Trapping of Vibrio cholerae Cytolysin in the Membrane-bound Monomeric State Blocks Membrane Insertion and Functional Pore Formation by the Toxin*

Background: Vibrio cholerae cytolysin (VCC) kills target eukaryotic cells by forming transmembrane oligomeric β barrel pores. Results: Alteration of key residue(s) in VCC arrests oligomerization and membrane insertion and compromises membrane pore formation. Conclusion: Trapping of VCC in its membrane-bound monomeric state blocks the membrane insertion step. Significance: This study provides novel insights regarding the membrane pore formation mechanism of VCC. Vibrio cholerae cytolysin (VCC) is a potent membrane-damaging cytolytic toxin that belongs to the family of β barrel pore-forming protein toxins. VCC induces lysis of its target eukaryotic cells by forming transmembrane oligomeric β barrel pores. The mechanism of membrane pore formation by VCC follows the overall scheme of the archetypical β barrel pore-forming protein toxin mode of action, in which the water-soluble monomeric form of the toxin first binds to the target cell membrane, then assembles into a prepore oligomeric intermediate, and finally converts into the functional transmembrane oligomeric β barrel pore. However, there exists a vast knowledge gap in our understanding regarding the intricate details of the membrane pore formation process employed by VCC. In particular, the membrane oligomerization and membrane insertion steps of the process have only been described to a limited extent. In this study, we determined the key residues in VCC that are critical to trigger membrane oligomerization of the toxin. Alteration of such key residues traps the toxin in its membrane-bound monomeric state and abrogates subsequent oligomerization, membrane insertion, and functional transmembrane pore-formation events. The results obtained from our study also suggest that the membrane insertion of VCC depends critically on the oligomerization process and that it cannot be initiated in the membrane-bound monomeric form of the toxin. In sum, our study, for the first time, dissects membrane binding from the subsequent oligomerization and membrane insertion steps and, thus, defines the exact sequence of events in the membrane pore formation process by VCC.

strains of the Gram-negative bacterium V. cholerae (1)(2)(3)(4)(5). In its mode of action, VCC belongs to the family of ␤ barrel poreforming protein toxins (␤-PFTs) (1, 6 -8). VCC is secreted by the bacteria in the form of a water-soluble, monomeric, inactive precursor molecule, termed Pro-VCC. Proteolytic removal of the N-terminal Pro-domain from this precursor generates the mature, active form of the VCC toxin (9 -12). VCC causes colloid-osmotic lysis of the target eukaryotic cells by forming transmembrane heptameric ␤ barrel pores/channels (6,13). High-resolution structural information is available for the water-soluble monomeric state (7) as well as for the transmembrane oligomeric pore form of VCC (6). Analysis of the structural models suggests that VCC follows the overall scheme of the archetypical ␤-PFT mode of action. However, the discrete intermediate events leading toward membrane pore formation by VCC have only been described to a limited extent.
Consistent with the generalized ␤-PFT mode of action, the mechanism of membrane pore formation by VCC is proposed to follow three distinct steps: binding of the toxin monomers onto the target cell membrane; formation of transient, metastable prepore oligomeric intermediates on the membrane; and conversion of the prepore oligomers into the transmembrane oligomeric ␤ barrel pores (6, 8, 14 -17). Studies on several ␤-PFTs, including VCC, also suggest that the formation of the functional transmembrane oligomeric pore structure involves membrane insertion of the pore-forming stem loop from each of the toxin protomers toward generation of the transmembrane ␤ barrel segments (18,19). However, it has not been tested experimentally, at least in the case of VCC, whether the membrane insertion of the stem loop could occur in the membrane-bound monomeric state before the prepore oligomer formation or whether the prepore oligomer formation precedes membrane insertion. Even in the case of generalized ␤-PFT mechanisms, such a sequence of events has not been established unambiguously. Previous studies have employed engineered ␤-PFT variants (for example, staphylococcal LukF and VCC) incapable of inserting their pore-forming stem loop into the membrane lipid bilayer (18,20). Such toxin variants, having their stem loop in a locked configuration via engineered disul-fide linkage, are found to remain trapped in their prepore oligomeric state (18,20). These observations, however, do not address the issue whether oligomerization is absolutely essential to initiate membrane insertion or whether membrane insertion could be initiated before prepore formation. Such notions can only be examined by trapping the ␤-PFT molecules in their membrane-bound monomeric state without allowing the formation of the oligomeric structures. In this direction, a direct correlation between oligomerization and membrane insertion has been shown in the case of staphylococcal ␣ toxin, an archetypical member of the ␤-PFT family (21). Staphylococcal ␣ toxin harboring a point mutation has been shown to display defective oligomerization of the membrane-bound toxins, with abortive membrane insertion of the pore-forming stem loop (21). This observation suggests that, in the case of staphylococcal ␣ toxin, the membrane insertion event depends critically on the prior oligomerization step. A similar mode of action has been documented in the case of perfringolysin O, a prominent member of the subclass of cholesterol-dependent cytolysins in the ␤-PFT family (22). Notably, streptolysin O, another important member in the cholesterol-dependent cytolysintype ␤-PFT category, highlights a distinct mechanism of membrane pore formation that may involve a different sequence of events (23). In the case of streptolysin O, it has been suggested that progressive assembly of the membrane-inserted monomeric units may act toward generation of the oligomeric pore structures of varying sizes. A similar assembly mechanism has also been observed in a recent study done on another cholesterol-dependent cytolysin class of ␤-PFT, pneumolysin (24).
Thus, it appears that the members of the ␤-PFT family may not necessarily follow a common generalized scheme toward exerting their membrane pore-forming activity. It is, therefore, critical to explore the sequence of the membrane insertion and oligomerization events for each individual member in the ␤-PFT family to elucidate the mechanistic details of their membrane pore-formation process.
A large number of studies have explored the mechanism(s) of membrane oligomerization process associated with the mode of actions of the ␤-PFT family members, including VCC. It is commonly proposed that the interactions of the ␤-PFT toxin monomers with the membrane components of the target cells act as the triggering mechanism to initiate the subsequent events leading toward membrane oligomerization, membrane insertion, and functional transmembrane ␤ barrel pore generation (1). In particular, membrane lipid components like cholesterol have been widely implicated in regulating the membrane pore formation process (25,26). In the case of VCC, the presence of cholesterol in the membrane lipid bilayer has been shown to be an obligatory requirement for efficient membrane oligomerization and functional membrane pore formation by the toxin (27)(28)(29)(30)(31)(32)(33). Cholesterol appears to regulate the mode of action of VCC by physically interacting with the toxin molecule and not by simply altering the physicochemical environment of the target membrane (32). It is important to note here that the VCC toxin molecule itself has been examined only to a limited extent, in the context of exploring its oligomerization mechanism. In particular, the interaction between the toxin monomers, which might be instrumental to mediate the oligomeri-zation process of VCC, has not been elucidated so far. Analysis of the VCC oligomer structure (6) highlights extensive interprotomer interactions between the neighboring subunits. The most prominent interactions are observed between the residues within the pore-forming stem region. Interestingly, trapping of the stem loop in its prestem configuration has been shown to block functional transmembrane oligomeric pore generation (SDS-stable oligomeric structures) without affecting prepore oligomer formation (SDS-labile oligomers) (18). It has also been shown that, even in the absence of the stem region, a truncated variant of VCC can form the prepore oligomer structure on the membrane (34). These observations clearly suggest that the interprotomer interactions involving the stem region of VCC are critically involved in generating the functional transmembrane oligomeric pore state without playing any significant role in initiating the oligomerization step of the membrane-bound toxin molecules. Therefore, it appears that the additional interactions that are not part of the poreforming stem region might be playing key role(s) in triggering membrane oligomerization of VCC.
In this study, to elucidate the details of the molecular mechanism of the oligomeric membrane pore-formation process of VCC, we identified the key residues in the VCC structure that are critical to trigger oligomerization of the membrane-bound toxin molecules. Alteration of such key residues blocks the membrane oligomerization step, arrests the protein in its membrane-bound monomeric state, and does not even allow membrane insertion of the pore-forming stem region from the toxin monomers. Our study, for the first time, separates the membrane binding step from the subsequent oligomerization and membrane insertion event for VCC as a prototype in the ␤-PFT family. This study also establishes that the membrane insertion indeed requires oligomerization of the membrane-bound VCC toxin protomers on the target membrane.

EXPERIMENTAL PROCEDURES
Purification of Recombinant VCC Variants-The recombinant form of WT VCC was generated as described previously (34,35). Recombinant VCC variants harboring single point mutation of D214A, W318F, R330A, or F581A were generated by a PCR-based method. Recombinant nucleotide constructs were verified by DNA sequencing. The VCC variants (D214A-VCC, W318F-VCC, R330A-VCC, and F581A-VCC) were purified following the method as described for WT VCC.
The homogeneity of the purified proteins was analyzed by SDS-PAGE and Coomassie staining. Protein concentrations were estimated by measuring the absorbance at 280 nm using the extinction coefficients of the proteins calculated from the corresponding amino acid compositions.
Intrinsic Tryptophan Fluorescence and far-UV CD-Intrinsic tryptophan fluorescence spectra were recorded using Fluromax-4 (Horiba Scientific, Edison, NJ) spectrofluorimeter upon excitation at 290 nm. The excitation and emission slits were set at 2.5 and 5 nm, respectively. All experiments were performed at 25°C using a protein concentration of 500 nM in 10 mM Tris-HCl buffer (pH 8.0). A Chirascan spectropolarimeter (Applied Photophysics, Leatherhead, Surrey, UK) was used to monitor the far-UV CD spectra of the VCC variants (ϳ400 nM).
Assay of Hemolytic Activity-Work with human blood was approved by the Institutional Bioethics Committee of the Indian Institute of Science Education and Research Mohali. The hemolytic activity of the VCC variants against human erythrocytes was determined as described previously (35). The kinetics of hemolysis were monitored by recording the decrease in turbidity of the human erythrocyte suspension in PBS (20 mM sodium phosphate buffer containing 150 mM NaCl (pH 7.4)) at 650 nm. The human erythrocyte concentration was adjusted in the reaction mixture corresponding to A 650 ϭ 0.9. The protein concentrations used were 100 nM.
Flow Cytometry-The binding of the VCC variants with human erythrocytes was monitored using a flow cytometrybased assay as described previously (33). Briefly, human erythrocytes (10 6 cells) were incubated with 75 nM protein for 30 min at 4°C in PBS. The cells were washed and treated with rabbit anti-VCC antiserum, followed by treatment with FITC-conjugated goat anti-rabbit antibody. The cells were analyzed for FITC fluorescence using a FACSCalibur (BD Biosciences) flow cytometer. Cells that were not incubated with VCC variants but stained with anti-VCC and anti-rabbit-FITC served as a control.
Calcein Release Assay-The membrane permeabilization activity of the VCC variants (1 M) against the membrane lipid bilayer of Asolectin-cholesterol liposomes (25 g/ml) was probed by monitoring the release of calcein trapped within the liposome vesicles, as described previously (33). Calcein fluorescence was recorded on a PerkinElmer Life Sciences LS 55 spectrofluorimeter at 520 nm upon excitation at 488 nm, using excitation and emission slit widths of 2.5 and 5 nm, respectively.
Pulldown Assay to Monitor Association of VCC Variants with Liposomes-The association of the VCC variants with Asolectin-cholesterol liposomes was monitored using a pulldownbased method. Liposomes (6.5 g) were incubated with 1 M protein in a 100-l reaction volume at 25°C for 30 min and subjected to ultracentrifugation at 105,000 ϫ g for 30 min at 4°C. After collecting the supernatant, liposome pellets were washed with PBS and then resuspended in 100 l of PBS. An equal volume of samples from the supernatant fraction and the resuspended liposome fractions were analyzed by SDS-PAGE/ Coomassie staining to probe for the free and liposome-bound VCC variants, respectively.
Detection of SDS-stable Oligomer Formation by Membranebound VCC Variants-For detection of SDS-stable oligomer formation in human erythrocyte cell membranes, cells (in PBS, corresponding to A 650 ϭ 0.9) were treated with the VCC variants (100 nM) in a reaction volume of 100 l at 25°C and subjected to ultracentrifugation at 105,000 ϫ g. Pellet fractions were washed with PBS and resuspended in 50 l of SDS-PAGE sample buffer. Dissolved pellets were divided into two equal parts. One half was incubated at room temperature, whereas the other half was boiled for 10 min and subsequently analyzed by immunoblotting. The sample without boiling would allow the detection of SDS-stable oligomers formed by the membrane-bound fractions of the VCC variants.
For detection of SDS-stable oligomers in Asolectin-cholesterol liposomes, liposome suspensions (6.5 g) were treated with the VCC variants (1 M) in a 100-l reaction volume as described above. Liposome-bound proteins were pelleted by ultracentrifugation at 105,000 ϫ g. Pellets were washed with PBS, and total pellet fractions were dissolved in 50 l of SDS-PAGE sample buffer. Dissolved pellets were divided into two equal parts. One half was incubated at room temperature, whereas the other half was boiled for 10 min and subsequently analyzed either by SDS-PAGE/Coomassie staining or immunoblotting. The sample without boiling would allow the detection of SDS-stable oligomers formed by the liposome-bound fraction of the VCC variants.
The immunoblot analyses were probed using rabbit anti-VCC antiserum, followed by reacting with horseradish peroxidase-conjugated goat anti-rabbit IgG. The immunoblot analyses were developed using an ECL Western blotting detection kit (GE Healthcare Life Sciences), and images were acquired on an ImageQuant LAS 4010 (GE Healthcare Life Sciences).
FRET-FRET from the tryptophan residue in the VCC variants to DPH embedded in the Asolectin-cholesterol liposome membranes was monitored on a PerkinElmer Life Sciences LS 55 spectrofluorimeter following the method described previously (34). Briefly, DPH-labeled Asolectin-cholesterol liposomes (50 g/ml) were treated with the VCC variants (1 M) at 25°C, and the FRET signal at 470 nm was recorded upon excitation at 280 nm, with excitation and emission slit widths of 2.5 and 5 nm, respectively. The tryptophan-to-DPH FRET signal was represented in terms of the relative change in DPH fluorescence at 470 nm using the following expression: [(fluorescence intensity at any time point Ϫ fluorescence intensity of the control)/Fluorescence intensity of the control)]. The DPH-labeled Asolectin-cholesterol liposomes without protein treatment served as a control.
BS 3 Cross-linking of Liposome-associated VCC Variants-Covalent cross-linking of SDS-labile prepore oligomers of the VCC variants in Asolectin-cholesterol liposomes was carried out using BS 3 (bis(sulfosuccinimidyl) suberate, Thermo Pierce) following methods described previously (34). Briefly, Asolectin-cholesterol liposomes (6.5 g) incubated with the VCC variants (1 M) in a 100-l reaction volume were subjected to ultracentrifugation at 105,000 ϫ g. Pellets were washed and resuspended in 50 l of PBS with or without 5 mM BS 3 and incubated for 30 min at 25°C. The reactions were quenched by 50 mM Tris-HCl (pH 8.0) at room temperature for 15 min. After ultracentrifugation at 105,000 ϫ g, the pellet fractions were analyzed by SDS-PAGE and Coomassie staining.
Surface Plasmon Resonance Measurements-Surface plasmon resonance (SPR) measurements were performed on a Biacore 3000 platform (GE Healthcare Life Sciences) at 25°C using an L1 sensor chip. The surface of the chip was conditioned with HBS (20 mM HEPES and 150 mM NaCl (pH 7.5)). The liposomecoated chip surface was prepared by injecting Asolectin-cholesterol liposome suspension (0.5 mM lipid concentration) at a flow rate of 1 l/min for 15 min and washed with one injection of 20 mM NaOH for 12 s at a flow rate of 100 l/min. Nonspe-cific binding was blocked by one injection of 0.1 mg/ml BSA for 5 min at a flow rate of 10 l/min.
For the binding experiments, proteins were injected for 10 min at a 5 l/min flow rate. Ten concentrations were used for each VCC variant. After each binding experiment, L1 chip was regenerated by stripping of the liposomes and bound proteins using one injection of 40 mM octyl ␤-D-glucopyranoside for 5 min at a flow rate of 10 l/min. No loss of liposome capturing efficiency was observed upon regeneration. The corrected sensogram plots were generated using BIAevaluation 4.1.1. software (GE Healthcare Life Sciences).
Amino Acid Sequence Alignment-Amino acid sequences were acquired from the NCBI server (http://www.ncbi.nlm. nih.gov/protein). The sequence alignment was generated using ClustalW (36) within the Biology Workbench server (37). The sequence alignment was rendered with the ESPript server (38).
Visualization of Structural Models-VCC protein structure coordinates (PDB codes 1XEZ and 3O44) were obtained from the Protein Data Bank. A structural model of the transmembrane oligomeric form of VCC was generated using the Orientations of Proteins in Membranes database server. Protein structural models were visualized using PyMOL.

Characterization of the VCC Variants Harboring Single Point
Mutations of D214A, R330A, and F581A-Analysis of the interprotomer interfaces of the VCC oligomer (6) shows the presence of an aspartate residue at position 214 (Asp-214), an arginine residue at position 330 (Arg-330), and a phenylalanine residue at position 581 (Phe-581) (Fig. 1, A and B). These residues, Asp-214, Arg-330, and Phe-581, are found to be highly conserved in the related cytolysin/hemolysin proteins of Vibrionaceae bacteria (Fig. 1A). Asp-214 is positioned within a unique loop structure in the cytolysin domain of VCC. It appears to participate in a salt bridge interaction with a conserved Lys-269 residue located in the neighboring protomer within the VCC oligomer structure (Fig. 1B). The Arg-330 residue is located within the first ␤ strand next to the membranespanning stem region. In the VCC oligomer structure, Arg-330 is involved in the hydrogen bond interactions with the side chain of a conserved residue, Ser-380, and the main chain carbonyl group of another conserved residue, Ala-218, from the neighboring protomer (Fig. 1B). The Phe-581 residue is located at the C-terminal boundary of the ␤-Trefoil domain of VCC, and it appears to participate in the van der Waals interaction with a conserved residue, Val-197, within the so called "cradle loop" of the adjacent VCC protomer (Fig. 1B). To examine the role of these three conserved residues, Asp-214, Arg-330, and Phe-581, in regulating the oligomerization mechanism of VCC, we generated three recombinant variants of the toxin harboring the single point mutation of D214A (D214A-VCC), R330A (R330A-VCC), and F581A (F581A-VCC) (Fig. 1C). All three mutants displayed intrinsic tryptophan fluorescence emission and far-UV CD spectra that overlapped with those of the WT VCC protein (Fig. 1, D and E). Overlapping intrinsic tryptophan fluorescence emission spectra of the wild type toxin and its mutants (Fig. 1D) indicated a similar environment for the 11 tryptophan residues within the molecular structures of the VCC variants. This, in turn, suggested overall similar global tertiary structures of D214A-VCC, R330A-VCC, and F581A-VCC compared with that of WT VCC. Likewise, nearly similar far-UV CD profiles of the wild type and the three VCC mutants (Fig. 1E) also indicated their similar secondary structural organization. Together, these data suggest that the mutations of D214A, R330A, and F581A in VCC did not noticeably alter the overall secondary and tertiary structural arrangements of the proteins.
Single Point Mutations of D214A, R330A, and F581A in VCC Abrogate Functional Pore Formation in the Membrane Lipid Bilayer of Human Erythrocytes and Synthetic Lipid Vesicles-VCC forms transmembrane oligomeric ␤ barrel pores in the membrane lipid bilayer of erythrocytes, thereby leading to colloid osmotic lysis of the cells. VCC-induced lysis of erythrocytes is considered to be the quantitative measure of the membrane pore-forming efficacy of the VCC toxin. Therefore, the functional membrane pore-forming properties of D214A-VCC, R330A-VCC, and F581A-VCC were tested by assaying their ability to trigger membrane-damaging cytolytic activity against human erythrocytes. We monitored lysis of human erythrocytes upon treatment with the VCC variants over a period of 1 h at 25°C. We observed that, at a concentration of 100 nM, F581A-VCC showed only about 30% of the wild-type lytic activity, whereas R330A-VCC could not display any noticeable lytic activity (Ͻ5% of the wild-type activity) ( Fig. 2A). D214A-VCC displayed nearly 55% of lytic activity under such conditions ( Fig.  2A). When tested over a prolonged period of up to 6 h, F581A-VCC showed ϳ60% of the wild-type activity, whereas the R330A mutant was still devoid of any lytic activity against human erythrocytes ( Fig. 2A). Notably, D214A-VCC started showing wild type-like activity at time points of 3-4 h ( Fig. 2A).
We tested whether the reduced activities were due to a compromised ability of the VCC variants to associate with the erythrocyte cells. For this, we monitored the binding of the VCC variants with human erythrocytes by using a flow cytometry-based assay. Our data showed that all three mutants, D214A-VCC, R330A-VCC and F581A-VCC, bound to the human erythrocytes with equal efficiency compared with that of the wild-type VCC toxin (Fig. 2B). These results, therefore, suggest that the single point mutations of D214A, R330A, and F581A in VCC could abrogate the membrane-damaging, pore-forming efficacy of the protein in human erythrocytes without affecting their ability to associate with the cells to any noticeable extent. The effect of the R330A mutation on the membrane pore-forming activity appeared to be more severe compared with that of the F581A mutation. The D214A mutation appeared to have a marginal effect on the pore-forming activity of VCC.
We also tested the functional pore-forming ability of D214A-VCC, R330A-VCC, and F581A-VCC in the membrane lipid bilayer of the Asolectin-cholesterol liposomes in terms of triggering the release of trapped calcein from within the liposome vesicles. As reported earlier, WT VCC displayed ϳ90% of calcein release activity within 30 min of incubation with the Asolectin-cholesterol liposomes (Fig. 3A). In contrast, the VCC mutants displayed severely compromised calcein release from the Asolectin-cholesterol liposome vesicles when tested over a period of 30 min (Fig. 3A). Prolonged exposure of liposomes with D214A-VCC and F581A-VCC, however, resulted in a sig-nificant extent of calcein release. When tested at the 6-h time point, D214A-VCC and F581A-VCC could induce ϳ80 and ϳ50% of calcein release from the Asolectin-cholesterol liposomes, respectively (Fig. 3A). Notably, the R330A mutant could not trigger any prominent calcein release, even after 6 h of treatment (Fig. 3A).
Consistent with our data regarding human erythrocytes, all three mutated variants of VCC displayed wild type-like binding with the membrane lipid bilayer of the Asolectin-cholesterol liposome vesicles. A pulldown-based assay showed that the D214A, R330A, and F581A variants of VCC could associate efficiently with the Asolectin-cholesterol liposomes (Fig. 3B), as observed with the WT VCC protein. We also employed a quantitative SPR-based assay to monitor the interactions of the VCC variants with the Asolectin-cholesterol membrane lipid bilayer. Steady-state binding sensograms showed that the D214A-VCC, R330A-VCC, and F581A-VCC proteins possessed wild typelike interaction efficacies with the Asolectin-cholesterol membrane lipid bilayer (Fig. 3C). Analysis of the end point response units (as obtained from the stable phase of the respective sensograms after completion of the protein injections) also revealed a nearly similar irreversible membrane association for the wild-type and mutant VCC variants (Fig. 3D).
Together, these data suggest that the single point mutations of D214A, R330A, and F581A in VCC abrogated the membrane pore formation ability of the toxin without significantly affecting the membrane interaction ability of the protein. Our results also suggest that, although the D214A and F581A mutations affected the membrane pore-forming activity to a moderate extent, the mutation of R330A had a drastic, deleterious effect on the process of membrane pore formation by VCC.
The Single Point Mutation of R330A in VCC Drastically Abrogates Membrane Oligomerization of the Toxin, whereas Oligomer Formation Is Only Affected Moderately by the D214A and F581A Mutations-Oligomerization of VCC in the membrane lipid bilayer is considered to be a key step toward gener-ating the transmembrane pore structure. Consistent with the generalized ␤-PFT mode of action, interaction of VCC with the target membrane lipid bilayer leads to the formation of transient, SDS-labile, prepore oligomeric intermediates, followed by their conversions into robust, SDS-stable transmembrane oligomeric pore structures. Therefore, to explore the mechanistic basis of the abortive membrane pore formation process caused by the D214A, R330A, and F581A mutations in VCC, we analyzed the oligomerization efficacy of D214A-VCC, R330A-VCC, and F581A-VCC in the membrane lipid bilayer of human erythrocytes and the Asolectin-cholesterol liposome system. For this, we monitored the ability of the membrane-bound VCC variants to generate an SDS-stable oligomeric assembly, a property commonly recognized as the signature of the transmembrane oligomeric pore structures of the archetypical ␤-PFT family members, including wild-type VCC. As reported earlier, WT VCC could form an SDS-stable oligomeric assembly in the membrane lipid bilayer of human erythrocytes (Fig.  4A). In comparison, F581A-VCC was found to display a reduced ability to generate SDS-stable oligomers in the human  erythrocyte membrane (Fig. 4A). F581A-VCC also displayed a markedly reduced oligomer formation when incubated in the presence of the Asolectin-cholesterol liposomes compared with that of WT VCC (Fig. 4B). F581A-VCC showed a noticeable extent of oligomerization in the membrane lipid bilayer of human erythrocytes and Asolectin-cholesterol liposomes (Fig.  4, A-C) that corresponded with its moderate extent of membrane-damaging, pore-forming activity. As reflected in its membrane pore-forming activity, the oligomerization efficacy of D214A-VCC was found to be only nominally reduced compared with that of the wild-type toxin (Fig. 4, A-C).
Notably, R330A-VCC did not form any detectable amount of SDS-stable oligomeric assemblies in the membrane lipid bilayer of human erythrocytes and Asolectin-cholesterol liposomes (Fig. 4, A and B). Even after 6 h of interaction with Asolectin-cholesterol liposomes, the membrane-bound fraction of R330A-VCC failed to generate SDS-stable oligomeric species (Fig. 4C). We also explored whether the R330A mutant of VCC could form any SDS-labile prepore oligomeric assemblies in the membrane lipid bilayer of Asolectin-cholesterol liposome vesicles. To trap any SDS-labile oligomers of the proteins generated in the membrane lipid bilayer of liposomes, we used the cross-linking agent BS 3 . BS 3 -mediated, covalent cross-linking could efficiently trap the oligomers of WT VCC generated in the liposome within 30 min of interaction (Fig. 4D). In contrast, BS 3 cross-linking could not arrest any such oligomeric species for R330A-VCC in the presence of the Asolectin-cholesterol liposomes within the said time frame (Fig. 4D). These data, therefore, suggest that the mutation of R330A in VCC not only abrogated the SDS-stable oligomeric assembly generation but also critically affected the SDS-labile prepore oligomer formation.
Taken together, our results establish that the mutation of R330A and F581A in VCC critically affected the oligomerization step of the membrane pore formation process. The D214A mutation affected oligomerization of VCC only to a nominal extent. Although the F581A mutation imposed only a modest oligomerization defect, the R330A mutation caused a severe blockade of the oligomerization step, thereby arresting the membrane-bound form of the protein in an abortive monomeric state.
The Mutations of D214A, R330A, and F581A in VCC Affects the Membrane Insertion Step of the Toxin-We examined whether the D214A-VCC, R330A-VCC, and F581A-VCC variants could insert their pore-forming stem region into the membrane lipid bilayer in the absence of efficient membrane oligomerization. For this, we monitored FRET from the tryptophan residue (Trp-318) located within the stem region of VCC to the DPH fluorophore embedded within the hydrophobic core of the membrane lipid bilayer, as described previously (34). An increased tryptophan-to-DPH FRET signal, upon incubation of the VCC variants in presence of the DPH-labeled Asolectincholesterol liposomes, would indicate the membrane insertion of the pore-forming stem region. It has been shown previously that a truncated variant of VCC lacking the stem region could not display any noticeable tryptophan-to-DPH FRET in the presence of the DPH-labeled Asolectin-cholesterol liposomes compared with that observed with the wild-type VCC toxin (34). To further confirm the specific role of Trp-318 in mediating such a FRET process, we generated a W318F mutant of VCC (W318F-VCC). W318F-VCC showed nearly overlapping intrinsic tryptophan fluorescence emission (Fig. 5A) and far-UV CD spectra (Fig. 5B) compared with that of the wildtype VCC, suggesting no major structural defect in the mutant.
More importantly, W318F-VCC displayed a wild type-like membrane pore-forming activity against human erythrocytes (Fig. 5C) and Asolectin-cholesterol liposomes (Fig. 5C, inset), suggesting that the mutation of W318F did not induce any defect in the toxin in terms of oligomeric membrane pore-formation efficacy. On the basis of our proposition, however, the alteration of Trp-to-Phe at position 318 within the pore-forming stem region would be expected to abrogate the tryptophanto-DPH FRET signal. Indeed, in our assay, we found that W318F-VCC did not show any time-dependent increase in the tryptophan-to-DPH FRET when incubated in the presence of the DPH-labeled Asolectin-cholesterol liposomes (Fig. 5D). As reported earlier, WT VCC showed a prominent increase in the tryptophan-to-DPH FRET signal under identical experimental conditions (Fig. 5D). Thus, these data validate that the abovementioned tryptophan-to-DPH FRET-based assay could be considered to be a robust method to monitor the membrane insertion of the VCC stem loop. Consistent with this notion, D214A-VCC and F581A-VCC showed a considerably reduced tryptophan-to-DPH FRET signal compared with that of WT VCC (Fig. 5D), suggesting that the mutations of D214A and F581A not only reduced the efficacy of membrane oligomerization but also affected the membrane insertion step to a moderate extent. Interestingly, no significant time-dependent increase in the tryptophan-to-DPH FRET signal was documented for the R330A-VCC variant (Fig. 5D). These data, FIGURE 5. Membrane insertion of the pore-forming stem loop of the VCC variants probed by a FRET-based assay. A-C, W318F-VCC was taken as a control for the FRET-based assay. W318F-VCC showed wild type-like intrinsic tryptophan fluorescence emission (A) and far-UV CD spectra (B). Fluorescence emission intensities are shown in terms of counts per second (cps). Far-UV CD signals are shown in terms of ellipticity measured in millidegrees (mdeg). Also, W318F-VCC displayed a similar extent of hemolytic activity (C) and liposome permeabilization (C, inset) compared with that of WT-VCC. D, incubation of WT-VCC in the presence of DPH-labeled Asolectin-cholesterol liposomes triggered a time-dependent increase in the tryptophan-to-DPH FRET signal, presumably because of an efficient FRET from the Trp-318 located within the stem region of the protein to the membrane-embedded DPH. This notion was confirmed by the observation that mutation of W318F in VCC completely blocked the time-dependent increase in the tryptophan-to-DPH FRET signal. For D214A-VCC and F581A-VCC, the efficiency of the process was compromised modestly. For R330A-VCC, no significant time-dependent increase in the FRET signal was observed, suggesting a severe blockade of the membrane insertion step for the mutant. therefore, suggest that the mutation of R330A could not only block membrane oligomerization of VCC but also arrest the membrane insertion of the pore-forming stem region from the membrane-bound monomeric toxin molecules.
Conclusion-The generalized mechanism of membrane pore formation by the ␤-PFTs involves three distinct steps: membrane binding, oligomerization, and membrane insertion of the pore-forming stem region from each of the toxin protomers to generate the transmembrane ␤ barrel pore. Despite this overall general scheme, individual members of the ␤-PFT family differ from each other in the intricate mechanistic details of the process. Although the oligomerization processes of the ␤-PFTs have been studied in some detail, the structural mechanism that drives oligomerization remains poorly described in most cases. Also, the sequence of events in the process of oligomeric membrane pore formation by the ␤-PFTs remains to be validated in every case because a generalization of such issue has not been established yet. In the case of VCC, the high-resolution structures have been determined for the water-soluble monomeric state as well as for the oligomeric pore form. However, the snapshots of the discrete intermediate events leading toward membrane pore formation by VCC have not been depicted earlier in mechanistic detail. The membrane interaction mechanisms of VCC have been studied previously in some detail. It has also been shown that VCC follows the archetypical ␤ barrel membrane pore-formation mechanism via formation of the prepore oligomeric intermediates. However, the critical interactions that govern the formation of the oligomeric structure on the membrane surface have not been elucidated earlier in mechanistic detail. Also, it has not been established conclusively before whether the membrane insertion step of VCC precedes prepore formation or whether it occurs only upon formation of the prepore intermediate. It has also not been tested before, for VCC, whether the membrane binding, oligomerization, and membrane insertion steps are strictly discrete events or whether they proceed concomitantly in a concerted manner.
In this study, we examined the role of the conserved Asp-214, Arg-330, and Phe-581 residues in regulating the oligomerization mechanism of VCC. Our study showed that the single point mutation of D214A, R330A, and F581A compromised the membrane oligomerization efficacy of VCC without significantly affecting the membrane-binding property of the toxin. Although mutations of D214A and F581A appeared to have modest effects, the R330A mutation exerted a more drastic blockade on the process. Our data demonstrated that the mutation of R330A in VCC not only blocked oligomerization of the membrane-bound toxins, but it also abrogated the membrane insertion of the pore-forming stem region from the membraneassociated toxin monomers. This is the first study that highlights the key residue(s) in VCC, the alteration of which blocks the oligomerization of the membrane-bound toxin monomers without even allowing the formation of the prepore oligomeric structures. Our study also enriches our insights regarding the membrane oligomerization mechanism of the ␤-PFT family of proteins in general. The oligomerization processes of a large number of ␤-PFTs have been explored previously. However, the exact structural mechanism that drives such processes has been elucidated only to a limited extent. One classic example is available in the case of staphylococcal ␣ toxin. It has been shown previously that alteration of a His-35 residue in the oligomerization surface of ␣ toxin abrogates the functional oligomeric pore formation and membrane insertion steps of the toxin (21,39). Interestingly, His-35 is located in a position within the ␣ toxin structure that matches that of Asp-214 in the VCC structure (6,15). In the case of VCC, however, Asp-214 appeared to play a nominal role in the oligomerization process. It is also important to note that the mutation of His-35 in ␣ toxin does not block prepore oligomer formation (39). Together, these observations clearly entail a diverse variation in the structural mechanisms associated with the oligomeric membrane pore formation by the ␤-PFT family members.
In sum, our study, for the first time, reveals the interactions and some of the associated key residues implicated in the membrane oligomerization process of VCC. Blockade of these interactions traps the protein in its membrane-bound monomeric state and arrests the critical membrane insertion step toward generation of the functional transmembrane oligomeric ␤ barrel pore structures by the toxin. The study provides novel insights regarding the membrane pore-formation mechanism of VCC and also enriches our insights in the broader context of the generalized ␤-PFT mode of actions.