Contaminated Dairy Plant To Establish The Prevalence

AEM Accepts, published online ahead of print on 23 May 2008 Appl. Environ. Microbiol. doi:10.1128/AEM.01828-07 Copyright © 2008, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved. 1 2 3 MFS/APPAD02.DOC May 12, 2008 4 A Retail Survey of Brazilian Milk and Minas Frescal Cheese and a 5 Contaminated Dairy Plant To Establish The Prevalence, 6 Relatedness, and Sources of Listeria monocytogenes† 7 8 D E Running Title: L. MONOCYTOGENES IN MINAS FRESCAL CHEESE FROM BRAZIL 9 10 J. Renaldi F. Brito1,2,3*, Emilia M. P. Santos2,4, Edna F. Arcuri2, Carla C. Lange2, Maria A. 11 V. P. Brito2, Guilherme N. Souza2, Mônica M. P. O. Cerqueira4, J. Marcela Soto Beltran5, 12 Jeffrey E. Call3, Yanhong Liu3, Anna C. S. Porto-Fett3 and John B. Luchansky3 13 14 Brazilian Agricultural Research Corporation (Embrapa), Labex, Beltsville, MD1; Embrapa, 15 Dairy Cattle National Research Center, Juiz de Fora, MG, Brazil2; U.S. Department of T P E C C A 16 Agriculture, ARS, Eastern Regional Research Center, MFSRU, Wyndmoor, PA, 19038, USA3; 17 Federal University of Minas Gerais, Veterinary School, Belo Horizonte, MG, Brazil4;5Centro de 18 Investigación en Alimentación y Desarrolo, Unidad Culiacán, Mexico 19 20 Key words: Listeria monocytogenes, Latin-style fresh cheese, food borne pathogen, dairy plant, 21 Minas Frescal Cheese. 22 23 * Corresponding author. Mailing address: Brazilian Agricultural Research Corporation 24 (Embrapa). Dairy Cattle National Research Center. Rua Eugenio do Nascimento, 610 - Dom 25 Bosco, CEP 36038-330 – Juiz de Fora, MG, Brazil. Phone: +55-32-3249-4703; E-mail: 26 jrfbrito@cnpgl.embrapa.br. 2 Mention of trade names or commercial products in this publication is solely for the purpose of 1 † 2 providing specific information and does not imply recommendation or endorsement by the U. S. 3 Department of Agriculture. D E T P C A E C 3 1 A study was designed to recover Listeria monocytogenes from pasteurized milk and 2 Minas Frescal cheese (MFC) sampled at retail and to identify the contamination source(s) 3 of these products in the corresponding dairy processing plant. Fifty milk samples (9 4 brands) and 55 MFC samples (10 brands) were tested from retail sites located in Juiz de 5 Fora, Minas Gerais, Brazil. All milk samples and 45 samples from 9 of 10 MFC brands 6 tested negative for L. monocytogenes; however, “Brand F” of MFC obtained from retail 7 establishments #119 and #159 tested positive. Thus, the farm/plant that produced Brand F 8 MFC was sampled; all samples from the milking parlor tested negative for L. 9 monocytogenes, whereas several sites within the processing plant and MFC samples tested D E T P E C 10 positive. All 344 isolates recovered from retail MFC, Plant F MFC, and Plant F 11 environmental samples were serotype 1/2a and displayed the same AscI or ApaI 12 fingerprints. Since these results established that the storage coolers served as the 13 contamination source of the MFC, the plant was closed so that corrective renovations could 14 be made. Following renovation, samples from sites that previously tested positive for the 15 pathogen were collected from the processing environment and MFC on multiple visits; all 16 tested negative for L. monocytogenes. In addition, on subsequent visits to retail 17 establishments #159 and #119, all MFC tested negative for the pathogen. Studies are 18 ongoing to quantify the prevalence, levels, and types of L. monocytogenes in MFC and 19 associated processing plants to lessen the likelihood of listeriosis in Brazil. 20 C A 4 1 Latin-style soft cheese, such as Queso Fresco (QF), have been identified in risk assessments by 2 U.S. regulatory agencies as a food of greater concern to public health 3 (www.foodsafety.gov/~dms/lmr2-toc.html) due to listeriosis. Listeria monocytogenes has also 4 been isolated from Latin-style soft cheese from South America, notably from Minas Frescal 5 cheese (MFC) in Brazil. For example, Destro et al. (11) found 2 of 20 samples (10%) of MFC 6 made with pasteurized milk as positive for L. monocytogenes. Silva et al. (38) recovered L. 7 monocytogenes from 7 of 17 samples (41%) of MFC made from raw milk and from 1 of 33 8 samples (3%) of MFC made from pasteurized milk. As another example, Carvalho et al. (6) 9 recovered L. monocytogenes from 3 of 93 samples (3%) of MFC made with pasteurized milk. All D E T P E C 10 positive samples were obtained from MFC manufactured using the traditional method plus the 11 addition of lactic acid. The Brazilian standard of identity of the Common Market of the South 12 (MERCOSUL; 29); defines MFC as a fresh cheese, obtained by the enzymatic coagulation of 13 milk with rennet and other appropriate coagulation enzymes; the use of lactic acid bacteria to 14 complement the coagulation step is optional. The relatively high moisture content (55 to 58%) 15 and pH levels (pH 5.0 to 6.3), low salt content (1.4-1.6%; addition of salt is optional), the 16 absence of defined starter cultures, and the considerable hand manipulation during manufacture 17 by small and very small processors are factors that may all contribute to providing a favorable 18 environment for contamination and survival/growth of L. monocytogenes in MFC (33, 38). 19 However, despite their compositional similarities in moisture, salt content, and pH levels, unlike 20 QF which has caused human illness in the U.S., listeriosis has not been linked to MFC or other 21 dairy products/food in Brazil (17, 18), presumably because it is not listed in the National 22 Notifiable Diseases Surveillance System as a disease of compulsory or immediate notification C A 5 1 and/or because there is not a comprehensive surveillance network in Brazil to survey foods for 2 this bacterium (4). 3 Numerous investigations on the presence, levels, and types of L. monocytogenes in 4 pasteurized milk have been conducted. For example, Frye and Donnelly (13) conducted a survey 5 for L. monocytogenes in pasteurized whole milk, nonfat milk, and chocolate milk produced in the 6 U.S. Of 5,519 samples tested over a five-week period beginning in June of 2000, only 1 of 1,846 7 (0.018%) nonfat milk samples tested positive for this pathogen. Jayarao et al. (19, 20) isolated 8 L. monocytogenes from 6 of 131 (5%) bulk tank milk (BTM) samples in Eastern South Dakota 9 and Western Minnesota, and in a separate survey of foodborne pathogens conducted with 248 D E T P E C 10 dairy herds in 16 counties in Pennsylvania they reported that 3 of 248 (1%) BTM samples tested 11 positive for this pathogen. In contrast, Casarotti et al. (7) did not find any species of Listeria in 12 20 samples each of raw milk, pasteurized milk, or MFC from retail establishments in the 13 municipality of Piracicaba, State of São Paulo, Brazil. Likewise, Nero et al. (31) did not recover 14 L. monocytogenes from single raw milk samples from each of 210 small- to medium-sized dairy 15 farms located in five States representing the three main milk producing regions in Brazil. Several 16 factors, such as the laboratory procedures performed, including their sensitivity and specificity, 17 and the nature and number of samples tested, could, at least in part, explain the differences in 18 prevalence of L. monocytogenes among these various studies. 19 C A L. monocytogenes is also widespread on farms (32, 44) and in dairy processing 20 environments (10, 21, 28, 30, 34, 43). For example, Kabuki et al. (21) examined environmental 21 and cheese samples from three Latin-style fresh cheese processing plants in New York City over 22 a total of 4 visits to each plant. They recovered L. monocytogenes from 27 of 246 (11%) 23 environmental samples from these processing plants. However, only finished products from one 6 1 plant (7 of 24; 29%) tested positive for L. monocytogenes. In a study conducted in two MFC 2 processing plants in the State of Bahia, Brazil, Silva et al. (37) isolated L. monocytogenes in one 3 of the plants from the raw milk (1 of 6 samples positive) and from the floor of the cheese 4 refrigeration room (1 of 2 samples positive) from among a total of 118 samples representing 5 environmental, food contact surfaces, and cheese. It was not possible to isolate the pathogen 6 from among 100 similar samples that were tested in the second plant. More importantly, L. 7 monocytogenes was not recovered from any of the cheese sampled from either of these two 8 facilities. 9 D E T P Based on the association of L. monocytogenes with raw and pasteurized milk, dairy farms, E C 10 cheese processing plants, and MFC and, in turn, on the potential threat of listeriosis, the primary 11 objective of this study was to determine the prevalence and levels of L. monocytogenes at retail 12 in pasteurized milk and MFC manufactured by a number of different processors in Brazil. A 13 secondary objective was to conduct molecular characterization of any isolates retained from this 14 survey to determine if there are persistent types, as well as to identify possible sources of 15 contamination in the corresponding dairy processing plants, so that corrective actions could be 16 taken. C A 7 MATERIALS AND METHODS 1 2 3 Retail sampling. A total of 50 pasteurized milk samples representing 9 different brands 4 (Brands 1-9; 5-7 samples per brand) were obtained from 37 separate retail establishments 5 (Table 1), chosen by stratified random sampling, located in five of eight areas of the city of Juiz 6 de Fora, in the state of Minas Gerais, Brazil. Fifty five samples of Minas Frescal cheese (MFC) 7 representing 10 brands (Brands A-J; 9 brands, 5 samples/brand; Brand F, 10 samples/brand) 8 were obtained from 24 of these same 37 retail establishments (Table 1). Both the milk and 9 cheese samples were obtained at retail between June and October of 2005 (dry/cold season). In D E T P E C 10 addition, MFC was also obtained on 2 visits to retail establishment #159 (1 sample per visit) and 11 4 visits to retail establishment #119 (6 samples total) between January and March 2006. Samples 12 were placed on ice and transported in cooler boxes to the Embrapa Dairy Cattle Microbiology 13 Laboratory (Juiz de Fora) for microbiological analyses where they were stored at 4ºC and 14 analyzed within 24 h of collection essentially as described by Marshall (26). 15 C A Dairy processing plant sampling. Plant F was surveyed for the pathogen during October of 16 2005 (Figure 1). Five of the 33 samples were from MFC and the remaining 28 were from raw 17 milk, pasteurized milk, and the plant environment. Following corrective renovations of Plant F, 18 an additional 126 environmental samples (29 per each of three visits in December of 2005, 19 March of 2006, and June of 2006, and 39 samples in October of 2006) were examined. An 20 additional 5 MFC samples (15 total MFC samples) obtained directly at Plant F on each visit in 21 March, June, and October of 2006 were also examined for the presence of the pathogen. 22 Manufacture of Minas Frescal cheese. In October of 2005, Plant F processed 23 approximately 400 liters of milk each day from a herd owned by the cheese manufacturer that 8 1 was located immediately adjacent to the plant. The herd contained 25 lactating crossbred cows 2 (Holstein X Zebu cattle). Milk was transferred in eight sanitized stainless steel vessels at ambient 3 temperature to the dairy plant immediately after milking that occurred twice each day. The MFC 4 was manufactured essentially as described by Silva et al. (37) with only slight modifications. 5 Briefly, milk was pasteurized (63oC, 30 minutes) in a double-walled tank and left to cool to 32oC 6 in the same tank (Figure 1A). Calcium chloride (20 g/100 liters of milk; Produtos Macalé Ltda., 7 Juiz de Fora, MG) was added, followed by addition of rennet (1%; Chymax, Chr. Hansen Ind. e 8 Com. Ltda, Valinhos, SP); the mixture was held at 35oC for 40 minutes to promote curd 9 formation. The curd was cut gently into small cubes with the aid of a cutter composed of a wire D E T P E C 10 string on a steel frame, before the whey was drained using a strainer made of plastic. The 11 resulting curd was placed into perforated circular cylindrical plastic molds (10 or 15 cm in 12 diameter; Produtos Macalé Ltda.) without the use of cheese cloth. The cheese was surface-salted 13 twice on each exposed side/face of the MFC (approximately 0.75% w/w NaCl on each side after 14 reversing/inverting the containers every 15 minutes each of two times). Next, and without 15 pressing, the molds containing curd were stored on trays that were placed inside three chest-type 16 coolers/refrigeration units that were covered and maintained at approximately 4oC for 18 to 24 h 17 to increase the firmness of the curd/cheese. The molds were removed on day 2 (Figure 1B) from 18 the coolers and placed onto the same sink where on day 1 the cheese was hooped, so that the 19 cheese could be removed from the molds and the excess whey could be drained. Lastly, the 20 cheese was placed into 18 cm ´ 24 cm (ca. 500 g MFC) or 20 cm x 30 cm (ca. 1,000 g MFC) 21 plastic bags that were sealed with a twist tie and stored in the chest coolers for up to 24 h or 22 transported directly to the marketplace for sale. C A 9 1 Isolation and identification of L. monocytogenes. Pasteurized milk (1,000 ml) and cheese 2 (500 g) samples were collected as finished packaged products at the above mentioned retail 3 establishments. Swabs (CB Products Ind. & Com. Ltd., São Paulo, SP, Brazil) were used to 4 sample teat cups and buckets from the milking parlor and associated equipment, as well as from 5 the floor, wall, sink, coolers, milk pasteurization vat, cheese molds and trays that had direct or 6 indirect contact with MFC, essentially as described (16). Two swabs were used to sample a 50- 7 cm2 area and this process was repeated five times by passing the same two swabs over the same 8 surface area of each sample. The swabs were placed in sterile tubes containing 10 ml of Brain 9 Heart Infusion broth (BHI; Difco Laboratories, Sparks, MD) for transport to the laboratory on D E T P E C 10 ice. Five ml of BHI were transferred to 45 ml of Listeria Enchrichment Broth (LEB; Difco) and 11 the contents were mixed at room temperature for 1 minute using a stomacher (model MK 1204, 12 ITR Instrumentos para Laboratórios Ltda., Esteio, RS, Brazil). Each sample was incubated at 13 30°C for 48 h. For MFC (25 g) and for test samples that were liquids, such as milk, whey, water 14 from the sinks and rinse water from the buckets used to transport milk (25 ml of each liquid 15 sample), the samples were aseptically added to 225 ml of (LEB) and mixed as described above 16 and incubated at 30°C for 48 h (detection limit = ≤1 CFU per sample). 17 C A After incubation, a loopful of each LEB enrichment sample was transferred to Palcam 18 selective agar (Difco) and to modified Oxford agar (MOX; Oxoid Ltd., Basingstoke, Hampshire, 19 UK) plates. The plates were incubated at 35°C and examined for the presence of Listeria-like 20 colonies after 24 and 48 h. From each selective agar plate, five presumptive Listeria colonies 21 were streaked onto trypticase soy agar (TSA, Acumedia®, Baltimore, MD) with 0.6% yeast 22 extract (Oxoid). The plates were incubated for 48 h at 30°C. Select isolates were examined for 23 Gram’s reaction, catalase production, nitrate reduction, and motility on SIM medium, as well as 10 1 for hemolysis and the CAMP reaction on 5% sheep blood agar (1). Isolates were further 2 confirmed as L. monocytogenes via PCR using primers targeting hlyA and genes 1 and 2 (2, 22). 3 Enumeration of L. monocytogenes from MFC. Levels of L. monocytogenes in all five 4 cheese samples obtained directly from Plant F in October 2005 were determined as follows: a 5 single 25 g portion was removed and transferred to a bag containing 225 ml of 2% sodium citrate 6 solution. Following homogenization for 2 minutes at room temperature in a stomacher, each 7 sample was serially diluted in 0.1% sterile peptone water and a total of 333 ml of the diluted 8 sample was spread plated onto each of three Oxford agar plates. The plates were incubated for 9 48 h at 35°C (detection limit = ≤10 CFU per 25 g of MFC). D E T P E C 10 Pulsed-field gel electrophoresis. Molecular subtyping was conducted using AscI and ApaI 11 (New England Biolabs Inc., Ipswich, MA) in accordance with the standardized CDC Pulse-Net 12 protocol (15), essentially as described by Gilbreth et al. (14). The AscI-digested DNA from a 13 laboratory control strain of L. monocytogenes (MFS1435; pulsotype 84, serotype 1/2a) was 14 included as a reference on all gels. Pattern images were acquired using a Bio-Rad Gel Doc 15 system with the Multi Analyst software program (Bio-Rad; v. 1.1; Hercules, CA) and compared 16 using the Applied Maths BioNumerics (version 4.0, Saint-Martins-Latem, Belgium) software 17 package. Pattern clustering was performed using algorithms within BioNumerics, specifically the 18 unweighted pair group method using arithmetic averages (UPGMA) and the Dice correlation 19 coefficient with a position tolerance of 1.0%. 20 C A Serotyping. Isolates were serotyped using both a multiplex PCR assay (12) designed to 21 identify serotype 4b, 1/2a, and 1/2b strains and a commercially-available antibody-based 22 serotyping kit (Denka Seiken Co., Tokyo, Japan). Strains MFS-53 (F2365; serotype 4b; Jalisco 23 cheese; 27), MFS-96 (H7858; serotype 4b; frankfurter; 9), MFS-98 (H7969; serotype 4b; 11 1 frankfurter; 9), MFS-110 (F6854; serotype 1/2a; turkey frankfurter; 8), and MFS-108 (ATCC 2 19116; serotype 4c; chicken; 3) were used as controls. D E T P C A E C 12 RESULTS 1 2 3 Our survey of 37 retail establishments in Juiz de Fora revealed that all 50 pasteurized milk 4 samples and cheese from 9 of 10 MFC brands tested negative by enrichment (≤1 CFU per 5 sample) for L. monocytogenes. However, between June and October of 2005, 60% (6 of 10 6 samples) of MFC from one brand, that being Brand F, tested positive. More specifically, 4 of 6 7 samples from retail establishment #119 and 2 of 4 samples from retail establishment #159 tested 8 positive. This corresponds to a brand prevalence of 10% (1 of 10 brands) and a MFC prevalence 9 of 11% (6 of 55 total MFC samples). Proximate composition analyses of the commercial MFC 10 showed on average a relatively high moisture content of 56.4% (54.9 to 58%) and pH levels of 11 6.09 (pH 4.91 to 6.76), a low salt content of 0.85% (0.41 to1.14%), a total solids content of 12 41.1% (35.4 to 45.1%), and lipids content of 19% (17.0 to 22.5%) as is typical/expected for this 13 type of cheese. 14 D E T P E C C A After finding that only one brand of cheese was contaminated with L. monocytogenes, our 15 team in Juiz de Fora worked through proper channels to share our findings with the owner of 16 Plant F and with authorities at the Minas Gerais State Service for Veterinary and Phytosanitary 17 Supervision (Instituto Mineiro de Agropecuária (IMA) / Inspeção e Fiscalização de Produtos de 18 Origem Animal). To their credit, the producer was keenly interested in working with us and with 19 the local authorities to identify the source(s) of contamination and make the necessary 20 modifications to correct the problem. During the initial visit to the Plant F farm/dairy in October 21 of 2005 (Fig. 1), all five samples taken from the milking parlor (bulk tank milk, teat cups, filters, 22 milk buckets, and rinse water obtained from buckets used to carry milk) tested negative by 23 enrichment for L. monocytogenes. However, all five samples of MFC obtained directly at Plant F 13 1 tested positive for the pathogen. L. monocytogenes were present at levels ranging from 2.48 to 2 4.28 log10 CFU/g (average = 3.64 log10 CFU/g) in these 5 samples. Perhaps more importantly, 10 3 of the 23 environmental samples taken from food contact and non-food contact areas in the 4 processing plant (floors, walls, sinks, refrigeration units, cheese molds, and liquid on trays, in 5 sinks, and in refrigerators) tested positive (Fig. 1). Therefore, we studied both the plant design 6 and cheese making protocol to narrow down the most likely source(s) of contamination. 7 Regarding the latter, as shown in Fig. 1, all samples which tested positive for the pathogen were 8 sites related to day 2 of manufacturing. Areas associated with contamination were the 9 coolers/refrigeration units wherein the cheese was stored overnight and sites that came in contact 10 with the cheese after it was taken from the coolers/refrigeration units, such as the molds and sink 11 area where the cheese was removed from the molds on day 2 of production. These data 12 suggested that the excess fluid/whey within the coolers served as a point source of 13 contamination. This hypothesis was further corroborated by the results obtained by PFGE 14 fingerprinting of the retained isolates from each positive sample (see below). 15 D E T P E C C A Plant F remained closed from October through December of 2005, and no cheese was 16 produced or sold, so that renovations could be made to this facility. As shown in Fig. 2, in 17 general, most of the changes made were neither extensive nor costly. For example, broken floor 18 and wall tiles were replaced, foot baths and sinks were installed both within and outside of the 19 plant (Fig. 2a), windows and doors were replaced and/or properly sealed, and new sinks (Fig. 2b) 20 and coolers/refrigeration units (Fig. 2c) were purchased and installed. In total, these 21 modifications/renovations cost the owners of Plant F less than $5000.00 USD. This is arguably 22 only a modest amount of monies, considering it eliminated the source/niche of L. monocytogenes 14 1 contamination and ultimately increased the amount of product made and profitability for the 2 producer, as well as enhanced the safety and wholesomeness of MFC produced at Plant F. 3 After the renovations were completed, 29 sites in the plant environment were tested in 4 December of 2005 and found to be negative for L. monocytogenes. Based on our data showing 5 the absence of L. monocytogenes in the plant environment, the local authorities gave approval for 6 the plant to reopen. For the purposes of this research project, these same 29 sites were tested in 7 March and June of 2006 and an additional 10 sites, for a total of 39 sites, were tested in October 8 of 2006; all 97 samples tested negative for the pathogen. In addition, 5 MFC samples were 9 obtained directly from Plant F during the visits in March, June, and October of 2006; all 15 D E T P E C 10 samples tested negative for the pathogen. Furthermore, between January and March of 2006 we 11 returned to retail establishments #119 (6 samples over 4 visits) and #159 (1 sample on each of 2 12 visits) and purchased MFC produced by Plant F; all 8 samples tested negative for the pathogen. 13 C A A total of 344 isolates of L. monocytogenes (5 to 20 isolates per positive sample) from retail 14 cheese (6 samples; 88 isolates), cheese obtained directly from Plant F (5 samples; 97 isolates), 15 and environmental sites within Plant F (10 samples; 159 isolates) were retained and further 16 characterized. All 344 L. monocytogenes isolates displayed the same AscI (and ApaI) restriction 17 endonuclease profile (REDP) which was designated as REDP 9 (Fig. 3). A subset of 24 isolates 18 representing each sample type testing positive for the pathogen were confirmed as being serotype 19 1/2a by both PCR and serological assays. As an aside, four representative serotype 1/2a isolates 20 displaying REDP 9 were confirmed as DUP-1046 via automated ribotyping with EcoRI 21 (RiboPrinter® Microbial Characterization System, DuPont, Qualicon, Wilmington, DE; data not 22 shown). 23 15 1 DISCUSSION 2 There is relatively little information about food borne diseases in Brazil. The most recent official 3 data (4) refers to the period from 1999-2004 wherein the etiology was not established in 2,989 of 4 3,737 (80%) of the reported outbreaks; L. monocytogenes was not included in these statistics. 5 Although L. monocytogenes is widespread in nature, as well as in the dairy environment, our 6 results and those of other investigators confirm that pasteurized milk is not a common source of 7 this pathogen in Brazil (7, 37). However, our data also confirmed that MFC can be a source of L. 8 monocytogenes to consumers even when it is prepared with pasteurized milk (6, 11, 38). The 9 observed brand prevalence of 10% and MFC prevalence of 11% pertain specifically to the Juiz D E T P E C 10 de Fora region of Brazil, and may or may not be reflective of the “true” prevalence of L. 11 monocytogenes in QF in other regions of Brazil or in other areas of the world. Our data 12 established Plant F as the origin of contamination, but it did not totally exclude the possibility 13 that MFC could be also contaminated at retail. Regardless, the fact that all 344 isolates displayed 14 an indistinguishable PFGE pattern indicated a point source of contamination. Moreover, this 15 specific pulsed field profile, namely REDP 9, is unique among some 700 strains and some 70 16 REDP in our database. It is also of interest that strains displaying REDP 9 were subsequently 17 ribotyped as DUP-1046. Sauders et al. (35) recovered DUP-1046 strains from RTE whitefish 18 salad in New York State and Manfreda et al. (25) recovered DUP-1046 strains from Gorgonzola 19 cheese in Lombardia, Italy. Although such strains have been associated with foods, it does not 20 appear that they have been responsible for listeriosis (36). Thus, it would be interesting to further 21 characterize such strains to determine their virulence potential and/or response to food 22 processing and storage conditions relative to their ability to persist in foods and cause human 23 illness. In addition, all isolates belonged to the same serotype, that being 1/2a, which is the C A 16 1 second most frequent serotype recovered from clinical listeriosis in humans, and the most 2 frequent serotype found in dairy products in Brazil (17, 18). Although no illnesses were linked to 3 the MFC produced by Plant F, with the possible exception of an outbreak of febrile 4 gastroenteritis in Sweden due to consumption of fresh cheese contaminated with a serotype 1/2a 5 strain (8), previous listeriosis outbreaks linked to Latin-style soft cheese have been caused by 6 serotype 4b strains of L. monocytogenes 7 D E Our study and others (28, 30, 44) show the importance of finding specific niches wherein L. T P 8 monocytogenes can survive in the food-processing environment to more specifically direct the 9 cleaning, disinfection, and renovation/repair efforts to these sites. No L. monocytogenes were 10 detected in raw milk, pasteurized milk, curd, or whey samples collected before the storage of 11 cheese in the coolers/refrigeration units. This indicates that raw milk was not the source of the 12 pathogen at the time of sampling, thus further confirming the low prevalence or absence of L. 13 monocytogenes in raw milk from Brazilian herds (7, 31, 37). However, even if L. 14 monocytogenes were present in raw milk, pasteurization would eliminate it. Likewise, Menendez 15 et al. (28) recovered L. monocytogenes from 1 of 20 raw milk samples, but not from 19 samples 16 of pasteurized milk, curd and whey. In this same study, L. monocytogenes was recovered from 8 17 of 31 samples obtained from several sites in the manufacturing area and in the ripening rooms 18 (28), indicating the possibility of post-pasteurization contamination. The introduction of 19 L. monocytogenes into cheese via contaminated raw milk was also considered as the cause of 20 contamination of fresh Mexican style cheese in California that resulted in 142 cases of human 21 listeriosis and 48 deaths (23), as well as another outbreak in Winston-Salem, North Carolina, in 22 2000 that resulted in 13 cases and 5 deaths (24). E C C A 17 1 The association of the production plant environment and equipment as a primary source of 2 L. monocytogenes contamination in foods is common (21, 34). Moreover, inoculation of milk 3 with L. monocytogenes confirmed that the pathogen survived the making of Mexican-style 4 cheese (40). Suggestions of post-processing contamination of MFC with L. monocytogenes have 5 also been made (38, 39). These data indicate that L. monocytogenes contamination may occur 6 both during and after cheese processing. In our case, contamination occurred during the cold 7 storage of MFC. This assumption was based on the recovery of the pathogen only after the 8 processed cheese was kept in the coolers/refrigeration units overnight. Moreover, all isolates of 9 L. monocytogenes recovered from MFC obtained from both retail and directly from Plant F D E T P E C 10 displayed the same PFGE profile as the isolates recovered from coolers/refrigeration units, the 11 sink, cheese molds, and excess liquid found on trays. One could also surmise that the persistence 12 and proliferation of the pathogen was facilitated by the cold storage conditions. 13 C A Our approach to control L. monocytogenes in the dairy plant was similar to that described by 14 others (28, 30, 42) in the sense that improved sanitary practices, and to some extent structural 15 changes to the plant, were made to eliminate any source of contamination or non-hygienic 16 handling during manufacturing, or storage. In our study, contamination was most likely confined 17 to the storage environment, since replacement of two of the three refrigeration units eliminated 18 the pathogen from within Plant F. Whereas we eliminated L. monocytogenes from the plant, 19 Unnerstad et al. (42) reported a decrease in the number of positive samples but not the complete 20 elimination of the problem. Based on the continuing recovery of strains displaying the same 21 serotype, that being serotype 3b, from both cheese and the dairy plant environment, they 22 suggested that some clones of L. monocytogenes survived in this environment for at least seven 23 years. In another study designed to characterize L. monocytogenes recovered from an ice-cream 18 1 plant, it was found that a predominate PFGE type persisted in this plant for seven years (30). In 2 our case, we were not able to recover L. monocytogenes from the finished product or from the 3 plant environment for 12 months following the interventions that were made. Either our strain 4 was transient in the dairy environment and/or the source of contamination was eliminated via 5 renovation and by improving hygienic measures. Our data also indicate that the point source of 6 contamination was the refrigeration units where cheese and other products were stored during 7 and after processing. Cotton & White (10) reported that 12 of 16 samples from fluid milk plants 8 that were positive for L. monocytogenes were obtained from the cooler, and they and others 9 reported that recovery of this pathogen from cheese kept at refrigerated temperatures is not D E T P E C 10 uncommon (41, 43). Success at controlling L. monocytogenes in dairy plants may vary, since it 11 depends on a multiplicity of factors as discussed herein and by others (21, 43). 12 Our results validated that systematic culturing and analyses of products and processing C A 13 facilities can identify areas that harbor L. monocytogenes. In addition, we also demonstrated that 14 it is possible, with appropriate interventions, to eliminate harborage points within a food 15 processing facility. Studies are ongoing to quantify the prevalence, levels, and types of L. 16 monocytogenes on dairy farms and in dairy processing plants to better manage the threat of 17 listeriosis in Latin-style fresh cheese. 19 ACKNOWLEDGMENTS 1 2 3 We gratefully acknowledge the contributions of the following individuals and organizations 4 who contributed significantly to the initiation or conduct of the project per se and/or the analyses 5 of the resulting data: Rosemary Martinjuk, Brad Shoyer, Ellen Sanders, and John P. Cherry 6 (USDA/ARS/ERRC; Wyndmoor, PA), Cristobal Chiadez and Jorge Sillier (CIAD; Culiacan, 7 Mexico), Carlos Rodriguez, Airdem Assis, Pedro Arrares, and Gretchen Flanley 8 (USDA/ARS/OIRP, Beltsville, MD), Paulo do Carmo Martins, Pedro Braga Arcuri, Marcos A. 9 S. Silva, Abiah N. I. Abreu, Alessandra P. Sant’Anna, Carolina F. S. Vasconcellos, and Gilmara 10 B. de Paula (EMBRAPA, Dairy Cattle National Research Center, Juiz de Fora, MG, Brazil), the 11 Minas Gerais State Service for Veterinary and Phytosanitary Supervision (Instituto Mineiro de 12 Agropecuária Inspeção e Fiscalização de Produtos de Origem Animal IMA-MG), and the 13 Brazilian Council for Technological and Scientific Development (CNPq) (Emilia M. P. Santos is 14 recipient of CNPq scholarship). This work was financially supported by PRODETAB (047- 15 02/99) and FAPEMIG (2411/05). D E T P E C C A 20 REFERENCES 1 2 3 1. Bille, J., J. Rocourt, and B. Swaminathan. 2003. Listeria and Erysipelothrix, p. 461-471. 4 In P. R. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual 5 of clinical microbiology, 8th ed. ASM Press, Washington, DC. 6 2. D E Border, P. M., J. J. Howard, G. S. Plastow, K. W. Siggens. 1990. Detection of Listeria 7 species and Listeria monocytogenes using polymerase chain reaction. Lett. Appl. Microbiol. 8 11:158-162. 9 3. T P Brosch, R., J. Chen, and J. B. Luchansky. 1994. Pulsed-field fingerprinting of listeriae: E C 10 identification of genomic divisions for Listeria monocytogenes and their correlation with 11 serovar. Appl. Environ. Microbiol. 60:2584-2592. 12 4. C A Berto, R. M. S. Alves, and E. H. Carmo. 2005. Vigilância epidemiológica das doenças 13 transmitidas por alimentos no Brasil, 1999-2004. Boletim Eletrônico Epidemiológico. Ano 14 5, no. 6, 7 p. 15 16 Carmo, G. M. I., A. A. Oliveira, C. P. Dimech, D. A. Santos, M. G. Almeida, L. H. 5. Carrique-Mas, J. J., I. Hokeberg, Y. Andersson, M. Arneborn, W. Tham, M.L. Danielsson-Tham, B. Osterman, M. Leffler, M. Steen, E. Eriksson, G. Hedin, and J. 17 18 Giesecke. 2003. Febrile gastroenteritis after eating on-farm manufactured fresh cheese – an 19 outbreak of listeriosis? Epidem. Infection. 130:79-86. 20 21 6. Carvalho, J. D. G., W. H. Viotto, and A. Y. Kuaye. 2007. The quality of Minas frescal cheese produced by different technological processes. Food Control 18:262-267. 21 1 7. Casarotti, V. T., C. R. Gallo, and R. Camargo. 1994. Ocorrência de Listeria 2 monocytogenes em leite cru, leite pasteurizado tipo C e queijo minas frescal comercializados 3 em Piracicaba, SP. Arch. Latin. Nutr. 44:158-163. 4 8. Centers for Disease Control and Prevention (CDC). 1989. Epidemiological notes and D E 5 reports listeriosis associated with consumption of turkey franks. Morb. Mortal. Wkly. Rep. 6 38:267-268. 7 8 9 10 11 12 13 14 15 16 17 18 19 9. Centers for Disease Control and Prevention (CDC). 1998. Multi-State outbreak of T P listeriosis – United States. Morb. Mortal. Wkly. Rep. 47:1085-1086. 10. Cotton, L. N. and C. H. White. 1992. Listeria monocytogenes, Yersinia enterocolitica, and E C Salmonella in dairy plant environments. J. Dairy Sci. 75:51-57. 11. Destro, M. T., A. M. Serrano, and D. Y. Kabuki. 1991. Isolation of Listeria species from some Brazilian meat and dairy products. Food Control 2:110-112. C A 12. Doumith, M., C. Buchrieser, P. Glaser, C. Jacquet, and P. Martin. 2004. Differentiation of the major Listeria monocytogenes serovars by multiplex PCR. J. Clin. Microbiol. 42:3819-3822. 13. Frye, C. and C. W. Donnelly. 2005. Comprehensive survey of pasteurized fluid milk produced in the United States reveals a low prevalence of Listeria monocytogenes. J. Food Prot. 68:973-979. 14. Gilbreth, S. E., J. E. Call, F. M. Wallace, V. N. Scott, Y. Chen, and J. B. Luchansky. 20 2005. Relatedness of Listeria monocytogenes isolates recovered from selected ready-to-eat 21 foods and listeriosis patients in the United States. Appl. Environ. Microbiol. 71:8115–8122. 22 1 15. Graves, L.M. and B. Swaminathan. 2001. PulseNet standardized protocol for subtyping 2 Listeria monocytogenes by macrorestriction and pulsed-field gel electrophoresis. Int. J. Food 3 Microbiol. 65:55-62. 4 16. Hickey, P. J., C.E. Beckelheimer, and T. Parrow. 1992. Microbiological tests for 5 equipment, containers, water, and air. In R. T. Marshal, (ed.), Standard methods for 6 examination of dairy products, 16th ed. APHA, Washington. 7 D E 17. Hofer, E., C. M. F. Reis, and C. B. Hoffer. 2006. Serovars of Listeria monocytogenes and T P 8 related species isolated from human clinical specimens. Rev. Soc. Brasil. Med. Trop. 9 39:32-37. 10 E C 18. Hofer, E., R. Ribeiro, and D. P. Feitosa. 2000. Species and serovars of the genus Listeria 11 isolated from different sources in Brazil from 1971 to 1997. Mem. Instit. Oswaldo Cruz. 12 95:615-620. 13 14 15 16 17 C A 19. Jayarao, B. M. and D. R. Henning. 2001. Prevalence of foodborne pathogens in bulk tank milk. J. Dairy Sci. 84:2157-2162. 20. Jayarao, B. M., S. C. Donaldson, B. A. Straley, A. A. Sawant, N. V. Hegde, and J. L. Brown. 2006. A survey of foodborne pathogens in bulk tank milk and raw milk consumption among farm families in Pennsylvania. J. Dairy Sci. 89:2451-2458. 18 21. Kabuki, D. Y., A. Y. Kuaye, M. Wiedmann, and K. J. Boor. 2004. Molecular subtyping 19 and tracking of Listeria monocytogenes in Latin-style fresh-cheese processing plants. 20 J. Dairy Sci. 87:2803-2812. 21 22. Lange, C., N. D. Peres, E. F. Arcuri, M. A. V. P. Brito, J. R. F. Brito, P. M. Garcia, and 22 M. M. O. P. Cerqueira. 2005. Identificação de Listeria monocytogenes pela reação em 23 cadeia da polimerase. Rev. Instit Lat Cândido Tostes. 60:150-153. 23 1 23. Linnan, M. J., L. Mascola, X. D. Lou, V. Goulet, S. May, C. Sainen, D. W. Hird, M. L. 2 Yonekura, P. Hayes, R. Weaver, A. Audurier, B. D. Plikaytis, S. L. Fannin, A. Kleks, 3 and C. V. Broome. 1988. Epidemic listeriosis associated with Mexican-style cheese. N. 4 Engl. J. Med. 319:823-828. 5 D E 24. MacDonald, P. D. M., R. E. Whitwam, J. D. Boggs, J. N. MacCormack, K. L. 6 Anderson, J. W. Reardon, J. R. Saah, L. M. Graves, S. B. Hunter, and J. Sobel. 2005. 7 Outbreak of listeriosis among Mexican immigrants as a result of consumption of illicitly 8 produced Mexican-style cheese. Clin. Infect. Dis. 40:677-82. 9 T P 25. Manfreda, G., A. De Cesare, S. Stella, M. Cozzi, and C. Cantoni. 2005. Occurrence and E C 10 ribotypes of Listeria monocytogenes in Gorgonzola cheeses. Int. J. Food Microbiol. 11 102:287-293. 12 13 14 15 16 17 18 19 20 21 26. Marshall, R. T. (Ed.). 1992. Standard methods for the examination of dairy products. 16. C A ed. Washington, DC: American Public Health Association, 546 p. 27. Mascola, L., Lieb, L., Chiu, J., Fannin, S.L., and Linnan, M.J. 1988. Listeriosis: and uncommon opportunistic infection in patients with acquired immunodeficiency syndrome. A report of five cases and a review of the literature. Am. J. Med. 84:162-164. 28. Menendez, S., M. R. Godinez, J. L. Rodrigues-Otero, and J. A. Centeno. 1997. Removal of Listeria spp. in a cheese factory. J. Food Safety 17:133-139. 29. Mercosul. 1996. Regulamento Técnico Mercosul de Identidade e Qualidade do Queijo Minas Frescal. Resolução Mercosul / GMC / Res. Nº 145/96. 30. Miettinen, M. K., K. J. Björkroth, and H. J. Korkeala. 1999. Characterization of Listeria 22 monocytogenes from an ice cream plant by serotyping and pulsed-field gel electrophoresis. 23 Int. J. Food Microbiol. 46:187-192. 24 1 31. Nero, L. A., M. R. de Mattos, V. Beloti, M. A. F. Barros, D. P. Netto, J. P. A. N. Pinto, 2 N. J. de Andrade, W. P. Silva and B. D. G. M. Franco. 2004. Hazards in non-pasteurized 3 milk on retail sale in Brazil: prevalence of Salmonella spp., Listeria monocytogenes and 4 chemical residues. Brazil. J. Microbiol. 35:211-215. D E 5 32. Nightingale, K. K., Y. H. Schukken, C. R. Nightingale, E. D. Fortes, A. J. Ho, Z. Her, 6 Y. T. Grohn, P. L. McDonough, and M. Wiedmann. 2004. Ecology and transmission of 7 Listeria monocytogenes infecting ruminants and in the farm environment. Appl. Environ. 8 Microbiol. 70:4458-4467. 9 T P 33. Pereira, M. M. G., M. T. Lima, and M. F. S Santana. 2006. Queijo Minas Frescal. E C 10 Universidade Federal do Piauí, Centro de Ciências Agrárias, Teresina, Piauí, Brasil. 11 Comunicado Técnico, n. 12, p. 1-4. 12 13 14 15 16 17 34. Pritchard, T. J., K. J. Flanders, and C. W. Donnelly. 1995. Comparison of the incidence C A of Listeria on equipment versus environmental sites within dairy processing plants. Int. J. Food Microbiol. 26:375-384. 35. Sauders, B. D., K. Mangione, C. Vicent, J. Schermerhorn, C. M. Farchione, N. B. Dumas, D. Bopp, L. Kornstein, E. D. Fortes, K. Windham, and M. Wiedmenn. 2004. Distribution of Listeria monocytogenes molecular subtypes among human and foods isolates 18 in New York state shows persistence of human disease-associated Listeria monocytogenes 19 strains in retail environments. J. Food Prot. 67:1417-1428. 20 36. Sauders, B. D., Y Schukken, L. Kornstein, V. Reddy, T. Bannerman, E. Salehi, N. 21 Dummas, B. J. Anderson, J. P. Massey, and M. Wiedmenn. 2006. Molecular 22 epidemiology and cluster analysis of human listeriosis cases in three U. S. States. J. Food 23 Prot. 69:1680-1689. 25 1 37. Silva, I. M. M., R. C. C. Almeida, M. A. O. Alves, and P. F. Almeida. 2003. Occurrence 2 of Listeria spp. in critical control points and the environment of Minas Frescal cheese 3 processing. Int. J. Food Microbiol. 81:241-248. 4 5 38. Silva, M. C. D., E. Hofer, and A. Tibana. 1998. Incidence of Listeria monocytogenes in D E cheese produced in Rio de Janeiro, Brazil. J. Food Prot. 61:354-356. 6 39. Silva, M. C. D., M. T. Destro, E. Hofer, and A. Tibana. 2001. Characterization and 7 evaluation of some virulence markers of Listeria monocytogenes strains isolated from 8 Brazilian cheeses using molecular, biochemical and serotyping techniques. Int. J. Food 9 Microbiol. 63:275-280. 10 T P E C 40. Solano-López, C. and H. Hernández-Sánchez. 2000. Behaviour of Listeria 11 monocytogenes during the manufacture and ripening of Manchego and Chihuahua Mexican 12 cheeses. Int. J. Food Microbiol. 62:149-153. C A 13 41. Uhlich, G. A., J. B. Luchansky, M. L. Tamplin, F. J. Molina-Corral, S. Anandan, and 14 A. C. S. Porto-Fett. 2006. Effect of temperature on viability of Listeria monocytogenes in 15 16 17 18 19 Queso Blanco. J. Food Safety 26:202-214. 42. Unnerstad, H., E. Bannerman, J. Bille, M. L. Danielsson-Tham, E. Waak, and W. Tham. 1996. Prolonged contamination of a dairy with Listeria monocytogenes. Neth. Milk Dairy J. 50:493-499. 43. Walker, R. L., L. H. Jensen, H. Kinde, A. V. Alexander, and L. S. Owens. 1991. 20 Environmental survey for Listeria species in frozen milk product plants in California. 21 J. Food Prot. 54:178-182. 22 23 44. Yoshida, T., Y. Kato, M. Sato, and K. Hirai. 1998. Sources and routes of contamination of raw milk with Listeria monocytogenes and its control. J. Vet. Med. Sci. 60:1165-1168. 26 1 TABLE 1. Retail establishments (RE) in Juiz de Fora from where pasteurized milk and Minas 2 Frescal cheese (MFC) samples were obtained (June-September, 2005). Region of Juiz de Total number of Fora RE per region 1) Central (Mid) 1) Central (North) 38 21 Milk samplesa/ RE number Cheese samples/RE number 7, 17, 90, 114, 115, 128, 7, 17, 90, 114, 128, 144, 144, 169 (n=8) 169 (n=7) 95, 142, 151, 154, 160, 142, 151, 154, 172 (n=4) D E T P 170, 171, 172, 173 (n=9) 1) Central (South) 24 45, 49, 147, 164, 165, 166, E C 167 (n=7) 2) North 3) Northeast C A 4) Northwest 45, 49, 147, 164 (n=4) 0 0 0 3 0 19 65, 158, 159, 161, 162, 163 65, 158, 159b, 162 (n=4) 0 (n=6) 5) West 11 32, 155, 156 (n=3) 155, 156 (n=2) 6) Southeast 0 0 0 7) South 12 119, 168 (n=2) 119b (1) 8) East 22 118, 157 (n=2) 118, 157 (n=2) Total 150 37 24 3 a 4 b 13 retail establishments were unique for milk samples. RE testing positive for L. monocytogenes in MFC. 27 FIGURE LEGENDS 1 2 FIG. 1. Diagram of Plant F cheese processing plant. Panel A: Arrows indicate the flow of 3 milk from the milking parlor, to the cheese vat to form the curd, to the sink/table to place the 4 curd in the molds, and to the coolers/refrigeration units for overnight storage on day 1 of 5 production. Panel B: Thin arrows show the flow of cheese from the coolers/refrigeration units, to 6 the sink/table for removal from the molds and subsequent packaging on day 2 of production. 7 Thick arrows show the flow of cheese after packaging either back to the coolers/refrigeration 8 units for storage or for shipment from the plant to retail establishments. 9 T P E C FIG. 2. Photographs of Plant F before and after renovations. C A 10 Panel A: Main entrance. 11 Panel B: Cheese processing area. 12 Panel C: Coolers used for storage of MFC. 13 D E FIG. 3. Schematic representation of AscI and ApaI-generated PFGE restriction patterns for 14 L. monocytogenes strains isolated from a sample of MFC and from the coolers/refrigeration unit 15 L. monocytogenes MFS1435 was used as the reference standard (Lm standard). 28 11-workbench 12-floor and wall 13-tap water 14-curd filter 15-hands of worker 16-hand wash bucket 17-curd bucket Cooler-A Cooler-B T P D E 10-cheese vat E C To Coolers A, B, C C A Cooler-C 6-raw milk 7-tap water 8-milk can 9-pasteurizer 1-teat cups 2-milk filters 3-rinse water 4-milk can 5-milk can 29 22-whey from cheese tray in cooler 23-condensate from cooler Cooler-A 19-cooler Cooler-B 20-cooler 25-whey from bench 26-whey from sink 27-cheese mold 28-cheese mold T P D E E C MFC from Plant F shipped to retail 29-cheese #1 30-cheese #2 31-cheese #3 32-cheese #4 33-cheese #5 C A 21-cooler Cooler-C To Retail and Coolers 30 Main entrance before renovation Fig 2a E C C A D E T P Main entrance after renovation 31 Fig 2b Cheese processing area before renovation Cheese processing area after renovation C A E C T P D E 32 Fig 2c Top-down view into cooler containing Minas frescal cheese on tray New storage coolers purchased after renovation C A E C T P D E 33 Fig. 3 D E T P C A E C