Sporulation environment of emetic toxin-producing Bacillus cereus strains determines spore size, heat resistance and germination capacity


  • M. van der Voort,

    1. Top Institute Food and Nutrition (TIFN), Wageningen, The Netherlands
    2. Laboratory of Food Microbiology, Wageningen University and Research Centre, Wageningen, The Netherlands
    3. Present address: Laboratory of Phytopathology, Wageningen University and Research Centre, Wageningen, The Netherlands
    Search for more papers by this author
  • T. Abee

    Corresponding author
    1. Laboratory of Food Microbiology, Wageningen University and Research Centre, Wageningen, The Netherlands
    • Top Institute Food and Nutrition (TIFN), Wageningen, The Netherlands
    Search for more papers by this author


Tjakko Abee, Laboratory of Food Microbiology, Wageningen University, Bomenweg 2, 6703 HD Wageningen, The Netherlands. E-mail: tjakko.abee@wur.nl



Heat resistance, germination and outgrowth capacity of Bacillus cereus spores in processed foods are major factors in causing the emetic type of gastrointestinal disease. In this study, we aim to identify the impact of different sporulation conditions on spore properties of emetic toxin-producing B. cereus strains.

Methods and Results

Spore properties of eight different emetic toxin-producing strains were tested, with spores produced in five different sporulation conditions: aerated liquid cultures, air–liquid biofilms, 1·5% agar plates, 0·75% agar plates and swarming colonies. Model food studies revealed spores from emetic toxin-producing strains to germinate efficiently on meat broth- and milk-based agar plates, whereas germination on rice-based agar plates was far less efficient. Notably, spores of all strains germinated efficiently when 0·1% meat broth was added to the rice plates. Analysis of spores derived from different environments revealed large diversity and showed biofilm spores for the strains tested to be the largest in size, the most heat resistant and with the lowest germination capacity.


Sporulation in complex conditions such as biofilms and surface swarming colonies increases heat resistance and dormancy of spores.

Significance and impact of the study

The results obtained imply the importance of sporulation conditions on spore properties of emetic toxin-producing B. cereus strains, as occur for instance in food processing.


The Bacillus cereus group of species can thrive in a great diversity of environments, such as soil, plant rhizosphere, insects and humans (Jensen et al. 2003; Vilas-Boas et al. 2007; Stenfors Arnesen et al. 2008; Carlin et al. 2010). To succeed in these conditions, members of the B. cereus group harbour a great diversity of traits, including the production of a large variety of toxins (Jensen et al. 2003; Vilas-Boas et al. 2007). A major factor enabling B. cereus to be present in different environments is its capacity to sporulate when conditions become unfavourable for growth, such as a lack of nutrients (Knaysi 1946). Spores are highly resistant survival structures that can survive food processing, especially when food is minimally processed (te Giffel 2001). Consequently, surviving spores of B. cereus can cause food spoilage after outgrowth (Adams and Moss 2000; Brown 2000). Next to this, food poisoning syndromes are related to B. cereus enterotoxin production in the human body after consumption or to emetic toxin (cereulide) production by B. cereus in foods before consumption, resulting in the diarrhoeal type and the emetic type of gastrointestinal disease, respectively (Ehling-Schulz et al. 2004; Schoeni and Wong 2005; Stenfors Arnesen et al. 2008).

To grow in processed foods, surviving B. cereus spores need to be reactivated in an event called germination (Abee et al. 2011). To start germination, the spore's sensing mechanism, consisting of germinant receptors, monitors the environment for the presence of nutrients, also called germinants. The nutrients trigger the germinant receptors and subsequently activate the germination process. In this germination process, the spore loses its heat resistance by rehydration and breaking down its protection layers, followed by outgrowth (Moir 2006; Paredes-Sabja et al. 2011). Subsequently, vegetative cells are able to multiply and for instance produce the emetic toxin cereulide in food (Ehling-Schulz et al. 2004). Cluster analysis showed emetic toxin-producing B. cereus strains to be distinct from nonemetic toxin-producing B. cereus strains; as the emetic toxin-producing strains are not able to degrade starch or to ferment salicin, they do not possess the genes encoding haemolysin BL and show only weak or no haemolysis (Ehling-Schulz et al. 2005). Furthermore, it has been shown that emetic toxin-producing strains grow at higher temperatures, as in general, they can grow at 48°C and not at 7°C. Next to this, spores of emetic toxin-producing strains possess higher heat resistance and germinate less efficiently, in comparison with nonemetic toxin-producing B. cereus strains (Carlin et al. 2006).

For both B. cereus and B. subtilis, the sporulation history, including temperature, nutrients and water activity, has been shown to influence spore properties such as heat resistance and germination capacity (Cazemier et al. 2001; de Vries et al. 2004; Hornstra et al. 2006; Rose et al. 2007; Nguyen Thi Minh et al. 2008; Moeller et al. 2011; Ramirez-Peralta et al. 2012). In contrast, the impact on spore properties of sporulation by B. cereus in air–liquid biofilms (Wijman et al. 2007) or swarming colonies (Ghelardi et al. 2007; Senesi et al. 2010) compared with sporulation on agar plates or in liquid cultures has not been specifically studied. In natural and in food-processing environments B. cereus may encounter a variety of conditions that next to growth and also may affect sporulation conditions and consequently spore characteristics. Therefore, the impact of sporulation conditions, including aerated liquid cultures, air–liquid biofilms, 1·5% agar plates, 0·75% agar plates and swarming colonies, on emetic toxin-producing B. cereus strain spore properties such as size, heat resistance, dipicolinic acid contents and germination capacity was determined. A selection of emetic toxin-producing B. cereus strains was used, including food- and outbreak-associated isolates.

Materials and methods

Strains and spore production

The eight emetic toxin-producing B. cereus strains used in this study are listed in Table 1. The strains were cultured shaking at 150 rpm in Luria broth [LB, Difco (BD), Breda, the Netherlands] at 30°C. For all sporulation conditions, the same sporulation medium based on Difco's sporulation medium, containing nutrient broth (Difco, 4 g l−1), (NH4)2SO4 (5 mmol−1), MgCl2 (1 mmol−1), Ca(NO3)2 (1 mmol−1), MnSO4 (66 μmol−1), ZnCl2 (12·5 μmol−1), CuCl2 (2·5 μmol−1), Na2MoO4 (2·5 μmol−1) and CoCl2 (2·5 μmol−1), and named nutrient broth sporulation medium (NBSM), was used to obtain spore batches. For the production of spores in aerated liquid cultures, 0·5% of overnight culture was inoculated in 50 ml liquid NBSM in a 250-ml Erlenmeyer flask. Subsequently, flasks were inoculated at 25°C with 200 rpm rotation for aeration. Spores from liquid cultures were harvested after 36 h by centrifugation in 50-ml tubes (Greiner, Alphen a/d Rijn, the Netherlands) at 5300 g for 7 min at 25°C using the Eppendorf 5804R, with rotor CL 007 F-34-6-38 07061 (New Brunswick, Nijmegen, the Netherlands). For production of spores on 1·5 and 0·75% NBSM agar plates, 100 μl of overnight culture was spread plated on 1·5 and 0·75% agar plates, respectively. The agar plates were subsequently incubated at 25°C. Spores from spread plated agar plates were harvested after 36 h by resuspending the cells from the agar plate in 40 ml 10 mmol−1 NaPO4 (NaPi) buffer (pH 7·4). For production of spores in swarming colonies, a droplet of 3 μl of overnight culture was placed on a 0·75% NBSM agar plate. Five droplets of one strain were placed per plate, which would finally make up one batch of spores per strain. Plates with five droplets were incubated at 25°C. Spores from swarming colonies were harvested after 7 days. By the use of a 10-μl loop, only the swarming area was taken from the swarming colony and not the cells that were present in the zone that was occupied by the initial droplet. Subsequently, the spores were suspended in 20 ml of 10 mmol−1 NaPi buffer (pH 7·4). For sporulation in biofilms, wells of 24-well plates filled with 2 ml of NBSM were inoculated with 0·5% of overnight culture. After 48 h of incubation at 25°C, the cell suspension was removed from the wells, and the wells were washed twice with 2 ml of 10 mmol−1 NaPi buffer (pH 7·4). Biofilm cells remaining on the sides of the well were harvested by swabbing and resuspension in 10 mmol−1 NaPi buffer (pH 7·4), and for one spore batch, the spores from 12 wells were pooled in a total of 2 ml NaPi buffer.

Table 1. Emetic toxin-producing strains used in this study
StrainsAlternative strain indicationsProvidedCesA PCROriginReference
  1. a

    Sequenced type strain. RIVM is the Dutch National Institute for Public Health and the Environment, Bilthoven, the Netherlands, FTC is the Food Technology Centre, Wageningen, the Netherlands.

AH187aF4810/72RIVM+Vomit (rice)Ehling-Schulz et al. (2004)
NC7401 FTC+Vomit (chow mein)Agata et al. (1994)
F5881/94 FTC+Food remnantsEhling-Schulz et al. (2005)
F3752A/8694-10FTC+Cooked riceEhling-Schulz et al. (2005)
F3351/8794-14FTC+Faeces (fried rice)Ehling-Schulz et al. (2005)
RIVM 3627 RIVM+Vomit (rice)Wijnands et al. (2002)
RIVM BC 330 RIVM+Dairy dessert

For each sporulation condition, biologically duplicate spore batches were obtained. In addition, all spore batches obtained from different backgrounds were washed twice by centrifugation at 7250 g and resuspension in NaPi buffer after the initial resuspension and finally resuspended in 10 mmol−1 NaPi buffer (pH 7·4). This was excluding the spores produced in biofilms, which were only washed once after resuspension, as washing was already applied in the well before harvesting. Sporulation in all conditions was similar for the strains used, with over 95% in liquid and 0·75% agar cultures, c. 90% in 1·5% agar cultures and complex colonies and c. 80% for biofilms (observed by microscopy analysis). Therefore, washing of the spore batches in NaPi buffer was sufficient to remove vegetative cells and cell debris from the spore suspension, subsequently, after washing, all spore batches used contained over 95% of spores.

The presence of cesA, necessary for emetic toxin production, was confirmed by PCR for all eight emetic toxin-producing strains, by the use of primers cesA_forw (GTTGGCGTGTTATGTGATCG) and cesA_rev (TCATCACCATGTCCAGAAT), which were designed based on the cesA nucleotide sequence (Rasko et al. 2007).

Size determination

The size (length) of spores was determined by use of flow cytometry and phase-contrast microscopy. Before running through a FACScalibur flow cytometer (Becton Dickinson, San Jose, CA, USA), spore batches were washed and resuspended in filter-sterilized (0·22 μm) physiological salt solution. By measuring the forward scatter, the surface of the beads with known diameters of the Flow Cytometry Size Calibration Kit (Molecular Probes Europe BV, Leiden, the Netherlands) was measured. The diameter (length) of the beads was correlated with forward scatter values, with a certain forward scatter value corresponding to a specific length. Subsequently, this correlation between the beads and the forward scatter was used to determine the length of the spores by measuring the forward scatter. Although spores are oval and not round like the calibration beads, this method resulted in spore length values that were in agreement with spore length values based on microscopy analysis. The diversity in spore lengths within one batch resulted in a normal distribution of the spore lengths, and therefore, the average length values presented are as determined by fitting a normal distribution of the forward scatter with Excel Solver (Frontline Systems, Inc., Incline Village, NV, USA). The variation in average lengths of the different spore batches was in line with phase-contrast microscopy analysis of the spore batches.

Heat resistance and dipicolinic acid content

To determine the heat resistance of the spore batches of the strains BC330 and F3752A/86, 20 μl of spore suspension at an OD600 of 0·01 was pipetted in six individual thin-walled 0·2-ml PCR tubes (Bio-rad, Veenendaal, the Netherlands) and heated in a water bath at 95°C. This temperature ensured a reduction in viability of the spores, whereas 90°C was shown not to be efficient in killing spores from several emetic toxin-producing B. cereus strains (Carlin et al. 2006). Appropriate dilutions were pour-plated after 1, 5, 10, 15, 20 and 30 min of heating, and colony-forming units on BHI plates, resulting from surviving spores, were determined after 48 h. Each biological duplicate spore batch was tested for its heat resistance, and values presented are averages of two spore batches. As the regression in survival was nonlinear, decimal reduction times (δ), which is the time required to kill 90% of the bacteria (min), were determined by fitting the Weibull model to the inactivation curves (Mafart et al. 2002). Nonlinearity was indicated by P values between 1 (linear regression) and 2·5. As, in our case, a higher P values corresponded to a higher δ, only the δ values are indicated. By use of a paired one-tailed student's t-test (MS Excel), a P value was determined comparing heat resistance of the spore batches. Observed differences were considered significant when the < 0·05.

Dipicolinic acid (DPA) content of spores was measured with a fluorescent assay described previously (Hindle and Hall 1999; de Vries et al. 2005), with minor modifications. Spores or cells were suspended in 10 mmol−1 NaPi buffer (pH 7·4) and destroyed by autoclaving (121°C, 15 min), a treatment known to release all DPA (Janssen et al. 1958), and pelleted at maximum speed in an Eppendorf centrifuge for 2 min. The supernatant was assayed for DPA as follows: 80 μl of supernatant was used in a final volume of 200 μl 10 mmol−1 NaPi buffer (pH 7·4) with freshly added TbCl3 at a final concentration of 100 μmol−1. Fluorescence was measured in 96-well UV microplates (Greiner) with the Safire microplate reader (Tecan, Salzburg, Austria), with the settings: excitation wavelength 272 nm, emission wavelength 547 nm, bandwidth excitation and emission wavelengths 7·5 nm, the lag time 10 μs and integration time 2000 μs, and in combination with Xfluor4 software ver. 4.40 (Tecan Benelux BV, Giessen, the Netherlands), as described previously (de Vries et al. 2005). The DPA content of the spores was related to the OD600 of the spore suspensions, as the number of spores per OD600 unit for the spore batches from different backgrounds was determined to be similar (data not shown).

Spore germination assays

Germination of the spore batches on food matrices was tested by plating appropriate dilutions of spore batches without and with an germination-activating heat treatment for 15 min at 70°C (Hornstra et al. 2005) on 1% rice powder (w/v) solution (Sambel Cap Jempol, Jakarta, Indonesia), semi-skimmed bovine milk (FrieslandCampina, Amersfoort, the Netherlands) and 1% (w/v) meat broth (Lab-Lemco, Oxoid, Badhoevedorp, the Netherlands) solidified with 1·5% agar. Colony forming units (CFUs) were determined after one, three and 7 days of incubation at 25°C. For all strains and all spore types, germination in liquid BHI at 25°C was observed to be c. 100% within 24 h by microscopy analyses. Furthermore, germination on BHI agar plates and milk and meat broth agar plates was observed to be the same after 7 days (data not shown). Therefore, germination on milk and meat broth after 7 days was regarded as 100% germination. And the CFU counts after 7 days were used to calculate the percentage of germination after 1 day of incubation. The number of CFUs, indicating germination, per spore batch on BHI agar plates observed after 7 days of incubation was used to determine the percentage of germination of spores on rice after 7 days, with the number of CFUs on BHI agar plates being regarded as 100% germination.

Germination capacity of the spore batches of strains BC 330, F3752A/86 and NC7401 in response to the combination of the amino acid l-alanine (5 mmol−1) and the purine riboside inosine (2·5 mmol−1) was monitored in 96-well plates by measuring OD600 values (Hornstra et al. 2005) by use of the Spectramax plus384 (Molecular Devices, Sunnyvale, CA, USA). This combination of nutrients resulted in over 95% of germination after 24 h for spores of all strains tested, as observed by microscopy analysis (data not shown). Germination was followed for duration of 90 min by measuring OD600 values every min at 25°C. To prevent sedimentation of the spores and lack of oxygen, the 96-well plates were shaken 30 s between every measurement. Nutrients were applied by adding 25 μl of a l-alanine (50 mmol−1 in water) solution and 25 μl of a inosine (25 mmol−1 in water) solution to 200 μl of spore suspensions, with an end volume of 250 μl. Spore suspensions used for germination experiments were either used directly or heat-treated for 15 min at 70°C before use. Subsequently, based on microscopy analyses, germination percentages were calculated by representing a drop in OD600 of 65% as 100% germination. Germination of the spores batches was statistically analysed by use of a paired one-tailed student's t-test (MS Excel) and considered significant when < 0·05.


Germination on food matrices

For this study, eight emetic toxin-producing B. cereus strains from different origins were selected, including both food- and outbreak-associated isolates (Table 1). For each strain, spores were generated in NBSM and tested for relevant characteristics.

First of all, germination of these spores was tested. Germination in response to meat broth (Lab-Lemco) showed that all spores derived from all eight strains germinated and grew out into colonies within 1 day of incubation at 25°C (Fig. 1). The spore germination response to the milk-based medium appeared slower than that in response to meat broth. This was most clearly shown for the nonheat-activated spores, for which between 45% (strain F3752A/86) and 98% (strain AH187) of the spores grew out into colonies after 1 day of incubation. After heat activation of the spores, outgrowth into colonies ranged between 79% (strain BC 296) and 98% (strain F3752A/86, Fig. 1). After 7 days of incubation for all strains, c. 100% of the heat-activated and nonheat-activated spores grew out into colonies on milk-based plates (data not shown). Spore germination for all strains was shown to be the lowest on rice. Monitoring germination and outgrowth after 7 days of incubation at 25°C showed that only spores of strain 94-10 and AH187 displayed significant germination and outgrowth after heat activation, with 18 and 5% of the spores growing out into colonies, respectively. Spore germination and outgrowth on rice for the other emetic toxin-producing strains were below 0·2% even after heat activation (Fig. 1). However, supplementing the rice with 0·1% of Lab-Lemco meat broth led to c. 100% of spore germination and outgrowth after 1 day for all strains tested (data not shown). Furthermore, for all strains, 100% of vegetative cells grew out into colonies on rice agar plates within 24 h, confirming that, once germinated, efficient growth can occur on rice medium-based agar plates (data not shown).

Figure 1.

Germination and outgrowth at 25°C of spores derived from emetic toxin-producing B. cereus strains on food matrices: on meat broth agar after 1 day (a), on milk agar after 1 day (b), on rice agar after 7 days (c). CFUs obtained after germination and outgrowth of heat-activated spores on BHI plates after 7 days were considered 100% germination, as determined for each strain individually. Black bars are nonheat-activated, and grey bars are heat-activated spores. Error bars represent the standard deviations of replicate measurements (n = 2).

Spore size

Spore size for the eight emetic toxin-producing strain spores produced in five different sporulation conditions was analysed using flow cytometry (Fig. 2a). Comparing the size of the liquid culture spores of the different strains shows that for two strains, the spores are (significantly) smaller than for the other six strains. The smallest average spore size of 0·76 μm was observed for strain F3752A/86, whereas the average spore size of liquid culture spores for strain F3351/87 was 0·94 μm. The size of liquid culture spores of the other strains ranged between 1·19 and 1·36 μm. In general, it was observed that spores produced on agar plates (1·5% agar, 0·75% agar, swarming colonies) were larger than spores from liquid cultures. Exceptions to this were observed for strain F3351/87 and strain NC7401, for which only the spores from swarming colonies were larger than spores from liquid cultures. Notably, for all eight strains, spores produced in biofilms were shown to be the largest in size. Differences in size between different sporulation conditions were also confirmed microscopically (Fig. 2b).

Figure 2.

Spore size as determined by flow cytometry for all strains (a) and illustrated by phase-contrast microscopy images for strain F3752/86 and BC 330 (b). In (a), averages of two independent spore batches are presented, white bars are liquid derived spores, from lightest grey to darkest grey are 0·75% agar, 1·5% agar and swarming colony–derived spores, and black bars are biofilm-derived spores. The error bars represent the standard deviation of two independent spore batches (n = 2). For each spore batch, c. 500 spores were measured. In (b), the bars indicate 1 μm.

Influence of sporulation conditions on germination of spores

Germination of spores from different sporulation conditions was assessed for three selected strains BC 330, F3752A/86 and NC7401 obtained from dairy dessert, rice and vomit. First, germination of the spores in response to the combination of 5 mmol−1 l-alanine and 2·5 mmol−1 inosine was studied, and germination was revealed to be stimulated by heat activation, as all spore batches tested germinated completely after heat activation within 30 min (Fig. 3). Interestingly, clear differences were observed comparing germination capacities of spores from the different sporulation conditions and from the heat- and nonheat-activated spores. For instance, nonheat-activated liquid culture spores appeared to be the slowest germinators, with spore germination efficiencies after 30 min of only 8, 10 and 37% for strain BC 330, F3752A/86 and NC7401, respectively (Fig. 4). In contrast, after heat activation, spores produced in liquid cultures were the fastest to germinate for strains F3752A/86 and NC7401, with spore germination efficiencies after 10 min already reaching 71 and 66%, respectively. For strain BC 330 also with heat activation, the liquid culture spores were amongst the slowest germinating spore batches. Furthermore, a comparison of the impact of the different sporulation conditions revealed, for both nonheat-activated and heat-activated spores from swarming colonies and biofilms, that these were amongst the slowest spore batches to germinate, whereas heat-activated and nonheat-activated 0·75% agar and 1·5% agar plate spores were the most efficient in germination (Fig. 4). In addition, for nonheat-activated spores of strain BC 330 and NC7401, the germination of swarming colony-derived and biofilm-derived spores was found to be statistically less efficient than germination of 0·75 and 1·5% agar plate-derived spores. Furthermore, for heat-activated spores of all three strains tested, it was observed that spores derived from biofilms germinated significantly less efficient than spores derived from 0·75% agar plates. The germination efficiencies determined by OD600 measurements were confirmed by microscopy analyses (data not shown).

Figure 3.

Germination of nonheat-activated (left) and heat-activated (right) spore batches of strains BC 330, NC7401 and F3752A/86 in a combination of the germinants l-alanine 5 mmol−1 and inosine 2·5 mmol−1. Germination of a single batch is presented (n = 1). The trend for the independent duplicate batch was similar as also indicated by the error bars in Fig. 5. Open diamonds are liquid-derived spores, closed light grey triangles are 1·5% agar-derived spores, open grey diamonds are 0·75% agar-derived spores, dark grey closed circles are swarming colony–derived spores, and black triangles are biofilm-derived spores. A decrease of 65% in OD600 represents 100% germination.

Figure 4.

Average germination efficiency for nonheat-activated spore batches after 30 min (a) and heat-activated spore batches after 10 min (b) in l-alanine 5 mmol l−1 and inosine 2·5 mmol l−1. Error bars represent the standard error of duplicate measurements (n = 2). White bars are liquid-derived spores, from lightest grey to darkest grey are 1·5% agar, 0·75% agar and swarming colony–derived spores, and black bars are biofilm-derived spores.

Influence of sporulation conditions on heat resistance of spores

Heat resistance of spores obtained at different sporulation conditions was assessed for two emetic toxin-producing strains, strain BC 330 and strain F3752A/86, both isolated from food. These strains displayed the sporulation condition-dependent differences in spore sizes and germination. By analysing the decimal reduction times (δ), it was clear that for strains BC 330 and F3752A/86, spores derived from liquid culture (δ of 13·3 and 5·6 min) and spores derived from 0·75% agar plates (δ of 10·8 and 6·8 min) were the least heat resistant. Spores produced on 1·5% agar plates (δ of 18·1 and 11·2 min) were more heat resistant; however, the difference was not significant. Notably, spores produced in swarming colonies (δ of 18·7 and 19·9 min) and in biofilms (δ of 23·0 and 19·3 min) showed the significantly highest heat resistance for both strains (Fig. 5).

Figure 5.

Spore heat resistance determined at 95°C. Average decimal reduction times of independent spore batches are represented (n = 2). Decimal reduction times δ are indicated as d1. White bars are liquid-derived spores, from lightest grey to darkest grey are 1·5% agar, 0·75% agar and swarming colony–derived spores, and black bars are biofilm-derived spores.

For all strains and all different sporulation conditions, the spore core DPA levels were determined. No significant differences were observed in the amount of DPA released from spore batches obtained under the different sporulation conditions. For strain AH187, the average DPA level per OD600 unit (de Vries et al. 2005) was found to be c. 60 μmol−1, whereas for the other seven strains, the average DPA level was around 40 μmol−1 (Fig. 6).

Figure 6.

Average dipicolinic acid (DPA) concentration per strain. Average values were determined for the biological duplicates of the five sporulation conditions. Error bars represent the standard error (n = 10).


Spores derived from different conditions of emetic toxin-producing strains were analysed and revealed to vary in spore size, heat resistance and germination-related properties. Comparing spore sizes of spores obtained from aerated liquid cultures, air–liquid biofilms, 1·5% agar plates, 0·75% agar plates and swarming colonies, for eight different emetic toxin-producing strains, showed liquid culture spores in general to be the smallest and air–liquid biofilm spores for all eight strains to be the largest in size. Furthermore, two groups of spores could be distinguished in spore size. The size of spores derived on 1·5% agar plates was clearly smaller for two strains (strain F3752A/86 and F3351/87), which produced spores with an average of 1·03 μm or smaller, whereas the other strains produced spores with averages ranging from 1·20 to 1·65 μm (Fig. 1). Previously, Carrera et al. (2007) showed that the mean lengths of B. anthracis spores can be categorized into two significantly different groups: one with mean spore lengths of 1·26 μm or shorter (three strains) and another group of four strains with mean spore lengths between 1·49 and 1·67 μm. In addition, spores of B. thuringiensis have been reported to swell and shrink in response to high and low relative humidity (Westphal et al. 2003). The factors that determine spore size in emetic toxin-producing B. cereus strains remain to be elucidated.

Although the medium components used for the five different sporulation conditions were the same, differences in for instance the pH, oxygen availability and/or primary and secondary metabolites in the medium at the time of sporulation will influence the sporulation process and the final spore properties. Differences in sporulation conditions, including differences in nutrient availability (de Vries et al. 2005; Hornstra et al. 2006; Setlow 2006; Moeller et al. 2011), have been reported to influence sporulation and consequently spore properties. Interestingly, no apparent difference in DPA content of the spores, one of the factors influencing wet heat resistance (Setlow 2006), was observed between the five different sporulation conditions. In addition, the DPA content of strain AH187 spores was shown to be higher than the DPA content of spores of the other seven strains tested. In agreement, strain AH187 was previously shown to be amongst emetic toxin-producing strains generating the most heat-resistant spores (Carlin et al. 2006). Spore resistance to wet heat is determined largely by the water content of the spore core, with a lower core water content generally resulting in better wet heat resistance (Setlow 2006). This low water content is partly due to the high concentration of DPA in the spore core (Setlow 2006; Coleman et al. 2010). In contrast, spores of B. subtilis strain ATCC 31324 produced at low water activity (aw of 0·950) were less heat resistant than spores produced at higher water activity (aw of 0·993); however, the spore core water content was not measured in that study (Nguyen Thi Minh et al. 2008). In addition, spore wet heat resistance is influenced by the level and type of mineral ions and the occurrence of small, acid-soluble spore proteins (SASP) in the spore core (Setlow 2006). Therefore, the underlying mechanisms contributing to heat resistance of spores derived from the five sporulation conditions used remain to be elucidated.

In the current study, the data suggest a positive correlation between heat resistance and dormancy (i.e. low germination rate) of spores, with the highest heat resistance and dormancy for biofilm-derived spores. A positive correlation between heat resistance and dormancy has been shown before for different Bacillus species, including B. megaterium, B. cereus and B. subtilis (Ghosh et al. 2009). Moreover, sporulation in biofilms has been studied in both B. cereus and B. subtilis (Branda et al. 2001; Wijman et al. 2007), and separate studies on biofilm-derived spores were performed to identify the resistance to sanitizing agents (Ryu and Beuchat 2005) and the germination capacity (Lindsay et al. 2006). However, this study is the first to compare the impact of sporulation in biofilms with the impact of sporulation in other conditions on a combination of spore characteristics such as germination, heat resistance and size. Now, it will be essential to further study the outer layers of the spore, including the cortex, the coat and the exosporium, and to study the germinant receptors, to see whether the germination diversity between spore batches from different backgrounds can be linked to altered outer layers and/or germinant receptor composition and/or activity (Moir 2006; Ghosh et al. 2008; Setlow 2012; Zhang et al. 2012).

Germination on rice was shown in this study to be inefficient for emetic toxin-producing strains (Fig. 1). This is somewhat surprising, because many emetic type gastrointestinal disease events are related to the consumption of farinaceous foods, such as rice and pasta (Kramer and Gilbert 1989; Shinagawa 1990). In the current study, strain F3752A/86 and AH187 showed germination on rice, and these strains were isolated from rice and a rice-associated outbreak, respectively. Three other strains isolated from rice or rice-associated outbreaks did not germinate on rice plates, together with three strains isolated from other foods. Notably, we have observed that the addition of a small amount of meat broth (0·1% w/w Lab-lemco) to rice was sufficient to induce 100% germination for all strains (data not shown). It will be interesting to analyse links between B. cereus-caused emesis and farinaceous foods that were either mixed or not mixed with meats and vegetables, because in the latter case, a range of germinants is provided leading to efficient germination of spores that survive the first processing steps.

In conclusion, the five different sporulation conditions were shown to influence size, heat resistance and germination capacity of spores derived from emetic toxin-producing B. cereus strains. In addition, clear differences in spore properties were observed between the eight strains tested, with no apparent relation with the origin of these strains. Finally, in industrial settings, the production of spores may occur in different settings, including biofilms, a condition shown to be associated with the production of spores with increased dormancy and heat resistance.