| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, PO Box 8146 Dep., N-0033 Oslo, Norway
Correspondence
Simon P. Hardy
simon.hardy{at}veths.no
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-helix bundle and a unique subdomain containing a hydrophobic β-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.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), the cellular mechanism of diarrhoea and identity of the responsible enterotoxin(s) remain unknown. Three putative enterotoxins of B. cereus have been proposed: haemolysin BL (Hbl), non-haemolytic enterotoxin (Nhe) and cytotoxin K (CytK), all of which are cytotoxic to epithelia in vitro. CytK is a haemolytic pore-forming toxin with homology to the β-barrel pore-forming toxins, including staphylococcal -haemolysin (Hardy et al., 2001a; Lund et al., 2000). Hbl and Nhe are tripartite toxins, in which all three components are necessary for maximal cytotoxic activity (Beecher & Macmillan, 1991; Lindbäck et al., 2004). Both are encoded by three genes co-transcribed as operons in which hblCDA encodes Hbl components L2, L1 and B, and nheABC encodes NheA, NheB and NheC. Sequence homology is apparent both between the three components in each complex and between the proteins of Nhe and Hbl (Granum et al., 1999; Ryan et al., 1997; Table 1). With no sequence homology to other toxins, Nhe and Hbl have remained unclassified as toxins.
|
). Nhe was identified in B. cereus following a large food-poisoning outbreak in Norway (Lund & Granum, 1996). The recovered isolate, NVH 0075/95, lacked both hbl and cytK (Ehling-Schulz et al., 2005) but was still cytotoxic, thus permitting the discovery of Nhe. The initial study indicated that the toxin was non-haemolytic.
Although the toxin profile may differ between strains, it appears that Nhe is a dominant cytotoxic component of B. cereus. In a study of over 100 B. cereus strains, cytotoxicity correlated well with the amount of Nhe produced but poorly with the concentration of Hbl (Moravek et al., 2006). The same group (Dietrich et al., 2005) found that mAb 1E11 raised against NheB prevented cytotoxicity in 20 out of 20 strains of B. cereus isolated from food. In a study of Bacillus thuringiensis strain 407 Cry– (indistinguishable from B. cereus), culture supernatants from hbl and cytK deletion mutants showed no impairment in cytotoxicity towards HeLa and Caco-2 cells (Ramarao & Lereclus, 2006), implying that Nhe was the major virulence factor.
Previously, we have shown that maximal cytotoxic activity of Nhe against Vero cell epithelia is dependent on all three Nhe components (Lindbäck et al., 2004). Cytotoxicity was quantified as inhibition of protein synthesis based on the impairment of leucine uptake (Lindbäck & Granum, 2006). With rapidly cytotoxic proteins this assay may be more a measure of uptake of radiolabelled leucine across the epithelial plasma membrane than of protein synthesis per se. Thus, following microscopic observations of increasing size of blebs in cells in response to B. cereus culture supernatants, we chose to investigate the nature of the plasma membrane damage using more direct markers. Using the B. cereus strain NVH 0075/95 and an isogenic nheBC mutant expressing only NheA, we show that Nhe acts as a pore-forming toxin to induce cell lysis. Additionally, we show structural and functional correlates between Nhe, Hbl and the pore-forming haemolysin cytolysin A (ClyA) from Gram-negative enteric bacteria.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), and 1C2, raised against NheB (Dietrich et al., 1999), were used for immunoblotting. mAb 1E11, raised against NheB, is able to neutralize the cytotoxic activity of Nhe, as described by Dietrich et al. (2005).
Creation of the nheBC mutant.
A truncation mutant in nheB of B. cereus NVH 0075/95, isolated from an outbreak of food poisoning in Norway (Lund & Granum, 1996), was created fortuitously whilst deleting nheC. Thus, the method described here is that devised to delete nheC alone. Subregions of the nheB and nheC genes were amplified by PCR using primer pairs GCGAAGCAATGGTTAGATGTA and CACAACTGAACAGGAGCTTC, and GCGATTGATCAAAAGGATAG and AAAAATAAACAGCAGAACTAGTCCAC, respectively. PCR was carried out for 30 cycles using an annealing temperature of 50 °C. PCR products of nheB and nheC were cloned into pCR T7/NT-TOPO and pCR 2.1-TOPO (Invitrogen), respectively, and transformed into Escherichia coli One Shot TOP10 (Invitrogen). nheC was excised from pCR2.1-TOPO using EcoRI, and ligated into the EcoRI site of pMAD (Arnaud et al., 2004). nheB was excised from pCRT7/NT-TOPO using BamHI and HindIII, and ligated into the BamHI site of pMAD together with the spectinomycin resistance cassette from pDG1726 (Guérout-Fleury et al., 1995) digested with BamHI and HindIII, producing pMAD
nheC. DNA sequencing confirmed that the plasmid did not contain any mutations in nheB introduced during PCR. This plasmid was electroporated into B. cereus NVH 0075/95 (Masson et al., 1989), and transformants were subjected to allelic exchange as described elsewhere (Arnaud et al., 2004). DNA sequencing of the obtained double-crossover mutant confirmed the deletion of nheC, while a single-base deletion of the 967th base (guanosine) was introduced into nheB. This deletion is likely to result in a truncated form of NheB due to a frameshift, as the last 80 amino acids were absent.
Crude toxin preparation (culture supernatants).
B. cereus NVH 0075/95 and the isogenic nheBC mutant were grown in a modified casamino acids/glucose/yeast extract (CGY) broth as defined by Lund & Granum (1997), but with the 1 % glucose replaced by 1 % sucrose (Ouhib et al., 2006). A 1 % inoculum of an overnight culture was incubated in 50 ml CGY (in a 250 ml flask) at 32 °C shaken at 100 r.p.m. for 5 h (early stationary phase). The supernatant was centrifuged and filtered through a 0.2 µm pore-size membrane filter and stored in aliquots at –80 °C.
Purification of Nhe components.
NheA and NheB were purified from 5 h culture supernatants of B. cereus NVH0075/95 as described previously (Lindbäck et al., 2004), with the following modifications: 1 l CGY was used to culture the organism rather than 3 l, and the proteins were precipitated with 50 % (w/v) ammonium sulphate. NheB was analysed by DEAE chromatography at pH 8. NheC was purified as a recombinant hexahistidine-tagged protein expressed in E. coli (Lindbäck et al., 2004). Protein concentrations were estimated using the absorbance method of Whitaker & Granum (1980), in conjunction with visual assessment of titrations of the purified Nhe components from band densities on SDS–PAGE gels.
SDS–PAGE and immunoblotting.
PAGE and Western immunoblotting were carried out as described previously (Lindbäck et al., 2004), using mAbs 1A8 (Dietrich et al., 2005) and 1C2 (Dietrich et al., 1999). Biotin-conjugated goat anti-mouse antibody was used as a secondary antibody (1 : 1000). A complex of streptavidin and biotinylated alkaline phosphatase was used at a dilution of 1 : 3000, prior to developing with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate.
Propidium iodide (PI) uptake by Caco-2 and Vero epithelia.
For microscopy studies, Vero and Caco-2 cells were grown on coverslips in 6-well microplates and tested after 3 days, having reached 80–90 % confluence. Cells were washed with 1 ml EC buffer solution and incubated with dilutions of culture supernatants. Cells were incubated in 5 % CO2 for up to 60 min. At various time points the medium was replaced with fresh medium containing 5 µg PI ml–1 (7.5 µM) and the coverslips were removed for examination by epifluorescence microscopy at the end of the experiment. A halogen light source was set to 570/620 nm excitation/emission wavelengths. Time-course experiments of PI fluorescence in epithelia were performed using a Wallac Victor3 fluorimeter (Perkin Elmer). Caco-2 cells were cultured in 24-well microplates and tested 10–14 days after seeding, while Vero cells were tested 2–4 days after seeding. After equilibrating cells in EC buffer for 10–15 min, microplates were transferred to the heated chamber of the Victor3 (held at 37 °C), the bathing solution (buffer) was replaced with one containing PI (5 µg ml–1), and dilutions of culture supernatant were added. Microplates were ‘bottom read’ to record fluorescence in cells attached to the base of the well using excitation/emission wavelengths of 575/595 nm. Wells were read at 1 min intervals for 1 s duration. Saponin (0.0025 %, w/v, in EC buffer) was used as a positive control to monitor fluorescence. Results are shown without subtraction of background fluorescence; hence baseline values increased with increasing volumes of culture supernatant. Note also that since divalent cations quench fluorescence (Gibbons et al., 2001), absolute values are not of any significance. This was particularly noticeable in calcium- and magnesium-free buffer, which gave higher basal readings than the standard EC buffer. The rate of PI uptake was expressed as the mean change in fluorescence per minute measured over a 10–15 min period, starting with the lowest reading obtained before fluorescence increased.
Epithelial ATP content and lactate dehydrogenase (LDH) release.
Cell culture medium was replaced with EC buffer and allowed to equilibrate for at least 10 min before exposure to culture supernatant in 1 ml final volumes. For measurement of LDH release following exposure of Vero and Caco-2 monolayers to the culture supernatant, duplicate samples of EC buffer were removed at various intervals and briefly centrifuged before analysis on an ADVIA 1650 autoanalyser (Beyer). Total cell monolayer ATP was measured by replacing the entire cell bathing solution after exposure to toxin at various time points with 1 ml buffer containing 1 % (v/v) Triton X-100. After 10 min incubation, duplicate 100 µl samples of the lysed cell suspensions were diluted 1 : 100 before mixing with 100 µl luciferase reaction buffer (BacTiter-Glo Microbial Cell Viability Assay, Promega) for 5 min and measurement in a Glomax 20/20 luminometer (Promega). Results were expressed in the original arbitrary lux units. Total monolayer LDH was measured at the end of the experiment after lysis with 1 ml 1.0 % Triton X-100 in EC buffer.
For osmotic protection of LDH release, glucose (30 mM, hydrodynamic radius 0.36 nm), PEG 400 (30 mM, 0.68 nm), PEG 1000 (28 mM, 0.94 nm) and PEG 8000 (10 mM, 3.78 nm) were dissolved in EC buffer. Osmotic protectants were added in reducing molarities so as to offset the increase in osmolarity in the final bathing solutions. Cells were preincubated for 10–15 min in the prewarmed solutions containing PEG before addition of 20 µl ml–1culture supernatants. Hydrodynamic radii were taken from Scherrer & Gerhardt (1971) and Planchot et al. (2000).
Planar lipid bilayer experiments.
Investigation of the ability of purified Nhe to form single channel-like pores in synthetic phosphatidylethanolamine–phosphatidylserine (PE–PS) lipid bilayers was carried out as described previously (Hardy et al., 2001b). Purified Nhe components in 0.1 M NaCl phosphate buffer in a ratio of 1 : 1 : 0.3 NheA : NheB : NheC, using approximately 40 ng NheA, were premixed before addition to the earthed bathing solution. Recordings were filtered at 500 Hz and acquired to computer disc using a Digidata 1200 AD converter at 5 kHz, and analysed off-line. Single-channel conductances were estimated individually using Win EDR (Strathclyde Electrophysiology Software).
Flow cytometric measurement of cell size and PI uptake.
Vero cells were detached from culture flasks by standard protocols with trypsin/EDTA inactivated by exposure to 10 % fetal calf serum in the growth medium. Prior to exposure to the toxin the cells were equilibrated at 37 °C in EC buffer with 5 µg PI ml–1 for 10 min. Vero cell suspensions were assayed for forward light scatter and PI uptake using a FACSCalibur flow cytometer (Becton Dickinson Biosciences) with a 488 nm wavelength argon laser with linear amplification of the forward- and side-scatter light signals and logarithmic amplification of the fluorescence signals of each cell. Data were collected from 10 000 cells per time point. Cells were held at room temperature for the duration of the experiment (15 min). Analysis was carried out using CellQuest Pro software (Becton Dickinson Biosciences). Following normalization of cell size at time zero, statistical comparison between the effect of 20 µl ml–1 culture supernatants of NVH 0075/95 and the nheBC mutant on cell size (mean forward scatter value) was carried out on the area under the curve for the two samples using the Mann–Whitney test.
Haemolysis assays.
Blood cells from different species were washed three to six times in PBS. Purified Nhe proteins were used at a ratio of 6 : 6 : 1 NheA : NheB : NheC, using approximately 100 ng NheA ml–1. Volumes of culture supernatants were incubated with 2 % (v/v) bovine blood, and freshly combined Nhe components were incubated with blood samples from different species diluted to concentrations obtaining the same OD630 as 1.5 % human blood. After incubation at 37 °C for 60 min on a roller incubator, samples were centrifuged and haemolysis was determined from the A540 of the supernatants. The percentage of haemolysis was calculated by comparing the A540 of the samples with positive (100 % lysis by 1 % Triton X-100) and negative controls. All experiments were performed at least twice.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
nheBC strain, in contrast to the wild-type strain NVH 0075/95 (Fig. 1). Growth curves were indistinguishable between the two strains, each yielding 108 c.f.u. ml–1 after 5 h incubation at 32 °C.
|
). In contrast, cells exposed to up to 50 µl ml–1 supernatant from the nheBC mutant for up to 60 min excluded PI and appeared morphologically indistinguishable from cells exposed to culture medium alone (Fig. 2b).
|
shows the increase in PI fluorescence from Caco-2 cells over 20 min following exposure to NVH 0075/95 culture supernatant, and the rate of propidium fluorescence increasing in a dose-dependent manner (Fig. 3b). In contrast, Caco-2 cells exposed to up to 50 µl ml–1 culture supernatant from the nheBC mutant exhibited no increase in fluorescence (Fig. 3a). PI uptake in Caco-2 cells exposed to 20 µl ml–1 NVH 0075/95 culture supernatant was not dependent on extracellular divalent cations, since cells bathed in extracellular buffer with no added calcium or magnesium also exhibited prompt increases in PI fluorescence (data not shown). The increase in PI fluorescence was abolished in cells exposed to toxin that had been held at 60 °C for 20 min (data not shown) or preincubated with the neutralizing mAb against NheB, 1E11 (Fig. 3a).
|
), cells that have been metabolically poisoned and in macrophages undergoing pyroptosis due to salmonellae (Fink & Cookson, 2006). In all three situations, PI uptake has been shown to be inhibited by 5 mM glycine but not by L-leucine. To determine whether the plasma membrane permeability to PI induced by Nhe was related to the activation of these undefined endogenous pores we exposed Caco-2 epithelia to NVH 0075/95 culture supernatants (20 µl ml–1) in the presence of 5 mM glycine and leucine. PI uptake was unaffected by either amino acid (data not shown).
Purinergic activation of certain cell types is also able to cause PI uptake (Pelegrin & Surprenant, 2006). However, PI uptake in Caco-2 cells was not detected following exposure of the cells to 0.3 M ATP (data not shown). Thus, PI fluorescence in Caco-2 cells exposed to Nhe in B. cereus culture supernatants is not due to the undefined endogenous death pore activated by pyroptosis or purinergic activation by ATP.
Purified Nhe forms pores in planar lipid bilayers
The above data are consistent with increases in plasma membrane permeability induced by Nhe rather than activation of the endogenous channels linked with the inflammasome and cell death. To support this distinction, we added purified Nhe components to planar synthetic lipid bilayer membranes, which by definition will lack endogenous channels or receptors. Fig. 4(a) (upper trace) shows the prompt step-like increases in membrane conductance observed following addition of the three purified Nhe components. Typically, multiple levels of current were observed in which openings and closings of the channels could be seen before the current exceeded the headstage or the bilayer was disrupted. The time taken for channels to insert in the bilayer was 1.5±0.2 min (mean±SE, n=14). The specificity of the pore formation being due to Nhe rather than co-purifying contaminants was confirmed by the prevention of the appearance of channels in the bilayer by addition of the neutralizing antibody against NheB (1E11) to the bathing solution prior to the addition of Nhe, at least within a period up to 15–17 min (Fig. 4a, lower trace). In contrast, addition of non-specific IgG (affinity-purified goat anti-rabbit antibody) did not significantly increase the time taken for channels to appear (2.3±0.6 min, mean±SE, n=7, P>0.05, unpaired t test). The range of channel conductances in 0.1 M NaCl is shown in Fig. 4(b). Channels of 200–700 pS dominated (mean 400±58 pS in 0.1 M NaCl), but both lower and higher conductances were observed. The pores were able to conduct both cations and anions; however, replacement of sodium with a larger cation (N-methyl-D-glucamine) reduced the mean conductance to 235±26 pS in 0.1 M sodium gluconate, whereas replacement of chloride ions with the larger gluconate ion had less effect, yielding channels with mean conductance of 400±33 pS (0.1 M sodium gluconate). Thus, cations appear to permeate preferentially with respect to anions in the large conductance channels formed by Nhe in planar lipid bilayers.
|
shows the prompt fall in intracellular ATP, in which over 90 % of ATP was lost after 5 min, in Caco-2 monolayers exposed to 20 µl ml–1 NVH 0075/95 supernatant. In contrast, supernatant from the nheBC mutant had no significant effect on ATP levels in the cell monolayers. An indication of the extent of membrane permeability was obtained by measuring the release of the high-molecular-mass LDH (140 kDa) from Caco-2 monolayers into the bathing solution. Fig. 5(b) shows that 88 % of total cell monolayer LDH was released at 60 min after exposure to NVH 0075/95 supernatant, whereas nheBC supernatant again failed to induce any loss of LDH. Similar release of LDH was observed in Vero cell monolayers exposed to supernatant from NVH 0075/95 but not in those exposed to supernatant from the nheBC mutant. LDH release indicates either complete breakdown of the membrane or the formation of large pores. In the latter case, LDH release should be delayed or prevented by osmotic protectants. To test this we used PEG of varying molecular size as an osmotic protectant of LDH release in Vero cells. Fig. 5(c) shows the delay in LDH release induced by the presence of the largest diameter osmotic protectant, PEG 8000 (approximate diameter 7–8 nm). The time taken to reach 50 % release of LDH increased from 13 min (control) to 28 min in the presence of PEG 8000. LDH release in the presence of additional glucose (0.9 nm diameter), PEG 400 (1.2–1.4 nm) and PEG 1000 (1.8–2.2 nm) was indistinguishable from that of cells without added osmotic protectant.
|
). The changes in cell size were significantly greater than those observed in cells exposed to supernatant from the nheBC strain, which did not take up PI (i.e. remained intact). Addition of mAb 1E11 to the culture supernatant prior to addition to the cells abolished both the increase in cell size and PI uptake, such that the changes in forward scatter overlaid those observed in the cells exposed to the nheBC mutant.
|
, Table 1
). Recently, the crystal structure of Hbl component B at 2 Å (0.2 nm) resolution was deposited in the Protein Data Bank (PDB; www.pdb.org) (PDB ID: 2nrj; Fig. 8a). NheB, NheC and Hbl L1 showed sufficient sequence similarities to Hbl B for generation of 3D homology models using the Hbl B structure as the template. The homology models of NheB and NheC created using SWISS-MODEL (Schwede et al., 2003; swissmodel.expasy.org) are shown in Fig. 8(b, c). The 3D structures are predominantly -helical, containing four-helix bundles that wrap around each other in left-handed supercoils, associated with a shorter fifth carboxy-terminal helix. A subdomain consisting of a β-hairpin flanked by two short -helices is located between the third and fourth helices of the main bundle folded up against the side of the larger domain. Hbl B, NheB and NheC all contain a predicted hydrophobic segment that correlates perfectly with the β-hairpin.
|
|
) revealed Hbl B to be structurally similar to a 34 kDa pore-forming haemolysin from E. coli named cytolysin A (ClyA) (Fig. 8d). (Wallace et al., 2000). Hbl B aligns to ClyA with a VAST structure-similarity score of 8.9 and a DaliLite (Holm & Park, 2000) Z-score of 13.8. The two proteins can be superimposed (Fig. 8e) with a root mean square deviation of 4.0 Å (0.4 nm) for 242 C residues, while the crystal structures of Hbl B and ClyA contain 331 and 301 residues, respectively. The subdomain containing the β-hairpin appears to have a different orientation in ClyA compared to Hbl B, as if tilted out in a more extended conformation. Although an alignment showed only weak similarity between the amino acid sequences of Hbl B and ClyA (Table 1, Fig. 7), they share strong structural similarities and several conserved residues.
Haemolytic activity of Nhe
In the light of the predicted structural and haemolytic activity of Hbl and ClyA we examined the activity of Nhe towards erythrocytes. Incubation of purified Nhe proteins with 1.5 % human erythrocytes at 37 °C for 1 h resulted in dose-dependent haemolysis, as determined by haemoglobin release measured at 540 nm (Fig. 9a). This activity was inhibited when the Nhe components were premixed with the 1E11 antibody before testing (data not shown). The haemolytic activity of Nhe varied between species, as shown in Fig. 9(b).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The use of an nheBC mutant in B. cereus is dependent on the fact that the strain used, NVH 0075/95, lacks genes encoding the other two dominant cytotoxic components, Hbl and CytK. The total loss of cytotoxicity in the nheBC mutant reveals the contribution of Nhe to cytotoxicity in culture supernatants. The nheBC mutant producing just the NheA component was employed as we have been unable to create a mutant with deletion of the entire nhe operon, as was also observed by Ramarao & Lereclus (2006).
The use of unpurified supernatants obviously needs to be judged cautiously, since B. cereus is known to produce a variety of additional exotoxins. However, they did not appear to have any primary ability to cause PI uptake, ATP loss or LDH release, since up to 50 µl ml–1 of supernatant from the nheBC strain had no demonstrable effect. Furthermore, the antibody neutralization effects strongly indicate that the formation of Nhe-specific pores is the cytotoxic determinant in NVH 0075/95 culture supernatants. Thus, it remains likely that the membrane damage is dependent on the activity of the three Nhe proteins and a cytotoxic contribution from other compounds need not be evoked.
The rapid time-course is consistent with direct pore formation rather than activation of inflammasome-mediated death pores. Cells undergoing caspase-dependent cell death resulting from salmonella infection permit influx of fluorescent dyes of up to 1 kDa, including propidium, a process that can be blocked by glycine. However this phenomenon is detectable in macrophages after hours rather than minutes (Fink & Cookson, 2006). More rapid is the membrane permeability to propidium dyes induced by ATP and the marine toxin maitotoxin (Estacion et al., 2003). However, the former appears to use pannexin-1 hemichannel pores triggered by ATP-gated P2X7 receptors (Pelegrin & Surprenant, 2006). We have excluded any dependence on exogenous ATP in our system as 300 mM ATP failed to induce PI uptake. The dye permeability induced by maitotoxin can be blocked by glycine, but similar inhibition of the Nhe pore was not observed. Thus, whilst we have not measured the caspase dependence of B. cereus culture supernatants, and the relationship between pannexin-1 pores and the inflammasome is far from straightforward (Pelegrin & Surprenant, 2007), we believe that the results can be explained without reference to pannexin-1 channels, pyroptosis and the undefined endogenous death pore blocked by glycine.
Given the complex geometry of endogenous eukaryotic ion channels there is no simple relationship between the electrical conductance in a planar lipid bilayer and the diameter of a transmembrane pore. Nevertheless, to help assess the size of the channels formed by Nhe (200–700 pS in 0.1 M NaCl) we note that the conductances of the other pore-forming toxins of B. cereus, CytK and Haemolysin II, are both 600 pS, although obtained in 10-fold higher concentrations of ions, i.e. 1 M NaCl and KCl (Hardy et al., 2001a; Miles et al., 2002). These toxins form pores by generating transmembrane β-barrels (Parker & Feil, 2005), and the archetypal β-barrel pore former of this group, Staphylococcus aureus -haemolysin, has a diameter of 2.8 nm at the pore entrance narrowing to 1.4 nm (Song et al., 1996). Thus, based on conductance values alone we expect the pore formed by Nhe to be substantially larger than this. The release of the high-molecular-mass LDH (140 kDa) is likely to be the result of cell lysis subsequent to pore formation and osmotic swelling. If so, then the delay in LDH release caused by PEG 8000 will be due to blocking of the preceding ionic fluxes through the Nhe pore.
Cytolysin A (ClyA), also known as HlyE and SheA, is a 34 kDa protein that was initially identified in E. coli K-12 as a haemolysin expressed during anaerobic growth (Green & Baldwin, 1997; Ludwig et al., 1999; Oscarsson et al., 1996). It is also present in enteropathogenic E. coli (Ludwig et al., 2004), Shigella flexneri (Wallace et al., 2000) and Salmonella enterica serovars Typhi and Paratyphi A (Oscarsson et al., 2002). ClyA is haemolytic and causes LDH release in nucleated cells (delCastillo et al., 1997; Oscarsson et al., 1996, 1999), which has led to the suggestion that it too mediates osmotic lysis through the formation of transmembrane homo-oligomeric pores (Eifler et al., 2006; Tzokov et al., 2006; Wallace et al., 2000). The narrowest pore diameter for ClyA has been measured in cryoelectron microscope images at 4 nm. This is significantly larger than the diameter for the β-barrel pore-forming toxins (see above) and consistent with the large conductance pores created by Nhe.
The available 3D crystal structure of Hbl component B showed tertiary structure resemblance to ClyA of E. coli (Wallace et al., 2000) in a near-global structural alignment, despite limited sequence similarity. Although four-helix bundles like those found in ClyA and Hbl B are quite common architectural features of proteins, the subdomain containing the hydrophobic β-hairpin represents a unique fold (Wallace et al., 2000). The different orientation of this subdomain in the two crystal structures (Fig. 8e) potentially represents crystallization in different conformations, reflecting alternative means of burying the hydrophobic β-hairpin from the solvent which, in ClyA, is achieved by forming homodimers. 3D crystal structures have not been determined for any of the Nhe proteins, but, judging by sequence homology, they are likely to adopt a fold similar to that of Hbl B, as structural similarities are usually highly conserved during protein evolution.
In addition to the common key structural features there are also clear functional correlates between Hbl, Nhe and ClyA: they are all bacterial pore-forming toxins with cytolytic and haemolytic activity. The pores formed by Nhe and ClyA are both weakly cation-selective in lipid bilayer experiments, and the pore conductances, 200–700 pS in 0.1 M NaCl for Nhe and 1000 pS in 0.1 M KCl for ClyA (Ludwig et al., 1999), are broadly comparable, given that potassium ions have greater ionic mobility than sodium. Comparisons between the pore sizes of ClyA and Nhe remain only very tentative, since Nhe requires three separate proteins for maximum cytotoxicity, whereas ClyA is a homooligomer, and we present no data to identify the extent to which Nhe needs to oligomerize in this process.
Structural comparison methods such as VAST and DaliLite used in the current study can detect evolutionary relationships that are not apparent when judged by sequence similarities alone (Gibrat et al., 1996). However, such methods may not always distinguish between remote homologues linked by divergent evolution, and analogous proteins obtained through parallel evolution, especially when only very limited sequence similarity is present. In such cases, determination of homology and classification into superfamilies is performed on combined sequence, structural and functional information, since divergent evolution from a common ancestor retains not only similar folding but also functional features (Dietmann & Holm, 2001). Thus, the strongest argument for homology between the ClyA and the Hbl/Nhe families of toxins stems from the association of similar unique structures with a common functional property, the formation of transmembrane pores in eukaryotic cell membranes.
In summary, through the use of a deletion mutant and purified proteins we have identified a mechanism of cytotoxicity for Nhe, namely osmotic lysis following pore formation in the plasma membrane. We propose that the structural and functional similarities among Nhe, Hbl and ClyA indicate that they constitute a superfamily of pore-forming toxins.
|
ACKNOWLEDGEMENTS |
|---|
Edited by: J. Green
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Beecher, D. J. & Macmillan, J. D. (1991). Characterization of the components of hemolysin-BL from Bacillus cereus. Infect Immun 59, 1778–1784.
del Castillo, F. J., Leal, S. C., Moreno, F. & del Castillo, I. (1997). The Escherichia coli K-12 sheA gene encodes a 34-kDa secreted haemolysin. Mol Microbiol 25, 107–115.[CrossRef][Medline]
Dietmann, S. & Holm, L. (2001). Identification of homology in protein structure classification. Nat Struct Biol 8, 953–957.[CrossRef][Medline]
Dietrich, R., Fella, C., Strich, S. & Märtlbauer, E. (1999). Production and characterization of monoclonal antibodies against the hemolysin BL enterotoxin complex produced by Bacillus cereus. Appl Environ Microbiol 65, 4470–4474.
Dietrich, R., Moravek, M., Burk, C., Granum, P. E. & Märtlbauer, E. (2005). Production and characterization of antibodies against each of the three subunits of the Bacillus cereus nonhemolytic enterotoxin complex. Appl Environ Microbiol 71, 8214–8220.
Ehling-Schulz, M., Svensson, B., Guinebretiere, M. H., Lindbäck, T., Andersson, M., Schulz, A., Fricker, M., Christiansson, A., Granum, P. E. & other authors (2005). Emetic toxin formation of Bacillus cereus is restricted to a single evolutionary lineage of closely related strains. Microbiology 151, 183–197.
Eifler, N., Vetsch, M., Gregorini, M., Ringler, P., Chami, M., Philippsen, A., Fritz, A., Müller, S. A., Glockshuber, R. & other authors (2006). Cytotoxin ClyA from Escherichia coli assembles to a 13-meric pore independent of its redox-state. EMBO J 25, 2652–2661.[CrossRef][Medline]
Estacion, M., Weinberg, J. S., Sinkins, W. G. & Schilling, W. P. (2003). Blockade of maitotoxin-induced endothelial cell lysis by glycine and L-alanine. Am J Physiol Cell Physiol 284, C1006–C1020.
Fink, S. L. & Cookson, B. T. (2006). Caspase-1-dependent pore formation during pyroptosis leads to osmotic lysis of infected host macrophages. Cell Microbiol 8, 1812–1825.[CrossRef][Medline]
Gibbons, S. J., Washburn, K. B. & Talamo, B. R. (2001). P2X7 receptors in rat parotid acinar cells: formation of large pores. J Auton Pharmacol 21, 181–190.[CrossRef][Medline]
Gibrat, J. F., Madej, T. & Bryant, S. H. (1996). Surprising similarities in structure comparison. Curr Opin Struct Biol 6, 377–385.[CrossRef][Medline]
Granum, P. E., O'Sullivan, K. & Lund, T. (1999). The sequence of the non-haemolytic enterotoxin operon from Bacillus cereus. FEMS Microbiol Lett 177, 225–229.[Medline]
Green, J. & Baldwin, M. L. (1997). The molecular basis for the differential regulation of the hlyE-encoded haemolysin of Escherichia coli by FNR and HlyX lies in the improved Activating Region 1 contact of HlyX. Microbiology 143, 3785–3793.
Guérout-Fleury, A. M., Shazand, K., Frandsen, N. & Stragier, P. (1995). Antibiotic-resistance cassettes for Bacillus subtilis. Gene 167, 335–336.[CrossRef][Medline]
Hardy, S. P., Lund, T. & Granum, P. E. (2001a). CytK toxin of Bacillus cereus forms pores in planar lipid bilayers and is cytotoxic to intestinal epithelia. FEMS Microbiol Lett 197, 47–51.[CrossRef][Medline]
Hardy, S. P., Ritchie, C., Allen, M. C., Ashley, R. H. & Granum, P. E. (2001b). Clostridium perfringens type A enterotoxin forms mepacrine-sensitive pores in pure phospholipid bilayers in the absence of putative receptor proteins. Biochim Biophys Acta 1515, 38–43.[Medline]
Hauge, S. (1955). Food poisoning caused by aerobic spore-forming bacilli. J Appl Bacteriol 18, 591–595.
Holm, L. & Park, J. (2000). DaliLite workbench for protein structure comparison. Bioinformatics 16, 566–567.
Koradi, R., Billeter, M. & Wuthrich, K. (1996). MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph 14, 51–55.[CrossRef][Medline]
Lindbäck, T. & Granum, P. E. (2006). Detection and purification of Bacillus cereus enterotoxins. In Methods in Biotechnology, Volume 21: Food-borne pathogens: Methods and protocols, pp. 15–26. Edited by C. C. Adley. Totowa, NJ: Humana Press.
Lindbäck, T., Fagerlund, A., Rødland, M. S. & Granum, P. E. (2004). Characterization of the Bacillus cereus Nhe enterotoxin. Microbiology 150, 3959–3967.
Ludwig, A., Bauer, S., Benz, R., Bergmann, B. & Goebel, W. (1999). Analysis of the SlyA-controlled expression, subcellular localization and pore-forming activity of a 34 kDa haemolysin (ClyA) from Escherichia coli K-12. Mol Microbiol 31, 557–567.[CrossRef][Medline]
Ludwig, A., von Rhein, C., Bauer, S., Huttinger, C. & Goebel, W. (2004). Molecular analysis of cytolysin A (ClyA) in pathogenic Escherichia coli strains. J Bacteriol 186, 5311–5320.
Lund, T. & Granum, P. E. (1996). Characterisation of a non-haemolytic enterotoxin complex from Bacillus cereus isolated after a foodborne outbreak. FEMS Microbiol Lett 141, 151–156.[CrossRef][Medline]
Lund, T. & Granum, P. E. (1997). Comparison of biological effect of the two different enterotoxin complexes isolated from three different strains of Bacillus cereus. Microbiology 143, 3329–3336.
Lund, T., De Buyser, M. L. & Granum, P. E. (2000). A new cytotoxin from Bacillus cereus that may cause necrotic enteritis. Mol Microbiol 38, 254–261.[CrossRef][Medline]
Masson, L., Prefontaine, G. & Brousseau, R. (1989). Transformation of Bacillus thuringiensis vegetative cells by electroporation. FEMS Microbiol Lett 60, 273–278.[CrossRef]
Miles, G., Bayley, H. & Cheley, S. (2002). Properties of Bacillus cereus hemolysin II: a heptameric transmembrane pore. Protein Sci 11, 1813–1824.[CrossRef][Medline]
Moravek, M., Dietrich, R., Buerk, C., Broussolle, V., Guinebretiere, M. H., Granum, P. E., Nguyen-The, C. & Märtlbauer, E. (2006). Determination of the toxic potential of Bacillus cereus isolates by quantitative enterotoxin analyses. FEMS Microbiol Lett 257, 293–298.[CrossRef][Medline]
Oscarsson, J., Mizunoe, Y., Uhlin, B. E. & Haydon, D. J. (1996). Induction of haemolytic activity in Escherichia coli by the slyA gene product. Mol Microbiol 20, 191–199.[Medline]
Oscarsson, J., Mizunoe, Y., Li, L., Lai, X. H., Wieslander, A. & Uhlin, B. E. (1999). Molecular analysis of the cytolytic protein ClyA (SheA) from Escherichia coli. Mol Microbiol 32, 1226–1238.[CrossRef][Medline]
Oscarsson, J., Westermark, M., Löfdahl, S., Olsen, B., Palmgren, H., Mizunoe, Y., Wai, S. N. & Uhlin, B. E. (2002). Characterization of a pore-forming cytotoxin expressed by Salmonella enterica serovars Typhi and Paratyphi A. Infect Immun 70, 5759–5769.
Ouhib, O., Clavel, T. & Schmitt, P. (2006). The production of Bacillus cereus enterotoxins is influenced by carbohydrate and growth rate. Curr Microbiol 53, 222–226.[CrossRef][Medline]
Parker, M. W. & Feil, S. C. (2005). Pore-forming protein toxins: from structure to function. Prog Biophys Mol Biol 88, 91–142.[CrossRef][Medline]
Pelegrin, P. & Surprenant, A. (2006). Pannexin-1 mediates large pore formation and interleukin-1β release by the ATP-gated P2X7 receptor. EMBO J 25, 5071–5082.[CrossRef][Medline]
Pelegrin, P. & Surprenant, A. (2007). Pannexin-1 couples to maitotoxin- and nigericin-induced interleukin-1β release through a dye uptake-independent pathway. J Biol Chem 282, 2386–2394.
Planchot, V., Roger, P. & Colonna, P. (2000). Suitability of starch granule porosity for biosynthesis and amylolysis susceptibility. Starch 52, 333–339.[CrossRef]
Ramarao, N. & Lereclus, D. (2006). Adhesion and cytotoxicity of Bacillus cereus and Bacillus thuringiensis to epithelial cells are FlhA and PlcR dependent, respectively. Microbes Infect 8, 1483–1491.[CrossRef][Medline]
Ryan, P. A., Macmillan, J. D. & Zilinskas, B. A. (1997). Molecular cloning and characterization of the genes encoding the L1 and L2 components of hemolysin BL from Bacillus cereus. J Bacteriol 179, 2551–2556.
Scherrer, R. & Gerhardt, P. (1971). Molecular sieving by Bacillus megaterium cell wall and protoplast. J Bacteriol 107, 718–735.
Schilling, W. P., Sinkins, W. G. & Estacion, M. (1999). Maitotoxin activates a nonselective cation channel and a P2Z/P2X7-like cytolytic pore in human skin fibroblasts. Am J Physiol Cell Physiol 277, C755–C765.
Schoeni, J. L. & Wong, A. C. L. (2005). Bacillus cereus food poisoning and its toxins. J Food Prot 68, 636–648.[Medline]
Schwede, T., Kopp, J., Guex, N. & Peitsch, M. C. (2003). SWISS-MODEL: an automated protein homology-modeling server. Nucleic Acids Res 31, 3381–3385.
Song, L., Hobaugh, M. R., Shustak, C., Cheley, S., Bayley, H. & Gouaux, J. E. (1996). Structure of staphylococcal -hemolysin, a heptameric transmembrane pore. Science 274, 1859–1866.
Tzokov, S. B., Wyborn, N. R., Stillman, T. J., Jamieson, S., Czudnochowski, N., Artymiuk, P. J., Green, J. & Bullough, P. A. (2006). Structure of the hemolysin E (HlyE, ClyA, and SheA) channel in its membrane-bound form. J Biol Chem 281, 23042–23049.
Wallace, A. J., Stillman, T. J., Atkins, A., Jamieson, S. J., Bullough, P. A., Green, J. & Artymiuk, P. J. (2000). E. coli hemolysin E (HlyE, ClyA, SheA): X-ray crystal structure of the toxin and observation of membrane pores by electron microscopy. Cell 100, 265–276.[CrossRef][Medline]
Whitaker, J. R. & Granum, P. E. (1980). An absolute method for protein determination based on difference in absorbance at 235 and 280 nm. Anal Biochem 109, 156–159.[CrossRef][Medline]
Received 27 October 2007;
revised 13 December 2007;
accepted 13 December 2007.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
| J MED MICROBIOL | ALL SGM JOURNALS | |
). The fluorescence increase (arbitrary units; a.u.) is abolished in cells bathed in buffer solution containing 10 µg mAb 1E11 ml–1 against NheB (75/95+mAb; 
). PI fluorescence did not increase when 20 µl ml–1 culture supernatant from the nheBC mutant (
) was added. Results are representative of six to eight wells for both 75/95 and 
) and 
). Values represent mean±SE, n=3, for each data point. At time zero the total cellular ATP concentration was approximately 10 µM as judged against a standard solution of ATP. The final concentration of cellular ATP at 30 min for the cells treated with NVH 0075/95 supernatant was approximately 0.2 µM. (b) Extracellular LDH was measured over time following exposure to NVH 0075/95 (
). Values are mean±SE, n=3. (c) Effect of PEG compounds on LDH release in Vero cell monolayers. LDH release induced by NVH 0075/95 supernatant was monitored in the presence of glucose, PEG 400, PEG 1000 and PEG 8000 (
, P<0.05). Cells exposed to EC buffer alone are shown as (
