J. Microbiol. Biotechnol.

(2012), 22(4), 530–534

http://dx.doi.org/10.4014/jmb.1107.07064
First published online February 3, 2012
pISSN 1017-7825 eISSN 1738-8872

Development of Real-Time PCR for the Detection of Clostridium perfringens

in Meats and Vegetables
Chon, Jung-Whan1†, Jong-Seok Park2†, Ji-Yeon Hyeon1, Chankyu Park3, Kwang-Young Song1,
Kwang-Won Hong4, In-Gyun Hwang5, Hyo-Sun Kwak6, and Kun-Ho Seo1*
1
KU Center for Food Safety, Veterinary Science Research Institute and College of Veterinary Medicine, Konkuk University, Seoul
143-701, Korea
2
Division of Research Planning and Management, Korea Food and Drug Administration, Cheong-won 363-951, Korea
3
Department of Animal Biotechnology, College of Animal Bioscience and Technology, Konkuk University, Seoul 143-701, Korea
4
Department of Food Science and Technology, Dongguk University, Seoul 100-715, Korea
5
Division of Microbiology, Korea Food and Drug Administration, Cheong-won 363-951, Korea
6
Division of Foodborne Diseases Prevention and Surveillance, Korea Food and Drug Administration, Cheong-won 363-951, Korea
Received: July 29, 2011 / Revised: December 6, 2011 / Accepted: December 7, 2011

A real-time PCR assay was developed and validated in- samples and foods because of its reliability, it is time and
house specifically for the detection of Clostridium perfringens labor intensive [4, 8]. Polymerase chain reaction (PCR) is
(Cl. perfringens) in meats and vegetables by comparing with one of the most commonly used methods for the detection
the culture method. The detection limit of the real-time of food-borne bacteria [10]. Real-time PCR using fluorescent
PCR assay in phosphate-buffered saline was 102 CFU/ml. reaction of amplified products is known to have high
When the two methods were compared in food samples sensitivity and efficiency among different types of PCR
inoculated with Cl. perfringens, the culture method assays that detects organisms within 24 h, including the
detected 52 positives, whereas real-time PCR detected 51 enrichment step [4, 6].
positives out of 160 samples. The difference was without Although detection of Cl. perfringens by real-time PCR
statistical significance (p>0.05). Real-time PCR assay is an has been applied in various studies, particularly in feces or
option for quality assurance laboratories to perform animal gut [1, 7, 16], few studies have been conducted on
standard diagnostic tests, considering its detection ability food samples. In this study, one primers/probe set was
and time-saving efficiency. designed to target α toxin, which is produced by all five
Keywords: Clostridium perfringens, real-time PCR, culture types of Cl. perfringens, for the development of a real-time
method, foods, detection PCR assay [13]. The real-time PCR was evaluated using
the primers/probe set in sensitivity and specificity tests.
Real-time PCR was also compared with the standard
culture method in the detection of Cl. perfringens in
Clostridium perfringens (Cl. perfringens) can cause necrotic various food types composed of different matrix and
enteritis, enterotoxemia, and hamorrhagic gastroenteritis in background microflora, meats and vegetables.
animals, while the bacterium is also a food-borne pathogen Five of Cl. perfringens and 11 of non-Cl. perfringens
that causes diarrhea and abdominal pain in humans [5, 13]. strains were used in this study (Table 1). All non-Clostridium
The bacterium is commonly found in vegetables and crops spp. were streaked onto nutrient agar (NA; Difco, Becton
as well as meat products [2, 9]. When cooked foods are Dickinson, Sparks, MD, USA) for two passages and
gradually cooled, heat-stable spores of Cl. perfringens may incubated in tryptic soy broth (Difco), whereas Clostridium
germinate and proliferate [14]. spp. strains were streaked onto 5% horse blood agar
Although the culture method is regarded as a standard (Oxoid, Hampshire, UK) for two passages and incubated
method for the detection of Cl. perfringens from clinical in a broth, a cooked meat medium (CMM; Oxoid) at 37oC
*Corresponding author for 24 h. Culturable Cl. perfringens counts were obtained
Phone: +82-2-450-4121; Fax: +82-2-450-3037; by plating 100 µl of the inocula onto 5% horse blood agar
E-mail: [email protected] and incubated at 37oC for 24 h under anaerobic condition.
†
These authors contributed equally to this work. Whenever necessary, cultures in CMM were serially diluted
531 Chon et al.

Table 1. Sensitivity and specificity of primers/probe. TaqMan probe (2.5 µl, 250 nM). The 96-microwell plate
Reference strain Results sealed with optical adhesive covers (Applied Biosystems)
was placed in an ABI-Prism 7500 sequence detector
Clostridium perfringens ATCC 3624 +
(Applied Biosystems). The reaction was run at 50oC for
Clostridium perfringens ATCC 13124 +
2 min and then 95oC for 10 min, followed by 40 cycles of
Clostridium perfringens PH1 +
Clostridium perfringens PH2 +
95oC for 15 s and 60oC for 60 s.
Clostridium perfringens PH3 + For sensitivity and specificity test of the designed
Clostridium difficile ATCC 9689 - sequences, the reaction of real-time PCR was examined in
Clostridium butyricum ATCC 19398 - 16 strains, 5 of Cl. perfringens and 11 of non-Cl.
Clostridium ramosum ATCC 25582 - perfringens. To determine the detection limit of real-time
Clostridium paraputrificum ATCC 25780 - PCR in PBS, genomic DNA was extracted as described
Escherichia coli O157:H7 ATCC 43888 - above from diluted overnight culture containing 105 CFU
Salmonella Enteritidis ATCC 10708 - per milliliter. The extracted DNA was then serially diluted
Bacillus cereus ATCC 11778 - (10-fold) in PBS, and the cycle threshold value (Ct value)
Yersinia enterocolitica ATCC 23715 - of the dilutions was measured by real-time PCR. The
Staphylococcus aureus ATCC 25923 - detection limit of real-time PCR using the primers/probe
Listeria monocytogenes ATCC 19111 - set was also measured in various foods. One strain producing
Vibrio parahaemolyticus ATCC 33844 - α toxin, ATCC 3624, was used. Inocula (1 ml each)
containing 1.8 × 102~1.8 × 108 CFU of ATCC 3624 were
serially spiked into 10 g of foods, making the final concentrations
in phosphate-buffered saline (PBS; pH 7.4; Sigma, St. of Cl. perfringens 1.8 × 101 to 1.8 × 107 CFU/g. Each
Louis, MS, USA) as needed, and Cl. perfringens numbers inoculated sample was put into 90 ml of 0.85% saline
were enumerated as above. water, followed by homogenizing for 30 s using a BagMixer
The cpa gene, an α toxin producing gene, was targeted stomacher (Interscience, St. Nom, France) and real-time
for the design of primers and probe. The sequence of the gene PCR was performed with genomic DNA extracted from 1 ml
was provided by GenBank (http://www.ncbi.nlm.nih.gov/ of each diluted sample (1.8 × 100 to 1.8 × 106 CFU/ml) as
Genbank/; Accession Number: X13608) and one primers/ described above. The lowest number of bacteria showing
probe set was designed using the Primer Express Software positive reaction was determined as the detection limit of
(Applied Biosystems, Foster City, CA, USA). The set was real-time PCR.
validated at the NCBI Web site (http://www.ncbi.nlm.nih.gov/ Foods with different matrix and background microflora,
blast/) using BLAST (Basic Local Alignment Search meats and vegetables, were used to determine the difference
Tool). The Cl. perfringens probe was labeled with 6- in detection ability of the real-time PCR and culture
carboxyfluorescein (FAM, the reporter dye) and 6- method. Meat samples included steamed pork and ground
carboxytetramethylrhodamine (TAMRA, the quencher beef, and vegetable samples included canned bean and
dye). The sequences were as follows: cpa (amplicon size vegetable mix composed of cabbage, cucumber, carrot,
85 bases); forward primer: 5'-AAA AGA AAG ATT TGT and lettuce. All samples were purchased from a local retail
AAG GCG CTT AT-3'; reverse primer: 5'-CCC AAG CGT market in Seoul, Korea. A mesophilic aerobic plate count
AGA CTT TAG TTG ATG-3'; probe: 5'-FAM TGC CGC was performed on uninoculated food samples according to
GCT AGC AAC TAG CCT ATG G -3'TAMRA. The a previously published study [4].
oligonucleotides and all reagents for PCR reaction used in One milliliter of the inoculum prepared by serial dilution
this study were synthesized and purchased from Bioneer of the overnight culture of CMM was spiked into 200 g of
Corporation (Daejeon, Korea). bulk samples evenly by pipetting to generate partial positives
DNA templates of the bacteria were extracted with and partial negatives for statistical comparison when divided
reference to Seo and Brackett [15]. One milliliter of into 20 subsamples. The concentrations of the inoculums
sample from pure cultures in PBS or food samples in ranged 820~3,600 CFU/ml and 1,140~15,140 CFU/ml of
enrichment broth was centrifuged at 14,000 rpm for 3 min. Cl. perfringens for bulk samples of meats and vegetables,
The pellets were resuspended in 200 µl of PrepMan Ultra respectively. The inoculated bulk sample was stabilized at
Reagent (Applied Biosystems) and boiled for 10 min. The 4oC for 24 h, and subsequently divided into 20 subsamples
samples were centrifuged at 14,000 rpm for 3 min. The of 10 g each. Two additional food samples (10 g each)
supernatant fluids were used for real-time PCR assay. were used as positive and negative controls. A positive
The extracted DNA fluids (5 µl) were transferred into control was prepared by spiking 10 g of the sample with
20 µl of PCR mix consisted of TaqMan Universal PCR approximately 107 CFU/ml of Cl. perfringens ATCC 3624.
Master Mix (Applied Biosystems, 12.5 µl), forward primer For negative control, uninoculated food (10 g) and sterilized
(2.5 µl, 900 nM), reverse primer (2.5 µl, 900 nM), and PBS (1 ml) were also prepared. All experiments were
 DETECTION OF CL. PERFRINGENS WITH REAL-TIME PCR 532

Table 2. Detection limits of Cl. perfringens ATCC 3624 using The number of background microflora and detection limit
designed primers/probe set in pure culture. of the real-time PCR in each food sample are presented in
Strain CFU/ml Results Table 3. As determined by aerobic plate counts, the number of
105
+ background microflora was 5.98 log CFU/g, 4.48 log CFU/g,
104 + and 3.44 log CFU/g in vegetable mix, ground beef, and
Cl. perfringens ATCC 3624 103 + steamed pork, respectively. From canned bean, less than
102 + 2 log CFU/g of background microflora was detected. For the
101 - detection limit test, greater than 1.8 × 103 CFU of bacteria
was required for positive reaction with real-time PCR in all
tested foods. Lee et al. [11] found that the detection limit
conducted twice per one sample type at different inoculation of real-time PCR was higher in vegetable than other types
levels. Samples were put into 90 ml of 0.85% saline water of foods for the detection of S. aureus. They reported that
and homogenized for 30 s. After stomaching, 1 ml of the detection limit of real-time PCR could be influenced by
sample was incubated in 9 ml of CMM at 37oC for 24 h. the matrix or background microflora of foods. In this
The incubated broth was streaked onto tryptose sulfite study, however, the detection limit of Cl. perfringens by
cycloserine (TSC; Oxoid) agar with 5% of egg yolk real-time PCR was identical in all types of foods (Table 3).
emulsion (Oxoid), followed by incubation at 37oC for 24 h It appears that the primers/probe sequences used in this
under anaerobic condition. Presumptive identification was study could be applied in various food samples with
based on colony morphology and aerobic/anaerobic growth. different matrixes and background microflora levels.
Suspicious colonies were finally confirmed by API 20A The performance of the real-time PCR and culture
strips (BioMérieux, Marcy l’Etoile, France). One milliliter method in food samples was compared in Table 4. No
of enrichment broth was collected and used for real-time positive reactions were obtained in the negative controls
PCR assay. Extraction of DNA templates and performance with the culture method and real-time PCR, whereas all
of real-time PCR was conducted as described above. positive controls were shown as positives with the two
The number of positives identified out of the total detection methods. Therefore, it was concluded that samples
number of samples was analyzed statistically by using the used in the experiments were not naturally contaminated
GraphPad Instat Software (GraphPad Software, San Diego, by Cl. perfringens, thereby excluding possible false
CA, USA). Differences between two methods were examined positives and negatives. In all food samples, the culture
for level of significance by using Fisher’s exact test. method detected 52 positives, whereas real-time PCR detected
Data providing the sensitivity and specificity of the 51 positives, out of 160 samples. There was no statistical
designated primers/probe set are presented in Table 1. difference between the two methods (p>0.05). It appears
With the exception of Cl. perfringens, there was no that the culture method and real-time PCR have similar
positive reaction with any unrelated bacteria. It provided capability in the detection of Cl. perfringens in foods
good sensitivity and specificity for the detection of Cl. regardless of the matrix and number of background
perfringens at the species level. microflora. Malorny et al. [12] detected Salmonella spp.
The detection limits of the primers/probe set in PBS are using the culture method and real-time PCR in various
listed in Table 2. In PBS, the detection limit of real-time foods. Real-time PCR showed 100% sensitivity and
PCR assay was approximately 102 CFU/ml. The detection specificity, compared with the culture method. Cai et al.
limits were also determined in artificially inoculated food [3] detected V. parahaemolyticus using the culture method
samples with different levels of background microflora.

Table 3. Detection limit of Cl. perfringens ATCC 3624 using a designed primers/probe set in experimentally inoculated food samples
without enrichment.
Number of cells Food samples (No. of background microflora, log CFU/g)
Strain
(CFU/ml) Steamed pork (3.44) Ground beef (4.48) Canned bean (< 2) Vegetable mix (5.98)
6
1.8 × 10 + + + +
1.8 × 105 + + + +
1.8 × 104 + + + +
Cl. perfringens
ATCC 3624 1.8 × 103 + + + +
1.8 × 102 - - - -
1.8 × 101 - - - -
1.8 × 100 - - - -
533 Chon et al.

Table 4. Comparison between the culture method and real-time PCR for the detection of Cl. perfringens in artificially inoculated foods.
Inoculation level Number of positive samples/total number of samples
Sample Trial p valuea
(CFU/g) Culture method Real-time PCR
Steamed 1 6.3 9/20 9/20 1.2488
pork 2 18.0 14/20 14/20 1.2689
Meat
Ground 1 4.1 3/20 3/20 1.3386
beef 2 8.2 5/20 5/20 1.2836
Canned 1 5.7 3/20 3/20 1.3386
bean 2 11.4 7/20 7/20 1.2589
Vegetables
Vegetable 1 27.0 1/20 0/20 1.0000
mix 2 75.7 10/20 10/20 1.2476
Total - 52/160 51/160 1.0000
a
There is statistical difference between the two methods if the p value is under 0.05.

and real-time PCR in various seafoods. Those outcomes 1. Albini, S., I. Brodard, A. Jaussi, N. Wollschlaeger, J. Frey, R.
correspond with that of the present study, suggesting real- Miserez, and C. Abril, 2008. Real-time multiplex PCR assays
time PCR’s superiority. for reliable detection of Clostridium perfringens toxin genes in
Previous studies reported that the cultured method’s animal isolates. Vet. Microbiol. 127: 179-185.
detection ability was profoundly diminished in fresh vegetable 2. Brynestad, S. and P. E. Granum. 2002. Clostridium perfringens
samples, compared with real-time PCR, owing to high and food borne infections. Int. J. Food Microbiol. 74: 195-202.
levels of competing background microflora in fresh 3. Cai, T., L. Jiang, C. Yang, and K. Huang. 2006. Application of
vegetable that could hinder the selective growth of target real-time PCR for quantitative detection of Vibrio parahaemolyticus
organism in traditional culture media [8, 11]. In the present from seafood in eastern China. FEMS Immunol. Med. Microbiol.
study, however, there was no difference in the isolation 46: 180-186.
4. Chon, J. H., J. Y. Hyeon, I. G. Hwang, H. S. Kwak, J. A. Han,
rate between the two detection methods, regardless of
Y. H. Chung, et al. 2010. Comparison of standard culture method
types of food. It suggests that the selectivity of TSC agar, and real-time PCR for detection of Vibrio parahaemolyticus in
the standard selective medium for Cl. perfringens, was seafoods and vegetables. Korean J. Food Sci. Technol. 42:
excellent for selectively detecting the bacteria in various 355-360.
types of food samples while excluding any possible 5. Erol, I., M. Goncuoglu, N. D. Ayaz, F. S. Bilir-Ormanci, and G.
interference from normal background microflora. Hildebrandt. 2008. Molecular typing of Clostridium perfringens
In this study, real-time PCR showed excellent performance isolated from turkey meat by multiplex PCR. Lett. Appl.
that generated more positives, albeit without statistical Microbiol. 47: 31-34.
difference. Moreover, real-time PCR significantly saves 6. Fukushima, H., K. Katsube, Y. Hata, R. Kishi, and S. Fujiwara.
assay time and labor, compared with the standard culture 2007. Rapid separation and concentration of food-borne pathogens
method [4, 8]. It appears that real-time PCR is an effective in food samples prior to quantification by viable-cell counting
and real-time PCR. Appl. Environ. Microbiol. 73: 92-100.
and sensitive presumptive screening tool for Cl. perfringens
7. Gurjar, A. A., N. V. Hegde, B. C. Love, and B. M. Jayarao.
in various types of food, complementing the standard 2008. Real-time multiplex PCR assay for rapid detection and
culture method. toxintyping of Clostridium perfringens toxin producing strains
in feces of dairy cattle. Mol. Cell. Probes 22: 90-95.
8. Hyeon, J. Y., I. G. Hwang, H. S. Kawk, J. S. Park, S. Heo, I. S.
Acknowledgments Choi, et al. 2009. Evaluation of an automated ELISA and real-
time PCR by comparing with a conventional culture method for
This research was supported by a grant from the Korea Food the detection of Salmonella spp. in steamed pork and raw
and Drug Administration in 2009 (09072_KFDA_029). broccoli sprouts. Korean J. Food Sci. Anim. Resour. 29: 506-
512.
Jung-Whan Chon and Ji-Yeon Hyeon were also partially
9. Jung, S. H., M. J. Hur, J. H. Ju, K. A. Kim, S. S. Oh, J. M. Go,
supported by the Brain Korean 21 (BK21) Project from the et al. 2006. Microbiological evaluation of raw vegetables. J.
Ministry of Education and Human Resources Development. Food Hyg. Safety 21: 250-257.
10. Lantz, P. G., R. Knutsson, Y. Blixt, W. A. Al-Soud, E. Borch,
and P. Rådström. 1998. Detection of pathogenic Yersinia
REFERENCES enterocolitica in enrichment media and pork by a multiplex PCR,
a study of sample preparation and PCR-inhibitory components.
Int. J. Food Microbiol. 45: 93-105.
 DETECTION OF CL. PERFRINGENS WITH REAL-TIME PCR 534

11. Lee, J. H., K. Y. Song, J. Y. Hyeon, I. G. Hwang, H. S. Kwak, 14. Sartory, D. P., M. Field, S. M. Curbishley, and A. M. Pritchard.
J. A. Han, et al. 2010. Comparison of standard culture method 1998. Evaluation of two media for the membrane filtration
and real-time PCR assay for detection of Staphylococcus aureus enumeration of Clostridium perfringens from water. Lett. Appl.
in processed and unprocessed foods. Korean J. Food Sci. Anim. Microbiol. 27: 323-327.
Resour. 30: 410-418. 15. Seo, K. H. and R. E. Brackett. 2005 Rapid, specific detection of
12. Malorny, B., E. Paccassoni, P. Fach, C. Bunge, A. Martin, and Enterobacter sakazakii in infant formula using a real-time PCR
R. Helmuth, 2004. Diagnostic real-time PCR for detection of assay. J. Food Prot. 68: 59-63.
Salmonella in food. Appl. Environ. Microbiol. 70: 7046-7052. 16. Wu, S. B., N. Rodgers, and M. Choct. 2011. Real-time PCR
13. Petit, L., M. Gibert, and M. R. Popoff. 1999. Clostridium assay for Clostridium perfringens in broiler chickens in a
perfringens: Toxinotype and genotype. Trends Microbiol. 7: challenge model of necrotic enteritis. Appl. Environ. Microbiol.
104-110. 77: 1135-1139.