OUP user menu

Expression of Bacillus cereus hemolysin II in Bacillus subtilis renders the bacteria pathogenic for the crustacean Daphnia magna

Elena V. Sineva, Zhanna I. Andreeva-Kovalevskaya, Andrey M. Shadrin, Yury L. Gerasimov, Vadim I. Ternovsky, Vera V. Teplova, Tatyana V. Yurkova, Alexander S. Solonin
DOI: http://dx.doi.org/10.1111/j.1574-6968.2009.01742.x 110-119 First published online: 1 October 2009


Hemolysin II (HlyII) is a pore-forming toxin of the opportunistic pathogen Bacillus cereus. Despite our understanding of the mechanism of HlyII cytotoxicity in vitro, many of its characteristics, including potential target cells, conditions of its action and expression, are not known. Here we report that the expression of hlyII in Bacillus subtilis renders the bacteria hemolytic and is able to kill the crustacean Daphnia magna. The hemolytic activity of hlyII-encoded B. subtilis strains in culture media is positively correlated with virulence in D. magna. Fluorescence microscopy reveals postinfection changes in the mitochondrial potential of intestinal tissue, suggesting that the formation of ionic pores leads to cell death. In the presence of the transcriptional regulator HlyIIR, HlyII expression decreases 200-fold, and B. subtilis expressing both hlyII and hlyIIR remains hemolytic, but not pathogenic to the crustacean.

  • Bacillus cereus
  • hemolysin
  • HlyII
  • fluorescence microscopy
  • daphnia
  • animal model


Microorganisms of the Bacillus cereus group are closely related (Helgason et al., 2000; Tourasse et al., 2006; Stenfors Arnesen et al., 2008) despite the fact that they adopt different ecological niches and display distinct virulence properties: Bacillus anthracis is the etiological agent of anthrax, Bacillus thuringiensis is an insect pathogen and B. cereus is an opportunistic, spore-forming microorganism capable of producing both intestinal and nonintestinal pathologies in humans (Jensen et al., 2003). The pathogenic properties of B. cereus can be attributed to a number of pathogenicity factors including proteins, protein complexes and nonribosomally synthesized peptides (Granum & Lund, 1997; Andreeva-Kovalevskaya et al., 2008; Stenfors Arnesen et al., 2008). One of the potential pathogenicity factors, hemolysin II (HlyII), was discovered several years ago (Sinev et al., 1993; Budarina et al., 1994). HlyII is a pore-forming protein belonging to the broad class of β-pore-forming toxins (Miles et al., 2002; Andreeva et al., 2007). HlyII is synthesized in bacteria as a precursor with a 31-aa signal peptide that allows the secretion of the 42-kDa soluble HlyII monomer into the culture media. HlyII can disrupt membranes of erythrocytes and other eukaryotic cells in vitro by forming membrane ionic oligomeric pores that induce cell lysis (Andreeva et al., 2006, 2007).

Currently, nothing is known about when or how bacteria use HlyII in vivo. Bacillus cereus strain VKM B-771 is nonpathogenic to warm-blooded organisms and originated from the B. cereus W (ATCC 11950) (McCloy, 1958) and it is the B. cereus isolate that is most similar to B. anthracis, based on recently available whole-genome shotgun sequencing. However, the low pathogenicity of VKM B-771 could be due to tight transcriptional control of the hlyII gene by several transcriptional regulators (Budarina et al., 2004). Nonetheless, the exact conditions that trigger hlyII expression are still unknown. The structure of the major HlyII negative transcriptional regulator HlyIIR was solved recently. It shows that that the protein has a large internal cavity, which suggests that the activity of the regulator is modulated by ligand binding that induces HlyII production (Kovalevskiy et al., 2007).

The study of opportunistic pathogens and its pathogenicity factors requires an appropriate animal model (Scully & Bidochka, 2006). The ecologies of B. cereus and related bacteria have been poorly studied (Jensen et al., 2003). For some, the invertebrate intestine was identified as one of several natural niches. The crustacean Daphnia magna, widely used in water toxicology, offers many advantages as a model for research on bacterial protein toxins, including low cost and easy handling, self-reproduction by parthenogenesis and a fully transparent body that allows microscopic observation. Daphnia could use different bacteria at the same rate as algae. Crustaceans possess both highly evolutionarily conserved mechanisms of innate immunity and a primitive adaptive immune response. The host range of B. cereus is not well known. Crustaceans could coexist in the environmental niches of B. cereus, but generally, these bacteria are not regarded as disease causing in invertebrates.

The goal of this study was to analyze the function and regulation of HlyII as an individual cytotoxic factor in daphnia.

Materials and methods

Bacteria and plasmids

The strains and plasmids used are listed in Table 1. Bacillus strains were grown in Luria–Bertani (LB) medium. Transformation of Bacillus subtilis organisms was performed according to previously described procedures (Cutting & Youngman, 1994).

View this table:
Table 1

Characteristics of microbial strains and plasmids used in this study

DescriptionReferences or sources
B. subtilis
BD170BD 168 thr-5 trpC2E.U. Poluektova
BD170-BD170-EH2BD 170 thr-5 trpC2 CmramyEhlyIIThis work
BD170-BD170-EH2RBD 170 thr-5 trpC2 CmramyEhlyII-hlyIIRThis work
B. cereus
VKM B-771Sinev et al. (1993)
E. coli Z85thi, Δ(lac-proAB), Δ(srl-recA), hsdR, supE, Tn10(Tcr), (F+traD, proAB, lacI, DM15)DH5α derivative
pUJ1pUC19 with 2.9-kb EcoRI fragment of VKM B-771, containing hlyII and hlyIIRSinev et al. (1993)
pSWEET- bgaBB. subtilis vector for integration at amyE, pDG364 derivativeBhavsar et al. (2001)
pEHBAnalog of pDG364This work
pBD170-EH2pEHB with hlyII geneThis work
pBD170-EH2RpEHB with hlyII and hlyIIR genesThis work

The vector plasmid pEHB was constructed using the following strategy. Plasmid pSWEET-bgaB (6), a derivative of pDG364, was digested by HindIII and its large fragment was religated. Plasmid pUJ2 (Sinev et al., 1993) was used as the source of the hlyII genes (Baida et al., 1999).

To generate pBD170-EH2R, a 2.9-kb EcoRI fragment of pUJ1, which contained the hlyII and hlyIIR genes in tandem, was cloned into the EcoRI site of pEHB. Plasmid pBD170-EH2 contains only the hlyII gene. To construct pBD170-EH2, a 1929-bp EcoRI–BssSI fragment of pUJ1 was treated with a Klenow fragment. Both hlyII fragments were then cloned into the EcoRI site of pEHB that carries the upstream and downstream segments of the amyE gene of B. subtilis and facilitates insertion of DNA cloned between these segments, in trans, at the amyE locus. The pEHB derivatives were linearized by PstI and introduced into strain BD170 using DNA-mediated transformation with chloramphenicol selection. The genotypes of the transformants were confirmed using PCR amplification, followed by DNA sequencing.

Bacterial culture preparation

Bacterial cells were grown in LB for 16 h at 24 °C to reach the late stationary phase. The cell pellet was harvested by centrifugation at 10 000 g, washed twice with phosphate-buffered saline (PBS; 74 mM Na-phosphate, pH 6.8, 74 mM NaCl) and resuspended in distilled water. Daphnia media were supplied with bacteria by the addition of an amount of the concentrated suspension. The bacterial cells were stained in the presence of 0.001% of Acridine orange, followed by extensive PBS washing. OD600 nm values used as a measure of bacterial cell concentration refer to a 1-cm optical path length.

Quantitative hemolytic assay

The quantitative hemolytic assay was modified for 96-well plates (Michel et al., 1990) from a previously described assay (Sinev et al., 1993). A detailed description can be found in the Supporting Information.

Culture and experimental conditions of D. magna

A clonal strain of D. magna was maintained at Samara State University (Russia). Daphnia magna cultures were maintained at room temperature (20±5 °C) with a daylight (12 : 12; light/dark) cycle. Before the in vivo experiments, D. magna juvenile females were placed randomly into 500-mL jars containing 450 mL of hard dechlorinated tap water and fed with dry baker's yeast (3 g L−1) every 2 days. Newborn individuals from the second brood were used as experimental animals.

Toxicity tests were performed to determine the survival of D. magna in the presence of our bacterial strains. Water-washed Bacillus culture serial dilutions (200 mL) were added to 20 jars containing individual neonates (24-h old), so that the final concentration was 104–106 CFU mL−1. Daphnia were checked at the same time each day for the presence of offspring or death events. Death events were recorded on the fifth day of exposure. The timing of broods and the number of neonates were recorded until the complete release of the next brood of neonates or up to 14 days. Uninfected animals were kept as a control group. Three replicates were used for each series and two generations of animals were followed. Infected hosts were examined under a light microscope at × 25 and × 50 magnification.

Fluorescence microscopy

Daphnia magna tissue damage was visualized using different fluorescent probes. The cell membrane permeability changes were visualized with the viability stain Hoechst 33342, which penetrates both intact and damaged cells, and ethidium bromide, which penetrates only damaged cells and tissues. Mitochondrial and plasma membrane potentials were visualized using 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1) (Jaaskelainen et al., 2003). Female daphnia were incubated with 6 × 105 CFU mL−1 of bacteria at 10–15 days of age. Images of at least three crustaceans were obtained at each time point for every bacterial culture. Staining of an individual crustacean was performed in 6 μM JC-1 for 45 min at room temperature in the dark. Free dye was removed by washing in water and daphnia were immediately imaged on an upright fluorescence microscope Leica DM6000B with a DFC350FX-R2 digital camera.

Statistical analysis

Statistical analysis was performed using sigmaplot 9 for Student's t-test assuming unequal variances. P-values <0.05 were accepted as statistically significant.


HlyII expression is regulated by HlyIIR in B. subtilis

Bacillus cereus is known to secrete a number of proteins with hemolytic activity, but their individual effects are very difficult to evaluate. The major hemolytic factor, cereolysin (Granum & Lund, 1997), was shown to be inhibited by cholesterol (Sinev et al., 1993). The activity of HlyII in B. cereus was determined both in the presence and in the absence of cholesterol. Cholesterol-independent activity was the sum of activity of HlyII and some other minor cytotoxic factors (Budarina et al., 1994). To study the hlyII regulation in Bacillus, we introduced a single copy of hlyII and its regulatory elements into B. subtilis genome using allelic replacement (Fig. 1a). Bacillus subtilis BD170 lacks known hemolytic or cytotoxic extracellular activities and shares some common regulatory circuits with B. cereus; therefore, it could serve as a suitable organism for studying the expression of hlyII in culture media or a host organism.

Figure 1

Expression of the Bacillus cereus hlyII gene in Bacillus subtilis is tightly regulated. (a) The diagram represents the structure of the genomic inserts in B. subtilis BD170. Strain B. subtilis BD170-EH2R bears both hlyII and the regulator hlyIIR, whereas B. subtilis BD170-EH2 has only hlyII. Construction of the plasmids and strains is described in Materials and methods. (b) Growth curves (for B. subtilis strains) and specific hemolytic activity (HU/OD600 nm) of the B. cereus and B. subtilis mutant strain culture supernatants. Bacillus cereus generally grows to a higher OD. Thus, for illustration purposes, we superimposed the points according to the stages of bacterial growth from exponential to early stationary phases to the growth curve of B. subtilis BD170. All experiments were repeated three times and representative curves are shown.

To analyze the function and expression regulation of hlyII, we created B. subtilis strains with (BD170-EH2R) and without (BD170-EH2) hlyIIR (Fig. 1a). No hemolytic activity was detected in crude cell extracts when hlyII was inserted into the B. subtilis chromosome; instead, all hemolytic activity was detected in the culture medium (Fig. 1b).

Firstly, the effects of the hlyII inserts in the B. subtilis genome on the hemolytic activity were evaluated by observing the strain phenotypes on blood agar plates. Both isogenic strains (further referred to as BD170-EH2R and BD170-EH2) were hemolytic. However, BD170-EH2 exceeded other strains in showing hemolysis even before bacterial growth became visible (Supporting Information, Fig. S1). The hemolytic activity was then measured in the culture supernatants during the entire growth phase. In B. cereus B-771 and BD170-EH2, hemolytic activity followed the same pattern during growth, with their maximum HlyII-specific activities (30 and 1500 HU/OD, respectively) occurring at the end of the exponential phase, followed by a gradual decrease in the stationary phase (Fig. 1b). The kinetics of HlyII expression pattern of BD170-EH2R was slightly different, with an onset in the early stationary phase and a low specific activity (7 HU/OD). The hemolytic activity of BD170-EH2 was almost the same as the total hemolytic activity of the parental B. cereus B771 strain and significantly exceeded the hemolytic activity of BD170-EH2R and the cholesterol-independent activity of B. cereus. No hemolytic activity was found in the cultural media of parental BD170. These results are the first evidence that HlyIIR is an efficient negative regulator (>200-fold) of hlyII expression in B. subtilis.

However, hemolytic activity in culture supernatants does not necessarily correlate with hemolytic activity on agar or with the pathogenic properties of the strains, because bacterial metabolism during infection can differ significantly from that in culture media. Therefore, we developed an animal model to study hlyII regulation in B. subtilis. In an attempt to establish an appropriate animal model, we studied the effect of B. cereus vegetative cells and spores on D. magna. The crustaceans were adversely affected by the presence of B. cereus in a concentration-dependent manner (Fig. 2a). We explored the effect of BD170-EH2R and BD170-EH2 on the longevity and fecundity of D. magna and compared this effect with that of B. cereus B-771.

Figure 2

Survival and fecundity of daphnia upon exposure to hlyII-expressed strains are dose dependent. (a) Mortality of the Daphnia magna population upon treatment with Bacillus cereus B-771. A series of survival curves is presented at 1.5 × 105 CFU mL−1 (•, solid line), 3 × 105 CFU mL−1 (○, short dashed line) and 6 × 105 CFU mL−1 (▾, long dashed line). (b) Expression of the hlyII gene leads to acute toxicity of Bacillus subtilis BD170-EH2. Fractions of surviving daphnia were plotted against incubation time. Control group survival was 100%. (c) LC50 determination for B. subtilis BD170-EH2. Fractions of D. magna that survived until day 5 of the experiment were plotted against the OD of the corresponding bacterial cultures. Data were fit using sigmaplot 9.0 to a four-parameter logistic curve, using the equation Y=min+(max−min)/(1+(X/LC50)Hillslope). The LC50 was equal to 5 × 105 CFU mL−1. (d) Fecundity of D. magna in the presence of sublethal concentrations of bacteria (6 × 104 CFU mL−1). The number of offspring was counted on day 16. Data are expressed as the means and SD from at least three independent experiments. *Significantly different from the starvation group (P<0.05). All experiments were repeated three times and representative curves or averages are shown.

HlyII expressed in B. subtilis is toxic to D. magna

It is well known that the freshwater crustacean D. magna is a highly efficient grazer of aquatic bacteria (Kamjunke et al., 1999), capable of filtering organisms as small as 0.4–2 μm. However, the ability of the crustacean to consume B. cereus and B. subtilis has, to our knowledge, never been tested. To test whether B. subtilis could be a tolerable feed for D. magna, we compared the longevity of four groups of daphnia: one fed normally (dry yeast, ∼105 CFU mL−1, once in 2 days), one without food, one incubated with B. subtilis BD170 (parental strain, 5 × 105 CFU mL−1; further referred to as BD170) and one fed yeast and incubated with B. subtilis. Mortality in the starvation groups was gradual and similar to the Bacillus-fed group, whereas the control yeast-fed group showed no mortality (data not shown). We did not observe any significant clearance of bacterial culture. However, when daphnia were also fed with yeast, the presence of parental B. subtilis had no negative impact on the animal function during the experiment (16 days). These results indicate that this strain of D. magna cannot use B. subtilis as a sufficient source of food under our experimental conditions and cannot clear the Bacillus bacterial culture.

Next, we exposed D. magna to our bacterial cultures for 16 days (using different concentrations) and determined the chronic effect on daphnia longevity. The presence of BD170-EH2R or wild-type BD170 (control) had no negative consequences for the crustacean; all animals survived the 16-day incubation at all the bacterial concentrations used (up to 1.8 × 106 CFU mL−1). In contrast, incubation of daphnia with B cereus or BD170-EH2 was fatal for the crustacean (Fig. 2a and b). Both bacilli had a very similar dose-dependent effect on D. magna longevity, despite the fact that B. cereus possesses other potential pathogenicity factors in addition to HlyII. The lethal concentration 50% (LC50) on the fifth day of the experiments for BD170-EH2 was 5.4 × 105 CFU mL−1 (Fig. 2c), whereas it was 4.5 × 105 CFU mL−1 for B. cereus. We calculated the sublethal concentration (6 × 104 CFU mL−1) and the lethal concentration (LC100, 1.2 × 106 CFU mL−1) for use in the subsequent study. It may be worth noting that daphnia motility was impaired upon placement in water containing either B. cereus or BD170-EH2 cells, even in sublethal doses. Hemolytic activity was not detected in the daphnia culture media containing any Bacillus strain at any day of the experiment. Thus, the toxicity of bacterial cultures is a direct consequence of toxin production inside the crustacean body. The absence of detectable hemolytic activity in water may be the result of rapid inactivation of the hemolytic factors when excreted in water (Andreeva et al., 2006). For the same reason, coincubation of daphnia with purified HlyII solutions has no effect on crustacean longevity. However, it is more likely that most cytotoxic factors are produced under certain physiological conditions, for example, by the actively grown bacterial culture only (Fig. 1b). The physiological state of the bacterial culture in water was not certain and it may have represented a mixture of living, cultivable bacteria and living, noncultivable bacteria (Byrd et al., 1991).

We then tested B. cereus, BD170-EH2 and BD170-EH2R spores in similar experiments. Addition of spore suspensions from all bacterial species at very high concentrations (up to 2 × 109 CFU mL−1) was harmless for D. magna (data not shown).

The results suggest that HlyII is a potent cytotoxin that is produced by growing bacterial cells, but is tightly regulated in culture and during infection.

Effect of HlyII on D. magna fecundity

The significant effects on daphnia mortality were observed at high bacterial concentrations. To investigate more subtle effects, we analyzed the fecundity of D. magna. Virulence is often associated with pathogen-induced mortality, even though loss of fecundity is also common (Jensen et al., 2006). The reason for declines in the number of brood could be both a simple loss of host resources for benefits of pathogen transmission and pathogen-induced sterility. To address this question, we monitored changes in the daphnia population at sublethal (1.5 × 104 CFU mL−1) concentrations in all the strains under study (Fig. 2d). Daphnia reproduction began between 11 and 16 days of age (5–10 descendants, under normal conditions), and so we monitored the number of offspring on day 16. First, we found that a short break in feeding, as required by the experimental conditions, resulted in a decline in the number of offspring, as daphnia generally produce fewer offspring when food is limited. No dramatic effect on the fecundity was observed upon the addition of BD170-EH2R to the host at the same concentration. In the presence of BD170-EH2, there was a further decrease in fecundity. Finally, daphnia almost ceased reproduction in the presence of B. cereus at the sublethal threshold. An increase in the bacterial concentration (to 3 × 105 CFU mL−1) had no significant effect on host fecundity for either the wild-type BD170 or the BD170-EH2R strain, although it drastically affected host survival in the presence of B. cereus and BD170-EH2 strains. The effect on fecundity was most likely due to changes in the vital function of the parent animals and not due to D. magna castration or direct killing of daphnia embryos.

Bacillus cereus and B. subtilis BD170-EH2 kill D. magna by intestinal disruption

To investigate the cause of daphnia lethality mediated by HlyII, we monitored the postinfection changes using a light microscope (Fig. S2). Juvenile daphnia images lacked sufficient contrast. Thus, to allow for a better view of the process, microscopic examinations were conducted in adult animals of about the same age (10–15 days). Both the control groups (B. subtilis BD170 and yeast-fed) seemed to be healthy and active during the experiment (Fig. 3a). Both B. subtilis BD170-EH2 and B. cereus cells killed daphnia under experimental conditions equally well (Fig. S2c and d).

Figure 3

In vivo light microscopy of daphnia reveals the distribution of bacteria through the crustacean body at different stages of infection. (a) Uninfected daphnia. (b) Representative image of a crustacean incubated with either Bacillus cereus B-771 or Bacillus subtilis BD170-EH2 (1.2 × 106 CFU mL−1, 12 h) prestained with Acridine orange. Primary accumulation of bacteria in the daphnia gut is visible. (c) Typical image of daphnia after 24 h exposure to the above conditions. (d) Pregnant daphnia incubated under conditions described for (b). Scale bar=1 mm.

To visualize the action of the hemolysin-expressing bacteria, we stained the Bacillus strains with Acridine orange. Daphnia were fed stained bacterial cells and examined under a microscope. First, we observed accumulation of the yellow dye in the intestine (Fig. 3b) and then a gradual lysis of the gut tissues. In the next stage, stained bacteria were found outside the intestine (Fig. 3c). These processes were followed by the death of daphnia. A similar result was obtained with pregnant daphnia (Fig. 3d). No accumulation of dye was observed in the brood pouch. However, due to the low sensitivity of the method, we cannot assertively conclude that bacteria had no access to the embryos. Accumulation of stained bacteria in the part of the intestine approaching the crustacean brain (under-eye area) could explain the rapid effect of HlyII-expressing bacteria on daphnia mobility. No significant changes were observed in the crustacean body during the same experiments with the control nonhemolytic strain of B. subtilis. However, prolonged incubation with any Acridine orange-stained bacteria decreased daphnia life span.

In vivo fluorescence microscopy

To further explore the cytotoxic effect of HlyII on the crustacean, we used fluorescent probes. Fluorescence microscopy has been used previously for the in vitro detection of microbial toxin toxicity on boar sperm and various human cells (Jaaskelainen et al., 2003). However, fluorescence probes were not used for in vivo studies in daphnia. Firstly, we confirmed that the attack of hemolysin-expressing bacteria on daphnia started in the intestine. At different times of infection, crustaceans were simultaneously stained with Hoechst 33342 and ethidium bromide. Incubation of daphnia with B. subtilis BD170-EH2 was found to change the fluorescence of the gut tissues from green (fluorescence of vital dye Hoechst 33342) to orange and red (fluorescence of ethidium bromide, which enters only damaged tissue) (Fig. S3). Thus, it can be concluded that HlyII produced by BD170-EH2 acts in the intestine.

Earlier, we reported that the HlyII efficiently lysed cultured human cells, which was accompanied by a decline in the mitochondrial transmembrane potential (Ψm) (Andreeva et al., 2006). To confirm that a similar mechanism operated in daphnia, we followed the changes of Ψm in whole crustaceans upon infection using JC-1 staining and fluorescence microscopy. The effect on Ψm of D. magna exposure to different bacterial cultures is shown in Figs 4 and 5 (× 5 images in Figs S4 and S5).

Figure 4

Intestinal tissues are affected during infection. Viable daphnia were exposed to Bacillus for 1 day, stained with JC-1, a potential-dependent dye, and visualized using a fluorescence microscope. Daphnia were stained in 6 μM JC-1 for 45 min in the dark. The image was taken immediately following a brief water wash. Regions with high membrane potential emit red fluorescence due to J-aggregate formation by concentrated JC-1. Depolarized regions emit green or yellow due to the fluorescence of JC-1 monomers. Scale bar=1 mm. (a) Uninfected daphnia; (b) Bacillus subtilis BD170; (c) B. subtilis BD170-EH2; and (d) Bacillus cereus B771.

Figure 5

Daphnia death is the result of bacterial toxin spread throughout the daphnia body. Potential-dependent JC-1 fluorescence of viable infected daphnia was visualized on day 4 using a fluorescence microscope, as described above. The observed deep disruption of the intestine was followed by propagation of the infection process to other tissues. Scale bar=1 mm. (a) Uninfected daphnia, control; (b) Bacillus subtilis BD170; (c) B. subtilis BD170-EH2; and (d) Bacillus cereus B-771.

The tissues of crustaceans not exposed to bacteria (Fig. 4a) or incubated with wild-type B. subtilis BD170 for 1 day (Fig. 4b) fluoresced orange-red, indicating that all tissues were intact and maintained a high membrane potential. However, after a 1-day incubation of daphnia with BD170-EH2 (Fig. 4c) or B. cereus (Fig. 4d), fluorescence emission of the gut tissues changed from red to green, indicating a decline of Ψm in the intestinal cells. Thus, during the first day of incubation, the changes were observed only in the gut tissues that could be seen clearly in images of higher magnification (Fig. S4). After a 4-day exposure, the effects developed further (Fig. 5c and d) and the changes affected the whole daphnia body, indicating the spreading of either bacteria or bacterial toxins. The control daphnia group, which was not exposed to bacteria, was intact after 4 days (Fig. 5a). Surprisingly, incubation with wild-type B. subtilis BD170 for 4 days led to the appearance of a weak green fluorescence in the gut, suggesting that these bacilli probably produce some substances that decrease mitochondrial membrane potential and may have a weak cytotoxic activity (Fig. 5b). This finding could explain the significant decline in fecundity in the presence of B. subtilis BD170 compared with the yeast-fed group (Fig. 2d). These results demonstrate that the attack of both BD170-EH2 and B. cereus on D. magna started in the intestinal epithelium and was followed by a decrease of the Ψm in the other tissues, which was seen more clearly on fluorescence patterns of larger sizes (Fig. S5).


Here, we report the first evidence that bacteria expressing HlyII can be lethal for D. magna even when acting without other B. cereus pathogenicity factors. It is still unknown whether crustaceans are actual targets for HlyII-expressing bacteria in the natural environment. However, the widespread application of B. thuringiensis spores in agriculture has made the numerous interactions of crustaceans and the microorganism possible.

Bacillus cereus and related species can be natural coinhabitants of crustaceans, which inhabit sands or soils. For example, a recent study (Hendriksen et al., 2006) demonstrated the abundance of B. cereus in sandy loams (about 105 CFU g−1 soil). Most importantly, many of these isolates carry a potentially functional hlyII gene. The expression of virulence factors is optimal at the temperature of their natural host. Taking into account that the optimum temperature for HlyII expression and action is between 15 and 28 °C (Sinev et al., 1993; Andreeva et al., 2007), it is not surprising that B. cereus could use this weapon in the crustacean gut. Thus, we cannot exclude that some crustaceans, perhaps including D. magna, can be a relevant host for B. cereus, particularly when saprophyte food is limited.

The study of opportunistic human pathogens and its pathogenicity factors requires a suitable animal system. Insects, nematodes and crustaceans are particularly useful because of the similarities between human and invertebrate immune systems (Scully & Bidochka, 2006) and a possibility that abundant invertebrates could serve as a reservoir of emerging diseases. Here, we report that D. magna can be exploited as an animal model investigating the toxicity of individual B. cereus pathogenic factors when expressed in the neutral carrier bacteria B. subtilis. Two major advantages of this model are the ability to test the activities and regulation of individual B. cereus pathogenic factors in an appropriate animal system and to study structure–function relationships of pore-forming toxins. Our results show that D. magna can serve as a novel model system for studying the effects of toxins on cell signaling and ion channels in situ.

In conclusion, the present fluorescence microscopy study showed that B. subtilis BD170-EH2 secreted HlyII into the intestine of D. magna that probably induced a decrease in ΔΨm, which was due to the pore-forming and membrane-associating properties of the HlyII protein. Secreted B. cereus HlyII from B. subtilis cells initially forms pores in intestinal cell membrane, which leads to a perturbation of ion homeostasis and finally to necrosis of the daphnia cells. The spreading toxins lead to changes in not only gut tissues but also other tissues of the daphnia body, resulting in animal death. It can be concluded that HlyII induces cell lysis that is accompanied by changes in the mitochondrial membrane potential.

Authors' contribution

E.V.S. and Z.I.A.-K. contributed equally to this work.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Comparative hemolysis of isogenic Bacillus strains on blood agar; (wt) – B. subtilis BD170; (EH2) – B. subtilis::hlyII; (EH2R) – B. subtilis::hlyII – hlyIIR. Bacterial cultures were grown overnight in LB and spread on agar using an inoculation loop. Plates were incubated at 24 °C and photographed at the indicated time points.

Fig. S2. Different stages of daphnia B. subtilis H2 and B. cereus infection analyzed by light microscopy. Microscopic examination of D. magna exposed to the corresponding Bacillus strains (0.1 OD) and monitored over time for viability. Photomicrograph of crustaceans: alive (A–B), after 2 days of exposure to the wild type B. subtilis BD170 (A) and H2 (B); postmortem examination of dead daphnia (C–D), after 3 days of incubation with B. subtilis H2 (C) and B. cereus B771 (D).

Fig. S3. In vivo imaging of daphnia during the incubation of EH2. Fluorescence microscopy of daphnia stained simultaneously with Hoechst 33342 and ethidium bromide. Daphnia were stained in 3 mM of Hoechst 33342 and 3 mM of ethidium bromide in the dark for 45 min, rinsed with water and immediately photographed on an upright fluorescence microscope Leica DM6000B. Scale bar is 1 mm. A. Control daphnia. Healthy tissues are bright green. Ethidium can enter only into the damaged cells. B. Daphnia incubated with 0.2 OD of EH2 for 24 h. Intestinal tissues are fluorescent red and the beginning of intestinal disruption is seen.

Fig. S4. Potential-dependent JC-1 staining of live daphnia visualized by fluorescence microscopy. Intestinal tissues are affected during the initial infection steps. Fluorescent images of JC-1-stained D. magna exposed with different bacterial strains (1 day, 0.2 OD). Regions of high mitochondrial polarization are indicated by red fluorescence due to J-aggregate formation by the concentrated dye. Depolarized regions are indicated by transition to green or yellow fluorescence of JC-1 monomers. Bar is 0.1 mm. A. Uninfected daphnia without any bacteria; B. B. subtilis BD170; C. B. subtilis EH2; D. B. cereus B771.

Fig. S5. Potential-dependent JC-1 staining of live daphnia visualized by fluorescence microscopy. Deep disruption of the gut was followed by the propagation of the infection process to other tissues. Fluorescent images of JC-1-stained D. magna exposed with different bacterial strains (4 days, 0.2 OD). Daphnia were stained in 6 μM JC-1 for 45 min in the dark. The image was taken immediately after a brief water wash. A. Uninfected daphnia without any bacteria; B. B. subtilis BD170; C. B. subtilis EH2; D. B. cereus.


This work was supported by grants from the Russian Foundation for Basic Research (07-04-01706 and 08-04-01424). We thank E.U. Poluektova (Vavilov Institute of General Genetics, Russia) for the strain B. subtilis BD170, A. Sorokin (Institut National de la Recherche Agronomique, France) for the plasmid pSWEET-bgaB, A.P. Bhavsar for providing the plasmid map, V.A. Yashin (Institute of Cell Biophysics, Russia) for the use of the Leica fluorescent microscope and V.V. Hauryliuk for critically reading the manuscript.


  • Editor: Marco Soria


View Abstract