Logo of iaiPermissionsJournals.ASM.orgJournalIAI ArticleJournal InfoAuthorsReviewers
Infect Immun. 2008 Dec; 76(12): 5677–5685.
Published online 2008 Sep 22. doi:  10.1128/IAI.00854-08
PMCID: PMC2583595

Evidence that Membrane Rafts Are Not Required for the Action of Clostridium perfringens Enterotoxin[down-pointing small open triangle]


The action of bacterial pore-forming toxins typically involves membrane rafts for binding, oligomerization, and/or cytotoxicity. Clostridium perfringens enterotoxin (CPE) is a pore-forming toxin with a unique, multistep mechanism of action that involves the formation of complexes containing tight junction proteins that include claudins and, sometimes, occludin. Using sucrose density gradient centrifugation, this study evaluated whether the CPE complexes reside in membrane rafts and what role raft microdomains play in complex formation and CPE-induced cytotoxicity. Western blot analysis revealed that the small CPE complex and the CPE hexamer 1 (CH-1) complex, which is sufficient for CPE-induced cytotoxicity, both localize outside of rafts. The CH-2 complex was also found mainly in nonraft fractions, although a small pool of raft-associated CH-2 complex that was sensitive to cholesterol depletion with methyl-β-cyclodextrin (MβCD) was detected. Pretreatment of Caco-2 cells with MβCD had no appreciable effect on CPE-induced cytotoxicity. Claudin-4 was localized to Triton X-100-soluble gradient fractions of control or CPE-treated Caco-2 cells, indicating a raft-independent association for this CPE receptor. In contrast, occludin was present in raft fractions of control Caco-2 cells. Treatment with either MβCD or CPE caused most occludin molecules to shift out of lipid rafts, possibly due (at least in part) to the association of occludin with the CH-2 complex. Collectively, these results suggest that CPE is a unique pore-forming toxin for which membrane rafts are not required for binding, oligomerization/pore formation, or cytotoxicity.

Bacterial pore-forming toxins (PFTs) create host plasma membrane lesions that produce permeability alterations (31). After these toxins bind in a monomer form to their receptors, the bound monomers typically oligomerize on the plasma membrane in a prepore. The prepore toxin oligomer then inserts into the host membrane, creating a functional transmembrane pore that is responsible for increasing ion permeability. A well-characterized example of PFTs is the family of cholesterol-dependent cytolysins, where ~50 monomers oligomerize to create a large pore of ~500 Å (33).

To improve the efficiency of oligomerization, many PFTs utilize host cell membrane rafts (14). For example, perfringolysin O, listeriolysin O, and aerolysin all use raft microdomains as platforms to cluster membrane-bound toxin monomers and accelerate the oligomerization process (2, 9, 34). Also known as lipid rafts or detergent-resistant membranes (DRMs), membrane rafts (using the current recommended nomenclature) (26) are small, dynamic microdomains of the eukaryotic plasma membrane that are rich in cholesterol, sphingolipids, and glycosylphosphatidylinositol-anchored proteins. These small, yet dynamic, domains coalesce into larger domains that cluster proteins involved in cellular processes such as cell signaling and vesicle trafficking (3).

Clostridium perfringens produces a 35-kDa pore-forming enterotoxin (C. perfringens enterotoxin [CPE]) that is responsible for the gastrointestinal symptoms of C. perfringens type A food poisoning, a leading food-borne illness in the United States (17). CPE employs a multistep mechanism of action whereby CPE first binds to receptors, which include several members of the claudin family of tight junction (TJ) proteins (7). This results in an sodium dodecyl sulfate (SDS)-sensitive complex of ~90 kDa, designated the small complex (SC). SC formation is not sufficient for cytotoxicity. Instead, CPE localized in the SC rapidly oligomerizes into an SDS-resistant hexameric prepore complex named CPE hexamer 1 (CH-1) (27, 30). Although ~450 kDa in size, CH-1 runs anomalously by SDS-polyacrylamide gel electrophoresis (PAGE) as a complex of ~155 kDa, possibly due to high SDS resistance and the compact nature of the complex (27, 30). CH-1 contains six copies of CPE, along with one or more claudins, and apparently assembles on the membrane surface.

After prepore assembly, CH-1 inserts into the plasma membrane of target cells, creating a cation-selective pore (12) that allows an influx of Ca2+ into the cell (19). Depending on the CPE dose applied, this Ca2+ influx induces cell death by either apoptosis or oncosis (4). An additional effect of this Ca2+ influx is morphological damage that disrupts the TJ between adjacent intestinal epithelial cells, thereby allowing free (unbound) CPE access to the basolateral surface of the cell, where additional SC and CH-1 complexes then form. At this point, a third CPE-containing complex, named CH-2, also assembles, possibly due to associations between the CH-1 complex and another TJ protein, occludin (29). Similar to CH-1, the occludin-containing CH-2 complex of ~600 kDa is SDS resistant and migrates anomalously by SDS-PAGE with an apparent size of ~200 kDa (27, 30).

Since lipid rafts are important for the action of many PFTs, including C. perfringens beta, epsilon, and iota toxins (11, 22, 23), the present study investigated whether membrane rafts also play a role in the action of CPE. A standard raft isolation protocol involving cold Triton X-100 extraction and sucrose density gradient centrifugation was used to evaluate the presence of CPE toxin complexes and CPE-associated host proteins within raft microdomains. This work provides the first evidence that CPE is a novel PFT with a mechanism of action that is apparently independent of membrane rafts.



CPE was purified to homogeneity from strain NCTC 8239, as described previously (21). Methyl-β-cyclodextrin (MβCD) was purchased from Sigma-Aldrich (St. Louis, MO). Claudin-4 and occludin antibodies were purchased from Invitrogen (Carlsbad, CA). Flotillin-1 antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and Na+,K+-ATPase antibody was obtained from Affinity Bioreagents (Golden, CO).

Cell culture.

Caco-2 cells were maintained in Eagle minimal essential medium (Sigma, St. Louis, MO) supplemented with 10% fetal bovine serum (Mediatech, Herndon, VA), 1% nonessential amino acids (Sigma, St. Louis, MO), 100 U/ml of ampicillin, and 100 μg/ml of streptomycin. All cultures were grown at 37°C with 5% atmospheric CO2.

SC formation.

Caco-2 cells were seeded into 100-mm dishes (Corning, Corning, NY) and grown to confluence (4 to 5 days). After treatment with CPE (2.5 μg/ml) in cold (4°C) Hanks' balanced salt solution without Ca2+ and Mg2+ (iHBSS) for 15 min, the toxin was removed, and cells were then washed twice with 10 ml iHBSS. Cells were then gently scraped and centrifuged for 3 min at 1,000 × g. Cell pellets were resuspended in 950 μl of Triton X-100 extraction buffer (see below) and loaded onto sucrose density gradients as described below. Visualization of the SC was acquired by immunoblotting with an anti-CPE antibody, as described below.

CH-1 and CH-2 complex formation.

Caco-2 cells were seeded into 100-mm dishes (Corning, Corning, NY) at a density of 2.5 × 105 cells/ml in 10 ml. Those cultures were then grown to confluence (4 to 5 days) and treated with 2.5 μg/ml of CPE in warm iHBSS for 60 min at 37°C. After this treatment, cells were washed twice with 10 ml iHBSS, harvested by gentle scraping, and centrifuged for 3 min at 1,000 × g to pellet material. Cell pellets were washed with 10 ml iHBSS, centrifuged, and resuspended with 950 μl of Triton X-100 extraction buffer and loaded onto sucrose gradients (see below). Visualization of CH-1 and CH-2 was acquired by Western blotting with an anti-CPE antibody, as described below.

Ib treatment.

Caco-2 cells were seeded into 100-mm dishes (Corning, Corning, NY) at a density of 2.5 × 105 cells/ml in 10 ml. Cultures were grown to confluence (4 to 5 days) and treated with 2.0 μg/ml of Ib for 15 min at 37°C.

Cholesterol depletion.

For experiments involving cholesterol depletion, Caco-2 cells were pretreated with 10 mM MβCD in iHBSS at 37°C. After 60 min of treatment, MβCD was removed, and cells were washed twice with iHBSS.

Cholesterol quantitation.

The cholesterol content of Caco-2 cells was determined before and after the depletion of cholesterol by MβCD treatment using the Amplex Red cholesterol assay kit (Invitrogen, Carlsbad, CA), with modifications. Briefly, Caco-2 cell monolayers with or without MβCD treatment (60 min) were washed two times, scraped, and lysed in 750 μl of the Amplex Red 1× reaction buffer. Cells were lysed for 15 min at room temperature, and insoluble material was pelleted by centrifugation for 1 min at 13,000 × g. A 50-μl aliquot of supernatant for each sample was transferred into wells of a clear 96-well microtiter plate. A 50-μl aliquot of Amplex Red working solution (cholesterol esterase omitted) was then added to each sample well. Samples were covered and incubated for 30 min at 37°C in the dark. Following this incubation, samples were read using Revelation 4.21 software (Dynax Technologies) by measuring the optical density absorbance at 570 nm.

Triton X-100 extraction of Caco-2 cells and sucrose density gradient centrifugation.

Caco-2 cells treated under the various conditions explained above were extracted with 950 μl of Triton X-100 extraction buffer (1% Triton X-100, 10 mM NaF, 25 mM Tris-HCl [pH 7.6], 150 mM NaCl, 5 mM EDTA, 30 mM Na-pyrophosphate, 10 mM β-glycerophosphate, 1× Complete protease inhibitor cocktail [Roche]) for 1 h at 4°C. Following this cold Triton X-100 extraction, lysates were overlaid with 950 μl of an 80% sucrose solution (80% sucrose in 25 mM Tris-HCl [pH 7.6], 150 mM NaCl, 5 mM EDTA, 30 mM Na-pyrophosphate, and 10 mM β-glycerophosphate). Subsequently, 1.9 ml of a 30% sucrose solution was overlaid on this sample, followed by 950 μl of a 5% sucrose solution. Samples were then spun at 200,000 × g for 18 h in a Beckman XL-90 ultracentrifuge. After centrifugation, 12 400-μl fractions were collected from the top of the gradient.


For the detection of claudin-4, occludin, flotillin-1, and Na+,K+-ATPase, each of the 12 gradient fractions was concentrated by trichloroacetic acid (TCA) precipitation. A one-fourth volume (100 μl) of 100% TCA was added to each fraction and incubated on ice for 1 h. Samples were then centrifuged (14,000 × g) for 15 min at 4°C. Pellets were washed twice with 200 μl of ice-cold acetone. Those pellets were resuspended in 50 μl of 1× Laemmli buffer and analyzed by SDS-PAGE using 10% (for occludin and Na,K-ATPase) or 12% (for claudin-4 and flotillin-1) acrylamide gels. Those separated proteins were transferred onto nitrocellulose and probed with the appropriate primary (rabbit anti-claudin-4, mouse anti-occludin, rabbit anti-flotillin-1, or mouse anti-Na+,K+-ATPase) and secondary (mouse or rabbit) antibodies conjugated with horseradish peroxidase (Sigma, St. Louis, MO).

For experiments to determine the localization of CPE complexes on sucrose gradients, 50-μl aliquots of each fraction were run on native 6% PAGE gels (for SC) or mixed with 10 μl of 5× Laemmli buffer and analyzed by SDS-PAGE using 4% acrylamide gels (for CH-1 and CH-2). After electrophoresis, separated proteins were transferred onto nitrocellulose and probed with a rabbit anti-CPE antibody followed by incubation with a goat anti-rabbit antibody conjugated to horseradish peroxidase (Sigma, St. Louis, MO). Densitometry of Western blots was performed using the program Quantity One (version 4.6.3; Bio-Rad).

Trypan blue staining.

Caco-2 cells were seeded into 35-mm dishes at a density of 5 × 104 cells/ml in 2 ml and grown to confluence (4 days). Control and MβCD-treated cells were washed and treated with CPE (2.5 μg/ml) for various times. After CPE treatment, cells were harvested by gentle scraping and centrifugation, and pellets were then resuspended in 500 μl of 0.4% trypan blue stain (Sigma). Live and dead cells were counted microscopically using a hemocytometer. Percent viability was calculated as the number of nonblue (alive) cells over the total number of cells (blue and nonblue).

Formation of CH-1 in the absence of CH-2.

For experiments requiring the formation of CH-1 in the absence of appreciable levels of the CH-2 complex, a previously described technique was used (27, 28). Caco-2 cells were seeded at a density of 2 × 105 cells/ml into two 75-mm transwell permeable supports (Corning) for each experimental condition (untreated or CPE treated). After confluence (5 days), cells were treated apically with CPE (1.0 μg/ml) for 30 min. After CPE treatment, cells were extracted and subjected to sucrose density gradient centrifugation as described above.


Detergent solubility and raft localization of the CPE SC.

To evaluate the involvement of membrane rafts in CPE action, we first investigated whether CPE complexes localize within DRMs. Since rafts are characteristically insoluble in Triton X-100 when extracted at 4°C (3), a series of sucrose density gradient experiments were performed using cold Triton X-100 extracts of Caco-2 cells treated with CPE (2.5 μg/ml) for 15 min at 4°C, conditions that allow the formation of the ~90-kDa SC without appreciable formation of the CH-1 or CH-2 complex (20). When fractions from the gradient were collected, subjected to native PAGE, and Western blotted with an anti-CPE antibody, the CPE SC was localized exclusively within the detergent-soluble fractions (fractions 8 to 12) of the gradient (Fig. (Fig.1A).1A). The marker protein flotillin-1 was used as a control for raft-associated proteins, and as expected, it migrated predominately toward the top of the gradient, mostly within fractions 2 to 5 (Fig. (Fig.1B).1B). In contrast, Na+,K+-ATPase, a nonraft protein marker (22), was present at the bottom of the gradient (Fig. (Fig.1B).1B). CPE treatment had no observable effect on the localization of either marker protein (data not shown).

FIG. 1.
The CPE SC is not localized within lipid rafts. (A) CPE Western blot of sucrose gradient fractions from Caco-2 cells containing SC. After treatment with CPE (2.5 μg/ml) for 15 min at 4°C, Caco-2 cells were cold extracted with Triton X-100 ...

Detergent solubility and raft localization of CH-1 and CH-2 complexes.

The results in Fig. Fig.11 indicated that SC formation, which corresponds to the initial binding of CPE to its receptor (35), does not occur within membrane rafts. However, rafts might still contribute to CPE oligomerization. To explore this possibility, experimental conditions that allow the formation of both the CH-1 and CH-2 SDS-resistant complexes in Caco-2 cells were used; i.e., monolayers of these cells were treated with CPE (2.5 μg/ml) for 60 min at 37°C, followed by a cold Triton X-100 extraction and sucrose density gradient centrifugation. Western blot analysis (Fig. (Fig.2A)2A) of those gradient fractions indicated that the CH-1 complex localized exclusively within the detergent-soluble fractions (fractions 8 to 12) at the bottom of the gradient. The CH-2 complex also localized predominately within those detergent-soluble fractions (Fig. (Fig.2A),2A), although a small pool of CH-2 complex was detected within the detergent-insoluble fractions (fractions 3 to 5).

FIG. 2.
Localization of the CH-1 and CH-2 complexes. (A) CPE Western blots of sucrose gradient fractions from Caco-2 cells containing CH-1 and CH-2. After treatment with CPE (2.5 μg/ml) for 60 min at 37°C, Caco-2 cells were cold extracted with ...

Since the data shown in Fig. Fig.22 suggested that, except for a small fraction of CH-2, both the CH-1 and CH-2 complexes are non-raft associated; a control was run to demonstrate that our methods could detect raft-associated toxin oligomers. Caco-2 cells were treated with Ib, the binding subunit of C. perfringens iota toxin, whose oligomers are localized, at least partially, in membrane rafts (11, 24). In agreement with those previous findings, Ib oligomers were detected (Fig. (Fig.2B)2B) within raft fractions.

Effect of cholesterol depletion on CPE complexes.

To evaluate whether membrane rafts are needed for CPE complex formation, these microdomains were disrupted by depleting cells of cholesterol using the compound MβCD, which extracts cholesterol directly from the plasma membrane of cells. When the total cholesterol content of Caco-2 cells was measured before and after MβCD treatment, MβCD had removed ~70% of the cholesterol from Caco-2 cell monolayers (data not shown).

The depletion of cholesterol in Caco-2 cells by MβCD pretreatment prior to the addition of CPE had no major effect on CH-1 complex formation (Fig. (Fig.3A).3A). Additionally, MβCD pretreatment did not affect the localization of the CH-1 complex in detergent-soluble fractions. Similarly, MβCD pretreatment had little or no effect on the formation or localization of the CH-2 complex pool that is non-raft associated (Fig. (Fig.3A).3A). However, the small raft-associated pool of the CH-2 complex (Fig. (Fig.2A)2A) in CPE-treated Caco-2 cells was absent from cells depleted of cholesterol with MβCD treatment prior to CPE treatment.

FIG. 3.
Effect of cholesterol depletion on CH-1 and CH-2 formation. (A) CPE Western blot of sucrose gradient fractions from CPE-treated Caco-2 cells with reduced cholesterol levels. Caco-2 cells were pretreated with 10 mM MβCD for 60 min, followed by ...

As a control to show the effectiveness of MβCD in disrupting the localization of normally raft-associated oligomers, Caco-2 cells were pretreated with MβCD prior to treatment with Ib. Figure Figure3B3B shows that cholesterol depletion effectively shifted Ib oligomers from raft to nonraft domains, confirming that MβCD treatment affects the presence of toxin complexes in membrane microdomains.

Effect of cholesterol depletion on CPE cytotoxicity.

Since the CH-1 complex is sufficient to obtain CPE-induced cytotoxicity (20), the results shown in Fig. Fig.33 suggested that membrane rafts do not play a major role in the cytotoxic effect of CPE action. To further address this question, we directly examined whether MβCD pretreatment can protect Caco-2 cell cultures from CPE by quantitatively measuring cell viability using trypan blue staining.

In these studies, Caco-2 cell monolayers were pretreated with 10 mM MβCD for 60 min or left untreated, followed by treatment with CPE (2.5 μg/ml) for various times (10, 20, 30, and 45 min). Figure Figure44 shows no significant effect of cholesterol depletion on the viability of Caco-2 cells in the absence of CPE (see time zero). Likewise, there was no inhibition of CPE-induced cytotoxicity with the pretreatment of Caco-2 cells with MβCD. There was, in fact, a slight increase in Caco-2 cell sensitivity to CPE after MβCD pretreatment. These data are consistent with the above-described Western blot analysis of CPE complex formation in which membrane rafts are not required for the formation of the CH-1 complex that triggers cytotoxicity.

FIG. 4.
Effect of cholesterol depletion on CPE-induced cytotoxicity. Caco-2 monolayers were left untreated (◊) or pretreated for 60 min with 10 mM MβCD (□). Cells were then treated for various times (10, 20, 30, or 45 min) with 2.5 μg/ml ...

Detergent solubility and raft localization of claudin-4.

One possible contributor to CPE complexes not being associated with membrane rafts would be for the toxin receptor to reside outside of raft domains in control (no CPE treatment) Caco-2 cells. The presence of claudins within raft domains has been somewhat controversial (15, 16, 18); therefore, we investigated whether the CPE receptor claudin-4 is associated with DRMs in our experimental system. Sucrose density gradient experiments followed by claudin-4 Western blots (Fig. (Fig.5A)5A) showed that when extracted from control Caco-2 cell monolayers, claudin-4 is normally localized to the detergent-soluble, nonraft fractions at the bottom of the gradient. CPE treatment of Caco-2 cell monolayers did not affect localization of claudin-4 into nonraft domains (Fig. (Fig.5B).5B). Similarly, MβCD treatment had no significant effect on the detergent solubility of claudin-4 in Caco-2 cells (Fig. (Fig.5C5C).

FIG. 5.
Raft localization of claudin-4. Caco-2 cells were untreated (A), treated with 2.5 μg/ml of CPE for 60 min (B), or treated for 60 min with 10 mM MβCD (C), followed by cold Triton X-100 extraction and sucrose density gradient centrifugation. ...

Detergent solubility and raft localization of occludin.

In addition to claudins, a second TJ protein named occludin is also present within the CH-2 complex (29). Existing literature is contradictory concerning the presence of occludin in membrane rafts, so we similarly assessed occludin localization in our experimental system. After sucrose gradient centrifugation of control Caco-2 cell Triton X-100 extracts and Western blot analysis using an anti-occludin antibody, occludin localized mostly in gradient fractions 3 to 5, suggesting that this TJ protein is mainly raft-associated (Fig. (Fig.6A).6A). In contrast, similar analysis of Caco-2 cells treated with CPE for 60 min revealed that most of the raft-associated occludin extracted from those cells had shifted to nonraft fractions in the sucrose gradient (Fig. (Fig.6B).6B). However, a small pool of occludin remained raft associated in CPE-treated cells. This CPE-induced shift in occludin to nonraft fractions was similar to that observed when Caco-2 cells were treated with MβCD alone, i.e., without CPE treatment (Fig. (Fig.6C).6C). The CPE-induced shift of occludin to the nonraft fraction was not a consequence of generalized membrane raft dissolution since flotillin-1 remained raft associated in these cells (data not shown).

FIG. 6.
Raft localization of occludin. Caco-2 cells were untreated (A), treated with CPE (2.5 μg/ml) for 60 min (B), or treated for 60 min with 10 mM MβCD (C), followed by cold Triton X-100 extraction and sucrose density gradient centrifugation. ...

Kinetics of CPE-induced occludin redistribution.

The CPE-induced shift of occludin from raft to nonraft fractions was further investigated by kinetic analysis. Densitometric scanning of occludin Western blots of pooled raft fractions (gradient fractions 2 to 4) or nonraft fractions (gradient fractions 9 to 12) from Caco-2 cells treated with CPE for various times showed that the shift of occludin from insoluble to soluble gradient fractions begins relatively quickly upon CPE treatment and increases with time thereafter up to 30 min (Fig. (Fig.7A).7A). Specifically, after only 10 min of CPE treatment, most occludin remained raft associated, as in untreated Caco-2 cells (Fig. (Fig.6A).6A). However, by 30 min of CPE treatment, the majority of occludin had shifted to nonraft fractions (Fig. (Fig.7A).7A). Additionally, CPE treatment caused a slight decrease in the total cellular levels of occludin, similar to that observed with MβCD treatment alone (Fig. (Fig.6C6C).

FIG. 7.
Kinetics of CPE complex formation and occludin distribution. (A) Caco-2 monolayers were treated with 2.5 μg/ml of CPE for the indicated times (10, 30, and 45 min), and sucrose gradients were then performed for Triton X-100 lysates of each time. ...

Since occludin is associated with the CH-2 complex (29), which forms mainly outside of raft microdomains (Fig. (Fig.2),2), densitometry analysis was performed on Western blot data to quantitatively determine the increase in CH-2 formation over time in order to test whether the kinetics of CH-2 complex formation could be a contributor to the CPE-induced shift of occludin from membrane rafts. As reported previously (29), this kinetic experiment demonstrated (Fig. 7B and D, diamonds) that CH-2 complex formation begins within 10 min of CPE toxin challenge. By 30 min of toxin treatment, significant amounts of the CH-2 had formed in Caco-2 cells under the experimental conditions used in these studies. Densitometry of occludin Western blots also revealed (Fig. 7C and D, squares) a time-dependent incorporation of occludin into the CH-2 complex, as reported previously (29). However, these analyses also suggested that not all occludin within Caco-2 cells may be associated with the CH-2 complex since considerable amounts of occludin were present at the gel dye front (Fig. (Fig.7C).7C). Coincident with CH-2 formation, the amount of occludin present at the gel dye front decreased with the time of CPE treatment under the experimental conditions used. These CPE blots showing the presence of CH-2 were stripped and reprobed with an occludin antibody to confirm the presence of occludin within CH-2 (data not shown). The data presented in Fig. Fig.7D7D are consistent with a relationship between CH-2 formation, occludin incorporation into CH-2, and the removal of occludin from lipid rafts (Fig. 7A to D, triangles).

To distinguish whether the shift of occludin from raft to nonraft domains seen in Fig. Fig.66 and and77 might correlate with CH-2 complex formation or is merely a by-product of CPE challenge, additional experiments that investigated the effects of SC and CH-1 on occludin distribution were performed. To assess the effect of SC formation on occludin distribution, the gradient fractions from Fig. Fig.11 were pooled and concentrated into raft and nonraft fractions, and Western blot analysis was performed using an anti-occludin antibody. Compared to untreated Caco-2 cells, results from this experiment (Fig. (Fig.8A)8A) showed that occludin remains in the insoluble raft fractions upon CPE treatment at 4°C, strongly suggesting that SC formation has no major effect on the localization of occludin in raft domains.

FIG. 8.
Effects of SC and CH-1 on occludin distribution. (A) After treatment with CPE (2.5 μg/ml) for 15 min at 4°C, Caco-2 cells were extracted with cold Triton X-100 and subjected to sucrose gradient centrifugation. Gradient fractions 2 to 5 ...

A second experiment was then performed to address whether CH-1 complex formation is sufficient to cause the redistribution of occludin observed after CPE treatment (Fig. (Fig.66 and and7).7). This experiment used Caco-2 cells grown on transwell supports, which produces polarized monolayers (28) with fully formed tight junctions; e.g., 4- to 70-kDa fluorescently labeled dextrans are impermeable to these monolayers (data not shown). As described previously (27, 28), formation of the CH-1 complex in the absence of any CH-2 formation was obtained by treating these cells apically with CPE (1.0 μg/ml) for 30 min.

Sucrose gradients and CPE Western blots were performed as described above, which confirmed the presence of CH-1 in nonraft fractions (data not shown) along with the absence of the CH-2 complex from these cells (Fig. (Fig.8C).8C). Western blots of these polarized transwell cultures containing TJs also confirmed the data in Fig. Fig.5,5, with the presence of claudin-4 in nonraft fractions (data not shown). Additionally, occludin Western blots (Fig. (Fig.8B)8B) of those same gradient samples revealed that, with or without CPE treatment, occludin remains raft associated in these CH-1-containing transwell cultures, indicating that CH-1 formation and CPE cytotoxicity from pore formation are not sufficient for the redistribution of occludin to nonraft fractions observed in the presence of CH-2.


Membrane rafts are now well-established contributors to the action of many bacterial toxins. With respect to PFTs, raft domains can promote toxin receptor binding, oligomerization, and oligomer insertion into membranes for pore formation. The involvement of membrane rafts in PFT oligomerization is attributable largely to their dynamic ability to coalesce into larger domains, allowing for protein clustering.

Since membrane rafts are involved in the action of many PFTs, the current study investigated whether these microdomains also contribute to the action of CPE, another PFT, whose action involves the sequential formation of three CPE-containing complexes, i.e., SC, CH-1, and CH-2 (27, 29, 35). Experiments assessing whether each of those CPE complexes is localized within membrane rafts showed that both the SC and CH-1 are found exclusively outside rafts in nonpolarized or polarized Caco-2 cells. Our assay system was capable of detecting toxin oligomers localized to membrane rafts since some Ib oligomers were detected in raft fractions, as reported previously (11, 24).

The localization of the CH-1 complex outside lipid rafts of polarized and nonpolarized cells was interesting since the formation of this complex, which appears to be sufficient for causing CPE-induced cytotoxicity, is thought to correspond to the CPE pore (19, 27, 28). Therefore, Caco-2 cells were pretreated with MβCD to deplete cells of their cholesterol and disrupt rafts to determine if these domains are required for CH-1 formation. This pretreatment had no detectable effect on the formation of CH-1 (Fig. (Fig.3)3) or on its presence within nonraft fractions. In contrast, MβCD pretreatment effectively shifted raft-associated Ib oligomers from raft into nonraft membrane domains in Caco-2 cells, confirming that this pretreatment was sufficient to disrupt membrane rafts in Caco-2 cells.

The inability of cholesterol depletion to affect CH-1 formation suggested that rafts are not required for CPE action. Further support for this suggestion was then provided by the inability of MβCD treatment to delay the onset of CPE-induced cytotoxicity in Caco-2 cells (Fig. (Fig.4).4). In fact, the pretreatment of cells with MβCD before CPE addition slightly increased CPE-induced damage to Caco-2 cell monolayers even though MβCD itself caused no cytotoxic effects in our system. One possible explanation for this increased CPE sensitivity is that cholesterol depletion might loosen the TJ so that CPE gains access to the basolateral side of the cell, where more CPE receptors are located (28). This possibility is supported by previous studies showing that MβCD treatment decreases transepithelial resistance in Caco-2 and MDCK cells (6, 15). It is also possible that removing an important structural component of the plasma membrane (i.e., cholesterol) compromised the overall integrity of the plasma membrane, increasing the susceptibility of the cell to CPE.

While there is less certainty regarding its contribution to CPE action, the CH-2 complex is clearly not required for CPE-induced cytotoxicity and forms only in damaged, dying cells (29). CH-2 may represent a second CPE pore that further contributes to increased membrane permeability alterations; it also appears to facilitate the observed removal of occludin from TJs of CPE-treated cells, with subsequent internalization into the cytoplasm (28). In the current study, most of the CH-2 complex formed by CPE-treated Caco-2 cells localized outside rafts, although a small pool of CH-2 in Caco-2 cells was raft associated (Fig. (Fig.2).2). MβCD-pretreated cells did not contain this small pool of raft-associated CH-2 (Fig. (Fig.3),3), supporting its association with raft microdomains.

CPE is an unusual PFT in that eukaryotic proteins are closely associated with the CPE oligomer (27). Therefore, we investigated whether the observed CPE oligomerization outside membrane rafts might involve CPE complex-associated eukaryotic proteins (especially receptor claudins) residing in nonraft membrane domains prior to CPE treatment. We first determined whether a TJ protein, claudin-4, is raft associated in our cell culture model system since (i) this CPE receptor is present in SC, CH-1, and CH-2 (27) and (ii) there are conflicting reports regarding the association of claudins with lipid rafts (15, 16, 18). The methods used in our study localized claudin-4 of Caco-2 cells exclusively within non-raft-associated fractions. Since claudin-4 is a CPE receptor and the SC forms immediately upon CPE binding, the presence of claudin-4 CPE receptors in nonraft lipid domains offers an explanation for the localization of SC to nonraft fractions.

The exclusive localization of CH-1 and claudin-4 in nonraft fractions of nonpolarized or polarized Caco-2 cells and the formation of CH-1 in MβCD-treated cells also indicate that, unlike most PFTs, CPE does not require rafts to promote its oligomerization. For CPE, raft-mediated clustering of receptors may be unnecessary since claudins already lie in close proximity to one another due to their TJ presence in multiple polymerized strands (8, 32). Furthermore, those strands also closely associate with similar claudin strands in the same and on apposing cells. However, membrane fluidity does appear to play some role in CH-1 formation since CPE binds and forms SC at 4°C, yet CH-1 and CH-2 formations are inhibited at this low temperature (29).

Although most PFTs oligomerize within rafts, CPE is not a unique PFT with respect to its ability to bind and oligomerize outside lipid rafts. Protective antigen (PA) of anthrax toxin also binds to receptors and oligomerizes outside of a raft. However, unlike CPE, oligomerization triggers the association of the PA prepore with membrane rafts (1). The presence of PA oligomers in rafts appears to be important for the eventual translocation of lethal and edema factors through the PA pore (37), i.e., for cytotoxicity, since MβCD inhibits this process. Thus, our results distinguish CPE from PA (and other PFTs) in that CPE not only binds and oligomerizes independently of raft domains but also forms active pores outside of these microdomains. The lack of movement of CPE oligomers into rafts might be attributable to claudin CPE receptors being localized in strands where movement is relatively restricted inside TJs.

The association of occludin with membrane rafts has also been controversial (5, 13, 16, 25). In contrast to the exclusive association of claudin-4 with nonraft fractions, our study found that most occludin localizes in raft fractions of control (no CPE treatment) nonpolarized or polarized Caco-2 cells. As expected, MβCD treatment of Caco-2 cells completely shifted occludin to nonraft membrane fractions. CPE treatment similarly shifted most of the occludin in polarized or nonpolarized Caco-2 cells from raft to nonraft membrane domains. This effect was time dependent (Fig. (Fig.7)7) and paralleled the appearance of most occludin-containing CH-2 in nonraft microdomains of CPE-treated Caco-2 cells. Establishing this correlation is consistent with the possibility that occludin is pulled from lipid rafts as it becomes associated with CH-2. Also consistent with CH-2 complex formation specifically contributing to the redistribution of occludin into nonraft domains, our results showed that the presence of the SC or CH-1 complexes (in the absence of CH-2) is not sufficient to shift occludin from rafts. This result also indicates that the removal of occludin from rafts is not merely a nonspecific consequence of CPE-induced cytotoxicity, nor is the removal of occludin from rafts due to nonspecific occludin internalization since our previous studies showed that occludin internalization occurs only in the presence of CH-2 (28). Additionally, the CPE-associated removal of occludin from raft domains is not a nonspecific consequence of raft breakdown, as we observed that CPE treatment does not remove the raft marker flotillin-1 from rafts.

However, the sequestration of occludin in CH-2 may not fully explain CPE effects on occludin membrane distribution since not all occludin present in CPE-treated Caco-2 cells is localized in the CH-2 complex, at least as assessed by Western blotting. CPE treatment also had other effects on occludin in Caco-2 cells. In particular, total occludin levels appeared to slightly decrease in CPE-treated cells. This effect was noticed in cells containing CH-1 in the absence of CH-2 but appeared to be stronger when both CH-1 and CH-2 complexes were present, possibly because cells containing both complexes together enhance cell death. The observed CPE-induced decrease in occludin could be a general response to cell death and also might involve, at least in part, the previously described CPE-induced removal of occludin from the TJ into the cytoplasm of Caco-2 cells, an effect that seemingly correlates with CH-2 formation (28).

It is also notable that even CPE-treated Caco-2 cells still contained a small pool of raft-associated CH-2 and occludin. This small amount of raft-associated CH-2 could represent interactions between SC or CH-1 and a distinct occludin pool. Different occludin pools could correspond to differences in occludin phosphorylation (18, 25, 36), polymerization (8), or internalization (28) or an occludin splice variant (10). Further studies will be needed to clarify the complicated relationship between occludin, rafts, and CH-2 formation.

In summary, the current study strongly suggests that CPE is a novel PFT that does not require lipid rafts for binding, oligomerization, or cytotoxicity. The presence of claudin receptors in polymerized strands may restrict the subsequent movement of receptor-bound CPE or CPE oligomers into rafts. Furthermore, CPE treatment causes a dramatic shift of most occludin in Caco-2 cells from raft to nonraft domains. This effect may be mediated, at least in part, by the sequestration of occludin in the CH-2 complex.


This work was supported by Public Health Service grant R37-AI019844-24 and T32 predoctoral training grant AI060525 from the National Institute of Allergy and Infectious Diseases.


Editor: S. R. Blanke


[down-pointing small open triangle]Published ahead of print on 22 September 2008.


1. Abrami, L., S. H. Leppla, and F. G. van der Goot. 2006. Receptor palmitoylation and ubiquitination regulate anthrax toxin endocytosis. J. Cell Biol. 172309-320. [PMC free article] [PubMed]
2. Abrami, L., and F. G. van Der Goot. 1999. Plasma membrane microdomains act as concentration platforms to facilitate intoxication by aerolysin. J. Cell Biol. 147175-184. [PMC free article] [PubMed]
3. Brown, D. A., and E. London. 1998. Functions of lipid rafts in biological membranes. Annu. Rev. Cell Dev. Biol. 14111-136. [PubMed]
4. Chakrabarti, G., X. Zhou, and B. A. McClane. 2003. Death pathways activated in CaCo-2 cells by Clostridium perfringens enterotoxin. Infect. Immun. 714260-4270. [PMC free article] [PubMed]
5. Duraisamy, Y., D. Lambert, C. A. O'Neill, and P. J. Padfield. 2007. Differential incorporation of docosahexaenoic acid into distinct cholesterol-rich membrane raft domains. Biochem. Biophys. Res. Commun. 360885-890. [PubMed]
6. Francis, S. A., J. M. Kelly, J. McCormack, R. A. Rogers, J. Lai, E. E. Schneeberger, and R. D. Lynch. 1999. Rapid reduction of MDCK cell cholesterol by methyl-beta-cyclodextrin alters steady state transepithelial electrical resistance. Eur. J. Cell Biol. 78473-484. [PubMed]
7. Fujita, K., J. Katahira, Y. Horiguchi, N. Sonoda, M. Furuse, and S. Tsukita. 2000. Clostridium perfringens enterotoxin binds to the second extracellular loop of claudin-3, a tight junction integral membrane protein. FEBS Lett. 476258-261. [PubMed]
8. Furuse, M., H. Sasaki, and S. Tsukita. 1999. Manner of interaction of heterogeneous claudin species within and between tight junction strands. J. Cell Biol. 147891-903. [PMC free article] [PubMed]
9. Gekara, N. O., T. Jacobs, T. Chakraborty, and S. Weiss. 2005. The cholesterol-dependent cytolysin listeriolysin O aggregates rafts via oligomerization. Cell. Microbiol. 71345-1356. [PubMed]
10. Gu, J. M., S. O. Lim, Y. M. Park, and G. Jung. 2008. A novel splice variant of occludin deleted in exon 9 and its role in cell apoptosis and invasion. FEBS J. 2753145-3156. [PubMed]
11. Hale, M. L., J. C. Marvaud, M. R. Popoff, and B. G. Stiles. 2004. Detergent-resistant membrane microdomains facilitate Ib oligomer formation and biological activity of Clostridium perfringens iota-toxin. Infect. Immun. 722186-2193. [PMC free article] [PubMed]
12. Hardy, S. P., M. Denmead, N. Parekh, and P. E. Granum. 1999. Cationic currents induced by Clostridium perfringens type A enterotoxin in human intestinal CaCO-2 cells. J. Med. Microbiol. 48235-243. [PubMed]
13. Kohler, H., T. Sakaguchi, B. P. Hurley, B. J. Kase, H. C. Reinecker, and B. A. McCormick. 2007. Salmonella enterica serovar Typhimurium regulates intercellular junction proteins and facilitates transepithelial neutrophil and bacterial passage. Am. J. Physiol. Gastrointest. Liver Physiol. 293G178-G187. [PubMed]
14. Lafont, F., L. Abrami, and F. G. van der Goot. 2004. Bacterial subversion of lipid rafts. Curr. Opin. Microbiol. 74-10. [PubMed]
15. Lambert, D., C. A. O'Neill, and P. J. Padfield. 2005. Depletion of Caco-2 cell cholesterol disrupts barrier function by altering the detergent solubility and distribution of specific tight-junction proteins. Biochem. J. 387553-560. [PMC free article] [PubMed]
16. Lambert, D., P. J. Padfield, J. McLaughlin, S. Cannell, and C. A. O'Neill. 2007. Ochratoxin A displaces claudins from detergent resistant membrane microdomains. Biochem. Biophys. Res. Commun. 358632-636. [PubMed]
17. Lynch, M., J. Painter, R. Woodruff, and C. Braden. 2006. Surveillance for foodborne-disease outbreaks—United States, 1998-2002. MMWR Surveill. Summ. 551-34. [PubMed]
18. Lynch, R. D., S. A. Francis, K. M. McCarthy, E. Casas, C. Thiele, and E. E. Schneeberger. 2007. Cholesterol depletion alters detergent-specific solubility profiles of selected tight junction proteins and the phosphorylation of occludin. Exp. Cell Res. 3132597-2610. [PMC free article] [PubMed]
19. McClane, B. A. 1994. Clostridium perfringens enterotoxin acts by producing small molecule permeability alterations in plasma membranes. Toxicology 8743-67. [PubMed]
20. McClane, B. A., and A. P. Wnek. 1990. Studies of Clostridium perfringens enterotoxin action at different temperatures demonstrate a correlation between complex formation and cytotoxicity. Infect. Immun. 583109-3115. [PMC free article] [PubMed]
21. McDonel, J. L., and B. A. McClane. 1988. Production, purification, and assay of Clostridium perfringens enterotoxin. Methods Enzymol. 16594-103. [PubMed]
22. Miyata, S., J. Minami, E. Tamai, O. Matsushita, S. Shimamoto, and A. Okabe. 2002. Clostridium perfringens epsilon-toxin forms a heptameric pore within the detergent-insoluble microdomains of Madin-Darby canine kidney cells and rat synaptosomes. J. Biol. Chem. 27739463-39468. [PubMed]
23. Nagahama, M., S. Hayashi, S. Morimitsu, and J. Sakurai. 2003. Biological activities and pore formation of Clostridium perfringens beta toxin in HL 60 cells. J. Biol. Chem. 27836934-36941. [PubMed]
24. Nagahama, M., A. Yamaguchi, T. Hagiyama, N. Ohkubo, K. Kobayashi, and J. Sakurai. 2004. Binding and internalization of Clostridium perfringens iota-toxin in lipid rafts. Infect. Immun. 723267-3275. [PMC free article] [PubMed]
25. Nusrat, A., C. A. Parkos, P. Verkade, C. S. Foley, T. W. Liang, W. Innis-Whitehouse, K. K. Eastburn, and J. L. Madara. 2000. Tight junctions are membrane microdomains. J. Cell Sci. 1131771-1781. [PubMed]
26. Pike, L. J. 2006. Rafts defined: a report on the Keystone Symposium on lipid rafts and cell function. J. Lipid Res. 471597-1598. [PubMed]
27. Robertson, S. L., J. G. Smedley III, U. Singh, G. Chakrabarti, C. M. Van Itallie, J. M. Anderson, and B. A. McClane. 2007. Compositional and stoichiometric analysis of Clostridium perfringens enterotoxin complexes in Caco-2 cells and claudin 4 fibroblast transfectants. Cell. Microbiol. 92734-2755. [PubMed]
28. Singh, U., L. L. Mitic, E. U. Wieckowski, J. M. Anderson, and B. A. McClane. 2001. Comparative biochemical and immunocytochemical studies reveal differences in the effects of Clostridium perfringens enterotoxin on polarized CaCo-2 cells versus Vero cells. J. Biol. Chem. 27633402-33412. [PubMed]
29. Singh, U., C. M. Van Itallie, L. L. Mitic, J. M. Anderson, and B. A. McClane. 2000. CaCo-2 cells treated with Clostridium perfringens enterotoxin form multiple large complex species, one of which contains the tight junction protein occludin. J. Biol. Chem. 27518407-18417. [PubMed]
30. Smedley, J. G., III, F. A. Uzal, and B. A. McClane. 2007. Identification of a prepore large-complex stage in the mechanism of action of Clostridium perfringens enterotoxin. Infect. Immun. 752381-2390. [PMC free article] [PubMed]
31. Tilley, S. J., and H. R. Saibil. 2006. The mechanism of pore formation by bacterial toxins. Curr. Opin. Struct. Biol. 16230-236. [PubMed]
32. Tsukita, S., M. Furuse, and M. Itoh. 2001. Multifunctional strands in tight junctions. Nat. Rev. Mol. Cell Biol. 2285-293. [PubMed]
33. Tweten, R. K., M. W. Parker, and A. E. Johnson. 2001. The cholesterol-dependent cytolysins. Curr. Top. Microbiol. Immunol. 25715-33. [PubMed]
34. Waheed, A. A., Y. Shimada, H. F. Heijnen, M. Nakamura, M. Inomata, M. Hayashi, S. Iwashita, J. W. Slot, and Y. Ohno-Iwashita. 2001. Selective binding of perfringolysin O derivative to cholesterol-rich membrane microdomains (rafts). Proc. Natl. Acad. Sci. USA 984926-4931. [PMC free article] [PubMed]
35. Wieckowski, E. U., A. P. Wnek, and B. A. McClane. 1994. Evidence that an approximately 50-kDa mammalian plasma membrane protein with receptor-like properties mediates the amphiphilicity of specifically bound Clostridium perfringens enterotoxin. J. Biol. Chem. 26910838-10848. [PubMed]
36. Wong, V. 1997. Phosphorylation of occludin correlates with occludin localization and function at the tight junction. Am. J. Physiol. 273C1859-C1867. [PubMed]
37. Young, J. A., and R. J. Collier. 2007. Anthrax toxin: receptor binding, internalization, pore formation, and translocation. Annu. Rev. Biochem. 76243-265. [PubMed]

Articles from Infection and Immunity are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...