Surveillance of Bacillus cereus Isolates in Korea from 2012 to 2014.

OBJECTIVES: To investigate the prevalence and toxin production characteristics of non-emetic and emetic Bacillus cereus strains isolated via the laboratory surveillance system in Korea.
METHODS: A total of 667 B. cereus strains were collected by the Korea National Research Institute of Health laboratory surveillance system from 2012 to 2014. The collected strains were analyzed by geographical region, season, patient age, and patient sex. Additionally, the prevalence rates of enterotoxin and emetic toxin genes were evaluated.
RESULTS: The isolation rate of B. cereus strains increased during the summer, but the isolation rate was evenly distributed among patient age groups. Emetic toxin was produced by 20.2% of the isolated strains. The prevalence rates of five enterotoxin genes (entFM, nheA, cytK2, hblC, and bceT) were 85.0, 78.6, 44.5, 36.6, and 29.7%, respectively, among non-emetic strains and 77.8, 59.3, 17.8, 11.9 and 12.6%, respectively, among emetic strains. Thus, the prevalence rates of all five enterotoxin genes were lower in emetic B. cereus.
CONCLUSION: The prevalence of enterotoxin genes differed between non-emetic and emetic B. cereus strains. Among emetic B. cereus strains, the prevalence rates of two enterotoxin genes (cytK2 and hblC) were lower than those among the non-emetic strains. In both the emetic and non-emetic strains isolated in Korea, nheA and entFM were the most prevalent enterotoxin genes.

INTRODUCTION

Foodborne diseases represent a serious threat to public health. Diarrheal disease is significantly increasing in prevalence worldwide year on year [1]. The bacterium Bacillus cereus is widely present in nature and can survive in harsh environments. B. cereus causes two types of gastrointestinal diseases: emesis and diarrhea. It produces one emetic toxin (cereulide) and several enterotoxins (hblC, hblD, hblA, nheA, nheB, nheC, cytK2, entFM, and bceT). Both the emesis and diarrhea caused by B. cereus are generally mild and self-limiting, although more serious and even lethal cases have occurred [2-4]. The diarrheal type is attributed to single or multiple enterotoxins. Particularly, a group of proteins including two heat-labile toxins and a three-component hemolysin (HBL; consisting of three proteins: B, L1, and L2) with enterotoxin activity have been purified and characterized [5]. Additionally, non-hemolytic enterotoxin (NHE, encoded by nheA, nheB, and nheC) is a key component contributing to B. cereus-mediated diarrhea [6,7]. Furthermore, single-component toxins, such as enterotoxin T (bceT) [8], enterotoxin FM (ent-FM) [9], and cytotoxin K (cytK) [10] are thought to be involved in B. cereus food poisoning. The pore-forming toxin, cytK, has two different forms, cytK1 and cytK2, which have 89% amino acid sequence homology [11,12]. Emesis is caused by a single heat-stable toxin, cereulide, which is produced in food [8,13,14]. This toxin is enzymatically synthesized by non-ribosomal peptide synthesis, and its genetic determinants are located within a 23-kb gene cluster (ces) on a large plasmid [2,15-17].

Several countries have reported B. cereus outbreaks [18-20]. Vomiting-type food poisoning is 10 times more prevalent than diarrheal-type food poisoning in Japan. However, in North America and Europe, diarrheal-type B. cereus infection is most frequent [21]. In Korea, 27 food poisoning outbreaks associated with B. cereus were reported from 2001 to 2008, but few cases of vomiting-type food poisoning caused by B. cereus were reported [22].

Emetic-type food poisoning caused by B. cereus occasionally includes symptoms of vomiting and diarrhea [7]. Therefore, the characterization of the enterotoxins produced by emetic B. cereus is necessary to obtain a better understanding of the food poisoning caused by this organism and to prevent misdiagnosis between diarrheal and emetic food poisoning.

EnterNet-Korea, an acute diarrheal laboratory surveillance system, was established in 2007 to improve laboratory activities and enhance reporting proficiency. This surveillance system is coordinated by the Korea National Research Institute of Health (NIH) and comprises 17 local environmental and health institutes and 70 participating hospitals. The target pathogens include 10 genera of bacteria and 5 types of viruses.

In this study, we analyzed the isolation trends and toxin gene profiles of B. cereus strains obtained via Enter-Net Korea from 2012 to 2014.

MATERIALS AND METHODS

1. Isolation of B. cereus from clinical samples

A total of 57,050 stool samples were collected by Enter-Net Korea from 2012 to 2014. The Enter-Net system is coordinated by the Korean NIH and comprises 17 local public health institutes and 105 participating hospitals. Stool samples were collected from patients who had diarrheal symptoms, and a total of 667 B. cereus strains were isolated from these samples.

We determined the isolation rate of B. cereus from the stool specimens during each 12-month period. Next, we divided patients into eight categories by age (< 10, 10–19, 20–29, 30–39, 40–49, 50–59, 60–69, and > 70 years) and surveyed the age and gender distributions of B. cereus isolation rates.

Mannitol-egg yolk-polymyxin B agar (Oxoid, Basingstoke, UK) was used as a selective medium for B. cereus isolation. For primary identification, the isolates were characterized by standard physiological and biochemical tests using the API® 50CHB and API 20E® bacterial identification systems (bioMérieux, Marcy-l’Étoile, France).

2. Polymerase chain reaction (PCR) amplification of enterotoxin and emetic toxin genes

DNA from each isolate was extracted using the Maxwell® 16 System Purification Kit (Promega, Madison, WI, USA) in accordance with the manufacturer’s instructions. All PCRs were performed using the Expanded High Fidelity Polymerase System (Roche, Basel, Switzerland) or Taq polymerase (Takara Bio, Otsu, Japan) according to the manufacturer’s instructions. The PCR primer sequences used in this study are shown in Table 1. hblC, nheA, entFM, and bceT were amplified according to the methodology described by Yang et al [7]; cytK2 was amplified as described by Guinebretiere et al [12]; and cereulide peptide synthetase (cer) was amplified as described by Kim et al [23].

Table 1

Primers used to amplify the target genes

Target gene	Sequence (5′ to 3′)	Size (bp)	Reference
entFM	CAAAGACTTCGTAACAAAAGGTGGT	290	Yang et al, 2005 [7]
	TGTTTACTCCGCCTTTTACAAACTT
nheA	ATTACAGGGTTATTGGTTACAGCAGT	475	Yang et al, 2005 [7]
	AATCTTGCTCCATACTCTCTTGGATGCT
cytK2	CAATCCCTGGCGCTAGTGCA	585	Guinebretirere et al, 2006 [12]
	GTGTAGCCTGGACGAAGTTGG
hblC	CCTATCAATACTCTCGCAACACCAAT	386	Yang et al, 2005 [7]
	TTTTCTTGATTCGTCATAGCCATTTCT
bceT	AGCTTGGAGCGGAGCAGACTATGT	701	Yang et al, 2005 [7]
	GTATTTCTTTCCCGCTTGCCTTTT
cer	ATCATAAAGGTGCGAACAAGA	188	Kim et al, 2010 [22]
	AAGATCAACCGAATGCAACTG

3. Statistical analysis

The collected data were analyzed using IBM SPSS Statistics ver. 21.0 (IBM Co., Armonk, NY, USA). The distribution of toxin genes from non-emetic and emetic isolated strains was analyzed using the chi-square or chi-square trend test. For statistical analysis, differences at p < 0.05 were considered to be significant.

RESULTS

1. B. cereus isolation rates

A total of 667 (6.9%) B. cereus strains were isolated from 57,050 stool samples by Enter-Net. The isolation rate of B. cereus slightly increased in the summers of 2012 and 2013, with strong seasonality in 2014 (Figure 1A). The age distribution of isolation rates showed no specific trend from 2012 to 2014 in any of the eight age groups (Figure 1B). Lastly, we divided the isolation rates by gender and found that females showed a slightly higher isolation rate than males, albeit with no statistically significant difference (Figure 1C).

Figure 1

Isolation rates of Bacillus cereus by (A) year of isolation, (B) patient age, and (C) patient gender in Korea, 2012–2014.

2. Detection of toxin genes by PCR

In this study, five enterotoxin genes (hblC, bceT, cytK2, entFM, and nheA) and one emetic toxin gene (cer) were detected using PCR (Table 1). Among all 667 strains, the prevalence rates of entFM, nheA, cytK2, hblC, bceT, and cer were 85.0, 78.6, 44.5, 36.6, 29.7, and 20.2%, respectively (Table 2). The emetic toxin-producing B. cereus strains comprised 20.2% of total strains isolated (n = 135). The prevalence rates of the assessed enterotoxin genes (entFM, nheA, cytK2, hblC, and bceT) among the emetic toxin-producing B. cereus strains were 77.8, 59.3, 17.8, 11.9, and 12.6%, respectively. Lastly, the prevalence rates of the five aforementioned enterotoxin genes among the non-emetic B. cereus strains were 86.8, 83.5, 51.3, 42.9, and 34.0%, respectively (Table 2).

Table 2

Presence of enterotoxin genes in emetic and non-emetic Bacillus cereus

Target gene	entFM*	nheA*	cytK2*	hblC*	bceT*	cer*
Total strains (n = 667)	567 (85.0)	524 (78.6)	297 (44.5)	244 (36.6)	198 (29.7)	135 (20.2)
Non-emetic strains (n = 532)	462 (86.8)	444 (83.5)	273 (51.3)	228 (42.9)	181 (34.0)	0 (0)
Emetic strains (n = 135)	105 (77.8)	80 (59.3)	24 (17.8)	16 (11.9)	17 (12.6)	135 (100)

Values are presented as number (%).

p < 0.05 by chi-square test.

3. Profiling of toxin genes

According to the presence or absence of enterotoxin genes, B. cereus strains harboring emetic and enterotoxin genes in our study could be divided into 29 and 20 groups, respectively (Table 3). In the B. cereus strains harboring the emetic toxin gene, the most common toxin gene profile was group F (entFM+, nheA+, hblC−, cytK2−, and bceT−). Group F was detected in 41.5% of these strains. The next most common gene profile was group G (entFM+), which was detected in 16.3% of emetic toxin-producing strains. Approximately 2% to 4% of B. cereus strains harboring the emetic toxin gene contained two to six toxin genes in total (Table 3). B. cereus strains harboring the emetic toxin gene, which lacked cer genes, were more diverse than B. cereus harboring the emetic toxin gene. Among non-emetic B. cereus strains, the group H toxin profile, which lacked cytK2, hblC, and bceT, was the most common (detection rate, 23.5%). The second most common toxin profile group was group A (14.5%, containing all five enterotoxin genes). Approximately 0.2% to 13% of non-emetic strains contained one to five toxin genes.

Table 3

Enterotoxin gene profiles of non-emetic and emetic toxin strains

Strains	Group	entFM	nheA	cytK2	hblC	bceT	cer	Strains, n (%)
Non-emetic strains (n = 532)	A	+	+	+	+	+	−	77 (14.5)
	B	+	+	+	+	−	−	71 (13.3)
	C	+	+	+	−	−	−	35 (6.6)
	D	+	+	+	−	+	−	33 (6.2)
	E	+	+	−	+	+	−	9 (1.7)
	F	+	+	−	−	+	−	20 (3.8)
	G	+	+	−	+	−	−	25 (4.7)
	H	+	+	−	−	−	−	125 (23.5)
	I	+	−	+	+	+	−	6 (1.1)
	J	+	−	+	−	+	−	10 (1.9)
	K	+	−	+	+	−	−	6 (1.1)
	L	+	−	+	−	−	−	10 (1.9)
	M	+	−	−	+	+	−	1 (0.2)
	N	+	−	−	−	+	−	7 (1.3)
	O	+	−	−	+	−	−	5 (0.9)
	P	+	−	−	−	−	−	22 (4.1)
	Q	−	+	+	+	+	−	3 (0.6)
	R	−	+	+	+	−	−	6 (1.1)
	S	−	+	+	−	−	−	9 (1.7)
	T	−	+	−	+	+	−	2 (0.4)
	U	−	+	−	−	+	−	1 (0.2)
	V	−	+	−	+	−	−	3 (0.6)
	W	−	+	−	−	−	−	25 (4.7)
	X	−	−	+	−	+	−	2 (0.4)
	Y	−	−	+	+	−	−	1 (0.2)
	a	−	−	+	−	−	−	4 (0.8)
	b	−	−	−	+	+	−	9 (1.7)
	c	−	−	−	+	−	−	4 (0.8)
	d	−	−	−	−	+	−	1 (0.2)
Emetic strains (n = 135)	A	+	+	+	+	+	+	3 (2.2)
	B	+	+	+	+	−	+	4 (3.0)
	C	+	+	+	−	+	+	4 (3.0)
	D	+	+	+	−	−	+	1 (0.7)
	E	+	+	−	+	−	+	1 (0.7)
	F	+	+	−	−	−	+	56 (41.5)
	G	+	+	−	−	+	+	4 (3.0)
	H	+	−	+	+	−	+	4 (3.0)
	I	+	−	+	+	+	+	2 (1.5)
	J	+	−	+	−	−	+	1 (0.7)
	K	+	−	+	−	+	+	1 (0.7)
	L	+	−	−	+	+	+	1 (0.7)
	M	+	−	−	−	+	+	1 (0.7)
	N	+	−	−	−	−	+	22 (16.3)
	O	−	+	+	−	−	+	1 (0.7)
	P	−	+	−	+	−	+	1 (0.7)
	Q	−	+	−	−	−	+	5 (3.7)
	R	−	−	+	−	+	+	1 (0.7)
	S	−	−	+	−	−	+	2 (1.5)
	T	−	−	−	−	−	+	20 (14.8)

DISCUSSION

In our study, 667 (6.9%) B. cereus strains were isolated from 57,050 stool samples. The isolation rate of B. cereus increased with rising temperatures and peaked from June to September, with an especially pronounced summer peak in 2014. Additionally, the isolation rate of B. cereus by age was evenly distributed.

We surveyed the distribution of toxin genes among non-emetic strains (n = 532) and emetic strains (n = 135). The prevalence rates of five enterotoxin genes (entFM, nheA, cytK2, hblC, and bceT) were quite different between groups at 86.8, 83.5, 51.3, 42.9, and 34.0%, respectively, amongst non-emetic strains and 77.8, 59.3, 17.8, 11.9, and 12.6%, respectively, among emetic strains. The most prevalent toxin genes in the non-emetic strains were entFM and nheA (23.5%, group H, 125 strains), and those in the emetic strains were entFM and nheA (41.5%, group F, 56 strains). Non-emetic strains showed highly diverse toxin gene profiles (29 patterns) while emetic strains showed less diversity (20 patterns). Based on our results, we deduced that B. cereus has a high level of genetic diversity in Korea. Toxin gene profiling studies of strains from the environment and food have produced some different results from those of our survey of clinical strains. In particular, cytK and emetic toxin were not found in environmental isolates from silo tanks, and isolates from food dominantly harbored the nhe toxin gene only [24]. In another report, the most prevalent toxin genes in isolates from food were nhe and entFM [17]. In addition, a severe foodborne outbreak of diarrheal disease caused by B. cereus strains harboring cytK was reported [25,26].

As our next steps, we plan to survey the molecular profiles of all isolates using multiple locus sequence typing and/or pulsed-field gel electrophoresis.

Our results are in accordance with those of previous studies conducted in other countries, which highlighted that nheABC and entFM were carried by emetic toxin-producing strains [16,17,22,23], as well as those of previous Korean studies [23,27,28]. The cytK gene is frequently detected in isolates from patients with diarrheal-type food poisoning caused by B. cereus [26]. However, the occurrence of cytK was lower than that of other enterotoxin genes in our study.

Given the public health importance of acute diarrheal disease, surveillance is performed in many countries. Examples include the Foodborne Diseases Active Surveillance Network (FoodNet, www.cdc.gov/foodnet) in the United States [29]; OzFoodNet (www.ozfoodnet.gov.au) in Australia [30]; and FoodNet-Canada (www.phac-aspc.gc.ca/foodnetcanada) in Canada. These surveillance systems mainly monitor gram-negative pathogenic bacteria. In Korea, Enter-Net monitors gram-negative and gram-positive bacteria, including B. cereus, Clostridium perfringens, Listeria monocytogenes, and Staphylococcus aureus.

Based on our findings, B. cereus was highly heterogeneous with a diverse genetic background. Due to such features, B. cereus could be an important emerging public health threat. Thus, as a preventive measure, hygiene education on diarrheal diseases should be addressed.