618

Journal of Food Protection, Vol. 47, No. 8, Pages 618-622 (August 1984) Copyright^, International Association of Milk, Food, and Environmental Sanitarians

Evaluation of the Microbiological Safety of Tofu

SUSAN K. KOVATS, MICHAEL P. DOYLE and NOBUMASA TANAKA*

The Food Research Institute, University of Wisconsin-Madison, Madison, Wisconsin 53706

(Received for publication March 23, 1984) Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/8/618/1651114/0362-028x-47_8_618.pdf by guest on 27 September 2021

ABSTRACT growth. For this reason, in Japan tofu has been consumed traditionally within a day or two after manufacture. How­ Studies were done to evaluate the safety of tofu inoculated ever, in the United States, tofu may be held at the retail with different bacterial pathogens and held at different tempera­ level many days before consumption, as some manufac­ tures. Clostridium botulinum (type A and/or B) toxin was pro­ turers predate their tofu for a 28-d shelf-life (2). Recent duced in tofu after 1 and 3 wk of storage at 25 and 15°C, reports indicate many retailers temperature abuse tofu by respectively, but not within 6 wk at 5 or 10°C. Compared with displaying the product on produce counters where temper­ water-packed tofu, vacuum-packaged tofu did not increase the risk of botulinal toxin production at 15°C. Staphylococcus au­ atures may be 10 to 15°C, and some retailers stack tofu reus and Salmonella typhimurium grew at similar rates at 10, on tables with no refrigeration. Rehberger et al. (13) 15 and 25°C, but decreased in number at 5°C. Staphylococcal found that tofu stored at 10°C until the manufacturer's enterotoxin was not produced within 4 wk at 10°C, 8 d at 15°C, pull date had high aerobic plate counts (up to 5.6xl06 or 2 d at 25°C, even though an S. aureus population of >107 CFU/g) and psychrotroph counts (up to 7.9 X 107 CFU/g), CFU/g was present in most samples analyzed. Staphylococcal indicating likely microbial spoilage. enterotoxin was detected in a 5-d sample held at 25°C. Toxin Because in the United States tofu is being marketed production may have been delayed because of the low amount under conditions quite different from those traditionally of oxygen in water-packed tofu. Yersinia enterocolitica grew at all temperatures evaluated (5, 10, 15 and 25°C), including used in the Orient, an evaluation of the microbiological safety of tofu under different stroage conditions was a 4- to 6-log,0 increase within 14 d at 5°C. Results of these studies substantiate the need for: (a) implementing a high level deemed necessary. This concern was strengthened by a of sanitary practices during tofu production, (b) pasteurizing recent outbreak of yersiniosis resulting from consumption tofu after packaging, and (c) properly refrigerating tofu during of tofu contaminated with Yersinia enterocolitica (3,12). storage and display. The purpose of this study was to determine the ability of foodborne bacterial pathogens to grow or produce toxin in tofu held at different temperatures.

MATERIALS AND METHODS Tofu, or soybean curd, originated in the Orient where this food has been a major source of protein for more Bacterial cultures and methods of enumeration than 2,000 years (20). It is the most popular soybean- Four foodborne bacterial pathogens were studied. These included: Clostridium botulinum, Staphylococcus aureus, Salmonella typhimurium based food in Japan (21), and is becoming increasingly and Yersinia enterocolitica. popular in the United States. A ten-strain spore mixture of C. botulinum, consisting of five strains Tofu is made from soybean milk which is the liquid each of types A and B (16) in water was used. Immediately before resulting from blended and heated soybeans previously inoculating tofu, the spores were heat-shocked at 80°C for 15 min. Im­ soaked in water (20). The blended and heated soybean mediately after inoculation, C. botulinum (three samples per batch) was enumerated using the 5-tube most probable number (MPN) method with slurry is filtered and precipitated with coagulant (calcium Trypticase-peptone-glucose-yeast extract broth as the growth medium or magnesium sulfate, or calcium or magnesium chloride) CO. to form a cheese-like curd which is molded into cakes. S. aureus strains 196 E (enterotoxin A and D producer) and 361 Many manufacturers package tofu cakes individually in (enterotoxin C2 producer) were kindly provided by M. S. Bergdoll, a water-packed, plastic container that is sealed with a University of Wisconsin-Madison. Cultures were stored in brain heart sheet of film or a tight-fitting lid. infusion (BHI; Difco) agar slants at 4°C. For each experiment, cultures The composition of tofu varies depending on the soy­ were grown overnight in BHI broth at 37°C, washed and resuspended in 0.01 M phosphate-buffered saline, pH 7.5 (PBS). The cell suspen­ bean variety used and the method of production. Tofu sions were diluted appropriately with PBS and mixed to give an equal is a very nutritious food typically containing 6 to 8.4% number mixture of the two strains. S. aureus was enumerated by plate protein (21). Hence it is quite susceptible to microbial count on Baird-Parker agar (Difco). Typical S. aureus colonies (black,

JOURNAL OF FOOD PROTECTION, VOL. 47, AUGUST 1984 MICROBIOLOGICAL SAFETY OF TOFU 619 shiny and convex surrounded by clear zone) were counted and randomly into Saran-coated mylar pouches (Curwood, Inc., New London, WI; chosen isolates were confirmed for coagulase activity. oxygen transfer rate less than 0.5 cm3/l atm/24 h/100cm2) and vacuum S. typhimurium strains 40252 and 40467 were obtained from the Wis­ packaged (Packaging Aid Corporation, San Francisco, CA, Model consin State Laboratory of Hygiene, Madison, WI. Each strain was 24 V). Some tofu from the same batch was packed in water, as de­ stored on Trypticase soy agar (TSA; BBL Microbiology Systems) slants scribed above, and incubated along with the vacuum-packaged tofu as at 4°C. For each experiment, cultures were grown overnight in Trypti­ the control. case soy broth (TSB; BBL Microbiology Systems) at 37°C. Cells of each strain were harvested, washed and resuspended in PBS, then di­ Analysis of samples luted and mixed to give an equal number mixture of the two strains. The entire contents of each container (tofu and water) were poured Hektoen enteric agar (Difco) was used for enumeration of S. into a sterile Stomacher bag and blended in a Stomacher (Lab Blender typhimurium. Randomly selected colonies were confirmed as S. 400) for 2 min. Vacuum-packaged samples were blended similarly, ex­ typhimurium by agglutination with the appropriate antisera (Difco). cept that 35 ml of double distilled water was added to each tofu sample Y. enterocolitica strains 34/1B and FRI-YE16 were used. Y. en- before stomaching. For samples inoculated with C. botulinum, the terocolitica 34/1B (serotype 0:8), kindly provided by C. C. G. Aulisio, blended tofu was centrifuged (5,000 x g, 5 min) and the supernatant Food and Drug Administration, Washington, DC, was originally iso­ fluid was used for the mouse toxicity assay (1). Samples inoculated lated from unchlorinated water used in the manufacture and packaging with other organisms were blended, diluted appropriately in PBS, and of tofu (3). Y. enterocolitica FRI-YE16 (serotype 0:3) is a virulent plated. Triplicate samples were analyzed for each treatment at each Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/8/618/1651114/0362-028x-47_8_618.pdf by guest on 27 September 2021 strain originally isolated from a porcine tongue (5). Cultures were stored sampling interval, and an average of these results is reported. on TSA slants at 4°C, and for each experiment, overnight cultures were Some samples containing over 106 cells of S. aureuslg were tested grown in TSB at 25°C. Cells of each strain were harvested, washed for the presence of staphylococcal enterotoxins (A, C2 and D) by and resuspended in PBS, then an equal number mixture of the two radioimmunoassay (11) and enzyme-linked immunosorbent assay (7). strains was made. Cells were enumerated using yersinia selective agar Tofu samples tested for enterotoxins were stored frozen (-20°C) until base containing yersinia selective supplement (Oxoid). Randomly the assays were performed. selected colonies were confirmed as virulent Y. enterocolitica by agglutination with antiserum (WA-SAA; 6). Biochemical characteristics of these isolates were determined by the API-20E diagnostic system (Analytab Products, Plainveiw, NY). RESULTS Preparation of tofu Dry soybeans (300 g; Noble Brand Seeds, Gibson City, IL) were C. botulinum toxin was produced in water-packed tofu washed with cold running water and soaked overnight (16 h) at room held at 15 and 25°C within 3 and 1 wk, respectively, temperature in 1,500 ml of water. Soaked beans were then rinsed, but not at 4 or 10°C within 6 wk (Table 1). Vacuum drained and blended in small portions (ca. 150 g of soaked beans per packaging did not increase the risk of botulinal toxin pro­ 250 ml of water) in a Waring Blendor for 2 min per portion. The duction, as toxin was first detected in a water-packed blended beans were combined and more water was added to a total 3,000 ml of water per 300 g (dry wt) of beans. Because the beans sample (1 of 3 samples tested was toxic) at 2 wk and had absorbed 390 ml of water during overnight soaking, a total of 2,610 in vacuum-packaged samples at 3 wk (3 of 3 tested were ml of water was added to the soaked beans. The blended beans were toxic) at 15°C (Table 2). then poured into an 8-L of stainless steel beaker, covered with S. aureus grew in tofu at 10°C and above (Table 3). aluminum foil, and heated in a steam chamber for 15 min at 96°C (in­ ternal temperature). Growth was most rapid at 25°C, with >5-log10 increase The heated slurry was filtered through a strainer lined with four occurring in 24 h. A similar increase occurred within 5 layers of sterile cheesecloth. The filtrate (soybean milk) was collected and 15 d at 15 and 10°C, respectively. S. aureus did not in a 4-L beaker and cooled to 70°C, then poured rapidly into an 8-L grow but rather decreased in number during storage at stainless steel beaker containing 250 ml of 0.2 M CaS0 suspension. 4 5°C. Staphylococcal enterotoxins A, C2 or D were not The mixture was stirred immediately with a ladle (only two strokes as detected (<1 ng/g) in tofu within 4 wk at 10°C, 8 d at the extent of stirring affects the hardness of the final product [20]), and left undisturbed for 15 min to coagulate. The bean curd was then 15°C, or 2d at 25°C. Enterotoxin A was detected in a transferred gently with a ladle into a press (19). The press was a box 5-d sample held at 25°C and enterotoxins A and C2 were (44.5x20.3x11.4 cm) constructed of 1.3-cm thick polysulfone, with detected in a 15-d sample held at 25°C. several slots and holes for drainage. A burlap cloth was fitted inside Survival and growth characteristics of S. typhimurium the box to hold the curd. The curd was placed on the cloth which was then folded over the top of the curd. A 1.3-cm thick polysulfone (Table 4) in tofu were similar to those of S. aureus at board which weighed 1.5 kg and fit snugly inside the box, was placed equivalent temperatures. S. typhimurium died off during on top of the wrapped curd for 1.5 h. The press and its accessories storage at 5°C but increased by 6 log10 within 24 h at were sterilized (autoclaved) before each use. The above procedure 25°C. yielded about 860 g of tofu from the 300 g of dry beans. Y. enterocolitica grew in tofu at all temperatures Inoculation and incubation of tofu evaluated (Table 5). A 4- to 6-log|0 increase occurred Twenty-gram portions of tofu were weighed into plastic containers within 14 d at 5°C, whereas >6-log10 increase resulted (55 ml capacity; presanitized by soaking ca. 15 min in a dilute hypoch­ within 24 h at 25°C. Y. enterocolitica isolates recovered lorite solution). Sterile tap water (30 ml) and 0.5 ml of bacterial culture were added to each container of tofu. The containers were covered with from tofu samples held at the four different temperatures snap-on lids and inverted gently twice. Tofu samples were separated agglutinated with WA-SAA following growth at 37°C in into four equal groups, and each group was incubated at one of four TSB, indicating the isolates continued to express their temperatures (5, 10, 15 or 25°C). virulence-associated antigenic determinants (and were Tofu that was vacuum packaged was inoculated with heat-shocked likely pathogenic) after growing in tofu. C. botulinum spores immediately before the soybean milk was added to the coagulant suspension. The amount of added inoculum was pre­ In general, all of the inoculated samples held for 3 wk pared considering that some spores would be lost in the whey that or longer at any of the temperatures studied were partially drained off. After the tofu was pressed and cooled, 35 g were weighed degraded and visually considered inedible.

JOURNAL OF FOOD PROTECTION, VOL. 47, AUGUST 1984 620 KOVATS, DOYLE AND TANAKA

TABLE 1. Clostridium botulinum toxin production in water-packed tofu held at different temperatures. Temperature No. of toxic samples/No. of samples tested at week: (°C) 0 1 2 3 4 5 6 5" 0/3 0/3 0/3 0/3 0/3 0/3 0/3 10b 0/3 0/3 0/3 0/3 0/3 0/3 0/3 15" 0/3 0/3 0/3 3/3 3/3 - - 25" 0/3 2/3 3/3 3/3 - - - "Inoculum of 93 C. botulinum spores per g of tofu. blnoculum of 52 C. botulinum spores per g of tofu.

TABLE 2. Clostridium botulinum toxin production in vacuum-packaged vs. water-packed tofu". Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/8/618/1651114/0362-028x-47_8_618.pdf by guest on 27 September 2021 Temperature Treatment No. of toxic sample/No. of samples tested at week: (°C) 0 1 2 3 4 5 6 10 VPb 0/3 0/3 0/3 0/3 0/3 0/3 0/3 WF 0/3 0/3 0/3 0/3 0/3 0/3 0/3

15 VP 0/3 0/3 0/3 3/3 - - _ WP 0/3 0/3 1/3 3/3 - - - aAverage inoculum of 240 C. botulinum spores per g of tofu. bVP, vacuum-packaged. CWP, water-packed.

TABLE 3. Staphylococcus aureus growth in water-packed tofu held at different temperatures. Temperature Logio S. aureusl% at: (°C) 0 h 16 h 24 h 48 h 5 d 8d 15 d 4 wk 6 wk 5 2.15 1.97 1.64 1.46 1.20 0.30 0.30 10 2.15 2.54 3.69 4.72 7.36b 6.85b 6.65 15 2.15 4.41 7.96b 8.34b 8.30 - - 25 1.81 6.41b 7.15b 7.73b - - - - - 25 2.15 8.41 8.69b>c 8.49 8.28b-c - - "-, not determined. 'Three samples were composited and tested for staphylococcal enterotoxins. cSamples positive for staphylococcal enterotoxin(s).

TABLE 4. Salmonella typhimurium growth in water-packed tofu held at different temperatures. Temperature Logio S. typhimuriumlg at: (°C) 0 h 16 h 24 h 48 h 5d 7 d 14 d 4 wk 6 wk 401 1.48 0.90 0.85 0.30 68 0.90 0.60 0.00 NDb

10 41 2.57 2.70 4.28 6.04 68 >3.70 <4.00 5.76 5.54

15 41 3.30 7.15 7.65 7.88 68 >3.70 8.04 8.49 25 86 6.28 7.85 8.32 41 7.96 8.26 8.18 8.62 68 >7.00 8.46 8.87 d-, not determined. bND, not detected.

JOURNAL OF FOOD PROTECTION. VOL. 47, AUGUST 1984 SAFETY OF TOFU 621

TABLE 5. Yersinia enterocolitica growth in water-packed tofu held different temperatures. Temperature Logio Y. enterocolitica/g at: CO Oh 16_h 24 h 48 h 5 d 7 d 14 d 4 wk 5 1.97 2.73 3.15 4.00 6.18

2.01 3.70 7.97 8.40

10 1.97 3.80 8.00 8.43 8.54

15 1.66 3.00 3.30 8.80 - 1.97 7.88 8.34 8.69 8.71 25 1.66 4.80 8.00 8.79 .... L97 - -__ 8.62 8.74 9.23 8.49 a-, not determined. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/8/618/1651114/0362-028x-47_8_618.pdf by guest on 27 September 2021

vacuum-packaged tofu, measures must be taken to pre­ DISCUSSION vent the growth of this organism. Our studies have shown that refrigerating tofu at less than 10°C will prevent Results of these studies substantiate the need for: (a) botulinal toxin production, and that holding tofu at 15°C implementing a high level of sanitary practices during or above is hazardous. Hence tofu should be held at less tofu production and packaging, (b) pasteurizing tofu after than 10°C to avoid the risk of botulism. However, even packaging, and (c) refrigerating tofu after packaging and though C. botulinum types A and B did not produce toxin during storage and display. At appropriate temperatures, in tofu at 10°C, both S. aureus and S. typhimurium grew tofu readily supported the growth of all pathogens to large populations within 15 d at this temperature. evaluated, indicating extreme care should be taken to Neither pathogen grew, but rather died during storage at avoid contamination of pathogens from environmental 5°C. Hence tofu manufacturers, distributors, retailers and sources by insanitary conditions and practices or from consumers should hold tofu at 5°C or less until used. contaminated ingredients, such as improperly treated Interestingly, staphylococcal enterotoxin was not de­ water. tected in many tofu samples that contained >107 S. au­ The importance of using properly treated water is reus!g. It is generally recognized that detectable enteroto­ exemplified by a 1981 outbreak of yersiniosis that invol­ xin is produced when S. aureus populations of >106 ved 87 cases in the state of Washington (3,12). The ill­ cells/g develop in foods (17). Water-packed tofu likely ness was associated with consumption of locally produced has a low oxygen content as evidenced by the ability of tofu packed together with untreated spring water. Two tofu packaged in this manner to support growth and toxin virulent strains of Y. enterocolitica serotype 0:8, which production by C. botulinum, an anaerobe that only grows had biochemical patterns similar to serotype 0:8 isolates in the absence of oxygen. It is well-established that alter­ recovered from hospitalized patients, were isolated from ation of the atmospheric composition of the environment water used in the processing of tofu (3). Our study indi­ of a food can influence the potential for staphylococcal cates that once present, Y. enterocolitica can grow to a enterotoxin production. This is particularly evident when large population within 7 to 14 d, even when tofu is oxygen is eliminated or reduced, in which case properly refrigerated at 5°C. The ability of Y. en­ staphylococcal enterotoxin production is reduced terocolitica to grow in foods at refrigeration temperature (4,8,14,18). Hence low oxygen content may be the factor has also been demonstrated by others (9,10,15). responsible for repressing staphylococcal enterotoxin pro­ A pasteurization treatment should not be used as a cor­ duction in water-packed tofu. rective measure for insanitary practices or the use of un­ The potential microbiological hazards of tofu produced treated water in tofu manufacture; however, sufficiently under insanitary conditions and/or improperly stored are heating tofu after packaging to inactivate Y. enterocolitica great. Hence it is imperative that tofu manufacturers, re­ would be a useful practice to eliminate the hazard of this tailers and consumers are made aware of this concern and psychrotrophic organism. take the appropriate measures to prevent microbiological Tofu should be relatively free of vegetative bacterial health risks. Additionally, as suggested by Rehberger et pathogens if it is prepared under proper sanitary condi­ al. (13), applying proper sanitary practices and storage tions. Cooking the soybean slurry at greater than 95°C temperatures will also improve the microbiological qual­ for 15 min should kill all vegetative bacteria, leaving ity of tofu, thereby reducing the risk of product spoilage. only microbial spores as survivors. C. botulinum spores may be among the surviving population. Since there is ACKNOWLEDGMENTS likely to be few bacterial competitors to inhibit the We thank Drs. C. W. Hesseltine and H. L. Wang, Northern Regional growth of C. botulinum and because the appropriate Research Center, U.S. Department of Agriculture, Peoria, IL, for their anaerobic conditions needed for this organism to grow helpful assistance and advice. We also thank Dr. M. S. Bergdoll and and produce toxin are present in both water-packed and Ms. Mary Evenson, Food Research Institute, University of Wisconsin-

JOVRNAL OF FOOD PROTECTION, VOL. 47, AUGUST 1984 622 KOVATS, DOYLE AND TANAKA

Madison, for the cultures of S. aureus and staphylococcal enterotoxin 10. Hanna, M. 0., J. C. Stewart, D. L. Zink, Z. L. Carpenter, and assays. C. Vanderzant. 1977. Development of Yersinia enterocolitica on This work was supported by U.S. Department of Agriculture agree­ raw and cooked beef and pork at different temperatures. J. Food ment No. 48-519B-2-1173; the College of Agricultural and Life Sci­ Sci. 42:1180-1184. ences, University of Wisconsin-Madison; and by contributions to the 11. Miller, B. A., R. F. Reiser, and M. S. Bergdoll. 1978. Detection Food Research Institute. of staphylococcal enterotoxins A, B, C, D, and E in foods by radioimmunoassay, using staphylococcal cells containing protein A REFERENCES as immunosorbent Appl. Environ. Microbiol. 36:421-426. 12. Nolan, C, N. Harris, J. Ballard, J. Allard, and J. Kobayashi. 1. Anonymous. 1976. Food and Drug Administration bacteriological 1982. Outbreak of Yersinia entercolitica - Washington State. Mor­ analytical manual for foods. Association of Official Analytical bid. Mortal. Weekly Rep. 31:562-564. Chemists, Washington, DC. 13. Rehberger, T. G., L. A. Wilson, and B. A. Glatz. 1984. Micro­ 2. Anonymous. 1983. Hinode Tofu Company. Soyfoods Winter:33- biological quality of commercial tofu. J. Food Prot. 47:177-181. 36. 14. Smith, J. L., R. L. Buchanan, and S. A. Palumbo. 1983. Effect 3. Aulisio, C. C. G., J. T. Stanfield, S. D. Weagant, and W. E. of food environment on staphylococcal enterotoxin synthesis: a re­ Hill. 1983. Yersiniosis associated with tofu consumption; serologi­ view. J. Food Prot. 46:545-555. cal, biochemical and pathogenicity studies of Yersinia en- 15. Stern, N. J., M. D. Pierson, and A. W. Kotula. 1980. Growth Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/8/618/1651114/0362-028x-47_8_618.pdf by guest on 27 September 2021 terocolitica isolates. J. Food Prot. 46:226-230. and competitive nature of Yersinia enterocolitica in whole milk. 4. Barber, L. E., and R. H. Deibel. 1972. Effect of pH and oxygen J. Food Sci. 45:972-974. tension on staphylococcal growth and enterotoxin formation in fer­ 16. Tanaka, N., E. Traisman, M. H. Lee, R. G. Cassens, and E. M. mented sausage. Appl. Microbiol. 24:891-898. Foster. 1980. Inhibition of botulinum toxin formation in bacon by 5. Doyle, M. P., M. B. Hugdahl, and S. L. Taylor. 1981. Isolation acid development. J. Food Prot. 43:450-457. of virulent Yersinia enterocolitica from porcine tongues. Appl. En­ 17. Tatini, S. R., B. R. Cords, and J. Gramoli. 1976. Screening for viron. Microbiol. 42:661-666. staphylococcal enterotoxins in foods. Food Technol. 30(4):64-74. 6. Doyle, M. P., M. B. Hugdahl, M. T. Chang, and J. T. Beery. 18. Thatcher, F. S., J. Robinson, and I. Erdman. 1962. The "vacuum 1982. Serological relatedness of mouse-virulent Yersinia en­ pack" method of packaging foods in relation to the formation of terocolitica. Infect. Immun. 37:1234-1240. the botulinum and staphylococcal toxins. J. Appl. Bacteriol. 7. Freed, R. C, M. L. Evenson, R. F. Reiser, and M. S. Bergdoll. 25:120-124. 1982. Enzyme-linked immunosorbent assay for detection of 19. Wang, H. L. 1967. Products from soybeans. Food Technol. staphylococcal enterotoxins in food. Appl. Environ. Microbiol. 21(5):115-116. 44:1349-1355. 20. Wang, H. L., and C. W. Hesseltine. 1982. Coagulation conditions 8. Genigeorgis, C, H. Riemann, and W. W. Sadler. 1969. Produc­ in tofu processing. Process Biochem. 17:7-12. tion of enterotoxin B in cured meats. J. Food Sci. 34:62-68. 21. Wang, H. L., G. C. Mustakas, W. J. Wolf, L. C. Wang, C. W. 9. Hanna, M. O., J. C. Stewart, Z. L. Carpenter, and C. Vanderzant. Hesseltine, and E. B. Bagley. 1979. Soybeans as human food - 1977. Effect of packaging methods on the development of Yersinia Unprocessed and simply processed. U.S. Department of Agricul­ enterocolitica on beef steaks. J. Food Safety 1:29-37. ture, Science and Education Administration, Washington, DC. 54 p.

JOURNAL OF FOOD PROTECTION, VOL. 47, AUGUST 1984