Rapid, High-Throughput Identification of Anthrax-Causing and Emetic Bacillus cereus Group Genome Assemblies via BTyper, a Computational Tool for Virulence-Based Classification of Bacillus cereus Group Isolates by Using Nucleotide Sequencing Data

Abstract

The Bacillus cereus group comprises nine species, several of which are pathogenic. Differentiating between isolates that may cause disease and those that do not is a matter of public health and economic importance, but it can be particularly challenging due to the high genomic similarity within the group. To this end, we have developed BTyper, a computational tool that employs a combination of (i) virulence gene-based typing, (ii) multilocus sequence typing (MLST), (iii) panC clade typing, and (iv) rpoB allelic typing to rapidly classify B. cereus group isolates using nucleotide sequencing data. BTyper was applied to a set of 662 B. cereus group genome assemblies to (i) identify anthrax-associated genes in non-B. anthracis members of the B. cereus group, and (ii) identify assemblies from B. cereus group strains with emetic potential. With BTyper, the anthrax toxin genes cya, lef, and pagA were detected in 8 genomes classified by the NCBI as B. cereus that clustered into two distinct groups using k-medoids clustering, while either the B. anthracis poly-␥-D-glutamate capsule biosynthesis genes capABCDE or the hyaluronic acid capsule hasA gene was detected in an additional 16 assemblies classified as either B. cereus or Bacillus thuringiensis isolated from clinical, environmental, and food sources. The emetic toxin genes cesABCD were detected in 24 assemblies belonging to panC clades III and VI that had been isolated from food, clinical, and environmental settings. The command line version of BTyper is available at https://github.com/lmc297/BTyper. In addition, BMiner, a companion application for analyzing multiple BTyper output files in aggregate, can be found at https://github.com/lmc297/BMiner. IMPORTANCE Bacillus cereus is a foodborne pathogen that is estimated to cause tens of thousands of illnesses each year in the United States alone. Even with molecular methods, it can be difficult to distinguish nonpathogenic B. cereus group isolates from their pathogenic counterparts, including the human pathogen Bacillus anthracis, which is responsible for anthrax, as well as the insect pathogen B. thuringiensis. By using the variety of typing schemes employed by BTyper, users can rapidly classify, characterize, and assess the virulence potential of any isolate using its nucleotide sequencing data.KEYWORDS Bacillus cereus group, taxonomy, virulence genes, whole-genome sequencing

Proposal of a Taxonomic Nomenclature for the Bacillus cereus Group Which Reconciles Genomic Definitions of Bacterial Species with Clinical and Industrial Phenotypes

“…Importantly, isolates which display phenotypic characteristics associated with "B. cereus" (e.g., motility and gamma bacteriophage resistance) can cause anthrax (2,30,31,(40)(41)(42). Despite the assertion that it is a clonal species with low diversity (46)(47)(48), the B. anthracis genomospecies cluster formed at 95 ANI encompasses lineages which fall outside the one most commonly associated with anthrax illness (Fig.…”

“…All genomes in the NCBI RefSeq Assembly database (59) which were submitted as one of 18 published B. cereus group species (35) were downloaded, along with the type strain genomes of three proposed effective B. cereus group species (60-62) (n ϭ 2,231, accessed 19 November 2018) (see Tables S1 and S6 in the supplemental material). QUAST version 4.0 (63) was used to assess the quality of each genome, and BTyper version 2.3.2 (31) was used to detect B. cereus group virulence genes in each genome, using default minimum amino acid sequence identity and coverage thresholds (50% and 70%, respectively) (Table S6) (31,64). Prokka version 1.12 (65) was used to annotate each genome, and the resulting coding sequences were used as input for the command-line implementation of BtToxin_scanner version 1.0 (BtToxin_scanner2.pl), which was used to identify Bt toxin genes in each genome using default settings (66).…”

“…Furthermore, genomic determinants responsible for some phenotypes are plasmid mediated, such as synthesis of anthrax toxin/capsular proteins (19)(20)(21)(22), bioinsecticidal crystal proteins (23)(24)(25), and emetic toxin (cereulide) synthetase proteins (26,27). These traits can be lost, gained, heterogeneous in their presence within a species, or present across multiple species (28)(29)(30)(31).…”

Cereulide synthetase acquisition and loss events within the evolutionary history of Group IIIBacillus cereus sensu latofacilitate the transition between emetic and diarrheal foodborne pathogen

“…species (21, 23–26, 53) were downloaded ( n = 2,231; accessed November 19, 2018). The ANI function in BTyper v. 2.3.3 (13) was used to calculate ANI values between each genome and the type strain/species reference genomes of each of the 18 published B. cereus s.l. species as they existed in 2019 (7).…”

“…(10–16). While cereulide-producing Group VI strains, also known as “emetic B. weihenstephanensis ”, have been isolated on rare occasions (14, 15, 17–19), the bulk of cereulide-producing strains belong to Group III (8, 10, 13, 16). Often referred to as “emetic B. cereus ", cereulide-producing Group III strains often harbor ces on plasmids (9, 10, 19), and have been linked to outbreaks around the world (5, 7, 8, 20).…”