The mechanism by which Bacillus cereus causes diarrhoea is unknown. Three putative enterotoxins have been proposed, haemolysin BL (Hbl), cytotoxin K and non-haemolytic enterotoxin (Nhe). Both Hbl and Nhe are three-component cytotoxins and maximal cytotoxicity of Nhe against epithelia is dependent on all three components. However, little is known of the mechanism of cytotoxicity. Markers of plasma membrane disruption, namely propidium iodide uptake, loss of cellular ATP and release of lactate dehydrogenase (LDH), were observed in epithelia exposed to Nhe from culture supernatants of B. cereus, but not in those exposed to supernatants from a mutant strain lacking NheB and NheC. Consistent with an exogenous cause of membrane damage, purified Nhe components combined to form large conductance pores in planar lipid bilayers. The inhibition of LDH release by osmotic protectants and the increase in cell size caused by Nhe indicate that epithelia lyse following osmotic swelling. Nhe and Hbl show sequence homology, and Hbl component B has remarkable structural similarities to cytolysin A (ClyA), with both structures possessing an a-helix bundle and a unique subdomain containing a hydrophobic b-hairpin. Correspondingly, we show that Nhe has haemolytic activity against erythrocytes from a variety of species. We propose that the common structural and functional properties indicate that the Hbl/Nhe and ClyA families of toxins constitute a superfamily of pore-forming cytotoxins.

CytK is a pore-forming toxin of Bacillus cereus that has been linked to a case of necrotic enteritis. PCR products of the expected size were generated with cytK primers in 13 of 29 strains. Six strains were PCR-positive for the related gene hly-II, which encodes haemolysin II, a protein that is 37 % identical to the original CytK. Five of the strains were positive for both genes. The DNA sequences of putative cytK genes from three positive strains were determined, and the deduced amino acid sequences were 89 % identical to that of the original CytK. The authors have designated this new cytK variant cytK-2, and refer to the original cytK as cytK-1. The CytK-2 proteins from these three strains were isolated, and their identity was verified by N-terminal sequencing. BLAST analysis using the cytK-2 gene sequences revealed very high homology with two cytK-2 sequences in the genomes of B. cereus strains ATCC 14579 and ATCC 10987. The differences between CytK-1 and the CytK-2 proteins were clustered to certain regions of the proteins. The isolated CytK-2 proteins were haemolytic and toxic towards human intestinal Caco-2 cells and Vero cells, although their toxicity was about 20 % of that of CytK-1. Both native and recombinant CytK-2 proteins from B. cereus 1230-88 were able to form pores in planar lipid bilayers, but the majority of the channels observed were of lower conductance than those created by CytK-1. It is likely that CytK-2 toxins contribute to the enterotoxicity of several strains of B. cereus, although not all of the CytK-2 toxins may be as harmful as the CytK-1 originally isolated.

The multidrug resistance P‐glycoprotein (P‐gp), which transports hydrophobic drugs out of cells, is also associated with volume‐activated chloride currents. It is not yet clear whether P‐gp is a channel itself, or whether it is a channel regulator. Activation of chloride currents by hypotonicity in cells expressing P‐gp was shown to be regulated by protein kinase C (PKC). HeLa cells exhibited volume‐activated chloride currents indistinguishable from those obtained in P‐gp‐expressing cells except that they were insensitive to PKC. HeLa cells did not express detectable P‐gp but, following transient transfection with cDNA encoding P‐gp, the volume‐activated channels acquired PKC regulation. PKC regulation was abolished when serine/threonine residues in the consensus phosphorylation sites of the linker region of P‐gp were replaced with alanine. Replacement of these residues with glutamate, in order to mimic the charge of the phosphorylated protein, also mimicked the effects of PKC on channel activation. These data demonstrate that PKC‐mediated phosphorylation of P‐gp regulates the activity of an endogenous chloride channel and thus indicate that P‐gp is a channel regulator.

No abstract

This study focuses on the interaction of the three components of the Bacillus cereus

Cell volume is frequently down‐regulated by the activation of anion channels. The role of cell swelling‐activated chloride channels in cell volume regulation has been studied using the patch‐clamp technique and a non‐invasive microspectrofluorimetric assay for changes in cell volume. The rate of activation of these chloride channels was shown to limit the rate of regulatory volume decrease (RVD) in response to hyposmotic solutions. Expression of the human MDR1 or mouse mdr1a genes, but not the mouse mdr1b gene, encoding the multidrug resistance P‐glycoprotein (P‐gp), increased the rate of channel activation and the rate of RVD. In addition, P‐gp decreased the magnitude of hyposmotic shock required to activate the channels and to elicit RVD. Tamoxifen selectively inhibited both chloride channel activity and RVD. No effect on potassium channel activity was elicited by expression of P‐gp. The data show that, in these cell types, swelling‐activated chloride channels have a central role in RVD. Moreover, they clarify the role of P‐gp in channel activation and provide direct evidence that P‐gp, through its effect on chloride channel activation, enhances the ability of cells to down‐regulate their volume.

Clostridium perfringens is an anaerobic bacterium that causes numerous important human and animal diseases, primarily as a result of its ability to produce many different protein toxins. In chickens, C. perfringens causes necrotic enteritis, a disease of economic importance to the worldwide poultry industry. The secreted pore-forming toxin NetB is a key virulence factor in the pathogenesis of avian necrotic enteritis and is similar to alpha-hemolysin, a β-barrel pore-forming toxin from Staphylococcus aureus. To address the molecular mechanisms underlying NetB-mediated tissue damage, we determined the crystal structure of the monomeric form of NetB to 1.8 Å. Structural comparisons with other members of the alpha-hemolysin family revealed significant differences in the conformation of the membrane binding domain. These data suggested that NetB may recognize different membrane receptors or use a different mechanism for membrane-protein interactions. Consistent with this idea, electrophysiological experiments with planar lipid bilayers revealed that NetB formed pores with much larger single-channel conductance than alpha-hemolysin. Channel conductance varied with phospholipid net charge. Furthermore, NetB differed in its ion selectivity, preferring cations over anions. Using hemolysis as a screen, we carried out a random-mutagenesis study that identified several residues that are critical for NetB-induced cell lysis. Mapping of these residues onto the crystal structure revealed that they were clustered in regions predicted to be required for oligomerization or membrane binding. Together these data provide an insight into the mechanism of NetB-mediated pore formation and will contribute to our understanding of the mode of action of this important toxin.

Caco-2 cells are models of absorptive enterocytes. The net transport of fluid from apical to basolateral surfaces results in ‘domes' forming in differentiated monolayers. Here, the effect of Campylobacter jejuni on this process has been examined. C. jejuni caused no changes in short-circuit current upon infection of Caco-2 cell monolayers in Ussing chambers. Thus, no active secretory events could be demonstrated using this model. It was therefore hypothesized that C. jejuni could inhibit the absorptive function of enterocytes and that this may contribute to diarrhoeal disease. C. jejuni infection of fluid-transporting (‘doming’) Caco-2 cells resulted in a significant reduction in dome number, which correlated with a decrease in tight junction integrity in infected monolayers, when measured as transepithelial electrical resistance. Defined mutants of C. jejuni also reduced dome numbers in infected monolayers. C. jejuni also altered the distribution of the tight junction protein occludin within cell monolayers. The addition to monolayers of extracellular gentamicin prevented these changes, indicating the contribution of extracellular bacteria to this process. Thus, tight junction integrity is required for fluid transport in Caco-2 cell monolayers as leaky tight junctions cannot maintain support of transported fluid at the basolateral surface of infected cell monolayers. Inhibition of absorptive cell function, changes in epithelial resistance and rearrangement of tight junctional proteins such as occludin represent a potential diarrhoeal mechanism of C. jejuni.