Effect of Epsilon Toxin–GFP on MDCK Cells and Renal Tubules In Vivo

Abstract

Epsilon toxin (∊-toxin), produced by Clostridium perfringens types B and D, causes fatal enterotoxemia, also known as pulpy kidney disease, in livestock. Recombinant ∊-toxin–green fluorescence protein (∊-toxin–GFP) and ∊-prototoxin–GFP were successfully expressed in Escherichia coli. MTT assays on MDCK cells confirmed that recombinant ∊-toxin–GFP retained the cytotoxicity of the native toxin. Direct fluorescence analysis of MDCK cells revealed a homogeneous peripheral pattern that was temperature sensitive and susceptible to detergent. ∊-Toxin–GFP and ∊-prototoxin-GFP bound to endothelia in various organs of injected mice, especially the brain. However, fluorescence mainly accumulated in kidneys. Mice injected with ∊-toxin–GFP showed severe kidney alterations, including hemorrhagic medullae and selective degeneration of distal tubules. Moreover, experiments on kidney cryoslices demonstrated specific binding to distal tubule cells of a range of species. We demonstrate with new recombinant fluorescence tools that ∊-toxin binds in vivo to endothelial cells and renal tubules, where it has a strong cytotoxic effect. Our binding experiments indicate that an ∊-toxin receptor is expressed on renal distal tubules of mammalian species, including human. (J Histochem Cytochem 52:931–942, 2004)

Keywords

clostridial toxins epsilon toxin MDCK cells renal tubules pulpy kidney disease

Epsilon toxin (∊-toxin) is the most potent clostridial toxin after botulinum and tetanus neurotoxins. It is produced by Clostridium perfringens types B and D and causes fatal enterotoxemia in sheep, goats, and occasionally calves and other animals, resulting in heavy economic losses (Payne and Oyston 1997).

C. perfringens can be found in the intestines of most animals. Bacterial numbers remain small because of peristalsis, and clinical disease does not occur unless the microbial balance in the gut is disrupted. Therefore, small amounts of ∊-toxin in the gut of healthy animals are considered innocuous. However, when the intestine is altered by sudden changes in diet or other factors, bacteria proliferate rapidly and produce large amounts of this toxin. Intestinal mucosal permeability is increased by ∊-toxin, thereby facilitating its absorption into the circulation (reviewed by Finnie 2004). Disease is principally manifested as severe and often fatal neurological disturbance; edema of lungs, heart, kidneys, and intestine is also a clinical sign related to microvascular damage (Songer 1996). This pathology is also known as pulpy kidney and overeating disease.

∊–Toxin is encoded by the etx gene (Hunter et al. 1992) and is synthesized and secreted as an inactive prototoxin of 311 amino acids (32.7 kD) that is converted to the fully active toxin by proteolytic removal of a basic N-terminal peptide of 14 residues and of 23 C-terminal residues (Bhown and Habeerb 1977; Minami et al. 1997). This proteolytic activation usually occurs in the gut of infected animals by the actions of trypsin and chymotrypsin, but it can also be achieved in vitro by controlled enzyme digestion.

The Madin–Darby canine kidney (MDCK) cell line, of epithelial origin from the distal convoluted tubule, is susceptible to ∊-toxin (Payne et al. 1994). The cytotoxic effects are very rapid and are enhanced by EDTA (Lindsay 1996). Cells undergo both cell cycle alterations (Borrmann et al. 2001) and morphological changes, including swelling and large bleb formation, in a process that is independent of the actin cytoskeleton and endocytosis (Petit et al. 1997; Borrmann et al. 2001). The intoxication process correlates with the formation of a membrane complex of ∼155 kD and the efflux of intracellular K⁺ and the influx of Na⁺, Cl⁻, and Ca²⁺ without entry of the toxin into the cytosol (Petit et al. 1997, 2001; Borrmann et al. 2001). The toxin decreases the trans-epithelial resistance of polarized MDCK cells without affecting intercellular junctions (Petit et al. 2003). The membrane complex is also formed in synaptosomes and corresponds to toxin heptamerization (Miyata et al. 2001). In artificial lipid membranes, ∊-toxin induces large pore formation, which appears to constitute its main cytotoxic mechanism (Petit et al. 2001). Furthermore, heptameric pore formation has recently been described within the detergent-insoluble microdomains of MDCK cells and rat synaptosomes (Miyata et al. 2002).

Animal models have also been used to study ∊-toxin effects. Vasogenic cerebral edema has been described in mice (Gardner 1973; Morgan et al. 1975; Finnie 1984), rats (Finnie et al. 1999; Ghabriel et al. 2000; Zhu et al. 2001), and calves (Uzal et al. 2002) injected with the toxin, and pulmonary and cardiovascular lesions have also been described elsewhere (Sakurai et al. 1983; Uzal et al. 2002). Specific receptor sites have been identified in the cerebral microvasculature (Buxton 1978; Nagahama and Sakurai 1991) as well as disruption of the blood-brain barrier, which causes a rapid and substantial increase in vascular permeability and leads to severe, diffuse vasogenic edema (reviewed by Finnie 2004). Moreover, morphological kidney alterations were described on infected domestic animals (pulpy kidney) as well as in IV-injected mice (Buxton 1978; Tamai et al. 2003).

Here we used recombinant techniques to develop green fluorescence protein-fused ∊-toxin and ∊-prototoxin that could be directly visualized by fluorescence microscopy. We demonstrate that both proteins bind to the MDCK cell membrane and that only the toxin is cytotoxic, thus retaining the binding properties and cytotoxicity of the non-fluorescent proteins. We analyzed the organ distribution, tissue localization, and histopathology of these proteins in IV-injected mice. Because the kidneys are, in addition to brain, critical target organs for ∊-toxin intoxication (pulpy kidney disease), we further analyzed these organs by ligand-receptor binding assays on cryostat slices.

Materials and Methods

Expression cDNA Constructs

∊-Toxin and ∊-prototoxin cDNAs were amplified from strain NCTC2062 of C. perfringens type D by polymerase chain reaction (PCR). Primers for ∊-prototoxin delimited the coding sequence from K46 to the end without the STOP codon and added the restriction sites HindIII and PstI at the 5′ and 3′ ends, respectively. Primers for ∊-toxin delimited the coding sequence from K46 to Y299 and added the same restriction sites as for the ∊-prototoxin. These two enzymes were used to clone both forms into pEGFP expression plasmid (Clontech; Palo Alto, CA), obtaining two fusion proteins with GFP: ∊-toxin–GFP and ∊-prototoxin'–GFP.

To improve the efficiency of protein purification, these constructs were used to reamplify and clone ∊-toxin–GFP and ∊-prototoxin–GFP into the pGEX-4T-1 expression plasmid (Amersham Biosciences; Freiburg, Germany) by another PCR with new primers. ∊-Toxin–GFP and ∊-prototoxin–GFP were cloned using EcoRI and NotI restriction enzymes. Thus, two new fusion proteins between glutathione-S-transferase (GST) and amplified products were obtained: GST–∊-toxin–GFP and GST–∊-prototoxin-GFP. These plasmids were transformed in Rosetta E. coli host strain (Novagen; Madison, WI).

Protein Expression and Purification

The expression of recombinant proteins was induced overnight at room temperature (RT) in 250 ml LB medium cultures with 0.4 mM isopropyl β-d-thiogalactopyranoside (IPTG).

Cells were pelleted and resuspended in ice cold buffer containing PBS, 1% Triton X-100, 0.1 mg/ml phenylmethyl-sulfonylfluoride (PMSF), 10 μg/ml aprotinin, and 10 μg/ml leupeptin, and were sonicated and centrifuged at 15,000 × g for 20 min. The resultant supernatant was incubated with 0.5 ml of previously PBS-equilibrated glutathione Sepharose 4B (Amersham Pharmacia Biotech) beads for 1 hr at 4C. Finally, recombinant proteins were eluted by thrombin elution in 10 mM Tris-HCl, pH 8.0, with 150 mM NaCl and 2.5 mM CaCl₂, according to the manufacturer's instructions. We routinely performed a further purification step and isolated ∊-prototoxin–GFP by anion exchange chromatography.

For MTT assay, ∊-prototoxin–GFP was activated by trypsin proteolysis with trypsin beads (Sigma–Aldrich; Madrid, Spain) according to the manufacturer's instructions.

Native ∊-toxin was purified by ion exchange chromatography as previously described (Petit et al. 1997).

Protein Analyses

Samples of ∊-toxin–GFP and ∊-prototoxin–GFP were loaded in a 12% acrylamide-bisacrylamide gel. After SDS-PAGE, the proteins were transferred onto nitrocellulose membranes (Bio-Rad; Hercules, CA) with a semi-dry unit (Sigma–Aldrich). After blocking using 5% non-fat dry milk in Tris-buffered saline (TBS) containing 0.1% Tween-20, membranes were incubated with 1:1000 concentration of anti-∊-toxin rabbit polyclonal antibody, prepared as previously described (Popoff 1987), for 1 hr at RT. After three washes in the same buffer, the membrane was incubated with anti-rabbit immunoglobulins conjugated to horseradish peroxidase (DAKO; Glostrup, Denmark). Finally, after five washes the blot was developed using the enhanced chemiluminescence method (ECL).

Cell Cultures

MDCK (Madin-Darby canine kidney) cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing glucose (4.5 mg/ml), sodium pyruvate (110 mg/liter), 2 mM glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin and supplemented with 10% fetal bovine serum (FBS), and were maintained at 37C in a humidified atmosphere containing 5% CO₂.

MTT Cell Death Assay

The assay is based on the capacity of living cells to reduce yellow tetrazolium MTT (3-(4, 5-dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide) and thus generate reducing equivalents such as NADH and NADPH. The resulting intracellular purple formazan can be solubilized and quantified by spectrophotometry and is proportional to the number of viable cells.

MDCK cells were grown to 80% confluence. Treated cells were incubated for 30 min at 37C with 5 nM, 50 nM, 100 nM, or 500 nM native ∊-toxin, ∊-toxin–GFP, trypsin-activated ∊-prototoxin–GFP, or ∊-prototoxin–GFP in PBS. Control cells were incubated with PBS. Thereafter, 0.5 mg/ml MTT (Sigma–Aldrich) was added and maintained for 30 min at 37C. The reaction was stopped by adding an equal volume of isopropanol:1 M HCl (24:1) and the cells were then lysed. Formazan was measured at 550 nm in a 96-microtiter plate. The experiment was done four times for each treatment, and means and standard errors were obtained and compared statistically by Student's í-test.

Binding Assays to Living MDCK Cells

Cells were grown to confluence on 24-well plates. Growing medium was replaced by a buffer containing 5 mM KCl, 140 mM NaCl, 5 mM NaHCO₃, 1 mM MgCl₂, 1.2 mM Na₂PO₄, 10 mM glucose, and 20 mM HEPES/NaOH, pH 7.4. Cells were maintained in this buffer for 1 hr, then incubated with 200 nM ∊-prototoxin–GFP or ∊-toxin–GFP in the same buffer for 1 hr at 4C and then brought to 37C. After the appropriate periods of time (0, 10, 30, and 60 min), cells were washed three times in PBS, fixed with 4% paraformaldehyde for 20 min, washed twice in PBS, and examined under a LEICA DMIRB/E inverted fluorescence microscope.

Binding Assays to Fixed MDCK Cells

Cells were grown on coverslips to confluence, washed three times in PBS, and fixed with 4% paraformaldehyde. Nonspecific binding was blocked by incubating the cells in PBS containing 0.2% gelatin and 10% FBS for 1 hr. Incubations with 200 nM ∊-toxin–GFP or ∊-prototoxin–GFP were performed in the same buffer for 1 hr at 4C or RT. After three washes in PBS, cells were fixed again with 4% paraformaldehyde for 10 min, then washed twice in PBS and mounted on slides using Immuno Floure Mounting Medium (ICN Biomedicals; Costa Mesa, CA), followed by examination under a Leica TCS 4D confocal microscope (Serveis Científico-Tècnics; Universitat de Barcelona, Barcelona, Spain).

To demonstrate the specificity of binding, competition experiments were carried out by co-incubating the cells with native ∊-toxin and ∊-prototoxin–GFP or ∊-toxin–GFP in a 20:1 molar ratio. For detergent experiments, blocking was done at 4C in the presence of 0.2% Triton X-100. The following steps were performed at RT in the absence of detergent.

Electron Microscopy

MDCK cells were grown on Thermanox to 100% confluence, washed three times in PBS, and fixed with 2% paraformaldehyde. Nonspecific binding was blocked by incubating the cells in a buffer containing PBS, 0.2% gelatin, and 10% FBS for 45 min. Incubations with 200 nM ∊-prototoxin-GFP were performed in the same buffer for 1 hr at RT. After three washes, another blocking step was done with PBS containing 1% BSA for 20 min. Next, cells were incubated with a 1:100 dilution of anti-∊-toxin polyclonal antibody in the same buffer for 1 hr. After three washes, the secondary antibody incubation was performed with 1:25 anti-rabbit immunoglobulins conjugated with 10-nm gold beads in the same buffer for 45 min. After three washes, cells were fixed in 2% paraformaldehyde, 2.5% glutaraldehyde in 100 mM PBS for 2 hr at RT. Primary fixation was followed by postfixation with 1.0% OsO₄ in 100 mM PBS for 1 hr. The samples were dehydrated in acetone and embedded in Spurr resin. After sectioning with a Reichert-Jung Ultracut E ultramicrotome, ultrathin sections were stained with uranyl acetate and lead citrate and examined under a Hitachi H-600 AB transmission electron microscope (Serveis Científico-Tècnics; Universitat de Barcelona, Barcelona, Spain).

In Vivo Studies in a Murine Model

Male OF1 Swiss mice weighing 20 g were anesthetized with sodium pentothal (35 mg/kg). ∊-Toxin–GFP and ∊-prototoxin–GFP were prepared in PBS with 1% BSA and injected IV. Tested doses of ∊-toxin–GFP were 7 ng/g, 14 ng/g, 70 ng/g, 750 ng/g, and 2500 ng/g per mouse. Animals injected with ∊-prototoxin–GFP and GFP alone (2.5 μg/g each) were sacrificed after 15 min. Organs (kidneys, brain, heart, lung, liver, spleen, testes, gut) were extracted and fixed by immersion in 4% paraformaldehyde for 12 hr. One half of each organ was embedded in paraffin, cut, and stained with hematoxylin and eosin (HE). The other half was immersed in 30% sucrose, frozen in isobutanol, cut in a cryostat to 6-μm slices, immediately mounted on slides using Immuno Floure Mounting Medium, and observed under a Leica TCS 4D confocal microscope (Serveis Científico-Tècnics).

Control experiments were also performed by injecting 1.65 μg/g native ∊-prototoxin simultaneously with 2.5 μg/g ∊-prototoxin–GFP (10:1 molar ratio). Organs were processed as described above.

Experiments were performed in accordance with the European Communities Council Directive of November 1986 (86/609/EEC) and were approved by the Institutional Review Board at the University of Barcelona.

∊-Toxin–GFP and ∊-Prototoxin–GFP Incubations on Kidney Slices

Kidneys from OF1 Swiss mice were processed to obtain cryostat slices as described above. Nonspecific binding was blocked by incubating the slices in PBS containing 0.2% gelatin and 10% FBS for 1 hr. The incubation with ∊-toxin–GFP or ∊-prototoxin—GFP was performed in the same buffer at a 200 nM concentration for 1 hr at RT. After three PBS washes, the samples were fixed in 4% paraformaldehyde for 10 min, mounted with Immuno Floure Mounting Medium, and examined under a Nikon Eclipse E800 fluorescence microscope.

In some experiments, blocking steps were performed in the presence of 0.2% Triton X-100.

For competition experiments, slices were co-incubated with native ∊-toxin and ∊-toxin–GFP or ∊-prototoxin–GFP in a 20:1 molar ratio.

The same incubation experiments were done with kidneys from rat, sheep, cow, and human. Sheep and cow samples were obtained from Mercabarna slaughterhouse in Barcelona. Slices from human healthy kidney were provided by the Servei d'Anatomia Patològica from Hospital Princeps d'Espanya, (L'Hospitalet de Llobregat, Spain). Samples were fixed by immersion in 4% paraformaldehyde for 12 hr and processed as described for mice to obtain cryostat slices and perform overlay experiments.

Results

Recombinant ∊-Toxin–GFP and ∊-Prototoxin–GFP Expression and Characterization

Recombinant ∊-toxin–GFP and ∊-prototoxin–GFP were cloned and successfully expressed in E. coli. After purification and SDS-PAGE, fluorescent bands were observed in a UV transilluminator. These bands had the expected size for recombinant proteins, as revealed by Coomassie Blue staining. Moreover, Western immunoblotting confirmed their ∊-toxin content (Figure 1A).

To functionally characterize ∊-toxin–GFP and ∊-prototoxin–GFP, we performed MTT death assays on MDCK cells (Figure 1B). The cells were exposed to a range of concentrations of native and recombinant ∊-toxin forms (0, 5, 50, 100, or 500 nM). ∊-Toxin–GFP showed similar cytotoxicity to native ∊-toxin, whereas no significant cytotoxic effect was found for ∊-prototoxin–GFP. MTT assay also confirmed that trypsin-activated ∊-prototoxin–GFP showed comparable cytotoxicity to ∊-toxin–GFP and native ∊-toxin.

Binding of ∊-Toxin–GFP and ∊-Prototoxin–GFP to Living MDCK Cells

To characterize the cell binding of GFP fusion proteins on living cells, we performed time-course experiments on MDCK cells. After incubation for 1 hr at 4C with ∊-toxin–GFP or ∊-prototoxin–GFP and without previous washing, cells were brought to 37C and maintained at that temperature for various time periods (0 to 60 min) before washing and fixation with paraformaldehyde (Figure 2). Figure 2A shows results for ∊-toxin–GFP and Figure 2B for ∊-prototoxin–GFP.

Figure 1

Characterization and cytotoxic effect of ∊-toxin–GFP and ∊-prototoxin–GFP on MDCK cells. (A) Both ∊-toxin–GFP and ∊-prototoxin–GFP were expressed as GST fusion protein, eluted with thrombin treatment, and analyzed by SDS-PAGE followed by Coomassie Blue staining, direct visualization on a UV transilluminator, or Western immunoblotting analysis. (B) Cytotoxic effect of ∊-toxin–GFP, ∊-prototoxin–GFP, and trypsin-activated ∊-prototoxin–GFP was compared with that of the native ∊-toxin purified from C. perfringens cultures. Cells were treated with 0, 5, 50, 100, or 500 nM concentrations of tested proteins. After MTT death assay, cytotoxicity was calculated as a percentage of cell survival referred to time 0 cells. Both ∊-toxin–GFP (□) and trypsin-activated ∊-prototoxin–GFP (▪) showed a similar cytotoxic effect to native ∊-toxin (▵), while ∊-prototoxin–GFP (△) showed no cytotoxic effect.

Incubation at 4C prevented both ∊-toxin–GFP and ∊-prototoxin–GFP cell binding (Figures 2A1 and 2B1) even with longer incubation periods (data not shown). However, 10-min incubation at 37C caused cell binding, indicating a temperature-dependent process (Figures 2A3 and 2B3). Longer incubation times with ∊-prototoxin–GFP increased fluorescence, which was mainly localized to cell membrane and peaked at 60 min (Figures 2B5 and 2B7). These cells were morphologically unaltered throughout the experiment, as shown in phase-contrast images (Figures 2B2, 2B4, 2B6, and 2B8), supporting the lack of cytotoxicity of ∊-prototoxin–GFP shown in Figure 1B.

On the other hand, cells incubated with ∊-toxin–GFP underwent the morphological alterations described for MDCK cells treated with ∊-toxin (Hambrook et al. 1995; Petit et al. 1997; Borrmann et al. 2001): nuclear condensation, vacuolation, and progressive swelling, leading to cell death (Figures 2A2, 2A4, 2A6, and 2A8). Dead cells finally detach from the cell culture substrate (Figure 2A8). Fluorescence images revealed time-dependent binding as for ∊-prototoxin–GFP (Figures 2A1, 2A3, 2A5, and 2A7).

Figure 2

Binding of ∊-toxin–GFP and ∊-prototoxin–GFP to living MDCK cells. Fluorescence and phase-contrast images of MDCK cells incubated with 200 nM of either ∊-toxin-GFP (A) or ∊-prototoxin–GFP (B). Cells were incubated for 1 hr at 4C, warmed to 37C, and maintained at this temperature for 0 (1, 2), 10 (3, 4), 30 (5, 6), or 60 min (7, 8). Note morphological alterations of ∊-toxin–GFP-treated cells (A8) and the lack of toxicity of ∊-prototoxin–GFP (B8). Insets are details of A8 and B8 images.

Moreover, no cytotoxic effect was detected when cells treated with ∊-toxin–GFP were washed after the 4C incubation period and then brought to 37C, confirming the absence of binding at 4C (data not shown).

Binding of ∊-Toxin–GFP and ∊-Prototoxin–GFP to Paraformaldehyde-fixed MDCK Cells

To further characterize the binding ability of GFP fusion proteins, we performed binding assays with ∊-toxin–GFP or ∊-prototoxin–GFP on paraformaldehyde-fixed MDCK cells. Both proteins bound to the MDCK cell membrane, as revealed by direct fluorescence detection on confocal microscopy (Figure 3A for ∊-prototoxin–GFP and Figure 3B for ∊-toxin–GFP). ∊-Prototoxin immunodetection by electron microscopy showed the binding to the external face of plasma membrane, in particular, to the apical side of polarized cells (a detail of this area is shown in Figure 3I). Specific binding was demonstrated by competition experiments in which fluorescence labeling was abolished when non-labeled ∊-toxin was co-incubated with GFP fusion proteins (20:1 molar ratio) (Figures 3C and 3D). Moreover, no fluorescence was detected when incubations were performed with GFP alone (data not shown).

Incubation of ∊-toxin–GFP (Figure 3E) and ∊-prototoxin–GFP (Figure 3F) at 4C confirmed the temperature dependence shown by the binding experiments on living cells.

The presence of 0.2% Triton X-100 in the blocking step diminished fluorescence and changed the uniform membrane labeling to a discontinuous spotted pattern (Figures 3G and 3H).

In Vivo Studies in a Murine Model

To determine in vivo cytotoxic effects and organ distribution of ∊-toxin–GFP and ∊-prototoxin–GFP, we injected mice IV with these recombinant proteins.

Figure 3

Binding of ∊-toxin–GFP and ∊-prototoxin–GFP on paraformaldehyde-fixed MDCK cells. Fluorescence images of fixed MDCK cells incubated with 200 nM of either ∊-prototoxin–GFP (left panels) or ∊-toxin–GFP (right panels). Specific binding was confirmed by adding 4 μM of non-labeled ∊-prototoxin (C) or ∊-toxin (D). The same incubations were done at 4C (E, F) or by preincubating with 0.2% Triton X-100 (G, H). (I) Electron micrograph of ∊-toxin immunodetection (10-nm gold beads) on the apical surface of an MDCK cell incubated with ∊-prototoxin–GFP.

Mice injected with ∊-toxin–GFP died after various periods of time in a dose-dependent manner (from 4 min for 2500 ng/g to 134 min for 7 ng/g). Animals injected with either ∊-prototoxin–GFP or GFP alone were sacrificed after 15 min.

After postmortem observation, organs were processed for both histological studies and fluorescence distribution. HE examination of toxin-injected kidneys revealed more congested and hemorrhagic medullae (Figure 4B) than in animals injected with either ∊-prototoxin–GFP or GFP alone (Figures 4E and 4H). Toxin-injected animals also showed severe alterations on renal distal tubules, including pyknotic nuclei and desquamated epithelia (Figure 4A). These alterations were observed in all animals injected with toxin but were more pronounced in those that received the highest doses. Morphological alterations were not found in animals injected with either ∊-prototoxin–GFP or GFP alone (Figures 4D and 4G). We did not find any significant morphological alterations in other organs analyzed, although an edematogenous process, especially in brain, is well documented after ∊-toxin intoxication (reviewed by Finnie 2004). We sporadically detected perivascular edema in cerebral microvasculature (data not shown), although the routine histopathological techniques used here are not the most appropriate to visualize such changes.

∊-Prototoxin–GFP and ∊-toxin–GFP were localized by direct fluorescence microscopy on cryostat slices of different organs. Fluorescence was detected on the luminal surface of the vascular endothelium of many blood vessels, most abundantly in brain (Figure 5C), but also in other organs such as intestine and testis (Figures 5A and 5B). Fluorescence accumulated mainly on kidneys, in both proximal and distal tubules (Figures 4F and 5E) and, to a lesser extent, in glomeruli and renal blood vessels (not shown).

These results indicate that ∊-toxin especially affects distal tubules, although it also reached and bound to proximal tubules.

Moreover, immunohistochemistry with anti-∊-toxin and anti-GFP antibodies revealed complete co-localization of both molecules (data not shown).

Fluorescence also accumulated at the luminal surface of proximal tubules but not at the distal tubules in GFP-injected animals (Figure 4I), nor was fluorescence detected on vascular endothelia in these control animals (data not shown).

In summary, the three recombinant proteins crossed the glomerular barrier and bound to the proximal tubules but only ∊-prototoxin–GFP and ∊-toxin–GFP bound to distal tubules and only ∊-toxin–GFP specifically degenerated them. Finally, both ∊-prototoxin–GFP and ∊-toxin–GFP, but not GFP alone, bound to endothelial cells.

Figure 4

Effect of IV injection of ∊-toxin–GFP and ∊-prototoxin–GFP on mouse kidney. Mice were injected with 7 ng/g ∊-toxin–GFP (A, B), 2.5 μg/g ∊-toxin–GFP (C), 2.5 μg/g ∊-prototoxin–GFP (D–F), or 2.5 μg/g GFP alone (G–I). Kidneys were divided and processed for HE staining (A, B, D, E, G, H) or for cryostat sectioning and examination under a fluorescence confocal microscope (C, F, I). Kidneys from animals injected with ∊-toxin–GFP (A) showed selective degeneration of distal tubules (black arrows) compared with kidneys from animals injected with ∊-prototoxin–GFP (D) or with GFP (G). Insets show magnified distal tubules, showing pyknotic nuclei in ∊-toxin–GFP injected mice. Medullae of animals injected with ∊-toxin–GFP were very hematic (B) compared with those of animals injected with ∊-prototoxin–GFP (E) or with GFP (H). Direct observation of kidneys from ∊-toxin–GFP (C) and ∊-prototoxin–GFP (F)-injected animals revealed accumulation of fluorescence at the luminal surface of proximal tubules (white arrowheads) and a whole membrane binding pattern on distal tubules (white arrow). Fluorescence of GFP-injected animals (I) was exclusively found at the luminal surface of proximal tubules (white arrowheads) but no fluorescence was detected on distal tubules (white arrow).

Figure 5

Localization of ∊-prototoxin–GFP in blood vessels of several organs (A–D) and renal tubules (E, F) after IV injection in mice. Mice were injected with ∊-prototoxin–GFP and organs were processed for cryostat sectioning and examined under a fluorescence microscope. We detected ∊-prototoxin–GFP (arrows) in blood vessels from intestine (A), testis (B), and brain (C), and in proximal (arrowheads) and distal (arrows) tubules of kidneys (F). Co-injection of excess native ∊-prototoxin with ∊-prototoxin–GFP drastically diminished vascular fluorescence in brain (D), abolished the label in distal renal tubules (F), and significantly increased label in proximal tubules (F, arrowheads). (A, B) Tissue autofluorescence is shown in yellowish brown; green fluorescence corresponds to ∊-prototoxin–GFP.

To show the specificity of toxin binding, displacement experiments were performed in which excess amounts of native ∊-prototoxin were injected simultaneously with ∊-prototoxin–GFP. Fluorescence label in cerebral endothelia diminished drastically (Figures 5C and 5D) and was abolished in distal renal tubules (Figures 5E and 5F), confirming specificity of binding in both structures. Vascular endothelia from other tissues were also displaced but to a lesser extent, which probably indicates that this binding has a nonspecific component (not shown). Interestingly, under these experimental conditions we detected increased label in the proximal tubules (Figure 5F), revealing that ∊-prototoxin–GFP was displaced from its specific binding sites and accumulated unspecifically.

Incubation Experiments of ∊-Toxin–GFP and ∊-Prototoxin–GFP on Cryostat Kidney Slices

We developed another approach, already used in ligand-receptor interaction studies, to analyze the specificity of ∊-toxin–GFP and ∊-prototoxin–GFP binding on kidneys, consisting of incubations of cryostat kidney slices of mice with these recombinant proteins. Fluorescence images revealed that both ∊-toxin–GFP and ∊-prototoxin–GFP bind to the luminal surface of distal tubule cells (Figure 6A for prototoxin; not shown for toxin). The specificity of this interaction was demonstrated by co-incubation with native ∊-toxin, which completely abolished fluorescence (Figure 6B).

Incubations in the presence of detergent clearly diminished fluorescence, indicating that the toxin receptor is mainly associated with detergent-soluble domains on the apical surface of distal tubule cells. However, some fluorescence was still detected in association with detergent-resistant areas (arrows in Figure 6C).

We obtained similar results in incubation experiments with slices from rat (Figure 6D), sheep (Figure 6E), cow (Figure 6F), and human kidneys (Figure 6G). Specificity of binding was also demonstrated in these species (Figure 6H for human).

Figure 6

Binding of ∊-prototoxin–GFP to kidney slices of different species. Fluorescence images of cryostat-processed mouse kidney slices incubated with 200 nM ∊-prototoxin–GFP alone (A) or in the presence of 10 μM non-labeled ∊-toxin (B). Cells from the distal tubule showed a specific binding of ∊-prototoxin–GFP in the apical pole (A, inset). The presence of 0.2% Triton X-100 in the blocking steps drastically diminished the binding (C), now confined to very few distal tubules (white arrows). The same binding pattern was obtained for kidneys from rat (D), sheep (E), cow (F), and human (G). (H) Displacement of ∊-prototoxin–GFP binding with an excess of non-labeled ∊-toxin in a human kidney slice. All the images show tissue autofluorescence in yellowish brown while the green fluorescence corresponds to ∊-prototoxin–GFP.

Discussion

In this study we generated recombinant ∊-toxin–GFP and ∊-prototoxin–GFP to improve their detection and to study their cell binding and tissue distribution profiles by direct fluorescence.

Among the cell lines initially examined in search of the lethal effects of ∊-toxin–GFP, only MDCK was sensitive, in agreement with previous reports indicating that these cells are the most appropriate cell model in which to study ∊-toxin toxicity (Lindsay 1996; Petit et al. 1997; Miyata et al. 2002). More recently, Shortt et al. (2000) identified another cell line (G-402), of human kidney origin, susceptible to effects of ∊-toxin exposure. It is relevant because of the effect of the toxin on the kidneys in vivo.

Here we show that ∊-toxin–GFP fusion protein retains its full toxic activity on MDCK cells, and no differences from the native ∊-toxin were detected. MTT assays revealed a maximum of ∼70% cell death, which was not affected by increasing the dose or the incubation period, consistent with a previously reported subpopulation of resistant cells in MDCK cultures (Shortt et al. 2000). Furthermore, both the time course and morphological changes on incubation with GFP recombinant ∊-toxin were similar to those also described here for native ∊-toxin. It has been described that the lethality of ∊-toxin in MDCK cells is drastically reduced by lowering temperature (Lindsay 1996; Petit et al. 2003). We show here that incubation at 4C prevents ∊-prototoxin–GFP and ∊-toxin–GFP binding and, in consequence, cytotoxicity, even at very long incubation periods.

Furthermore, co-incubation of ∊-prototoxin–GFP in excess with ∊-toxin–GFP prevented cell death in MDCK cells, thus revealing that both proteins compete for the same binding sites but that only ∊-toxin–GFP kills the cells (data not shown). Direct fluorescence studies confirmed that both ∊-toxin–GFP and ∊-prototoxin–GFP have similar binding properties to paraformaldehyde-fixed MDCK cells. The labeling showed a peripheral cell pattern, as reported elsewhere using indirect immunofluorescence (Petit et al. 1997). Moreover, according to our electron microscopy observations, most of the toxin was found in the apical cell membrane of polarized MDCK cells, where binding has a stronger cytotoxic effect (Petit et al. 1997).

Interestingly, pretreatment of fixed cells with detergent at 4C clearly changed the fluorescence distribution on the cell membrane to a punctate pattern. These observations represent the visual evidence of detergent-resistant membrane domains (DRMs) recently described on MDCK cells in association with toxin heptamerization (Miyata et al. 2002).

At this point we confirm that ∊-prototoxin–GFP is a convenient tool for studies of toxin binding mechanisms, since it binds to the same sites as active toxin but has a less damaging effect.

We injected mice IV with the above reported recombinant GFP proteins and analyzed their organ distribution and histopathology. Macroscopically, we found that kidneys of mice injected with ∊-toxin–GFP were more hematic, especially in the medullary area, than those injected with ∊-prototoxin–GFP or GFP alone. Coincident with these morphological observations, histological examination revealed hemorrhagic medullae. Moreover, we show images of selective degeneration found on distal tubules, mainly located in the cortical area, pyknotic nuclei and disorganized epithelia being the main observations. These results may help to understand the soft, friable, and very congested kidneys of infected livestock (pulpy kidney disease). On the other hand, no histopathological alterations were observed in other organs at the ∊-toxin doses administered here.

Direct fluorescence images revealed that fluorescence accumulated on two structures: vascular endothelia and renal distal tubules. Several organs of animals injected with either ∊-prototoxin–GFP or ∊-toxin–GFP showed fluorescence on the luminal surface of blood vessel endothelia, coinciding with the increased vascular permeability associated with this pathology and the edematogenous renal medullae described here. Vascular binding was, however, detected mainly in brain, where receptor sites for the toxin have been already described (Buxton 1978; Nagahama and Sakurai 1991). Experiments in which non-labeled ∊-prototoxin was injected in excess and simultaneously with ∊-prototoxin–GFP further confirmed specificity of toxin binding to brain vascular endothelia because fluorescence label drastically decreased under these experimental conditions.

On kidneys, fluorescence was mainly associated with the luminal surface of proximal tubules of all the mice analyzed, including those injected with ∊-toxin–GFP, ∊-prototoxin–GFP, or GFP alone, which indicates that these proteins cross the glomerular filter barrier and accumulate, somehow nonspecifically, on the luminal surface of these tubules. Fluorescence was detected to a lesser extent on the distal tubules of ∊-toxin–GFP- and ∊-prototoxin–GFP-injected mice, but never in the case of GFP alone. The control experiments confirmed our hypothesis about the coexistence of two binding components in renal tubules, one related to binding to proximal tubules, which was not specific and therefore non-displaceable with non-labeled ∊-prototoxin, and a specific component related to distal tubules, which was displaceable. These findings elucidate a previous report (Nagahama and Sakurai 1991) in which radioactive ∊-prototoxin was accumulated mainly in kidneys of injected mice but only a small fraction of radioactivity was displaced by ∊-prototoxin.

To complete this study of ∊-toxin pathology on kidneys, we incubated cryostat-obtained mice kidney slices with recombinant ∊-prototoxin–GFP, ∊-toxin–GFP, or GFP alone as control. Both forms of toxin, but not GFP alone, bound to distal tubules in a displaceable way, reinforcing our hypothesis about binding specificity. Furthermore, experiments in which samples were pretreated with 0.2% Triton X-100 revealed the presence of an ∊-toxin receptor mainly associated with detergent-sensitive areas. Further experiments are under way to identify this receptor.

Because pulpy kidney disease mainly affects livestock, we also performed the incubation experiments on cryoslices of sheep and cows and obtained the same distribution pattern as that of mice and rats.

Interestingly, incubations performed using human kidney cryoslices further confirm this distribution and could explain the sensitivity to ∊-toxin of a cell line of human kidney origin recently described (Shortt et al. 2000). This is the first evidence that ∊-toxin binds specifically to distal tubules of kidneys of different species, including those such as sheep and cows, in which pulpy kidney disease can be fatal, or those such as humans, in which ∊-toxin-associated pathology has not been reported.

In conclusion, we have generated useful recombinant toxins to study by direct fluorescence the effects and distribution of ∊-toxin. We have shown for the first time the specific binding on distal renal tubules of different species. In addition, we hypothesize that although several animals including mice, rats, and also humans could be sensitive to the toxin, intoxication occurs only in animals in which Clostridia overgrow, which leads to an excess of toxin production.

Footnotes

Acknowledgements

Supported by DGESIC from the Spanish Goverment (PM98–0194), and DURSI from the Generalitat de Catalunya.

We are grateful to Serveis Científic Tècnics of the Universitat de Barcelona for confocal and electron microscopy assistance and to the Mercabarna slaughterhouse for providing biological samples. We also thank Prof Dr Marta Carreras (Anatomia Patològica, Hospital Princeps d'Espanya, Bellvitge, Barcelona) for providing human samples. We thank Robin Rycroft for careful reading of the manuscript. A.S.-J. is a recipient of a predoctoral fellowship from Instituto de Salud Carlos III (BEFI N/ref: CPC/CLC).

References

Bhown

Habeerb

(1977) Structural studies on epsilon-prototoxin of Clostridium perfringens type D. Localization of the site of tryptic scission necessary for activation to epsilon-toxin. Biochem Biophys Res Commun, 78:889–896.

Borrmann

Gunther

Kohler

(2001) Effect of Clostridium perfringens epsilon toxin on MDCK cells. FEMS Immunol Med Microbiol, 31:85–92.

Buxton

(1978) The use of an immunoperoxidase technique to investigate by light and electron microscopy the sites of binding of Clostridium welchii type-D epsilon toxin in mice. J Med Microbiol, 11:289–292.

Finnie

(1984) Histopathological changes in the brain of mice given Clostridium perfringens type D epsilon toxin. J Comp Pathol, 94:363–370.

Finnie

Blumbergs

Manavis

(1999) Neuronal damage produced in rat brains by Clostridium perfringens type D epsilon toxin. J Comp Pathol, 120:415–420.

Finnie

(2004) Neurological disorders produced by Clostridium perfringens type D epsilon toxin. Anaerobe, 10:145–150.

Gardner

(1973) Pathology of Clostridium welchii type D enterotoxaemia. II. Structural and ultrastructural alterations in the tissues of lambs and mice. J Comp Pathol, 83:509–524.

Ghabriel

Zhu

Reilly

Blumbergs

Manavis

Finnie

(2000) Toxin-induced vasogenic cerebral oedema in a rat model. Acta Neurochir Suppl, 76:231–236.

Hambrook

Lindsay

Hughes

(1995) Morphological alterations in MDCK cells induced by exposure to Clostridium perfringens epsilon-toxin. Biochem Soc Trans, 23:44S.

10.

Hunter

Clarke

Kelly

Titball

(1992) Cloning and nucleotide sequencing of the Clostridium perfringens epsilon-toxin gene and its expression in Escherichia coli. Infect Immun 60:102–110.

11.

Lindsay

(1996) Assessment of aspects of the toxicity of Clostridium perfringens epsilon-toxin using the MDCK cell line. Hum Exp Toxicol, 15:904–908.

12.

Minami

Katayama

Matsushita

Okabe

(1997) Lambda-toxin of Clostridium perfringens activates the precursor of epsilon-toxin by releasing its N- and C-terminal peptides. Microbiol Immunol, 41:527–535.

13.

Miyata

Matsushita

Minami

Katayama

Shimamoto

Okabe

(2001) Cleavage of a C-terminal peptide is essential for heptamerization of Clostridium perfringens epsilon-toxin in the synaptosomal membrane. J Biol Chem, 276:13778–13783.

14.

Miyata

Minami

Tamai

Matsushita

Shimamoto

Okabe

(2002) Clostridium perfringens epsilon-toxin forms a heptameric pore within the detergent-insoluble microdomains of MDCK cells and rat synaptosomes. J Biol Chem, 277:39463–39468.

15.

Morgan

Kelly

Buxton

(1975) Vascular leakage produced in the brains of mice by Clostridium welchii type D toxin. J Comp Pathol, 85:461–466.

16.

Nagahama

Sakurai

(1991) Distribution of labeled Clostridium perfringens epsilon toxin in mice. Toxicon, 29:211–217.

17.

Payne

Oyston

(1997) The Clostridium perfringens ɛ-toxin. In Rood

, ed. The Clostridia: Molecular Biology and Pathogenesis. San Diego, Academic Press, 439–447.

18.

Payne

Williamson

Havard

Modi

Brown

(1994) Evaluation of a new cytotoxicity assay for Clostridium perfringens type D epsilon toxin. FEMS Microbiol Lett, 116:161–167.

19.

Petit

Gibert

Gillet

Laurent-Winter

Boquet

Popoff

(1997) Clostridium perfringens epsilon-toxin acts on MDCK cells by forming a large membrane complex. J Bacteriol, 179:6480–6487.

20.

Petit

Gibert

Gourch

Bens

Vandewalle

Popoff

(2003) Clostridium perfringens epsilon toxin rapidly decreases membrane barrier permeability of polarized MDCK cells. Cell Microbiol., 5:155–164.

21.

Petit

Maier

Gibert

Popoff

Benz

(2001) Clostridium perfringens epsilon-toxin induces a rapid change in cell membrane permeability to ions and forms channels in artificial lipid bilayers. J Biol Chem., 276:15736–15740.

22.

Popoff

(1987) Purification and characterization of Clostridium sordellii lethal toxin and cross-reactivity with Clostridium difficile cytotoxin. Infect Immun, 55:35–43.

23.

Sakurai

Nagahama

Fujii

(1983) Effect of Clostridium perfringens epsilon toxin on the cardiovascular system of rats. Infect Immun, 42:1183–1186.

24.

Shortt

Titball

Lindsay

(2000) An assessment of the in vitro toxicology of Clostridium perfringens type D epsilon-toxin in human and animal cells. Hum Exp Toxicol, 19:108–116.

25.

Songer

(1996) Clostridial enteric diseases of domestic animals. Clin Microbiol Rev, 9:216–234.

26.

Tamai

Ishida

Miyata

Matsushita

Suda

Kobayashi

Sonobe

(2003) Accumulation of Clostridium perfringens epsilon-toxin in the mouse kidney and its possible biological significance. Infect Immun, 71:5371–5375.

27.

Uzal

Kelly

Morris

Assis

(2002) Effects of intravenous injection of Clostridium perfringens type D epsilon toxin in calves. J Comp Pathol, 126:71–75.

28.

Zhu

Ghabriel

Blumbergs

Reilly

Manavis

Youssef

Hatami

(2001) Clostridium perfringens prototoxin-induced alteration of endothelial barrier antigen (EBA) immuno-reactivity at the blood-brain barrier (BBB). Exp Neurol, 169:72–82.