Clostridium sordellii Lethal-Toxin Autoprocessing and Membrane Localization Activities Drive GTPase Glucosylation Profiles in Endothelial Cells

Clostridium sordellii is a bacterium that can infect humans and cause serious disease and death. The principle virulence factor associated with clinical symptoms is a large protein toxin known as lethal toxin. The mechanism of lethal-toxin intoxication is assumed to be similar to that of the homologous toxins from C. difficile, but very few studies have been done in the context of endothelial cells, a relevant target in C. sordellii infections. This study was designed to test the role of the lethal-toxin enzymatic activities and membrane localization in endothelial cell toxicity and host substrate modification.

mechanism of toxin action, especially that of TcsL, since not all clinical isolates contain TcsH (11).
While much of what we understand about the TcsH/TcsL mechanism comes from analogy to the homologous TcdA and TcdB toxins from Clostridium difficile, there are several reports validating the functional similarities. Toxin activity relies upon binding to a receptor(s) on the cell surface and clathrin-mediated endocytosis into the host cell (12). Maturation of the endosome causes a conformational change in the pore-forming domain of TcsL, causing it to form a pore in the endosomal membrane (13,14). The autoprocessing domain is activated by host inositol hexakisphosphate and cleaves the glucosyltransferase domain (GTD) (15)(16)(17)(18), presumably to permit access to substrates residing at the plasma membrane. The GTD glucosylates small GTPases, predominately Rac, Ras, Ral, and Rap (19)(20)(21). The glucosylation leads to cytoskeletal rearrangement and rounding of the cells and also causes the induction of apoptosis (22)(23)(24)(25).
Previous studies investigated the different roles of host GTPases in the cellular response to intoxication and showed that glucosylation of H, K, and NRas leads to the apoptosis seen in cells by causing cell cycle arrest (26)(27)(28). Studies have also shown that Rac1 glucosylation drives a change in cell morphology through actin cytoskeletal rearrangement, which exerts the cytopathic effects (29). While work has been done to understand the relationship between specific GTPase glucosylation and the cellular effects induced by TcsL, an in-depth look at the role of the toxin functional activities in the context of cellular intoxication has not been reported. We introduced mutations into TcsL to inhibit the glucosyltransferase and autoprocessing enzymatic functions and studied the changes these mutations exhibit in lung endothelial cells.

TcsL induces cytotoxicity in mPMVECs.
TcsL cytotoxicity is driven by glucosylation of Ras, and many Ras-related pathways are mutated in standard cell lines to induce immortality. Much of the work done to study the impact of TcsL has used either transformed cell lines or cell lines that do not accurately represent the tissue specificity shown during C. sordellii infections. C. sordellii infection leads to edema and hypotension, and TcsL induces vascular permeability in the lungs of mice (8). This suggests that lung endothelial cells are a physiologically relevant model for studies of TcsL function. We chose conditionally immortalized murine pulmonary microvascular endothelial cells (mPMVECs), which behave similarly to primary cells when grown at 37°C but are permissive for expression of the simian virus 40 (SV40) large T antigen at 33°C (30). We used a recombinant system to allow expression and purification of TcsL with specific point mutations (12,18). Cells were treated with TcsL across a range of concentrations at both the permissive temperature (33°C) and the nonpermissive temperature (37°C). TcsL induced significantly higher levels of cytotoxicity at 37°C (Fig. 1A). A comparison between recombinant TcsL and TcsL purified from C. sordellii indicates that both forms of TcsL induce similar effects on endothelial cells, with no statistical difference in cytotoxicity levels (Fig. 1B).
TcsL autoprocessing and glucosyltransferase activities are important for cytotoxicity. To determine the importance of the enzymatic activities of TcsL, mutations were introduced into the active sites of the GTD and the autoprocessing domain. The glucosyltransferase activity was ablated by the introduction of the mutations D286N and D288N (DxD), and the autoprocessing activity was eliminated by introducing a C698A mutation (18,31,32) (see Table S1 in the supplemental material). The endothelial cells were treated with TcsL as well as the glucosyltransferase and autoprocessing mutants (DxD and C698A, respectively). Cell viability was determined at 24 and 48 h postintoxication using CellTiter-Glo. When cells were treated with glucosyltransferasedeficient TcsL, no cell death was seen except at the highest concentration tested (Fig. 2). The autoprocessing mutant was attenuated but still induced cytotoxicity. These findings suggest that glucosyltransferase activity is required for cytotoxicity, while the autoprocessing activity is important but not required for induction of cell death.
We next assayed the impact of enzyme mutation on the glucosylation of Rac1 and Ras GTPases in cells. We analyzed endothelial cell lysates that had been treated with 10 nM TcsL, TcsL DxD, and TcsL C698A. Cell lysates were analyzed by Western blots using Rac1 and Ras antibodies that are specific to unglucosylated GTPase (21,27,33). Once Rac1 and Ras are modified by TcsL, the antibody can no longer recognize its epitope, and the signal is lost. The cells that were treated with TcsL showed modification after 1 h, and glucosylation of both GTPases was nearly complete by 2 h (Fig. 3). When TcsL DxD was used to intoxicate cells, the loss of the glucosyltransferase activity rendered the mutant unable to glucosylate both Rac1 and Ras. Interestingly, TcsL C698A was able to quickly glucosylate Rac1, similar to wild-type TcsL, but was attenuated in its ability to glucosylate Ras GTPases. The introduction of the autoprocessing mutation did not impact the glucosylation of Rac1 or H-Ras in an in vitro assay (see Fig. S1 in the supplemental material). The difference in cellular Rac1 and Ras glucosylation when the autoprocessing activity was ablated suggests that there is a localization difference for Rac1 and Ras.

Mutations in the GTD membrane localization domain inhibit TcsL cytotoxicity. Previous reports identified a membrane localization domain (MLD) on the GTD
that is involved in the localization of the GTD to the plasma membrane (34,35). To test the importance of the MLD in TcsL cytotoxicity, F17N and R18A mutations (analogous to mutations made in the study reported in reference 35) were created, as well as a double mutation (F17N R18A). These residues are found on the surface of the MLD and were previously implicated in membrane localization (35), and the double mutant (F17N R18A) shows a defect in membrane association in a liposome binding assay (see Fig. S2 in the supplemental material). A triple mutation (F17N R18A C698A) was also created to determine the impact of both autoprocessing and MLD mutations on cytotoxicity. Endothelial cells were treated with TcsL and TcsL MLD mutants, and cytotoxicity was assessed. Relative to treatments with wild-type TcsL, cytotoxicity was significantly attenuated when cells were treated with the MLD mutants (Fig. 4). The level of attenuation for TcsL MLD mutants is similar to that of TcsL C698A and suggests that TcsL cytotoxicity is dependent upon GTD localization to the cell membrane. To understand how MLD mutations affect the mechanism of cellular response, we next assayed the glucosylation of Rac1 and Ras by TcsL MLD mutants. Cell lysates were obtained from endothelial cells treated with TcsL and TcsL MLD mutants and analyzed by Western blotting for Rac1 and Ras glucosylation. While TcsL was able to quickly glucosylate Rac1 and Ras, the MLD mutants were attenuated in their capacity to glucosylate substrates in the cell (Fig. 5). Both TcsL F17N and TcsL R18A were delayed in their glucosylation of Rac1 and Ras, and the double mutant, TcsL F17N R18A, was the most attenuated in its glucosylation of Rac1 and Ras. The triple mutant (TcsL F17N R18A C698A) was also inhibited in both Rac1 and Ras glucosylation. These observations indicate that the MLD interaction with the membrane is important for the glucosylation of both Rac1 and Ras and that the MLD mutation prevents the autoprocessing-deficient TcsL from efficiently glucosylating Rac1.

DISCUSSION
The purpose of our study was to understand the impact of the different TcsL enzymatic activities in the context of lung endothelial cells. Although many cell lines have been used in the study of TcsL-induced effects, we wanted to find a cell line that would be a good model for the effects seen during infection. Symptoms seen commonly in C. sordellii infection are edema, hypotension, and multiorgan failure, indicating that the toxins act strongly upon host microvasculature. Our use of conditionally immortalized murine pulmonary microvascular endothelial cells allowed us to induce a primary cell-like state that removed the potential interference of common mutations associated with immortalization. Figure 1 shows that TcsL is a potent cytotoxin in these cells, which allowed us to assess the impact of enzymatic activities in the cell death mechanism.
When endothelial cells were treated with TcsL and the DxD and C698A mutants, we saw that cytotoxicity was inhibited for the glucosyltransferase mutant and impaired in the autoprocessing mutant. The inhibition of cytotoxicity when treating cells with TcsL

FIG 4 TcsL membrane localization domain is important for cytotoxicity. Endothelial cells were treated with various concentrations of TcsL, TcsL membrane localization mutants (TcsL F17N, TcsL R18A, and TcsL F17N R18A), TcsL C698A, or the triple mutant TcsL F17N R18A C698A (C ؉ F ؉ R). Cell viability was determined with Glo luciferase 24 h (A) and 48 h (B) after intoxication. Relative viability was calculated by comparing the signal to cells that were mock treated and represents the average of three replicates. Error bars represent standard errors of the means. ****, P < 0.0001.
DxD supports previous research showing that glucosylation of the host cell GTPases leads to arrest of the cell cycle and induction of apoptosis (28). The decrease in cytotoxicity seen with TcsL C698A suggests that the autoprocessing activity is important in TcsL cytotoxicity (Fig. 2).
Previous studies investigating the role of autoprocessing in the C. difficile toxins suggested that preventing GTD release through inactivation of the autoprocessing activity has only modest effects on cytopathic responses (36) and, in the case of TcdB, no impact on toxin-induced necrosis (37). The observations raise the question of why the autoprocessing activity has been retained in this family of toxins. A key difference between TcsL and the C. difficile toxins is that TcsL is more active in the modification of Ras GTPases (21,33), and Ras inactivation has been linked to TcsL-induced cell death (28). We therefore assayed whether TcsL and the TcsL C698A mutant were capable of modifying both Rac and Ras. The experiments in Fig. 3 reveal that while autoprocessing is not required for the modification of Rac, it is important for the efficient modification of Ras.
The differential ability to glucosylate the GTPases when the GTD is not cleaved from the holotoxin suggests a difference in the localization of Rac and Ras GTPases. It has been reported that Rac cycles to the endosomes where it is activated before trafficking back to the membrane (38). TcsL C698A does not release the GTD, and it remains bound to the endosome. The GTD may then encounter Rac that has been trafficked to the endosome for activation. Ras GTPases, however, are trafficked to and found in abundance at the plasma membrane after translation (39). We therefore propose that the GTD that remains tethered to the endosome does not immediately encounter and glucosylate Ras proteins. We propose that the slow Ras glucosylation is due to endosomal membranes, with the tethered GTD, recycling back to the cell surface. The delay in Ras glucosylation is enough to cause the decrease in cytotoxicity.
Previous studies identified an MLD on the TcsL GTD that is conserved across all large clostridial toxins and important in membrane localization. The MLD may be important for GTD to be inserted into the plasma membrane and may help tether it to the membrane (40,41). We looked at the impact of MLD point mutations on both cytotoxicity (Fig. 4) and the glucosylation of host GTPases (Fig. 5). Our work shows that the introduction of MLD mutations inhibits the ability to induce cytotoxicity and delays the glucosylation of both Rac and Ras GTPases, supporting the importance of the MLD in tethering GTD to the cell membrane, where it can interact with and glucosylate the GTPases. Interestingly, when we combined the MLD mutations with the autoprocessing mutation, we saw a decrease in cytotoxic ability as well as glucosylation of both Rac and Ras. While TcsL C698A was able to efficiently glucosylate Rac, the loss of efficient Rac glucosylation with the introduction of MLD mutations suggests that it is not enough for the GTD to be tethered to the endosome but also relies on the ability of the GTD to interact with the endosomal membrane through the MLD.
Our studies have shown the impact of mutating the enzymatic domains of TcsL. While the glucosyltransferase activity is needed for Rac modification and cytotoxicity, autoprocessing-deficient TcsL is impaired in cytotoxicity but efficient in its modification of Rac. We also show that the interaction with the cell membrane, not just proximity, is needed for efficient glucosylation of GTPases. The increased understanding of each toxin domain during host intoxication provides a foundation for more targeted approaches to study toxin-induced cellular events.

MATERIALS AND METHODS
Toxin cloning, expression, and purification. TcsL was amplified from C. sordellii strain JGS6382 and inserted into a BMEG20 vector (MobiTec) using BsrGI/KpnI restriction digestion sites in the vector, as reported previously (18). Point mutations were introduced into the glucosyltransferase domain (F17N, R18A, D286N, and D288N) and autoprocessing domain (C698A) from the wild-type recombinant TcsL sequence using a QuikChange mutagenesis protocol. Construct names and primers used are found in Table S1 in the supplemental material. Recombinant toxins were expressed in Bacillus megaterium using 35 ml of overnight culture seeded into 1 liter of LB. Expression of toxin was induced at an optical density at 600 nm (OD 600 ) of roughly 0.5 with 5 g D-xylose and grown for 4 h at 37°C with shaking at 220 rpm. The cells were pelleted and resuspended in lysis buffer (20 mM KP i [pH 7], 500 mM NaCl, DNase, and protease inhibitors [Sigma]). The bacteria were passed through an Emulsiflex homogenizer for lysis, and centrifugation was performed at 48,000 ϫ g for 30 min. Supernatants were run over Ni affinity anion exchange columns with a gradient from 20 mM Tris (pH 8) to 20 mM Tris (pH 8)-600 mM NaCl and on Superdex 200 size exclusion columns, and toxins were eluted in 20 mM HEPES (pH 7)-50 mM NaCl.
Native TcsL was purified from C. sordellii strain JGS6382, obtained from David Aronoff (Vanderbilt University). Expression and purification were done as previously described (37).
Cell culture. Conditionally immortalized murine pulmonary microvascular endothelial cells (mPM-VECs) were obtained from the laboratory of Ambra Pozzi (Vanderbilt University). The cells were grown at 33°C using EGM-2 (Lonza) supplemented with 10 ng/ml gamma interferon (IFN-␥) in 5% CO 2 . For use, mPMVECs were transferred to 37°C for overnight growth in EGM-2 medium without IFN-␥.
Viability assays. mPMVECs were plated at 2,500 cells/well in black-wall 96-well plates. The cells were treated with toxin the following day and incubated at 37°C in 5% CO 2 for either 24 or 48 h. After incubation with toxin, viability was assessed by measuring the amount of ATP present by addition of CellTiter-Glo (Promega), and the luminescence was read using a BioTek Synergy 4 plate reader. Relative viability was calculated by setting the readings for mock-treated samples as 100%.
Western blot analysis. Cell lysates were prepared from mPMVECs that had been plated at 200,000 cells/ml in 10-cm dishes the day before. Cells were intoxicated with 10 nM toxin and incubated at 37°C for 0, 1, 2, and 3 h before manual lifting from the dish. Cells were washed and resuspended in lysis buffer (250 mM sucrose, 10 mM Tris [pH 7.4], 3 mM imidazole) and passed through a 27-gauge needle 25 times. The lysates were clarified by centrifugation. Samples were run on SDS-PAGE gels (Bio-Rad) and transferred to polyvinylidene difluoride (PVDF) for Western analysis. Blots were probed with antibodies for unmodified Rac1 (610651; BD), total Rac1 (clone 23A8; Millipore), unmodified Ras (ab52939; Abcam), total HRas (sc-520; Santa Cruz), and GAPDH (sc-25778; Santa Cruz). Horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit antibodies (7076 and 7074, respectively; Cell Signaling) were applied as secondary antibodies, and the blots were visualized using Pierce enhanced chemiluminescence Western blotting substrate (Thermo) and exposure to film. Film was scanned using the Odyssey Licor imaging system and analyzed using Image Studio Lite.
Statistics. Statistics were performed using two-way analysis of variance (ANOVA), and P values were determined using Dunnett's multiple comparisons test on GraphPad Prism.