Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1229-1153(Print)
ISSN : 2465-9223(Online)
Journal of Food Hygiene and Safety Vol.32 No.4 pp.249-253

Comparison of TEMPO BC and MYP Plate Methods for the Enumeration of Bacillus cereus in Various Foods

Da Yeon Lee, Hee Yeon Kim1, Yong Sun Cho*
Korea Food Research Institute, Food Analysis Center, Seongnam, Korea
1KOTITI Testing & Research Institute, Microorganism Analysis Team, Seongnam, Korea
Correspondence to: Yong Sun Cho, Korea Food Research Institute, Food Analysis Center, 516, Backyeon-dong, Bundang-gu, Seongnam- si, Gyeonggi-do 13539, Korea 82-31-780-9242,
20170322 20170510 20170719


This study aimed to compare the automated most-probable-number (MPN) TEMPO BC and the quantitative mannitol-egg yolk-polymyxin (MYP) plate methods for enumeration of Bacillus cereus in food samples known to be frequently contaminated. Food products that were naturally or artificially contaminated with B. cereus were analyzed by both methods. A difference of less than 1 log (CFU/g) between the two methods was noted in 95.3% samples. There were no significant differences in artificially contaminated products between the two methods in terms of R2 values for sauce products, jorim products, fish products, etc. However, a significant difference was noted for sunsik, fermented soybean products, and products. The linear equation of naturally versus artificially contaminated food was log (TEMPO BC) = 0.8453 × log (MYP plate agar) + 0.1642. Statistical analysis of the results showed good agreement between the two methods. Due to growing interest in food safety, the use of the TEMPO BC method may increase. In response to this trend, the results from this study will offer valuable comparative data on the feasibility of existing methods and help develop new approaches for food safety testing.


    Ministry of Science, ICT and Future Planning

    Bacillus cereus is a well-known food pathogen that causes emesis and diarrhea. Though both forms of food poisoning are normally self-limiting, occasional fatal outcomes have been reported1). The Bacillus cereus group consists of six different closely related species: B. anthracis, B. cereus, B. mycoides, B. pseudomycoides, B. thuringiensis, and B. weihenstephanensis. Of these, B. cereus is most commonly associated with foodborne illness. B. cereus is a facultative anaerobic, spore-forming, motile microorganism2).

    Traditional methods for enumeration of quality indicators such as B. cereus in food are laborious and materialintensive. In addition, quality assurance in the food industry requires rapid testing methods. Moreover, since existing tests consider a particular growth phase of the candidate microorganisms, factors such as dilution factor need to be considered as well, making the process time-consuming and requiring technical skill. This has given rise to several alternative and rapid methods3). The most-probable-number (MPN) method requires inoculation of multiple tubes with serially diluted sample, and the statistical basis for this test allows quantification of even a low number of target microorganisms 4). However, this method is particularly complicated; therefore, a computer operated automated system (TEMPO; Biomériux, Marcy L’Etoile, France) was based on 16 MPN technique and reduce the effort required to estimate bacterial populations5). The instrument employs a unique microchanneled card into which the sample is introduced under vacuum, automatically creating the serial dilutions necessary for MPN estimation. Samples are then added to vials of prepared dehydrated medium and introduced onto cards in an automated vacuum chamber. The cards are then removed from the vacuum chamber and incubated as instructed. After incubation, the cards are placed in a reading chamber equipped with a detector capable of determining fluorescence associated with growth of organisms in specific channels on the cards. Subsequently, MPN is automatically calculated using computer software6,7). The most important parameters for feasibility of quantifying quality-related microorganisms in various foods are short turn-around time and ease of use, particularly due to large numbers of samples to be collected and processed each day. This has driven further studies on the TEMPO system to monitor and control the hygienic quality of food products.

    Therefore, the aim of this study was to compare and analyze results of quantitative estimation of B. cereus in various food samples currently available using the TEMPO B. cereus (BC) system and the mannitol-egg yolk-polymyxin (MYP) plate method accoding to the Korea Food Code8).

    Materials and Methods


    In this study, both naturally contaminated samples (samples without inoculation) and artificially contaminated food samples were used. A total of 616 samples were collected from different retail outlets in Korea and were immediately transported to the laboratory in insulated cooler boxes. Samples were stored at 4°C until analysis. All samples were analyzed by both the TEMPO BC and ISO 7932 MYP plate methods. Naturally contaminated samples consisted of fish products (n = 96), meat products (n = 15), vegetable products (n = 33), sunsik (a powdered mixture of roasted grains and other foods; n = 24), fermented soy products (n = 4), red pepper products (e.g., red pepper, red pepper paste, etc.; n = 109), products containing oil (n = 18), bakery and rice cakes (n = 45), ready-to-eat foods (n = 21), and prepared foods (n = 19), as listed in Table 1. Artificially contaminated samples (n = 232) consisted of special purpose products (infant formula), sauce products, red pepper products, bean paste products, fermented soybean products, fish products, pickled products, sunsik, kimchi products, and jorim products. Descriptions of individual types of food used in this study are given below.

    Artificial contamination of samples

    B. cereus (ATCC 21772) was used for artificial contamination of samples. Stock cultures of B. cereus were grown on nutrient agar (Merck Darmstadt, Germany) for 24 h, and a suspension was prepared in Butterfield’s phosphate buffered dilution water (BPD; Difco, Detroit, MI, USA). The dilutions included blank (0 CFU/g), low level (1-10 CFU/g), medium level I (10-100 CFU/g), medium level II (100-1000 CFU/g), and high level (1000-10000 CFU/g). Food samples were prepared by homogenizing 25 g food in 225 mL phosphate-buffered saline in a stomacher (Seward, London, UK) at 260 rpm for 1 min, followed by quantitative analysis by both TEMPO BC and MYP plate methods9). The sample of 25 g for the test using MYP plates was diluted in 225 mL BPD. For the MYP plating method, homogenized samples were serially diluted, and 0.2 mL was spread on MYP growth medium (Merck Darmstadt) for a total of five times to achieve a total inoculum of 1 mL. Plates were incubated for 24 h at 30°C, and pink colonies with a hazy zone surrounding them due to lecithinase production were counted. From these plates, more than five typical colonies were selected and transferred to nutrient agar medium (Merck Darmstadt) and incubated at 30°C for 18-24 h. Colonies were identified using VITEK II (bioMerieux). The enumeration of B. cereus in the food sample was based on the percentage of colonies that were morphologically consistent with B. cereus. For example, in addition to an average count of 65, obtained with 10−4 dilution of the sample, if four out of five colonies tested were confirmed as B. cereus, the B. cereus CFU/g of food was calculated as = 65 × 4/5 × 10000 × dilution factor (10) = 5,200,00010).

    TEMPO BC method

    The samples were processed for the TEMPO BC method by mixing 1 mL serially diluted sample with 3 mL buffer in a TEMPO BC medium vial. The mixture was then injected onto a BC card using the TEMPO preparation system and incubated at 30°C for 24 h. Results were obtained from the TEMPO read system.

    Statistical analysis

    Bacterial counts were converted to logarithmic form. Values obtained within the enumeration ranges for both methods were subjected to Pearson correlation coefficient and linear regression analysis. All statistical analyses were performed using Microsoft Excel software. The significance level was defined as p < 0.05. If the absolute value of the correlation coefficient, R2 was > 0.9, the values measured by the two methods were considered as correlated. Concordance analysis was performed using the method described in the Campden and Chorleywood Food Research Association (CCFRA) Guideline number 2911) and agreement indicated that the difference between the two methods was lower than 1 log base 10.

    Results and Discussion

    Quantitative test results of samples

    Contamination by B. cereus in samples (Table 1) of various commercially available food products was quantitatively compared using the TEMPO BC and ISO 7932, MYP plate methods. Of the 384 samples tested, 364 samples (95.3%) showed a difference of less than 1 log CFU/g, while 20 samples (5.2%) showed a difference of greater than 1 log CFU/g (Table 2). The latter consisted of five rice cake samples (23.8%) out of a total of 21 rice cakes samples, as shown in Table 3. The difference in quantitative results from TEMPO BC and MYP plate methods was between log 2.1 CFU/g and log 3.6 CFU/g. Additionally, the MYP plate method showed a higher count of B. cereus. Red pepper products (food products containing red pepper paste or red pepper powder) also showed discrepancies (Table 3). The principle underlying the TEMPO BC testing method is measuring fluorescence generated by enzymes produced due to proliferation of bacteria on a TEMPO card. It is thought that the matrix contained in rice cakes may have influenced the results from the TEMPO BC method. According to the manufacturer instructions7,12), the recommended dilution ratio for food products containing buckwheat flour, bean flour, or dietary fiber is 1:200. Additionally, it is not recommended to use TEMPO BC for enumeration of the B. cereus group in buckwheat flour, soy flour, and external pea fiber. Therefore, quantitative analysis using these products should be performed with consideraion of the dilution ratio.

    Quantitative test results of artificially contaminated samples

    Quantitative tests were performed on food products (10 types) artificially contaminated with B. cereus using the MYP plate and TEMPO BC methods, and the results were analyzed using paired t-tests. The results showed that statistically no significant (p > 0.05) for sauce products (R2= 0.98), jorim products (R2= 0.98), fish products (R2 = 0.96), special purpose products (R2= 0.95), pepper paste products (R2 = 0.95), kimchi products (R2 = 0.92), and pickled products (R2= 0.90). In contrast, the results for sunsik (R2= 0.94), fermented soybean products (R2 = 0.86), and bean paste products (R2 = 0.71) showed significant differences between the two test methods (p < 0.05). A comparison of the two test results for each food type could yield more precise results by further correction with linear regression equation (Table 4).

    Sunsik, fermented soybean products, and bean paste products may provide false-positive results in the fluorescence reaction. Fermented soybean products often harbor diverse Bacilli accumulated after fermentation and aging13), which makes it difficult to morphologically select and enumerate B. cereus using the MYP plate method. The TEMPO BC method quantitatively analyzes B. cereus groups, and in cases where the sample is heavily contaminated, only B. cereus would be identified. This may have contributed to discrepancies between the results from these two methods. The default dilution used in this study was 1:4 (detection limit < 10 CFU), and all food types were compared and analyzed according to the type. The reason we diluted to 1: 4 in our study was to minimize detection limit. However, the results showed a lot of difference from the MYP method. Therefore, in order to apply the automated TEMPO testing method on sunsik, fermented soybean products, and bean paste products, it is necessary to perform an additional test with a dilution ratio of 1:200 (detection limit < 50 CFU), as described in the manufacturer’s instructions.

    Statistical analysis of various food products using TEMPO BC and MYP plate methods

    Linear regression analysis was performed on the comparison of the two testing results applied on various natural food product samples and artificially contaminated samples (n = 616). The linear equation was log(TEMPO BC) = 0.8453 × log(MYP plate method) + 0.1642, and the R2 value was 0.803. The results from the paired t-test showed no significant difference (correlation coefficient R2 > 0.9, and p > 0.05). A high correlation was obtained between the two methods except for a few food products.

    Due to the recent surge in the number of food samples received for testing for contamination with B. cereus to determine food safety levels, prompt test results are required. However, current standard methods, such as the MYP plate method, require a minimum of 4 days for the results to be available. Based on the results from this study, using the automated TEMPO BC apparatus will not only shorten the turn-around time required to just 1 day but also reduce the requirement for trained personnel. This method also offers the benefit of increased reliability of results due to automation and subsequent reduction in human error. However, additional analysis and testing are needed based on the characteristic of testing products because the two quantitative test methods showed a low level of agreement for Korean traditional food products with high levels of contamination by B. cereus.

    With growing consumer interest in food safety, the Korean Food Code and international standards are gaining importance. Consequently, the use of TEMPO BC testing is expected to increase as well. The results from this study will offer valuable comparative data on the feasibility of existing methods and help develop new approaches for food safety testing.


    This research was supported by the Main Research Program (E0152202-02) of the Korea Research Food Institute (KFRI) funded by the Ministry of Science, ICT & Future Planning.



    Category and number of samples

    Agreement between TEMPO BC and MYP plate methods

    1)Difference of less than 1 log between the methods.
    2)Data for which at least one of the two methods gave results of “less than or more than”.
    3)Numerical data that could be transformed in log10 for both methods.

    Discrepancies between TEMPO BC and the MYP plate methods

    Statistical results of samples artificially injected with Bacillus cereus per food product type


    1. Fricker M , Reissbrodt R , Ehling-Schulz M (2008) Evaluation of standard and new chromogenic selective plating media for isolation and identification of Bacillus cereus , Int. J. Food Microbiol, Vol.121 ; pp.27-34
    2. Ankolekar C , Rahmati T , Labbé RG (2009) Detection of toxigenic Bacillus cereus and Bacillus thuringiensis spores in US rice , Int. J. Food Microbiol, Vol.128 ; pp.460-466
    3. Torlak E , Akan IM , Gokmen M (2008) Comparison of TEMPO EC and TBX medium for the enumeration of Escherichia coli in cheese , Lett. Appl. Microbiol, Vol.47 ; pp.566-570
    4. Zitz U , Domig KJ , Hoehl A , Weiss H , Wilrich PT , Kneifel W (2011) Evaluation of three application of a semi-automated most-probable-number method for the assessment of microbiological parameters in dairy products , Accred. Qual. Assur, Vol.16 ; pp.299-309
    5. Katase M , Tsumura K (2011) Enumeration of micro-organisms in processed soy products with an automated most probable number method compared with standard plate method , Lett. Appl. Microbiol, Vol.53 ; pp.539-545
    6. Line JE , Stern NJ , Oakley BB , Seal BS (2011) Comparison of an automated most-probable-number technique with traditional plating methods for estimating populations of total aerobes, coliforms Escherichia coli associated with freshly processed broiler chickens , J. Food Prot, Vol.74 ; pp.1558-1563
    7. Kim YJ , Wee SH , Yoon HC , Heo EJ , Park HJ , Kim HJ , Moon JS (2012) Comparison of an automated most-probable- number technique TEMPO® TVC with traditional plating methods PetrifilmTM for estimating populations of total aerobic bacteria with livestock products , J. Food Hyg. Saf, Vol.27 ; pp.103-107
    8. (2016) Available from: Accessed November 14,
    9. Schulten SM , in't Veld PH , Nagelkerke NJD , Scotter S , de Buyser ML , Rollier P , Lahellec C (2000) Evaluation of the ISO 7932 standard for the enumeration of Bacillus cereus in foods , Int. J. Food Micriobiol, Vol.57 ; pp.53-61
    10. (2016) Bacteriological Analytical Manual, Available from: Accessed November 14
    11. (2001) Guideline 29. Guideline establishing the suitability of food microbiological methods, Campden and Chorleywood Food Research Association Group,
    12. Owen M , Willis C , Lamph D (2010) Evaluation of the TEMPO® most probable number technique for the enumeration of Enterobacteriaceae in food and dairy products , J. Appl. Microbiol, Vol.109 ; pp.1810-1816
    13. Yun SH , Kim YS , Jeong DY , Hahn KS , Uhm TB (2009) Reevaluation of enumeration of Bacillus cereus grown on mannitol-egg york-polymyxin B agar , Korean J. Microbiol, Vol.42 ; pp.208-214