Comparison of NheA toxin production and doubling time between Bacillus cereus and Bacillus thuringiensis

In this study, we compared the toxin gene expression, NheA toxin production, doubling time, and viable cell number for several strains of the food poisoning bacteria Bacillus cereus and the microbial pesticide Bacillus thuringiensis. The two B. cereus and six B. thuringiensis strains evaluated were confirmed to possess and transcribe the nheABC, hblCDA, and cytK genes using polymerase chain reaction (PCR) and reverse-transcription PCR. NheA toxin production was compared based on the absorbance at 414 nm using a Tecra BDE-VIA kit. The NheA-specific production (absorbance/viable cell number) values indicated that the two B. thuringiensis var. kurstaki isolates from microbial pesticide produced the highest amount of toxin (0.66–0.95) than other B. thuringiensis (0.14–0.45) and the B. cereus strains (0.19–0.31). However, the B. thuringiensis strains had longer doubling time (20–26 min) than the B. cereus strains (18–19 min). Interestingly, two B. thuringiensis var. kurstaki isolates produced the highest amount of NheA toxin, and their doubling times (20–22 min) were close to those of the B. cereus strains tested.


Introduction
Bacillus cereus and Bacillus thuringiensis, gram-positive, motile rod-shaped, and spore-forming bacteria, are classified into the same Bacillus cereus group due to their high genetic and biochemical similarities. Although the two have many features in common, B. cereus is known to cause food poisoning, whereas B. thuringiensis has been used in formulations for microbial pesticides.
B. cereus can cause food poisoning in two different ways: emetic or diarrheal. The emetic syndrome is caused by the intake of an emetic toxin called cereulide, which is stable against heat, acidic pH, and digestive enzymes [1]. Vomiting usually occurs with this syndrome [1][2][3]. In contrast, diarrheal syndrome is caused by the intake of bacteria or spores that germinate inside the human intestine, creating enterotoxins [4,5]. Hbl, Nhe, and CytK are major enterotoxins that cause the diarrheal syndrome. The Hbl toxin consists of three protein subunits, Hbl L2, L1, and binding protein B [6,7], which are encoded by the hblC, hblD, and hblA genes, respectively [8,9]. Likewise, the Nhe toxin is made of three proteins, NheA, B, and C, which are the products of the nheA, nheB, and nheC genes, respectively [10]. CytK is a single protein [11] encoded by the cytK gene.
During sporulation, B. thuringiensis produces various insecticidal proteins [12], encoded by different kinds of cry genes [13]. To date, more than 500 types of cry genes have been discovered and among them, the cry1 subfamily is known as a major group [14]. Since these crystal proteins show highly effective pesticidal properties on insects including libythea celtis and diptera [15,16], B. thuringiensis has been widely adopted as an eco-friendly pesticide material [17]. Due to recent increasing demand, the market share for microbial pesticides in the crop protection market has increased consistently every year, reaching $300 million in 2013. This figure is expected to increase to $450 million by 2023 [18].
Many studies have been conducted to evaluate the safety of B. thuringiensis. The results revealed no association between excessive exposure of humans, mice, rats, rabbits, and sheep to B. thuringiensis and the outbreak of food poisoning [19][20][21][22][23][24][25][26][27][28]. Another study reported that although some B. thuringiensis strains were found in the feces of workers using pesticides, there was no direct correlation between the use of pesticides and food poisoning [29,30]. Microbial pesticides made of B. thuringiensis have been assessed to be safe. However, some studies have shown that B. thuringiensis strains contain some enterotoxigenic genes [30][31][32][33] and do produce enterotoxins [34]. In addition, there have been a few reports on the outbreak of food poisoning caused by B. thuringiensis, not B. cereus, although it was unclear whether the strains that caused food poisoning were derived from microbial pesticides [35,36].
Several studies have been conducted on the presence of enterotoxigenic genes in B. cereus and B. thuringiensis and on enterotoxin production by the bacteria [37][38][39]. However, little is known about the amount of toxin produced or the rate of bacterial growth with regard to doubling time. Comparing doubling time and toxin production in pesticide isolates would be helpful to understand the safety of using these bacteria commercially. In this study, we used polymerase chain reaction (PCR) and RT-PCR (reverse-transcription PCR) to investigate the presence and expression of hblCDA, nheABC, cytK, and cry genes in B. cereus and B. thuringiensis strains. We also measured and compared the amount of NheA toxin produced, along with the number of viable cells and their doubling time.

B. cereus and B. thuringiensis strains
The two B. cereus and six B. thuringiensis strains used in this study are listed in Table 1. Two B. thuringiensis var. kurstaki and two B. thuringiensis var. aizawai were isolated from commercial microbial pesticides. Each sample was diluted with sterile phosphate buffer and incubated in mannitol egg yolk polymyxin agar selective medium for B. cereus. After 24 h of incubation, pink colonies were collected and confirmed with API 50CH and API 20E systems (bioMerieux, Inc., Marcy I'Etoile, France). PCR was used to classify the colonies into three different groups: emetic toxin-producing B. cereus (positive for groEL and ces genes), enterotoxin-producing B. cereus (positive for groEL gene), or crystal toxin-producing B. thuringiensis (positive for groEL and cry genes).

Nucleic acids extraction, primers, and PCR conditions
For the incubation of bacterial strains, Luria-Bertani (LB) and Brain-Heart-Infusion (BHI) media (Difco, Detriot, MI, USA) were used. All bacteria were plated on LB agar and incubated overnight at 37°C. A single colony from each strain was inoculated in 5 mL of LB broth and incubated aerobically at 37°C for 8 h. For the extraction of chromosomal DNA, the Power Prep TM DNA Extraction kit (KogeneBiotech, Seoul, Korea) was used. Total RNA was extracted using the Minibest Universal RNA Extraction kit (Takara, Otsu, Shiga, Japan), and the Quantitect Reverse-Transcription kit (Qiagen, Hilden, Germany) was used for cDNA synthesis. Conventional PCR and RT-PCR were performed using the Step One Plus real-time system (AB, Foster, CA USA). The primers used in this study were synthesized by Cosmogenetech (Seoul, Korea) and are listed in Table 2.
The PCR mixture is comprised of 1 lL template DNA, 2.5 lL of 10 9 Taq buffer, 0.5 lL of dNTP stock solution (containing 10 mM of dNTP), 0.125 lL Taq polymerase (Solgent, Daejeon, Korea), and 1 lL of primer stock solution (10 pmol/lL) in a total volume of 25 lL. The following PCR procedure was used: 95°C for 5 min, 40 cycles consisting of 95°C for 30 s, 60°C for 30 s, and 72°C for 1 min, with a final extension cycle at 72°C for 5 min. PCR products were analyzed in 1.0% agarose gel in a Tris-acetate EDTA (TAE) buffer supplemented with ethidium bromide (0.5 lg/mL TAE), with 100 bp plus (TIANGEN, Bejing, China) used as a molecular mass marker. The RT-PCR mixture is comprised of 1 lL of cDNA, 2.5 lL of 10 9 Taq buffer, 0.5 lL of dNTP stock solution (containing 10 mM of dNTP), 0.125 lL Taq polymerase, and 1 lL of primer stock solution (10 pmol/ lL) in a total volume of 25 lL. During RT-PCR, one sample with total RNA was used as the negative control and the overall procedure proceeded with the same.

Measurement of NheA toxin
A commercial kit, the Tecra BDE-VIA (3 M, St. Paul, MN, USA), was used for the detection of enterotoxin. The Tecra kit has been reported to detect NheA toxin in the Nhe complex. For the immunoassay, overnight cultures originating from a single colony were diluted 1:100 in fresh BHI broth and incubated at 37°C with moderate shaking (150 rpm). The supernatant was isolated by centrifugation and subjected to enterotoxin immunoassays in accordance with the manufacturers' instructions. Positive or negative readings of the results were carried out by referring to the color card supplied with the kit. To compare the relative toxin productivity, the amount of NheA toxin produced was measured at a wavelength of 414 nm using a 680 XR microplate reader (Bio-Rad, Hercules, CA, USA). Although the manufacturer did not provide a standard curve, the effective range of the measured absorbance values was 0.2-3.0. According to the manufacturer's instruction, the larger the absorbance within the effective range, the higher was the amount of Nhe toxin. This experiment was repeated twice, and the average values are shown in Table 5.

Measurement of doubling time and viable cell number
Each B. cereus and B. thuringiensis strain was plated on LB agar and incubated overnight at 37°C. A single colony of each strain was inoculated in 5 mL of BHI broth and incubated aerobically at 37°C with shaking at 80 rpm for 12-18 h. Afterward, 100 lL of the overnight culture was re-inoculated in 9.9 mL of fresh BHI broth and incubated at 37°C with moderate shaking (150 rpm). Viable cells were counted every 20 min, and colony forming units (CFUs) were calculated by multiplying the average number of colonies by the dilution factor. Doubling time during the log phase was determined from the linear part of a semilogarithmic plot of the number of CFUs per milliliter over time.

Results
Expression of various toxin genes in B. cereus and B. thuringiensis strains To examine whether the strains used in this study produced emetic toxin, we confirmed the presence of the ces gene, an emetic toxin synthesis gene, using PCR. All eight strains showed negative results, confirming that they were not emetic toxin-producing strains ( Table 3). The groEL gene-specific amplicon found only in the Bacillus cereus group was detected in all eight strains [40]. Next, we confirmed the presence of the cry gene by using PCR to distinguish between B. cereus and B. thuringiensis [41]. As shown in Table 3, the DNA fragment specific to the cry gene was amplified only in the six B. thuringiensis strains, including the reference strains and microbial pesticide isolates. In addition, we investigated the presence of the enterotoxigenic genes nheABC, hblACD, and cytK, for which all eight strains showed positive results ( Table 3). The expression of each gene was then determined using RT-PCR, and as shown in Table 4, all eight strains exhibited positive results for the nheABC, hblACD, and cytK genes. Thus, we confirmed that the B. cereus and B. thuringiensis strains used in this study had several enterotoxigenic genes, which were all expressed at the mRNA level.

Comparison of NheA toxin production and doubling time
The Tecra BDE-VIA kit used to identify Nhe toxin production by the B. cereus and B. thuringiensis strains showed that all eight strains produced NheA toxin (

Discussion
Biochemical and genetic similarities between B. cereus and B. thuringiensis make it difficult to correctly distinguish them. The examination of crystal toxins under a microscope is commonly used to distinguish the two, but the process is laborious and time-consuming [42]. In this study, we conducted PCR assays with primers specific to the cry1 gene to distinguish the two strains [41], resulting in the detection of cry1 gene-specific amplicon in only six B. thuringiensis strains. Some of the major enterotoxins related to the diarrheal syndrome induced by B. cereus include Hbl, Nhe, and CytK. These enterotoxins are known to be destabilized or destroyed as they pass through digestive organs such as the stomach and intestines, but the diarrheal syndrome caused by B. cereus persists after the enterotoxins have passed through the digestive system. This suggests that the diarrheal syndrome is caused by an enterotoxin or enterotoxins produced by the B. cereus that proliferate in the intestine. At present, the safety of B. thuringiensis as a microbial pesticide has been verified in many animal experiments [43]. However, it is still unclear why excessive intake of B. thuringiensis microbial pesticides does not cause food poisoning, even though like B. cereus, they have several diarrhea-causing enterotoxin genes and produce enterotoxins [43]. Our PCR and RT-PCR results correspond well with previous studies reports indicating that similar to B. cereus, B. thuringiensis also has several enterotoxigenic genes, some of which are expressed [30][31][32][33]. The amount of NheA toxin produced by each strain evaluated was measured based on the absorbance after immunoassays, and all eight strains showed positive results. This is consistent with the results of previous studies showing that most B. cereus isolated from nature and from various samples produce the Nhe toxin [44,45]. However, research comparing the amount of toxin produced by each B. cereus and B. thuringiensis strain and the difference in doubling time over the same period has yet to be conducted. Although the results of this study were not quantitative, the microbial pesticide B. thuringiensis produced more toxin than the food poisoning bacteria B. cereus. Moreover, among the microbial pesticides evaluated, B. thuringiensis var. kurstaki exhibited higher toxin production than B. thuringiensis var. aizawai. However, most enterotoxins are destroyed as they pass through the stomach; therefore, the number of spores and vegetative cells that survive in the intestines, their growth rate, and their ability to produce toxins seem to affect the pathogenicity. The infective dose and the incubation period necessary for the diarrheal syndrome caused by B. cereus to occur are 10 5 -10 7 CFU and 8-16 h, respectively [4,46]. Therefore, it is estimated that the greater the toxin production ability and the faster the growth rate of B. cereus or B. thuringiensis vegetative cells proliferating in the intestine, the greater the pathogenicity. The doubling time for the microbial pesticide B. thuringiensis is 20-26 min (mean 23.3 min), while that of B. cereus is 18-19 min (mean 18.5 min), with an average difference of 4.8 min. Thus, the growth of B. thuringiensis was about 26% slower than that of B. cereus. The slower growth of B. thuringiensis may result from its large sized plasmid, which would correspond to the finding that E. coli carrying a plasmid has a slower growth rate than one without a plasmid due to the metabolic burden [47]. The cry genes in B. thuringiensis are usually present in plasmids, and when these plasmids are removed, the plasmidlost B. thuringiensis cannot be distinguished from B. cereus [48]. Although both B. cereus and B. thuringiensis produce food poisoning-causing enterotoxins, the slower growth rate of the latter may partly explain why it is safe to include B. thuringiensis in microbial pesticides. The difference in doubling time between the two strains in the intestine may affect the time it takes to reach the infective dose. However, the growth rate for the microbial pesticide B. thuringiensis was slower than that for B. cereus, and B. thuringiensis var. kurstaki (mean 21 min) grew faster and produced more Nhe toxin than B. thuringiensis var. aizawai (mean 25.6 min). Therefore, when a large amount of the former is used as a microbial pesticide in crops, more attention should be paid to the safety of the crops and the resulting food. Further studies on the enterotoxin production and growth rate of clinical isolates of B. cereus and B. thuringiensis are needed.