Bacillus cereus (sensu lato (s.l.)) is a frequent cause of foodborne illnesses, which can be classified as the diarrheal and the emetic type. The emetic type, caused by the uptake of the toxin cereulide, is of special concern, due to the intensity of the symptoms (Rouzeau-Szynalski et al., 2020). Cereulide is an extremely heat, pH, and proteolysis resistant toxin consisting of the cyclic amino acid sequence [D-O-Leu D-Ala L-O-Val D-Val]3, while also different isoforms of this dodecadepsipeptide are described (Marxen et al., 2015a; Walser et al., 2022). This toxin is produced under suitable conditions in food upon growth of so-called emetic Bacillus cereus group strains. Hitherto, emetic strains were only reported within the B. cereus group species B. paranthracis (mesophilic, panC clade 3) and very rarely within the species B. mycoides (synonym B. weihenstephanensis; psyochrotolerant, panC clade 6) (Carroll and Wiedmann, 2020; Parte et al., 2020; Thorsen et al., 2006). Ingesting food contaminated with cereulide, causes an intoxication with symptoms of nausea and vomiting which usually start within 0.5 to 6 h after intake. Mostly, the symptoms are self-limiting within 24 h. However, severe cases including fatal ones have been reported with cereulide-intoxications leading to organ failures, brain oedema and rhabdomyolysis (Dierick et al., 2005; Mahler et al., 1997; Naranjo et al., 2011; Posfay-Barbe et al., 2008; Rouzeau-Szynalski et al., 2020; Schreiber et al., 2022; Shiota et al., 2010). These severe symptoms are attributed to cell damage caused by disturbance of the mitochondrial transmembrane potential by cereulide acting as a K+ ionophore.

Cereulide is formed in a non-ribosomal peptide synthetase (NRPS) process, involving multiple enzymes encoded in the cesHPTABCD gene cluster. This gene cluster is usually located on a large plasmid of 270 kb (designated e.g. ces plasmid, pCER270, pBCE, pBCE4810 or pNCcld), which shares its backbone with the pXO1 plasmid of B. anthracis (Dommel et al., 2010; Ehling-Schulz et al., 2015; Ehling-Schulz et al., 2006; Takeno et al., 2012). The gene cesH is separately controlled by its own promotor (PH), whereas the cesPTABCD genes are polycistronically transcribed and controlled by one main ces promotor (P1). CesA and CesB are peptidyl carrier proteins (also referred to as actual NRPS or cereulide-synthetase), which assemble the amino acids to the cyclic peptide. CesP is a phosphopantetheinyl transferase. Transfer of 4′-phosphopantetheine (4´-PP) to the cereulide-synthetase (a process that is also called priming) is a prerequisite for binding the amino acid monomers and intermediates. CesT is a type II thioesterase, which removes acyl residues from acyl-4´-PP mistakenly transferred to the cereulide-synthetase and thereby deblocks misprimed enzymes (Ehling-Schulz et al., 2006; Schwarzer et al., 2002). CesC (ATPase domain) and CesD (transmembrane domain) form an ABC transporter that binds the cereulide-synthetase to the cell membrane, which is a further prerequisite for effective toxin formation (Gacek-Matthews et al., 2020). The cesH gene encodes a putative hydrolase of yet unknown function (Ehling-Schulz et al., 2015).

Cereulide synthesis is not only controlled by the ces gene cluster but also by chromosomally located genes encoding for transcriptional regulators such as AbrB, Spo0A or CodY. These regulatory proteins link toxin production to intrinsic and extrinsic factors like cell growth status, nutrient availability and oxygen availability (Ehling-Schulz et al., 2015). AbrB interacts with the ces promotor and represses the transcription of the cesPTABCD operon. Phosphorylated Spo0A represses the transcription of abrB, which in turn enhances cereulide synthesis. CodY also interacts with the ces promotor. Binding of branched chain amino acids or GTP to CodY enhances its affinity to the ces promotor leading to repression of the cesPTABCD operon (Ehling-Schulz et al., 2015; Frenzel et al., 2012; Kaiser and Heinrichs, 2018; Lücking et al., 2009; Lücking et al., 2015). Recently, the plasmid encoded PagRBc was described as a further transcriptional repressor, which interacts with the ces promotor (Kalbhenn et al., 2022). In spite of the multitude of prior research, the complex regulation of cereulide production, including post-transcriptional and post-translational processes, is far from being completely understood (Ehling-Schulz et al., 2015; Kranzler et al., 2016).

The amount of cereulide produced in a food matrix depends on various factors. These factors can be divided into external and strain dependent factors. External factors, as reviewed by Rouzeau-Szynalski et al. (2020), include the temperature, the oxygen availability and the composition of the food matrix. The food matrix determines the pH-value, the water activity, the availability of nutrients (including the glucose and amino acid contents as well as vitamins and trace elements), the NaCl concentration and further compounds, which may inhibit cereulide production such as high concentrations of branched chain amino acids or long chain polyphosphates. Focussing on cereulide production under laboratory conditions, equivalent external factors apply to the bacterial growth medium and incubation conditions. For instance, Agata et al. (1999), Apetroaie-Constantin et al. (2008) and Rajkovic et al. (2006a) observed huge variations in cereulide production depending on the composition and physical condition (solid or liquid) of the growth medium. Moreover, incubation under static or shaken conditions influences cereulide production. Depending on the strain and the growth medium, optimum temperatures between 20 °C and 37 °C were observed for cereulide production (Apetroaie-Constantin et al., 2008; Ellouze et al., 2021; Häggblom et al., 2002; Kranzler et al., 2016; Rouzeau-Szynalski et al., 2020).

Regarding strain dependent factors, several studies reported variations in the cereulide production capacities of different strains (Apetroaie-Constantin et al., 2008; Apetroaie et al., 2005; Carlin et al., 2006; Finlay et al., 2000; Häggblom et al., 2002; Jääskeläinen et al., 2003; Kranzler et al., 2016; Phat et al., 2017; Rajkovic et al., 2006b; Shaheen et al., 2006; Stark et al., 2013). However, only very few studies focus on a comparison of strains, and even less studies try to associate the cereulide producing phenotype to a certain genotype. For example, Apetroaie et al. (2005) and Shaheen et al. (2006) determined varying cereulide production levels and ribopatterns of 13 and six strains, respectively. However, the ribopatterns did not match the cereulide phenotype consistently. Studies of Carlin et al. (2006) and Stark et al. (2013) classified B. cereus (s.l.) strains carrying the ces gene cluster (17 and 76 strains, respectively) as low, medium and high cereulide producers, but without further genetic characterization.

With the goal to assess the pathogenic potential of emetic strains more precisely based on their genotype, we attempted to associate cereulide production capacities with genetic characteristics. For this, cereulide levels in liquid cultures of 31 emetic B. cereus (s.l.) strains were determined after incubation for 24 h at 24 °C, 30 °C and 37 °C. The tested strains underwent whole genome sequencing, which allowed for an in-depth characterization of gene sequences related to cereulide production. The taxonomy, population structure and phylogenetic relationships of the strains were evaluated based on average nucleotide identity, multi-locus sequence typing (MLST), core genome MLST and single nucleotide polymorphism analyses. Although the number of strains analysed in our study was comparably low for a genome wide association study, we also tested the GWAS approach to link genetic variation with cereulide levels.