Tracing the Contamination Origin of Coliform Bacteria in Two Small Food-Processing Factories

1910 Journal of Food Protection, Vol. 71, No. 9, 2008, Pages 1910 1914 Copyright, International Association for Food Protection Research Note Tracing the Contamination Origin of Coliform Bacteria in Two Small Food-Processing Factories TATSUYA TOMINAGA, 1 * MASAHIRO SEKINE, 1 AND HIROSHI OYAIZU 2 1 Saitama Industrial Technology Center North Institute, 2-133 Suehiro, Kumagaya-shi, Saitama 360-0031, Japan; and 2 Biotechnology Research Center, The University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, Japan MS 07-354: Received 11 July 2007/Accepted 13 April 2008 ABSTRACT The objective of this study was to trace contamination sources of coliform bacteria by comparing the types of coliforms between food samples and the processing environments in two small food-processing factories (factories A and B). Fermentation tests of five sugars enabled the successful classification of 16 representative type strains into eight distinct groups. The grouping procedure was then applied to comparison of the coliform flora between food products and various locations in their processing environments. The consistency between each food and the tested locations was evaluated using the Jaccard index. The air conditioner and refrigeration room floor in factory A showed an index of 1.00, while the shaping machine in factory B showed an index of 0.98, indicating that these locations could be contamination sources. The validity of our results was confirmed by randomly amplified polymorphic DNA, which showed 100% matched profiles between the air conditioner and the food in factory A, and highly matched profiles between the machine and the food in factory B. This method for comparing the coliform flora between food and environments has the potential to be a reliable tracing tool for various food industries. Prepared ready-to-eat foods are popular; however, there is a risk of bacterial contamination during the production of such foods. If a foodborne pathogen such as Staphylococcus aureus, Listeria monocytogenes, or Escherichia coli O157:H7 happens to contaminate the food, then consumers can become ill, and the fallout can severely damage the producer (14, 16). Therefore, producers continually monitor the sanitary state of their food products and factory environments by enumerating coliforms, which are used as indicator organism for safety and quality (23). In cases where coliform bacteria are isolated from a certain food, it is necessary to trace the contamination source and clean up the contamination. However, coliform bacteria are naturally distributed in the food-processing environment (12, 13), and can be isolated from food-processing machines, floors, and exhaust vents. Thus, it can be difficult to determine the true source of contamination. The term coliform is not a taxonomical word, but rather a food sanitarian word (23). A coliform is defined as a gram-negative, aerobic, or facultative anaerobic non sporeforming rod capable of fermenting lactose within 48 h at 37 C (20). Various bacterial species are included in this group. Their physiological characteristics, such as optimum temperature, ph, and nutrient conditions, differ among strains and species. Temperature, humidity, and nutrient conditions found within a food-processing factory also differ from location to location, meaning that the coliform flora will tend to differ by location. Since the flora of food * Author for correspondence. Tel: 81-485-210614; Fax: 81-485- 256052; E-mail: tominaga@saitama-itcn.jp. isolates should be expected to coincide with those of the contamination source, it should be possible to identify the source by comparing coliform flora between the contaminated food and its processing environment. The objective of this study was to trace contamination sources of coliform bacteria by comparing the types of coliforms between food samples and the processing environments in two small food-processing factories (factories A and B). MATERIALS AND METHODS Bacterial strains. The bacterial type strains used in this study were as follows: Citrobacter freundii IAM 12471 T, Enterobacter aerogenes IAM 12348 T, Enterobacter amnigenus JCM 1237 T, Enterobacter cloacae subsp. cloacae IAM 12349 T, Enterobacter sakazakii JCM 1233 T, Escherichia vulneris JCM 1688 T, Hafnia alvei JCM 1666 T, Kluyvera cryocrescens JCM 7580 T, Leclercia adecarboxylata JCM 1667 T, Moellerella wisconsensis JCM 5895 T, Raoultella ornithinolytica JCM 6096 T, Raoultella planticola JCM 7251 T, Raoultella terrigena JCM 1687 T, Serratia fonticola JCM 1242 T, Serratia liquefaciens JCM 1245 T, and Serratia rubidaea JCM 1240 T. All strains were grown aerobically at 37 C. Sample isolation. Coliform bacteria were isolated from a bean-based processed food produced in factory A, and a potatobased processed food produced in factory B. One gram of food was suspended in 10 ml of sterile water containing 0.85% NaCl, and the sample was mixed vigorously using an Ecan TubeMixer TM-2000 (Asahi Techno Glass, Chiba, Japan). The resulting mixture (100 l) was spread on X-Gluc Magenta-Gal agar (Nissui Pharmaceutical, Tokyo, Japan) and incubated aerobically at 37 C for 24 h, and the red-purple colonies grown on the plate were

J. Food Prot., Vol. 71, No. 9 TRACING THE ORIGIN OF COLIFORM BACTERIA 1911 TABLE 1. Fermentation test of type strains Fermentation a Species SBE TRE DARL MEL MLZ Group b Citrobacter freundii IAM 12471 T 1 Enterobacter aerogenes IAM 12348 T 2 E. amnigenus JCM 1237 T 3 E. cloacae subsp. cloacae IAM 12349 T 3 E. sakazakii JCM 1233 T 3 Escherichia vulneris JCM 1688 T 3 Hafnia alvei JCM 1666 T 4 Kluyvera cryocrescens JCM 7580 T 3 Leclercia adecarboxylata JCM 1667 T 2 Moellerella wisconsensis JCM 5895 T c 5 Raoultella ornithinolytica JCM 6096 T 6 R. planticola JCM 7251 T 6 R. terrigena JCM 1687 T c 7 Serratia fonticola JCM 1242 T 2 S. liquefaciens JCM 1245 T c 2 S. rubidaea JCM 1240 T c 8 a SBE, sorbose; TRE, trehalose; DARL, D-arabitol; MEL, melibiose; MLZ, melezitose. b Bacterial strains showing identical profiles were grouped together. c The result differed from that reported in a previous study (1). isolated as coliform bacteria. Within each factory, areas measuring 10 by 10 cm of various food-processing locations were swabbed with Fuki-Fuki check (Eiken Chemical, Tokyo, Japan). The cotton-tipped swabs were rinsed in 10 ml of sterile water containing 0.85% NaCl, and 100 l of each sample was cultured on an X- Gluc Magenta-Gal agar plate as above. Approximately 100 redpurple colonies per swabbed location in plant A, or approximately 200 colonies per swabbed location in plant B were picked randomly for further analysis. Fermentation test. A mixture of 1.5% of purple broth base (Difco, Becton Dickinson, Sparks, Md.) and 2.5% agar was sterilized, mixed with 1% of sugar (D-arabitol, melibiose, melezitose, sorbose, or trehalose, filter sterilized with a 0.45- m-pore-size membrane), and allowed to solidify. The plates were subjected to aerobic incubation at 37 C for 16 h, and sugar fermentation was considered positive when the plate changed color from purple to yellow around the coliform colony. Calculation of the Jaccard index. The Jaccard index was calculated essentially as described by Barnes and Gordon (3). The percentage of each coliform group isolated from food was assigned to x 1, x 2, x 3,...,x n, while that from the food-processing environment was assigned to y 1, y 2, y 3,...,y n. The equation used was as follows: Jacaard index i xy i x y 2 2 i i The resulting values were between 0 and 1, with values approaching 1 as the coliform flora became increasingly similar between the food and swabbed samples. RAPD analysis. Genomic DNA was isolated using an ISOPLANT kit (Nippon Gene, Tokyo, Japan) according to the manufacturer s instructions. Random amplified polymorphic DNA (RAPD) analysis was performed essentially as described by Pacheco et al. (18, 19). In short, RAPD reactions were carried out with 10-mer primers 1252 (5 -GCG GAA ATA G-3 ) 1254 (5 - CCG CAG CCA A-3 ) and 1290 (5 -GTG GAT GCG A-3 ). PCR was performed in a final volume of 50 l containing 1 amplification buffer, 0.2 mm mixed dntps, 2.5 U of Ex Taq DNA polymerase (Takara Bio, Shiga, Japan), 50 pmol of a single primer, and 500 ng of template DNA. The amplification was carried out in a Mastercycler gradient thermocycler (Eppendorf, Hamburg, Germany) under the following conditions: 4 cycles of 94 C for 5 min, 37 C for 5 min, and 72 C for 5 min; followed by 30 cycles of 94 C for 1 min, 37 C for 1 min, and 72 C for 2 min; and a final incubation at 72 C for 10 min. The reaction products were separated by electrophoresis on 1.2% agarose gels, stained with ethidium bromide (0.5 g ml 1 ), and examined under UV transillumination. A 100-bp DNA ladder (Takara Bio) was used as a molecular-weight marker. Different polymorphic band patterns were assigned to the different RAPD types. RESULTS Method of investigating coliform flora. The fermentation ability of certain sugars is known to differ among various coliform species (1). For the present studies, type strains of 16 coliform species frequently isolated from food were selected (15, 17, 22), and sugars that could be used to discriminate among the strains were identified. Five sugars (D-arabitol, melibiose, melezitose, sorbose, and trehalose) were selected as having distinguishable fermentation patterns among the 16 coliform strains (Table 1). Strains showing identical fermentation abilities were grouped together and assigned arbitrary group numbers; this yielded eight groups. To examine the practicality of this phenotypebased typing method, coliform strains isolated from actual foods were subjected to fermentation testing, and the number of colonies representing each bacterial group was tallied, in order to directly investigate the coliform flora. Table 2 shows the flora from the two different tested food products. All of the coliform bacteria isolated from the food processed in factory A showed the same fermentation abil-

1912 TOMINAGA ET AL. J. Food Prot., Vol. 71, No. 9 TABLE 2. Coliform flora of foods Product of factory: Coliform flora (%) Group 2 Group 3 Group 6 Others a tested (n) b Strains A 0 100 0 0 153 B 16 33 39 12 223 a The number is the sum of the percentages of strains belonging to groups other than groups 2, 3, and 6. b The numbers of total strains isolated from each food. ity (corresponding to group 3), suggesting that they were homogeneous. In contrast, the coliform flora from food processed in factory B was heterogeneous, consisting of group 6 (39%), group 3 (33%), group 2 (16%), and others (12%). Thus, our method could be used to distinguish coliform flora from one another. Tracing the contamination origins of coliform bacteria. It seemed that the coliform flora difference between factories A and B was due to that difference of the contamination sources. To examine whether the coliform flora differed by location/environment within each factory, coliform bacteria were isolated from various locations in factories A and B, and subjected to the sugar fermentation tests (Fig. 1). As might be expected, most of the swabbed points from factories A and B had different coliform flora. We then calculated the Jaccard index of each food-location sample pair, testing for similarities (Table 3). In factory A, the flora from the air conditioner and the refrigeration room floor matched exactly (Jaccard index 1.00) with that from the TABLE 3. The Jaccard index of coliform flora between food and swabbed locations in factory A and in factory B Swabbed point Jaccard index Strains tested (n) a Factory A Air conditioner 1.00 98 Refrigeration room floor 1.00 99 Drain 0.50 98 Soaking water 0.02 100 Work area floor 0.02 100 Plant floor 0.01 97 Factory B Shaping machine 0.98 278 Raw material stock room floor 0.75 238 Refrigeration room floor 0.69 237 Blending machine 0.55 188 Work area floor 0.54 257 Chiller 0.43 255 Raw material 0.21 230 a The numbers of total isolated strains from each swabbed point. food, suggesting that these two points could be candidates for the origin of contamination. Similarly, in factory B, the flora from the shaping machine samples showed a very high match (Jaccard index 0.98) with the food, suggesting that the machine could be a candidate for the origin of contamination. Confirmation of contamination origin by RAPD analysis. Comparing the coliform flora between food sam- FIGURE 1. The coliform flora of the food-processing environments in factory A (A) and factory B (B). The band chart under the machine or environment name shows the coliform flora at each point. The designation of the flora is as follows: group 2 (s), group 3 (r), group 4 (v), group 6 (p), and group 7 (u). Each group name corresponds to that given in Table 1. Designations t and g represent patterns outside the analyzed groups; they had fermentation patterns (sorbose, trehalose, D-arabitol, melibiose, melezitose) of (,,,, ) and (,,,, ), respectively.

J. Food Prot., Vol. 71, No. 9 TRACING THE ORIGIN OF COLIFORM BACTERIA 1913 TABLE 4. RAPD analysis of coliform strains isolated from food and the candidates for its contamination origin in factory A a Group 3 Product RAPD type (%) Air conditioner Refrigeration room floor 3-1 100 100 0 3-2 0 0 100 Jaccard index 1.00 0.00 a The consistency of RAPD profiles between food and the candidates for its contamination origin were analyzed with the Jaccard index. ples and various food-processing points within each factory enabled us to identify candidate contamination sources. To confirm whether the potential candidates were really the contamination origins, RAPD analysis was used to compare the isolates at the strain level (Tables 4 and 5). The RAPD profile obtained using primer 1254 for the strains isolated from the air conditioner in factory A completely matched that from the food product, confirming that the air conditioner was the origin of contamination (Table 4). The same result was obtained using primers 1252 and 1290 (data not shown). RAPD analysis with primer 1254 showed the greatest number of amplified bands in samples from factory A, so this primer was used for analysis of the samples from factory B. Various RAPD profiles were obtained from both the food and the shaping machine samples (Table 5); analysis of the RAPD profiles corresponding to members of group 2 (Jaccard index 0.98), group 3 (Jaccard index 0.84), and group 6 (Jaccard index 0.87) from each of the samples suggested that the shaping machine was the origin of contamination. DISCUSSION Coliform is a general name given to various bacterial aggregates that can be classified by differences in sugar assimilation or antimicrobial susceptibility (1, 4, 10). In the present study, we discriminated coliform groups based on differences in their ability to ferment a panel of five sugars. The five sugars were selected based on the following criteria: (i) the fermentation abilities of the 16 tested typestrains were clearly positive or negative, depending on species, with sugars yielding plus or minus results excluded to avoid errors of interpretation, and (ii) whether the combined sugar panel could be used to differentiate the type strains in as much detail as possible. The 16 type strains could be differentiated into eight distinct groups, using the five-sugar panel developed herein. Some strains, such as E. amnigenus, E. cloacae, E. sakazakii, E. vulneris, and K. cryocrescens, showed identical profiles; the addition of more sugars to the panel in the future could be used to better discriminate among these strains. Considering that the coliform flora analyzed in the present study showed differences between the two examined foods, we feel that the five selected sugars were adequate for the purposes of this study and could be useful in the field. We found location-specific differences among the flora TABLE 5. RAPD analysis of coliform strains isolated from food and the candidates for its contamination origin in factory B a Group Product RAPD type (%) Shaping machine Group 2 2-1 73 87 2-2 13 7 2-3 7 0 2-4 7 0 2-5 0 7 Jaccard index 0.98 Group 3 3-3 27 13 3-4 13 27 3-5 20 27 3-6 7 0 3-7 33 20 3-8 0 7 3-9 0 7 Jaccard index 0.84 Group 6 6-1 27 27 6-2 53 40 6-3 7 33 6-4 7 0 6-5 7 0 Jaccard index 0.87 a The consistency of RAPD profiles between food and the candidates for its contamination origin were analyzed with the Jaccard index. of samples swabbed from different areas of each factory. Coliform bacteria are an assemblage of various species; differences in optimum growth conditions may mean that specific species predominate in certain environments. Thus, the detection of primarily group 3 coliforms from the refrigeration room floor in both factories A and B may suggest that group 3 coliforms are resistant to low temperatures. Alternatively, the coliform bacteria living in certain foodprocessing environments may have been initially transmitted by humans, animals, insects, vegetables, and so on (8, 9). Differences in the initial input of a certain coliform strain may lead to static coliform flora. Regardless of the initial source, we herein found that the coliform flora differed by location in both of the studied factories. Numerical analysis using the Jaccard index allowed us to identify concordance between the coliform flora of machines/locations with that found in the food samples from each factory. In factory A, the air conditioner and refrigeration room floor were identified as potential sources of contamination. RAPD analysis showed 100% matched profiles between the air conditioner and the food, indicating that the air conditioner was likely to be the contamination source. Although our phenotype-based typing method could not narrow the contamination source down to a single place, the concordance in coliform flora indicated that the source seemed to be included in our candidates. In the fu-

1914 TOMINAGA ET AL. J. Food Prot., Vol. 71, No. 9 ture, additional optimization of the culture medium composition or the sugar panel components may enable identification of a single contamination source. In factory B, the shaping machine was identified as the contamination source candidate. RAPD analysis of the food and the machine showed various RAPD profiles, most of which coincided. Thus, we herein successfully used a phenotype-based typing method to compare coliform flora. DNA-based typing methods like RAPD and pulsed-field gel electrophoresis have been used to trace the contamination source of foodborne pathogens such as L. monocytogenes and E. coli O157:H7 (2, 5, 7, 21). DNA-based typing methods have higher discrimination power than do phenotype-based typing methods. Higher discrimination is necessary for tracing specific foodborne pathogens, as it is often necessary to discriminate between very similar strains belonging to the same species. In contrast, coliform bacteria are an assemblage of various species, making them more amenable to phenotype-based typing methods. In general, the more bacteria analyzed, the easier it is to trace the contamination (11). Analysis of numerous bacterial strains, using DNAbased methods has the disadvantages of large labor and time requirements, whereas phenotype-based analysis of multiple strains can be performed with minimal increases in labor and time. Thus, it seems appropriate to use phenotype-based methods to analyze coliform flora. Previously, Garcia et al. (6) reported tracing potential sources of coliform contamination of apple cider by counting and comparing the total coliform numbers from potential sources. If the authors had examined the coliform flora of the apple cider and its potential sources, as reported herein, then the contamination source could have been identified more concretely. Thus, our method of tracing coliform contamination by comparing the flora between food and its processing environment could prove useful in many food industries in which food poisoning outbreaks are of great concern. REFERENCES 1. Anonymous. 2004. Instruction for API 50CHB/E medium. bio- Mérieux, Inc., Hazelwood, Mo. 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 analyses of Escherichia coli O157:H7 and O157 nonmotile isolates recovered from beef cattle and carcasses at processing plants in the Midwestern states of the United States. Appl. Environ. Microbiol. 67:3810 3818. 3. Barnes, B., and D. M. Gordon. 2004. Coliform dynamics and the implications for source tracking. Environ. Microbiol. 6:501 509. 4. Burnes, B. S. 2003. Antibiotic resistance analysis of fecal coliforms to determine fecal pollution sources in a mixed-use watershed. Environ. Monit. Assess. 85:87 98. 5. Destro, M. T., M. F. F. Leitão, and J. M. Farber. 1996. Use of molecular typing methods to trace the dissemination of Listeria monocytogenes in a shrimp processing plant. Appl. Environ. Microbiol. 62:705 711. 6. Garcia, L., J. Henderson, M. Fabri, and M. Oke. 2006. Potential sources of microbial contamination in unpasteurized apple cider. J. Food Prot. 69:137 144. 7. Giovannacci, I., C. Ragimbeau, S. Queguiner, G. Salvat, J. L. Vendeuvre, V. Carlier, and G. Ermel. 1999. Listeria monocytogenes in pork slaughtering and cutting plants: use of RAPD, PFGE and PCR- REA for tracing and molecular epidemiology. Int. J. Food Microbiol. 53:127 140. 8. Gordon, D. M. 2001. Geographical structure and host specificity in bacteria and the implications for tracing the source of coliform contamination. Microbiology 147:1079 1085. 9. Gordon, D. M., and F. FitzGibbon. 1999. The distribution of enteric bacteria from Australian mammals: host and geographical effects. Microbiology 145:2663 2671. 10. Harwood, V. J., J. Whitlock, and V. Withington. 2000. Classification of antibiotic resistance patterns of indicator bacteria by discriminant analysis: use in predicting the source of fecal contamination in subtropical waters. Appl. Environ. Microbiol. 66:3698 3704. 11. Johnson, L. K., M. B. Brown, E. A. Carruthers, J. A. Ferguson, P. E. Dombek, and M. J. Sadowsky. 2004. Sample size, library composition, and genotypic diversity among natural populations of Escherichia coli from different animals influence accuracy of determining sources of fecal pollution. Appl. Environ. Microbiol. 70:4478 4485. 12. Kaneko, K., H. Hayashidani, K. Takahashi, Y. Shiraki, S. Limawongpranee, and M. Ogawa. 1999. Bacterial contamination in the environment of food factories processing ready-to-eat fresh vegetables. J. Food Prot. 62:800 804. 13. Leclerc, H., D. A. A. Mossel, S. C. Edberg, and C. B. Struijk. 2001. Advances in the bacteriology of the coliform group: their suitability as markers of microbial water safety. Annu. Rev. Microbiol. 55:201 234. 14. Le Loir, Y., F. Baron, and M. Gautier. 2003. Staphylococcus aureus and food poisoning. Genet. Mol. Res. 2:63 76. 15. Lindberg, A.-M., Å. Ljungh, S. Ahrné, S. Löfdahl, and G. Molin. 1998. Enterobacteriaceae found in high numbers in fish, minced meat and pasteurized milk or cream and the presence of toxin encoding genes. Int. J. Food Microbiol. 39:11 17. 16. Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607 625. 17. Olsson, C., S. Ahrné, B. Pettersson, and G. Molin. 2004. DNA based classification of food associated Enterobacteriaceae previously identified by Biolog GN Microplates. Syst. Appl. Microbiol. 27:219 228. 18. Pacheco, A. B. F., B. E. C. Guth, K. C. C. Soares, L. Nishimura, D. F. D. Almeida, and L. C. S. Ferreira. 1997. Random amplification of polymorphic DNA reveals serotype-specific clonal clusters among enterotoxigenic Escherichia coli strains isolated from humans. J. Clin. Microbiol. 35:1521 1525. 19. Pacheco, A. B. F., K. C. Soares, D. F. Almeida, G. I. Viboud, N. Binsztein, and L. C. S. Ferreira. 1998. Clonal nature of enterotoxigenic Escherichia coli serotype O6:H16 revealed by randomly amplified polymorphic DNA analysis. J. Clin. Microbiol. 36:2099 2102. 20. Rompre, A., P. Servais, J. Baudart, M.-R. de-roubin, and P. Laurent. 2002. Detection and enumeration of coliforms in drinking water: current methods and emerging approaches. J. Microbiol. Methods 49:31 54. 21. Shere, J. A., K. J. Bartlett, and C. W. Kaspar. 1998. Longitudinal study of Escherichia coli O157:H7 dissemination on four dairy farms in Wisconsin. Appl. Environ. Microbiol. 64:1390 1399. 22. Stiles, M. E., and L.-K. Ng. 1981. Biochemical characteristics and identification of Enterobacteriaceae isolated from meats. Appl. Environ. Microbiol. 41:639 645. 23. Tortorello, M. L. 2003. Indicator organisms for safety and quality uses and methods for detection: minireview. J. AOAC Int. 86:1208 1217.