331 Journal of Food Protection, Vol. 68, No. 2, 2005, Pages 331±335 Copyright Q, International Association for Food Protection Levels and Enterotoxigenicity of Clostridium perfringens in Pozole, Tamales, and Birria V. NAVARRO-HIDALGO,1,2 E. CABRERA-DIÂAZ,1 H. ZEPEDA,2 L. MOTA DE LA GARZA,2 A. CASTILLO,3 AND R. TORRES-VITELA1* 1Laboratorio de MicrobiologõÂa Sanitaria, Centro Universitario de Ciencias Exactas e IngenierõÂas, Universidad de Guadalajara, Marcelino GarcõÂa BarragaÂn 1451, Guadalajara 44430, MeÂxico; 2Deptartamento de MicrobiologõÂa, Escuela Nacional de Ciencias BioloÂgicas, Instituto PoliteÂcnico Nacional, Carpio y Plan de Ayala, MeÂxico, D. F.; and 3Department of Animal Science, Texas A&M University, College Station, Texas 77843-2471, USA Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/2/331/1675907/0362-028x-68_2_331.pdf by guest on 28 September 2021 MS 03-419: Received 29 September 2003/Accepted 24 September 2004 ABSTRACT A quantitative survey of Clostridium perfringens in typical foods served at local restaurants was conducted for 18 months in Guadalajara, Mexico. A total of 151 samples, including goat's birria (50), pozole (50), and beef tamales (51), were collected from small restaurants in Guadalajara. Samples were tested for C. perfringens by the most probable number (MPN) method and for mesophilic aerobic plate counts (MAPCs) and coliform, yeast, and mold counts by plate count methods. Isolates con®rmed as C. perfringens were further sporulated and tested for cytotoxic or cytotonic effect against Vero cells as an indication of enterotoxin production. C. perfringens was detected in 78 (52%) of all samples at concentrations that ranged from 2.3 to 5.4 log MPN/g. Average MAPCs were 1.3 to 2.7 log CFU/g, depending on the type of dish. Coliform counts ranged from less than 1.0 to 1.5 CFU/g, and yeast and mold counts were less than 1.0 log CFU/g in all cases. A total of 118 isolates of C. perfringens were tested for enterotoxic effect on Vero cells; 82 (70%) showed activity against Vero cells. Of them, 31 isolates induced cell lysis, indicating cytotoxic effect; 41 induced cell elongation, indicating cytotonic effect; and 10 produced both cytotoxic and cytotonic effect. Dilution of the bacterial ®ltrates that were still producing an effect on Vero cells ranged from 1:80 to 1:5,120. These results underscore the importance of determining enterotoxigenicity when testing for C. perfringens in foods. The reported frequency of cases and outbreaks of food- the high protein and water content and a pH suitable for borne infections and intoxications is increasing in industri- bacterial growth. Other local foods where C. perfringens alized and developing countries. According to the Centers may be a signi®cant hazard if involving temperature abuse for Disease Control and Prevention, most outbreaks are are pozole and tamales. Pozole is a broth that contains hom- caused by biological hazards, especially pathogenic bacte- iny, and, immediately before serving, cooked pork and ria, with Salmonella, Staphylococcus aureus, and Clostrid- shredded cabbage or lettuce are added. The pork is usually ium perfringens being among the most common etiologic held at room temperature to speed the tempering with the agents of foodborne disease (2). In The Netherlands, 2,621 hot broth, which may be a contributing factor for C. per- incidents of foodborne illness caused by different agents fringens spores to germinate and grow. Pozole with meat were reported between 1991 and 1994, involving 7,567 ill and vegetables is usually taken home for further serving, people. Of the outbreaks where the etiologic agent was and temperature abuse may promote the growth of C. per- identi®ed, 11% were caused by C. perfringens (15). Most fringens (3). of the foodborne outbreaks that were caused by C. perfrin- An enterotoxin released from C. perfringens during gens between 1988 and 1992 in the United States were sporulation is responsible for diarrhea (5). The ability of C. related to eating meat products, including beef, chicken, perfringens isolates to produce enterotoxin has been inves- and turkey; at least eight of these outbreaks were linked to tigated using serological or molecular methods or by tissue Mexican food (9). culture assays. African green monkey kidney cells (Vero Two large outbreaks of C. perfringens toxic infection cells) are sensitive to the effect of C. perfringens entero- were recently reported in Mexico, implicating 1,777 cases toxin, and the Vero cell cytotoxin assay has been described (1, 3). In one of them, birria, a typical Mexican dish made as a convenient test for C. perfringens enterotoxin detection with cooked goat or beef and a unique tomato gravy, was (7). The effects induced by C. perfringens enterotoxin on identi®ed as the vehicle of the pathogen. This dish is com- Vero cells include morphological changes, detachment from monly served in gatherings. The meat is cooked overnight glass surfaces, decreased viability and plating ef®ciency, and the gravy boiled for several hours. However, if tem- and altered macromolecular synthesis (8). perature abuse occurs, pathogens can grow rapidly due to Laboratory testing for C. perfringens in foods is usu- * Author for correspondence. Tel: 52(33) 3650-0374; Fax: (0133) ally performed during outbreak studies and not for routine 36500374; E-mail: [email protected]. food testing. Therefore, few studies are available on the 332 NAVARRO-HIDALGO ET AL. J. Food Prot., Vol. 68, No. 2 prevalence or counts of this pathogen in foods. The objec- Becton Dickinson) for further determination of their effect on tive of this work was to determine the frequency and con- Vero cells. centration of C. perfringens in birria, pozole, and tamales Another portion of 15 g of sample was weighed, placed in a sold in Guadalajara City, Mexico, and to determine the abil- sterile stomacher bag, added to 135 ml of sterile 0.1% peptone ity of C. perfringens strains isolated from these foods to water, and then pummeled for 1 min in a stomacher 400 mixer (Tekmar Co.). Decimal dilutions were then plated for mesophilic produce enterotoxin as determined by the observation of aerobic plate count (MAPC) by pour plate in a standard method cytotoxic or cytotonic activity against Vero cells. agar (BBL, Becton Dickinson), incubating at 358C for 48 h; total MATERIALS AND METHODS coliform count by pour plate in violet red bile agar (BBL, Becton Dickinson), incubating at 358C for 24 h; and yeast and mold Sample collection and preparation. One hundred ®fty-one counts by pour plate method in potato dextrose agar (BBL, Becton samples were collected from small restaurants in Guadalajara City, Dickinson) added to ampicillin (200 mg/liter, Sigma Chemical Mexico, including birria (50 samples), pozole (50 samples), and Co., St. Louis, Mo.) and rose bengal (600 mg/liter, Sigma), in- beef tamales (51 samples). These foods are usually served at res- cubating at 208C for 72 to 196 h. The pH of the samples was Downloaded from http://meridian.allenpress.com/jfp/article-pdf/68/2/331/1675907/0362-028x-68_2_331.pdf by guest on 28 September 2021 taurants dedicated only to each particular food commodity; there- determined in a homogenate of 5 g of sample in 10 ml of recently fore, 25 restaurants were sampled twice for each commodity. boiled water at room temperature using a Conductronic 10 pH These restaurants were randomly selected for sampling between meter (Conductronic, Puebla, Mexico). October 1998 and April 2000. The temperature at purchase time was recorded for each sample using a K-type thermocouple con- Vero cell assay for enterotoxin. Each C. perfringens isolate nected to a Fluke 50 thermometer (Fluke Corp., Everett, Wash.). was thawed at room temperature for 1 h and then reactivated by Each sample was placed inside a sterile stomacher bag, sealed growing two consecutive times in tryptic soy broth (BBL, Becton with a rubber band, and transported to the laboratory at room Dickinson) with 0.6% yeast extract, incubating at 378C for 24 h temperature for analysis within 1 h after collection. The likelihood in an anaerobic jar with the gas-pak system. The isolates then for C. perfringens to grow in the samples during 1 h at an esti- were transferred to Duncan and Strong's sporulation medium mated room temperature of 308C was modeled using the U.S. modi®ed (11) and incubated at 378C for 24 h in an anaerobic Department of Agriculture's Pathogen Modeling Program V6.1 atmosphere. A supernatant of these isolates was obtained by cen- and determined to be negligible. In the laboratory, samples were trifuging for 10 min at 1,700 3 g in a Centra CL-2 centrifuge weighed as follows: birria, 23 g of meat and 2 ml of gravy; pozole, (Termo IEC, Needham Heights, Mass.) and ®ltering through a 15 g of hominy, 7.5 g of pork, and 2.5 ml of broth; and tamales, 0.22-mm membrane ®lter (Millipore, Bedford, Mass.). The spore- 25 g of the inside meat. Each 25-g sample was placed in a sterile and cell-free supernatants were stored at 48C for 24 to 48 h. Before stomacher bag and added to 225 ml of sterile 0.1% peptone water testing against Vero cells, the supernatants were centrifuged again and homogenized by manual rubbing and shaking for 90 s. This for 10 min at 4,000 3 g in a Microfuge E centrifuge (Beckman homogenization procedure was chosen after preliminary studies Instrument, Inc., Palo Alto, Calif.) and ®ltered through a 0.22-mm showed that manual sample mixing yielded higher counts than membrane ®lter (Millipore). A strain of enterotoxigenic Esche- pummeling with a stomacher 400 mixer (Tekmar Co., Cincinnati, richia coli (provided by Dr.