished Experiments:
Pilot study 1. Prevalence of L. monocytogenes, S. aureus, Salmonella and E. coli O157:H7 in pasteurized milk and Minas Frescal Cheese (MFC). Pasteurized milk and MFC samples were obtained between June and September 2005 (the dry/cold season/winter) and between December 2005 and March 2006 (wet/warm season/summer) from up to 37 retail sites, located in seven regions of Juiz de Fora, Minas Gerais, Brazil. S. aureus was the only pathogen recovered from milk (1 of 9 brands positive in winter and 2 of 9 brands positive in the summer). S. aureus was also recovered from 14 of 50 (28%) cheese samples, from 6 of 10 brands in winter, and from 22 of 50 (44%) samples representing 6 of 10 brands in summer.
L. monocytogenes was recovered from 4 of 50 samples (8%), from 1 of 10 (10%) brands of MFC in winter. An intervention was made in the dairy plant (see next item). No isolation of L. monocytogenes was made from cheese samples tested in summer. Salmonella and E. coli O157:H7 were not recovered from any sample of milk or cheese.
The isolation and identification of all isolates were conducted at CNPGL. Molecular characterization is being conducted at ERRC – MFSRU. Further phenotypic and genotypic characterizations of L. monocytogenes, S. aureus and other staphylococci isolates are being conducted at MFSRU.

Pilot study 2. Identification of sources of L. monocytogenes contamination at a dairy plant and implementation of corrective measures.

Analysis of data of Pilot study 1 indicated that the source of contamination of Minas Frescal Cheese with L. monocytogenes was the dairy plant. Moreover, although only one brand of cheese was contaminated, a total of 6 of 10 cheese samples obtained from this plant at retail tested positive for L.. monocytogenes (4 of 5 during the survey and 2 of 5 after). Thus in October of 2005 the farm/dairy that produced brand F cheese was sampled. All 5 samples from the milking parlor tested negative for L. monocytogenes, whereas 10 of 23 sites/samples from the processing plant environment and 5 of 5 MFC samples tested positive for the pathogen. All 344 isolates (5-20 isolates/sample) selected from the 21 positive retail and Plant F samples were serotype 1/2a and displayed the same AscI pulsed-field restriction profile.
The results of pulsed-field fingerprinting (Figure 1) also established that the coolers/refrigeration units served as the point source within Plant F that contaminated MF. Based on these findings, and in collaboration with both the producer and the State Service for Veterinary and Phytosanitary Supervision (IMA, MG), failures in the hygienic process and plant design were identified and subsequently corrected. The plant remained closed from October through December of 2005, until the corrective renovations of the facility were completed and 29 samples collected from the processing environment tested negative for L. monocytogenes. Moreover, on 3 subsequent visits to Plant F between March and October of 2006 an additional 97 environmental and 15 MF samples from Plant F and MF from retail establishments all tested negative for L. monocytogenes. These findings confirm: (a) the environment of the processing plant plays an important role in preventing Listeria contamination and should be a matter of primary concern for the dairy sector and health authorities; (b) post-processing contamination is probably a common event in dairy establishments; (c) the benefit of proper sanitization and design of small dairy plants, thus reducing the risk of contamination with L. monocytogenes and probably other pathogens; (d) PFGE is a useful tool for identification of the genetic similarity among isolates, and for identifying the source of contamination.
These two pilot studies are part of a PhD thesis (Emilia M. P. dos Santos, UFMG). Preliminary studies to standardize the PCR reaction to identify/confirm the identity of food pathogens were conducted by Paula M. Garcia and Nadia D. Peres (MSc students/UFMG).
Ongoing experiments:
Prevalence and molecular characterization of S. aureus, Salmonella spp. and L. monocytogenes recovered from Minas Frescal cheese, Coalho cheese, and goat cheeses produced in Brazil.
This survey was planned to confirm the extension of the contamination of L. monocytogenes, S. aureus and Salmonella spp. from cheese (summer/2006-2007 and winter/2007). In this survey we increased the sample size (740 cheese samples in total) and the number of brands (16) in Minas Gerais State. Surveys are being conducted in two dairies to identify the source of contamination of S. aureus and L. monocytogenes, using the same approach described in Pilot Study 2.
Three Embrapa Units (CNPC, CNPAT and CTAA) are conducting similar surveys in Sobral/Fortaleza (goat cheese), Fortaleza (“coalho” cheese) and Rio de Janeiro (Minas Frescal cheese). The molecular characterization of the isolates will be conducted in Embrapa laboratories with the MFSRU/ERRC support.

Prevention of L. monocytogenes contamination in Minas Frescal cheese by chemical/biological intervention.

Our results show that contamination of MF cheese by L. monocytogenes can be attributed in part to post processing contamination. L. monocytogenes may be present even when pasteurized milk is used for cheese making. Since L. monocytogenes is able to grow in food kept at refrigeration temperature (4-8oC), its presence in read to eat food, as is the case of MF cheese, is always a concern. . Thus, we are conducting an experiment to test an intervention procedure (use of nisin) to prevent L. monocytogenes growth in MFC. The experiment is being conducted in the Challenge Facility / Pilot Plant of the MFSRU-ERRC (April – June 2007). Experimentally produced MFC are being artificially contaminated with a set of L. monocytogenes isolates. Cheese samples are kept at 4°C and at 10°C (abusive temperature), and examined on days 0, 3, 6 and 10 post-processing. We are also examining three different possibilities of cheese contamination (during manufacturing, immediately after cheese molding, and the following day before packaging). The results obtained so far (the first of three trials) indicate that addition of nisin (a food grade bacteriocion) is effective in reducing L. monocytogenes numbers in artificially contaminated cheese (2 to 3 log difference). The number of bacteria (total bacterial counts and L. monocytogenes counts) increased significantly (1 to 3 log difference) when cheese samples were kept at abusive temperature (10°C). We need more data to confirm some of the results such as the influence of the timing of contamination. These results indicate that the addition of nisin is a promising tool to eliminate or prevent L. monocytogenes contamination of MFC. They also indicate that sanitary quality of MFC would benefit if they were kept at 4°C instead of 8°C to 10°C (or even higher temperatures) usually found in many retail establishments (data from our project).
Future collaboration (new projects):
A proposal on food safety of goat and lamb meat products has been discussed by Francisco Selmo F. Alves and other scientists from CNPC with John B. Luchansky (ERRC) and J. R. F. Brito (Embrapa Labex).

Validation of the effect of interventions and processes on persistence of pathogens in foods (ERRC / MFS CRIS 1935-41420-012-001
Participating Institutions: USDA Food Safety and Inspection Service (USDA/FSIS), USDA Animal Plant Health Inspection Service (USDA/APHIS), U.S. Food and Drug Administration (FDA), CIAD (Mexico), IFR (United Kingdom), Teagasc (Ireland), Purdue University, Rutgers University, University of Wisconsin, Hatfield Quality Meats, Tropical Cheese Industry, Food Products Association, National Cattlemen’s Beef Association, National Pork Board, and Embrapa Network of Food Safety Laboratories (CNPGL, CNPC, CTAA, CNPSA, CNPAT).

Objectives: Develop technical information and technologies needed by federal regulatory agencies, the food industry, consumers and the international scientific community on pathogenic bacteria and virus to ensure a safe food supply.
Approach: Two experiments were conducted as part of the above mentioned CRIS project. The first experiment was designed to characterize Enterobacter spp isolated from shell eggs using PFGE. The second one was designed to study the fate of Bacillus anthracis (Sterne) in pasteurized whole liquid egg stored at different temperatures and cooked using a commercial grill.
Experimental Results:
Experiment 1. The prevalence of Enterobacter associated with shell eggs was determined previously to assess the sanitation practices of three commercial facilities in the Southeastern United States. Enterobacter were recovered from 117 of 837 (13.97%). In the present study we established the relatedness of 176 isolates (1-5 isolates per positive sample) recovered from 96 of these 117 positive samples. Pulsed-field fingerprinting generated 121 XbaI pulsotypes (34-96% related) distributed as follows: Enterobacter cloacae (104 isolates; 73 pulsotypes, 39-100% related); E. amnigenus (32 isolates; 25 pulsotypes, 33-100% related); E. sakazakii (27 isolates; 12 pulsotypes, 33-100% related); and E. taylorae (9 isolates; 7 pulsotypes, 38-100% related). There was no predominate or persistent pulsotype among isolates from a given plant for any of the 3 visits; the 101 plant X isolates displayed 73 pulsotypes, the 14 plant Y isolates displayed 8 pulsotypes, and the 61 plant Z isolates displayed 41 pulsotypes. Multiple isolates were recovered from 47 of the 96 (49%) positive samples and multiple pulsotypes were displayed by isolates from 34 of these 47 samples (72%). For the majority of visits to each plant, more Enterobacter isolates/pulsotypes were recovered before washing/processing than during or after washing. Pulsed-field fingerprinting of isolates from all three stages indicated that processing was effective at reducing the prevalence and types of Enterobacter, and also demonstrated the possibility of multiple sources of contamination. These data establish that new/unique isolates of Enterobacter are introduced into the shell egg environment with each batch of eggs and that there is considerable heterogeneity among isolates and species associated with shell eggs.
Experiment 2. Commercial pasteurized whole liquid egg (WLE) was inoculated with heat- shocked (80°C, 10 min) or non-heated shocked spores of Bacillus anthracis (Sterne) (ca. 4.0 log10 spores/ml), stored at 4, 15 or 25°C for up to 14 days, and cooked using a commercial “clam-shell” type grill set at 149°C. For each sampling interval, two 52-ml portions of WLE inoculated with heat-shocked or non-heat shocked spores were placed in an aluminum mold to form a WLE patty with standardized size (21.6 x 7.7 cm) and thickness (2.5 mm) and cooked for either 2, 5 or 8 min. The average internal temperature of the WLE during cooking was 92°C±6°C. When inoculated WLE was stored at 4°C pathogen numbers decreased by 0.5 log spores/ml (heat-shocked) and 2.0 log spores/ml (non-heat shocked). In contrast, when inoculated WLE was stored at 15°C for 3 days or 25°C for 1 day, pathogen numbers increased by 3.0 log spores/ml (heat shocked) and 2.0 log spores/ml (non-heat shocked), and 0.2 log spores/ml (heat shocked) and 1.9 log spores/ml (non-heat shocked) respectively. When inoculated WLE was stored at 4°C and cooked for 2, 5 or 8 min, pathogen numbers decreased ca. 1.0 log spores/g for both heat shocked and non-heat shocked spore preparations. However, after storage at 15 or 25°C and cooking for up to 8 minutes, pathogen numbers decreased by ≥ 6.0 log spores/g for both heat-shocked and non-heat shocked spores. Thus, storage at abusive temperatures of 15 or 25°C allowed for germination and growth of B. anthracis in WLE and consequently much greater susceptibility to killing by a subsequent heat treatment. Storage of WLE at 4°C did not allow for spores germination. These data reinforce the need for additional research to decrease the potential risk due to contamination of WLE with a threat agent.