INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Clostridium sordellii, is an important veterinary pathogen, causing abosomal bloat, hemorrhage, ulcers, liver disease, myositis, and sudden death in sheep (14, 21). C. sordellii has also been reported in human maternal deaths with a disease presenting as a toxic shock-like syndrome (4, 15). Unfortunately, the etiology of C. sordellii disease remains poorly understood. Myonecrosis is apparently rather infrequent, and death may be due to septicemia, as well as the expression of several exotoxins. Like many disease-causing clostridia, a single major virulence factor from C. sordellii has not been proposed, and most of the C. sordellii toxinsincluding phospholipase C, neuraminidase, and hemolysinare largely uncharacterized (20). In contrast, two other C. sordellii extracellular toxins, hemorrhagic toxin (TcsH) and lethal toxin (TcsL), have garnished interest in recent years (6). Intrigue with TcsH and TcsL stems from both their unique mechanisms of action and their potential use as tools in cell biology (8, 16).

TcsH and TcsL are members of the large clostridial toxins (LCTs), which represent a novel group of exceptionally large (270 to 304 kDa) bacterial virulence factors with the capacity to inactivate multiple target substrates. Along with TcsH and TcsL, LCTs include Clostridium difficile toxins A and B (TcdA and -B) and Clostridium novyi alpha toxin (Tcn). (For recent reviews of LCTs, see references 1, 5, and 6.) LCTs target members of the Ras superfamily of small GTP-binding proteins and, using UDP-glucose as a cosubstrate, TcsL, TcsH, TcdA, and TcdB glucosylate Ras proteins with various substrate specificities. TcsH, TcdA, and TcdB glucosylate Ras superfamily members Rho, Rac, and Cdc42, whereas TcsL glucosylates Ras, Rac, and Ral. Unlike other LCTs, Tcn utilizes UDP-n-acetylglucosamine, rather than UDP-glucose, as a cosubstrate for glycosylation of Rho, Rac, and Cdc42. In all cases, the LCT glycosylates a threonine (T37 or T35) residue in loop 1 of the effector-binding region, thus blocking important downstream signaling events. For Rho, Rac, and Cdc42, this inactivation may prevent effective membrane-to-cytosol cycling, regulation of actin polymerization, vesicular trafficking, and transcriptional activation, which ultimately causes extensive actin condensation, cell rounding, and cell death. Glucosylation of Ras prevents effective downstream signaling, such as phosphorylation of ERK1 and ERK2 mitogen-activated protein kinase (17).

As intracellular bacterial toxins, LCTs must enter or contact the cytosol to modify their target substrates. TcdB's mechanism of cell entry is the best characterized among LCTs and appears to require endosomal acidification and translocation from endocytic vesicles (10, 18). TcdB appears to translocate to the cytosol within 30 min following cell treatment, and at low pH the toxin exposes hydrophobic domains and forms ion-conducting channels (3, 18). Still, very little is known about the steps in cell entry for these remarkably large toxins or whether these steps in entry play any part in localization with particular substrates. The substrate targets of most LCTs cycle from the cytosol, where they are maintained in the GDP-bound form, to the plasma membrane, where the process of nucleotide exchange and activation can occur (11). In contrast, Ras, a target of TcsL, remains anchored at the cell membrane and does not cycle to the cytosol (2). Whether cellular localization plays a role in recognition of GTPase substrates by LCTs is not known. In addition, it is not clear if the cellular glucosylation of Ras by TcsL is driven solely by substrate specificity or also involves a mechanism of cell entry different from those of other LCTs. In this study, we investigated the mechanisms of cell entry by TcsL and report that TcsL cytopathic effects (CPE) are dramatically enhanced by an extracellular acid pH and blocked by lysosomotropic inhibitors. The acid-pulsed entry increased the rate of TcsL intoxication, lowered the cytopathic dose, allowed complete substrate modification, and caused ultrastructural effects similar to those of TcdB.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture and media. CHO-K1 cells, human epitheloid carcinoma (HeLa) cells, and mouse macrophage RAW cells were all obtained from the American Type Culture Collection (ATCC), Manassas, Va. These cell lines were maintained in RP-10 medium supplemented with 10% fetal bovine serum. Cultures were grown at 37°C in the presence of 6% CO₂. Unless otherwise noted, all reagents and chemicals were purchased from Sigma Chemical Co., St. Louis, Mo.

Purification of TcsL and TcdB. Cells of C. sordellii strain 9714 (ATCC) or C. difficile strain 10463 (ATCC) were grown in cellulose ester dialysis tubing with a 10,000- to 12,000-molecular-weight cutoff (Spectrum Medical Industries, Houston, Tex.), suspended in 1 liter of 0.5× brain heart infusion broth (Becton Dickinson, Sparks, Md.). Following growth at 37°C for 72 h, the culture was centrifuged at 12,000 × g for 30 min, and the supernatant was collected. TcsL and TcdB were subsequently purified at 4°C by sequential steps of high-resolution liquid chromatography as previously described (18). Each step in the purification was followed by cytotoxicity assays with CHO cells, Western blot analysis with TcdB polyclonal antiserum (a generous gift from Rodney Tweten), and visualization by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Additionally, the activity of purified TcsL was confirmed based on its substrate recognition profile (glucosylation of Rac, but not Rho and Cdc42). Following the final step of purification, the protein concentration was determined by Bradford assay (Bio-Rad Laboratories, Hercules, Calif.), and the sample was frozen at 80°C in 100-µl aliquots. Prior to use, samples of each toxin were thawed on ice and used immediately. Since repeated freeze-thawing caused increased degradation, aliquots of TcsL and TcdB were thawed once, used in the appropriate experiment, and discarded.

Inhibitor experiments. For inhibitor assays, CHO, HeLa, or RAW cells from a confluent monolayer were plated at 5 × 10⁴ cells/well in a 96-well plate and incubated for 18 h. Cells were then preincubated for 30 min with 1 × 10² M ammonium chloride, 1 × 10⁵ M monensin, or 1 × 10⁷ M bafilomycin A1 at 37°C. One picomole of TcsL or TcdB was added to the inhibitor-treated cells, and the CPE were monitored by visualization. For the time course assay, CHO cells from a confluent monolayer were plated at 5 × 10⁴ cells/well in a 96-well plate and incubated for 18 h, at which point, 1 pmol of TcsL was added to the cells. At the indicated time points, bafilomycin A1 was added to the treated cells to a final concentration of 5 × 10⁷ M. Each sample was monitored for 16 h, and CPE were determined by visualization.

Acid pulse experiments. CHO, HeLa, or RAW cells from confluent monolayers were plated at 5 × 10⁴ cells/well in a 96-well plate, incubated for 18 h, and then treated with either 5 × 10⁷ M bafilomycin A1 or 100 mM ammonium chloride for 30 min. Following treatment with the lysosomotropic agents, TcsL was added to cells across 10-fold dilutions in amounts ranging from 10 pmol to 1 fmol in a total volume of 100 µl. One hour following toxin treatment at 37°C, an acid pulse was performed by exposing the cells to acidified medium (pH 4.0) for 10 min and then replacing it with neutralized medium (pH 7.5). To determine the optimal pH for activation, cells were incubated with TcsL for 1 h at 37°C, and buffered medium was added to cells for 10 min and then removed and replaced with neutralized medium (pH 7.5). The buffers and pHs are as follows: 100 mM ammonium acetate for pHs 4.0, 4.5, 5.0, and 5.5; 100 mM morpholineethanesulfonic acid (MES) for pHs 6.0 and 6.5; and 100 mM Tris for pHs 7.0 and 7.5. The treated cells were monitored for 11 h at 37°C, and CPE was determined by visualization.

Fluorescent analysis of TcsL. Fluorescent analyis of TcsL was performed as previously described for TcdB (18). Briefly, 2-(p-toluidinyl) naphthalene-6-sulfonic acid, sodium salt (TNS) (Molecular Probes, Eugene, Oreg.) solutions were prepared in the appropriate buffers for each of the pHs to be analyzed and used at a final concentration of 150 µM. Twenty picomoles of TcsL was added to each buffer in a final volume of 2 ml and incubated at 37°C for 20 min. Each sample was analyzed on an SLM 8100 photon counting fluorimeter (Spectronic Instruments, Rochester, N.Y.) with an excitation of 366 nm and an emission scan of 380 to 500 nm with a slit width of 2.0. For the pH shift experiments, 20 pmol of TcsL was incubated with a mixture of 150 µM TNS, 50 mM Tris, 1 mM EDTA, and 100 mM NaCl (pH 7.5) for 20 min, and fluorescence was determined. The pH was adjusted to 4.0 by the gradual addition of 1 N HCl, and the emission spectrum was generated. The pH was then adjusted back to 7.5 by gradual addition of 1 N NaOH, and the fluorescence spectrum was determined again.

TcsL neutralization assays. CHO cells and HeLa cells from a confluent monolayer were plated at 5 × 10⁴ cells/well in a 96-well plate and incubated for 18 h, at which point, TcsL or TcdB was added to cells at a final concentration of 2 × 10⁹ M. At the indicated times, anti-TcdB antibody was added to the TcsL- and TcdB-treated cells, or cells were washed four times with 100 µl of phosphate-buffered saline (PBS). Cells were then monitored for 11 h, and CPE were determined by visualization.

SEM. Scanning electron microscopy (SEM) analysis was carried out at the Samuel Noble electron microscopy facility (The University of Oklahoma). CHO cells from a confluent monolayer were plated at 10⁵ cells/well in a 24-well plate and incubated for 18 h, at which point, cells were treated with TcsL, acid-pulsed TcsL, or TcdB. After 11 h, cells were incubated with 2.5% gluteraldehyde-0.1 M cacodylate buffer at pH 7.3 for 30 min. Cells were then incubated with 0.1 M cacodylate buffer-1% osmium tetroxide (pH 7.3) for 1 h. The cells were then washed three times with 1 ml of cacodylate buffer per well at pH 7.3. Samples were then dehydrated in a graded series of ethanol ranging from 30% to 100%. Critical point drying with an Autosamdri-814 device (Tousimis Research Corporation, Rockville, Md.) followed, at which point, samples were coated with gold up to a thickness of 400 Å in a Hammer VI sputter coating unit (Anatech, Ltd., Alexandria, Va.). Observations were performed in a JEOL JSM-880 SEM.

Glucosylation assay. Extracts from treated CHO cells were used as a source of substrate for TcsL glucosylation. To prepare these extracts, CHO cells were grown in 75-cm² tissue culture flasks until confluent at a density of ~1 × 10⁷ cells/flask. The cells were treated with 20 pmol of TcsL in the presence of bafilomycin A1, with or without an acid pulse. Two hours after treatment, the cells were then washed three times in ice-cold PBS, followed by mechanical removal in the presence of lysis buffer (1 mM MgCl₂, 1 mM MnCl₂, 0.1 mM phenylmethylsulfonyl fluoride, 10 µg of leupeptin per ml, 25 mM triethanolamine-HCl [pH 7.5]), similar to a previously described method (13). Cells were sonicated on ice five times for 30-s intervals, and the resulting extract was centrifuged at 40,000 × g for 8 h. The supernatant was removed and concentrated in a Centricon concentrator (Millipore, Bedford, Mass.) with a 10-kDa molecular mass cutoff until the extract reached a final volume of 0.5 ml.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of lysosomotropic agents and time course of cell entry by TcsL. The steps in cell entry for TcdA and TcdB have been documented by others and our group (9, 10, 18). While TcsL is genetically similar to TcdA and TcdB, the toxin targets different substrates, and the extent to which cell entry contributes to this difference in target recognition is not known. Since TcsL is known to interact with cytosolic targets, the toxin may exploit vesicular trafficking to gain access to substrates. In the first set of experiments, we directly compared the impact inhibitors of endosomal acidification or inhibiting conditions had on TcsL and TcdB entry. Using the CHO, HeLa and RAW cell lines, we tested monensin, bafilomycin A1, and ammonium chloride for the ability to block TcsL CPE and found, similar to previous reports (3, 17), that TcsL-induced CPE were blocked by the inhibitors of endosomal and lysosomal acidification. Similar to earlier reports on TcdB, cellular intoxication was blocked by monensin, bafilomycin A1, and ammonium chloride (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Time course of TcsL cytosolic entry. In a 96-well plate, CHO cells (5 × 104 cells/well) were incubated with TcsL (1 pmol) in a final volume of 100 µl, and bafilomycin A1 (1 × 107 M) was added at 10-min intervals from 0 to 180 min. Each sample was tested in triplicate, and CPE were determined at 16 h. The error bars mark the standard deviation from the mean. Similar levels of inhibition were found in two subsequent repetitions of the same experiment. B, PBS control; LT, TcsL; I, bafilomycin.

Extracellular pH pulse bypasses the bafilomycin A1 block. To determine whether the blocking effects of bafilomycin A1 could be bypassed, cells pretreated with bafilomycin A1 were treated with TcsL for 1 h at 37°C and then acid pulsed (pH 4.0) for 10 min. Interestingly, as shown in Fig. 2 and 3, the acid pulse entry of TcsL not only relieved the bafilomycin A1 block, but also dramatically increased the rate of intoxication, as well as significantly reducing the minimal cytopathic dose. Treatment with acid-pulsed TcsL resulted in toxic effects within less than 1 h and reached 100% within 2 h. In contrast, the first CPE from TcsL treatment at neutral pH were not evident until 6 h and did not reach their maximum until after 10 h. At neutral pH, the amount of TcsL yielding a 50% cytopathic dose was 500 fmol, whereas the acid pulse 50% cytopathic dose was less than 1 fmol. The CPE was not a reflection of cell necrosis, since at 2 h, all of the cells appeared viable by trypan blue exclusion assay (data not shown).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   pH range of acid pulse-induced entry of TcsL. In a 96-well plate, CHO cells (5 × 104 cells/well) were incubated with bafilomycin A1 (5 × 107 M) in a final volume of 100 µl for 30 min. The cells were then treated with TcsL (1 pmol) for 1 h at 37°C, subjected to a 10-min pH pulse across a range of pHs, and finally returned to pH 7.8 by addition of neutralized medium. Samples were observed for 11 h, and CPE were determined by visualization. Similar effects were determined for CHO, HeLa, and RAW cells. Curves: A, pH 4.0; B, pH 4.5; C, pH 5.0; D, TcsL (no pulse); E, pH 5.5; F, pH 6.0; G, pH 6.5; H, pH 7.0; I, pH 7.5.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   pH-enhanced rate of CHO cell intoxication by TcsL. In a 96-well plate, CHO cells (5 × 104 cells/well) were treated with either 1 pmol of TcsL or acid-pulsed TcsL (pH 4.0) in the presence of bafilomycin A1 (5 × 107 M) in a final volume of 100 µl and then observed for 11 h. CPE were determined by visualization. Curves: A, acid-pulsed TcsL; B, TcsL. Similar rates were obtained for HeLa and RAW cells.

Acid pH stimulates hydrophobic transitions in TcsL. TNS is a convenient probe for determining the exposure or sequestering of hydrophobic domains under various conditions. In an earlier report, we had shown that TcdB goes through low-pH-induced structural changes that stimulate an increase in TNS fluorescence (18). To determine whether acid pH was having a similar impact on TcsL structure, we carried out a set of TNS fluorescence experiments. As shown in Fig. 4, and in line with our previous report on TcdB, we found that as TcsL is exposed to lower pH, the protein begins to expose more hydrophobic regions. Furthermore, and also similar to TcdB, this hydrophobic transition is reversible, since raising the pH back to neutrality results in a decrease in TNS fluorescence.

View larger version (31K): [in this window] [in a new window]   FIG. 4.   TNS fluorescent analysis of TcsL. The analysis of pH-induced conformational changes in TcsL was carried out with the fluorescent probe TNS. Each spectrum represents the experimental sample with background (TNS and buffer alone) subtracted. For the pH shift condition, the toxin was treated with TNS at pH 7.5. The solution was then titrated to pH 4.0 by gradual addition of 1 N HCl and returned to neutrality by addition of 1 N NaOH. (A) TcsL hydrophobicity across a range of pHs. TcsL was incubated with TNS at pHs 4.0, 5.0, 6.0, 7.0, and 8.0, and the fluorescent spectrum of each sample was obtained. TNS fluorescence for samples above pH 5.0 were not above background levels. Emission from pHs 6.0, 7.0, and 8.0 were not detectable above background levels. (B) TNS analysis of TcsL hydrophobicity following pH shift.

Neutralization of extracellular TcsL. While TcdB and TcsL appear to be similar in some aspects of cell entry and acid-induced structural changes, the two proteins are set apart by the long delay in cytosolic entry by TcsL at neutral pH. This slow entry might be explained by a slow rate of effective cell binding or relatively slow entry under test conditions. To address these possibilities, we used a washing procedure and antibody neutralization assay to determine at which time point TcsL has been endocytosed under neutral-pH conditions. In one experiment, cells were treated with TcsL and washed three times vigorously at specific time points to remove accessible toxin, then monitored for 18 h to determine CPE. As shown in panel A of Fig. 5, TcsL CPE could be depleted by washing the cells up to 3 h after treatment, indicating a majority of TcsL was still accessible at the cell surface and apparently not tightly bound. For comparison, TcdB was subjected to the same wash experiment, and, as shown in Fig. 5, by 20 min, the CPE could not be reduced by washing the cells.

View larger version (11K): [in this window] [in a new window]   FIG. 5.   Time course of TcsL extracellular neutralization. In a 96-well plate, CHO cells (5 × 104 cells/well) were treated with 1 pmol of TcdB, TcsL, or acid-pulsed TcsL in a final volume of 100 µl. At the indicated time points, cells were then subjected to neutralizing antiserum or wash treatments. (A) Treated cells were washed vigorously with PBS four times at the indicated time points. (B) Neutralizing antiserum (10 µl) was added to cells at the indicated time points. In both experiments, cells were then observed for 11 h, and CPE were determined by visualization. Curves: solid triangles, TcdB; solid circles, TcsL.

SEM analysis of CHO cells treated with acid-pulsed TcsL. In addition to adopting a time course of intoxication similar to that of TcdB, acid-pulsed TcsL appeared to confer morphological changes similar to those of TcdB. When examined under an inverted light microscope, cells treated with acid-pulsed TcsL demonstrated a different morphology from cells treated with TcsL at neutral pH (data not shown). Furthermore, the cells treated with acid-pulsed TcsL looked remarkably similar to TcdB-treated cells (data not shown). To better characterize these changes in morphology, we analyzed the treated cells by SEM. Since the time courses of intoxication between acid-pulsed TcsL and TcsL at neutral pH are different, we attempted to normalize the effect by examining the cells at time points at which rounding was complete, but cell death, assayed by trypan blue exclusion, had not occurred. For TcdB and acid-pulsed TcsL, this time point was 2 h, and for TcsL at neutral pH, this time point was 11 h. As shown in Fig. 6, the results of SEM analysis of CHO cells treated with acid-pulsed TcsL appeared similar to those with TcdB-treated cells, with rounding and smooth cell surfaces. This morphology was clearly different from that of cells treated with TcsL at neutral pH, in which rounding and extensive cell surface blebbing had occurred.

View larger version (106K): [in this window] [in a new window]   FIG. 6.   SEM analysis of TcdB-, TcsL-, and acid-pulsed TcsL-treated CHO cells. CHO cells were grown on coverslips and treated with 1 pmol of TcdB, TcsL, or acid-pulsed TcsL in a final volume of 100 µl. After detection of changes in morphology, cells were fixed, dried, and mounted for SEM analysis and visualization. (A) PBS control (magnification, ×1,600). (B) TcsL (magnification, ×3,300). (C) TcdB (magnification, ×7,500). (D) Acid-pulsed TcsL (magnification, ×7,500).

Differential glucosylation in extracts from cells treated with acid-pulsed TcsL. Recently, Barth et al. (3) reported on the pH-induced channel-forming activity of TcdB and TcsL. Therefore, it was possible that the CPE of cells treated with acid-pulsed TcsL were due to the formation of channels at the cell surface and not related to cytosolic entry. To address this possibility, we examined lysates from cells treated with TcsL at neutral pH or with an acid pulse to determine the relative amount of glucosylated substrate. CHO cells were treated with TcsL under the two conditions and incubated for 2 h, at which point, the cells were collected and lysates were prepared. The lysates from these samples were then used in glucosylation assays to determine whether acid-pulsed entry allowed TcsL access to the substrate targets. As can be seen in Fig. 7, lysates from cells treated with acid-pulsed TcsL did not have accessible substrate, whereas cells treated with TcsL at neutral pH were still able to present substrate for glucosylation. These results suggest that while channels may be formed during the acid pulse entry, the observed CPE are more likely a result of substrate modification.

View larger version (107K): [in this window] [in a new window]   FIG. 7.   TcsL differential glucosylation of extracts from TcsL- and acid-pulsed TcsL-treated cells. Extracts from CHO cells that had been treated with 1 pmol of TcsL or acid-pulsed TcsL were used in the glucosylation assay to determine if pretreatment under these conditions blocks substrate. Glucosylation assays were carried out with [14C]UDP-glucose and TcsL with the extracts as target substrates. Incorporation of the radiolabeled glucose was determined by autoradiography. (A) Glucosylation of extracts from PBS-treated CHO cells. (B) Glucosylation of extracts from TcsL-treated CHO cells. (C) Glucosylation of extracts from TcsL-treated (acid pulsed; pH 4.0) CHO cells.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Although TcsL and TcdB are genetically similar (83% identity), these toxins target different sets of substrates (5). Interestingly, TcsL targets Ras, which, unlike other LCT substrates, remains anchored at the inner plasma membrane, and, following farnesylation, does not cycle from this site to the cytosol. Whether TcsL also differs from TcdB in its mechanism of entry has not been investigated. Based on the results from our inhibitor assays, TcsL entry appeared to be similar to TcdB entry. Furthermore, acid pH appears to have a similar effect on TcdB and TcsL structure, stimulating the exposure of hydrophobic regions. The possibility that these two toxins are temporally different in their entry pathways became evident to us from the bafilomycin A1 time course studies. In a previous study, we found that treatment with bafilomycin A1 was ineffective at blocking TcdB activity when bafilomycin A1 was added to cells 30 min after toxin treatment (18). In the current study of TcsL, results from the washing and antibody neutralization experiments suggest the lag in cell entry could be accounted for, at least in part, by a slow initiation of endocytosis or low affinity for the tested cells at neutral pH. All of these effectsslow entry, low rate of intoxication, and high cytopathic dosecould be alleviated by providing a brief acid pulse at the cell surface.

The increase in cytopathic activity for acid-pulsed TcsL differs from findings in our previous report on TcdB (18). In the earlier TcdB experiments, the blocking effects of lysosomotropic agents could be bypassed by using an extracellular acid pulse; however, acid pulse entry did not dramatically reduce the cytopathic dose. While different from TcdB, acid-pulsed TcsL entry does appear to be similar to the activity reported for diphtheria toxin almost 20 years ago (19). In the case of diphtheria toxin, while presumably not the natural mode of entry, bypassing time-consuming steps in vesicular trafficking with an acid pulse increases the rate of inhibition of protein synthesis. As a further comparison, Helicobacter pylori VacA requires extracellular acid pH for activation and is optimally active under these conditions (7). For VacA, the acid-enhanced activity appears to be part of the natural mode of action for this toxin, since H. pylori may face an acidified environment. In the case of TcsL, whether the acid pH conditions reflect a natural mechanism of activation or simulation of the endocytic vesicle is unclear.

Finally, there is only limited information on C. sordellii disease, and very little is known about the tissue or cell type that this organism targets. In this study, we selected three cell lines (CHO, HeLa, and RAW) to use as targets in our assays, and each of these cell lines gave similar data. Since the receptor for TcsL is not known, we cannot dismiss the possibility that low receptor number might account for the slow entry, as well as accounting for the results from the wash and antibody neutralization experiments. Low receptor number would not account, however, for the blocking activity of lysosomotropic inhibitors and the acid pulse effects. Taken together, these data certainly suggest endosomal trafficking and acid pH are important to the entry and cytopathic activity of TcsL.