Bacillus cereus from the environment is genetically related to the highly pathogenic B. cereus in Zambia.

To follow-up anthrax in Zambia since the outbreak in 2011, we have collected samples from the environment and the carcasses of anthrax-suspected animals, and have tried to isolate Bacillus anthracis. In the process of identification of B. anthracis, we collected two isolates, of which colonies were similar to B. anthracis; however, from the results of identification using the molecular-based methods, two isolates were genetically related to the highly pathogenic B. cereus, of which clinical manifestation is severe and fatal (e.g., pneumonia). In this study, we showed the existence of bacteria suspected to be highly pathogenic B. cereus in Zambia, indicating the possibility of an outbreak caused by highly pathogenic B. cereus.

Bacillus cereus is mainly known as a cause of emetic and diarrheal food poisoning (typical B. cereus) [6]. In the past few decades, in humans, there have been reports of severe manifestations and fatal cases (e.g., pneumonia) resembling anthrax due to B. cereus (highly pathogenic B. cereus) [5, 10]. B. cereus belongs to the B. cereus group that includes B. anthracis, which is a well-known causative agent of anthrax worldwide; B. thuringiensis, which is a source of effective insecticide; and B. mycoides, B. pseudomycoides and B. weihenstephanensis [6]. B. cereus, B. anthracis and B. thuringiensis are considered to belong to a single species on the basis of the genome similarity [4]. Highly pathogenic B. cereus harbors B. anthracis genes in its genome [2, 5, 7]. Accordingly, it is highly pathogenic and is genetically closer to B. anthracis than typical B. cereus and B. thuringiensis. Indeed, it is difficult to distinguish highly pathogenic B. cereus from typical B. cereus, B. anthracis and B. thuringiensis, and these genetic situations have raised questions regarding the species to which the origin is and the species to which it should be classified.

In 2011, there was an outbreak of human anthrax in Zambia [3]. We have followed-up anthrax in Zambia since 2011. In the process of B. anthracis isolation during the surveillance study, we isolated a large number of Bacillus spp. from soil in Lower Zambezi National Park (15°40.931′S, 29°27.635′E). Bacillus spp. were isolated using the standard procedure. In brief, an aliquot of 1 g of specimen suspended in 10 ml of sterilized saline was incubated at 75°C for 20 min and then spread and cultured on 10% (v/v) sheep blood agar. In the process of identification of B. anthracis, several flat, “medusa head” and dry colonies formed by gram-positive spore-forming large-rod bacteria were harvested. Hemolysis on blood agar helps in the differentiation of B. anthracis, which is almost always nonhemolytic, and B. cereus, which is usually strongly hemolytic. In the hemolysis test, B. anthracis CZC5, which was isolated in Zambia [8], and B. cereus JCM2152 were used as standards. We collected two interesting colonies, named LZ77-2 and LZ78-8, which displayed weak hemolysis at 35°C for 48 hr, but were similar to B. anthracis at 37°C for 24 hr. In the detection of B. anthracis genes by conventional PCR [1], it is conceivable that LZ77-2 and LZ78-8 were B. anthracis isolates lacking plasmids or other bacteria that have similar properties to B. anthracis (Table 1). Furthermore, we performed the following: (i) determination of the 640th nucleotide of PlcR, which is a transcriptional regulator of extracellular virulence factors; (ii) PCR detection of four lambda phage genes specifically integrated into the chromosome of B. anthracis; and (iii) phylogenetic analysis of the dnaJ sequence encoding heat shock protein 40. These are known as useful methods to differentiate B. anthracis from the B. cereus group at the genetic level [1, 10]. In B. cereus and B. anthracis, the 640th nucleotides are guanine/cytosine and thymine, respectively. In B. anthracis, a nonsense mutation caused by thymine inactivates the function of PlcR and ensuing nonhemolytic activity [9]. The 640th nucleotide in plcR of LZ77-2 and LZ78-8 was guanine (Table 1). Furthermore, four lambda prophage genes (lambda phage 01–04) specific to B. anthracis were not detected in LZ77-2 and LZ78-8 (Table 1). These results indicated that both the strains were not B. anthracis lacking plasmids. The dnaJ sequence is a potential molecular marker to discriminate highly pathogenic B. cereus strains from typical B. cereus strains [10]. As a result of phylogenetic analysis, LZ77-2 and LZ78-8 were grouped into the B. anthracis group, and they clustered with highly pathogenic B. cereus strains, including B. cereus GTC02891, GTC02896, GTC03221 and GTC03222, which were isolated from severe infection outbreaks [10]; B. cereus 03BB102, which was isolated from a fatal pneumonia case [5]; and B. cereus var. anthracis CI, which was isolated from lethal anthrax in a chimpanzee [7] (Fig. 1). Taken together with previous reports [2, 5, 7], B. cereus and B. thuringiensis classified in the B. anthracis group are highly pathogenic B. cereus and B. thuringiensis isolated from humans. Therefore, it is suggested that LZ77-2 and LZ78-8 are highly pathogenic B. cereus because of its genetic properties.

Table 1. Characterization of Bacillus spp. analyzed in this study

Species	Strain	Hemolysis	Conventional PCRb)			plcR nt640	Lambda phage				dnaJ groupc)
			pag	cap	Ba813		01	02	03	04
Bacillus cereus	LZ77-2	Hemolytica)	–	–	+	G	–	–	–	–	BA
Bacillus cereus	LZ78-8	Hemolytica)	–	–	+	G	–	–	–	–	BA
Bacillus cereus	JCM2152	Hemolytic	–	–	–	G	–	–	–	–	BC 2
Bacillus anthracis	CZC5	Nonhemolytic	+	+	+	T	+	+	+	+	BA

a) After incubation at 35°C for 48 hr, weak hemolytic activity was displayed. b) pag, cap and Ba813 indicate protective antigen encoded by plasmid pXO1, capsule protein encoded by plasmid pXO2 and chromosome gene in Bacillus anthracis, respectively. c) BA and BC indicate Bacillus anthracis and Bacillus cereus, respectively.

Fig. 1.

Partial dnaJ phylogeny among the Bacillus cereus group. DnaJ sequences of B. cereus strains LZ77-2 and LZ78-8 as well as published data for 107 isolates were used to construct a neighbor-joining tree. The tree was constructed according to a published method [10]. The strains shaded with grey are Bacillus strains used in this study. The scale bar represents evolutionary distance in substitutions per site.

We show here that bacteria suspected to be highly pathogenic B. cereus are present in Zambia. It is generally assumed that animals are infected with B. anthracis by ingesting its spores from the environment. Humans are infected on contact with infected animals. Indeed, human anthrax caused by zoonotic transmission occurred in 2011 on contact with anthrax-infected hippopotamus carcasses [3]. Cases of anthrax-suspected deaths in wildlife have been reported every year in Lower Zambezi National Park; however, our examinations showed that most cases were not caused by anthrax. Therefore, Zambia may be a high-risk country for highly pathogenic B. cereus infection outbreaks. Our results indicate the possibility of an outbreak caused by a highly pathogenic B. cereus in Zambia.