SEARCH

SEARCH BY CITATION

Keywords:

  • bacterial protein toxin;
  • cholesterol;
  • membrane;
  • pore-forming toxin;
  • Vibrio cholerae cytolysin

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Vibrio cholerae cytolysin (VCC) belongs to the family of β-barrel pore-forming protein toxins. VCC is secreted by the bacteria as water-soluble monomers, which upon binding to target eukaryotic cells form transmembrane heptameric β-barrel channels. High-resolution 3D structures are described both for the water-soluble monomeric form and the transmembrane oligomeric pore; albeit that our understanding of the mechanistic details of the membrane pore-formation process remains incomplete. Here, we report the characterization of a nonfunctional VCC variant harboring a single point mutation of Ala425Val positioned within a potential membrane-interacting loop in the VCC structure. The mutation appears to affect interaction of the toxin with erythrocytes as well as cholesterol-containing liposome membrane, without affecting the oligomerization ability of the membrane-bound toxin molecules. The membrane-bound oligomers formed by this VCC mutant do not appear to represent the functional pore assembly of the toxin; rather, such assembly could be considered as being trapped in an abortive, nonfunctional oligomeric state. Our results suggest that the Ala425Val mutation in VCC critically compromises its cholesterol-dependent membrane-interaction mechanism and also abrogates the process of functional membrane pore formation by the toxin.


Abbreviations
dansyl-PE,

dansylated phosphatidylethanolamine

FRET,

fluorescence resonance energy transfer

GMF,

geometric mean fluorescence

PFT,

pore-forming toxin

VCC,

Vibrio cholerae cytolysin

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial β-barrel pore-forming toxins (β-PFTs) constitute a unique class of membrane-damaging cytolytic proteins, and are implicated in the virulence mechanisms of a wide spectrum of pathogenic bacteria [1, 2]. In general, bacterial β-PFTs are secreted by the organisms as water-soluble monomeric molecules, and in contact with their target eukaryotic cell membranes they convert into transmembrane oligomeric β-barrel channels, thus inducing colloid-osmotic lysis of the target cells [3-6]. In the majority of cases, formation of a transient ‘pre-pore’ oligomeric state is speculated and/or detected on the membrane surface [7, 8], followed by considerable conformational change(s) in the molecule, so that each subunit donates β-strand(s) to form the transmembrane β-barrel [2, 9-12]. Despite the general overall scheme, β-PFTs deviate significantly from each other in terms of the fine detail of the mechanism involved in membrane pore formation, namely: (a) β-PFT interactions with the target cell membranes, (b) conversion of the membrane-bound monomers into the oligomeric structure, and (c) generation of the transmembrane β-barrel pore.

Vibrio cholerae cytolysin (VCC) is a prominent member of the β-PFT family [11, 13]. It is produced by many pathogenic strains of V. cholerae, the causative agent of the severe diarrheal disease cholera [14]. VCC displays potent cytolytic activity against a range of eukaryotic cells, and has also been shown to possess enterotoxicity in terms of inducing fluid accumulation in ligated rabbit ileal loops [15-18]. Based on these observations, VCC has been proposed as a potential virulence factor of V. cholerae. Consistent with the generalized mode of action displayed by other members of the β-PFT family [3-6], VCC is secreted by the bacteria as water-soluble monomers, which in contact with the target cell membranes form transmembrane oligomeric β-barrel channels [19-23]. High-resolution 3D structures have been determined for both the transmembrane oligomeric pore [11] and the water-soluble monomeric form of the VCC protein (in the form of a precursor called Pro-VCC) [13]. Detailed analysis of the structural models demonstrates significant structural and organizational changes in the VCC molecule during transition from the water-soluble monomeric state to the transmembrane oligomeric pore; however, the exact mechanism(s) that triggers such alterations in the VCC structure remains elusive. It is commonly believed that interactions of VCC with the membrane (or specific membrane components) play critical role(s) in triggering the multiple steps of the membrane pore-formation process, namely: (a) membrane binding, (b) oligomerization of membrane-bound monomers, and (c) membrane insertion of the ‘stem region’ from each monomeric subunit to generate a functional transmembrane β-barrel pore. In particular, the involvement of membrane cholesterol has been implicated in modulating the efficacy of some of these events [24-29]. Nevertheless, the detailed structural mechanism associated with the VCC–membrane interaction(s) that regulates formation of the functional transmembrane pore is only vaguely elucidated. Moreover, the implications of the critical features present in the VCC molecule itself in modulating distinct steps of the membrane pore-formation event are only partly understood.

In this study, we report the characterization of a nonfunctional VCC variant that harbors a single point mutation of Ala425Val. The mutation is located within a potential membrane-interacting loop in the core cytolysin domain of the VCC structure. The mutation appears to affect interaction of the toxin with the erythrocyte membrane without affecting the oligomerization ability of the membrane-bound toxin molecules. Interestingly, the oligomers formed by this mutant variant in the erythrocyte membrane do not represent the functional transmembrane pore assembly capable of triggering the colloid-osmotic lysis of the target cells. Thus, it appears that the Ala425Val mutation in VCC traps the toxin in a membrane-bound abortive, nonfunctional oligomeric state. The mutation is found to disrupt interaction of the toxin with cholesterol-containing liposome membranes without affecting its basal, low-level association with liposome that lacks cholesterol. This indicates that the Ala425Val mutation affects a cholesterol-dependent mechanism of interaction of the toxin with the liposome membrane. Such a compromised cholesterol-dependent membrane interaction, coupled with an abortive pore-forming ability, suggest that the Ala425Val mutation abrogates a critical step toward the formation of a functional membrane channel by the VCC toxin.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Recombinant variant of VCC with single point mutation of Ala425Val

In the process of cloning, expression and purification of recombinant wild-type VCC (WT-VCC), we obtained a VCC variant (Ala425Val-VCC) that harbored a point mutation of Ala425Val within a potential membrane-interacting loop in the toxin (Fig. 1A,B) [11]. The point mutation resulted from a change in the nucleotide sequence that was presumably introduced during the PCR amplification step of the cloning process. This mutant variant of recombinant VCC did not show any detectable difference in terms of its expression and solubility compared with the wild-type VCC molecule. The mutant variant displayed a similar tryptophan fluorescence emission and far-UV CD profile as the wild-type VCC toxin (Fig. S1), suggesting that the Ala425Val mutation did not induce any major change in the overall tertiary and secondary structural organization of the protein. Furthermore, Ala425Val-VCC exhibited an unfolding profile (as revealed by the intrinsic tryptophan fluorescence emission and far-UV CD spectra) similar to that of WT-VCC, when subjected to thermal denaturation (Fig. S1). In order to explore further the possible effect of the Ala425Val mutation on the VCC molecular structure, we constructed a homology-based structural model of the Ala425Val-VCC molecule using the structure of the monomeric water-soluble form of WT-VCC as a template [13]. We compared the modeled structure of the mutant protein with that of the wild-type VCC. The WT-VCC structure showed the presence of a distinct surface-exposed hydrophobic patch encompassing the region corresponding to the above-mentioned ‘potential membrane-interacting loop’, a site also harboring the residue Ala425 (Fig. 1C). Notably, incorporation of the Ala425Val mutation was found to increase the extent of this surface hydrophobic patch to a significant degree (Fig. 1C). This result is consistent with the increased hydrophobicity of the side chain of valine residue compared with that of the alanine side chain. Taken together, these data indicated that the presence of the Ala425Val mutation could have introduced marked alteration in the local microenvironment within the potential membrane-interacting loop of the VCC protein structure.

image

Figure 1. Structural models showing the location of the Ala425Val mutation in the membrane-interacting region of the VCC structure. (A) Structural model of the mature form of the VCC protein [13]. The ‘prestem’ region is shown in orange. The location of the Ala residue at position 425 is shown in red. The inset shows a zoomed view of the location of the Ala residue at position 425. (B) Model of the VCC heptameric assembly [11]. The location of the Ala425Val mutation within a potential membrane-interacting loop in VCC structure is shown in red. The transmembrane β-barrel channel is shown in orange. (C) Comparison of the structural models of wild-type and mutant VCC. A structure-based homology model of Ala425Val-VCC was constructed using the wild-type VCC monomer structure (PDB entry: 1XEZ) as the template. The structural models were analyzed and compared for the presence of surface-exposed hydrophobic patches. The WT-VCC structure showed presence of a prominent surface-exposed hydrophobic patch contouring the potential membrane-interacting loop harboring the Ala425 residue. A homology-based structural model of Ala425Val-VCC suggested that introduction of the mutation could notably extend this hydrophobic patch on the protein surface. The surface-exposed hydrophobic patches are colored from red (hydrophobic) to blue (hydrophilic) gradient.

Download figure to PowerPoint

Single point mutation of Ala425Val in the VCC molecule critically impairs the hemolytic activity of the toxin

The single point mutation of Ala425Val in the VCC molecule critically impaired the hemolytic activity of the toxin. Recombinant WT-VCC analyzed in the present study, displayed strong hemolytic activity against human erythrocytes; WT-VCC induced 100% lysis of the human erythrocytes at a concentration of ~ 6.5 μg·mL−1 (Fig. 2A). In contrast, the Ala425Val-VCC mutant displayed no detectable hemolytic activity even at fivefold higher concentration (Fig. 2A). These data therefore suggested that the Ala425Val mutation critically affected the functional pore-formation ability of VCC in the human erythrocyte cell membrane.

image

Figure 2. Effect of the Ala425Val mutation on binding, oligomerization and functional pore-forming activity of the VCC toxin in the human erythrocyte membrane. (A) Comparison of the hemolytic activity of WT-VCC (●) and Ala425Val-VCC (■). Error bars indicate the standard deviations determined from three measurements. The data are representative of at least three independent assays each carried out in triplicate. (Inset) SDS/PAGE/Coomassie staining profile of purified proteins (10 μL load of 65 μg·mL−1 protein samples). Lane 1, Ala425Val-VCC; lane 2, WT-VCC. (B) Comparison of binding and oligomerization propensities of WT-VCC (lanes 3 and 4) and Ala425Val-VCC (lanes 1 and 2) as probed by western blotting using anti-VCC serum. Human erythrocyte cells in NaCl/Pi (corresponding to D650 = 0.8) were incubated with the wild-type or mutant variant of the VCC toxin (65 μg·mL−1) for 1 h at 25 °C in a final reaction volume of 200 μL. Subsequently, reaction mixtures were subjected to ultracentrifugation at 105 000 g, and the pellets were washed twice with NaCl/Pi, dissolved in SDS/PAGE sample buffer, subjected to SDS/PAGE and western blotting using anti-VCC serum. Boiling in the presence of SDS/PAGE sample buffer led to dissociation of the oligomeric assembly into the monomeric form, thus showing only the total amount of the membrane-bound fractions of the proteins. Unboiled samples allowed detection of the SDS-stable oligomeric assembly of the VCC proteins formed in the human erythrocyte membranes. Band intensities were estimated using the Gel Analysis tool within the imagej software (US National Institutes of Health, Bethesda, MD, USA, http://imagej.nih.gov/ij/, 1997–2011). Normalized band intensities are indicated at the bottom for qualitative comparison. The data shown are representative of at least three independent experiments. (C) Comparison of the hemolytic activity and corresponding oligomer load in the human erythrocyte membrane. Human erythrocytes were treated with the appropriate concentrations of Ala425Val-VCC and WT-VCC, and membrane-bound oligomer load was probed by western blotting using anti-VCC serum as described in (B). For simplicity, only the oligomer bands are shown. Lane 1, 65 μg·mL−1 Ala425Val-VCC; lane 2, 6.5 μg·mL−1 WT-VCC; lane 3, 3.25 μg·mL−1 WT-VCC; lane 4, 1.625 μg·mL−1 WT-VCC. Band intensities were compared as described in (B). Normalized band intensities are indicated at the bottom for qualitative comparison. Data are representative of at least three independent experiments. (D) Binding of the VCC variants to human erythrocytes determined by flow cytometry. Protein concentrations are indicated at the top of each panel. Shaded peak, control; solid line, WT-VCC; dashed line, Ala425Val-VCC.

Download figure to PowerPoint

Binding and oligomerization ability of the Ala425Val-VCC variant in the erythrocyte membrane

In order to explore the mechanism by which the Ala425Val mutation impaired the hemolytic activity of the VCC toxin, we compared the binding and oligomerization activities of the wild-type and mutant proteins to the erythrocyte membranes (Fig. 2B–D). We monitored targeting of the wild-type and the mutant VCC proteins to the human erythrocyte membrane by immunoblotting using anti-VCC serum. Quantification of the immunoblotting data revealed that the Ala425Val-VCC mutant possessed approximately half of the wild-type interaction propensity with the erythrocyte cell membrane (Fig. 2B). These data indicated that the interaction of VCC with the erythrocyte cell membrane was reduced, at least to some notable extent, because of the incorporation of the single point mutation of Ala425Val. A more quantitative estimation of the binding ability of the VCC variants toward the human erythrocyte membrane was obtained using a flow cytometry-based assay (as described in Materials and Methods). The result of this assay clearly showed that the Ala425Val mutation critically affected binding of the VCC molecule with the erythrocyte membrane (Fig. 2D).

When probed for the efficiency to form the SDS-stable oligomeric assembly on the erythrocyte cell membrane (a property exhibited by the archetypical β-PFTs including VCC), the membrane-bound fraction of the Ala425Val-VCC mutant showed almost equivalent oligomerization ability compared with the wild-type protein (comparing the ratio of the total membrane-bound fractions with ratio of the oligomer fractions; Fig. 2B). The data suggested that the oligomerization propensity of the Ala425Val-VCC mutant was not affected to any significant extent. It must be noted that, as observed with the WT-VCC protein, the Ala425Val-VCC mutant could also form the typical ring-like oligomeric structures on the erythrocyte cell membrane (Fig. 3).

image

Figure 3. Transmission electron micrograph showing the ring-like oligomeric structures formed by (A) WT-VCC and (B) Ala425Val-VCC on the human erythrocyte membrane. Ring-shaped oligomeric structures are marked with arrow.

Download figure to PowerPoint

Ala425Val mutation traps the VCC toxin into an abortive oligomeric state in the erythrocyte membrane

Characterization of the Ala425Val-VCC protein in comparison with WT-VCC indicated that the mutant variant was defective in terms of its interaction with the human erythrocyte membrane. However, it must be emphasized that the Ala425Val-VCC protein displayed significant binding, and formed a considerable amount of SDS-stable oligomeric assembly in the erythrocyte membrane, particularly at a concentration of 65 μg·mL−1 (Fig. 2C); albeit that at this concentration the Ala425Val-VCC protein did not induce any considerable hemolytic activity against human erythrocytes. One possible explanation would be that the extent of the membrane-bound oligomers generated by the mutant protein might not be sufficient (due to the decreased membrane binding) to trigger lysis of the target cells. An alternative explanation would be that the SDS-stable oligomeric forms of the Ala425Val-VCC mutant may not represent the functional pore of the toxin. To explore these possibilities, we compared the membrane-bound oligomer load for Ala425Val-VCC and WT-VCC at two extreme concentrations: 65 μg·mL−1 for Ala425Val-VCC, i.e. the concentration, at which the mutant could not induce any lysis of human erythrocytes; and 6.5 μg·mL−1 for WT-VCC, which induced ~ 100% hemolyis. In this experimental setup, the membrane-bound oligomer load was found to be almost equivalent in the mutant and the wild-type toxin (Fig. 2C); even so, this amount of SDS-stable oligomer of the mutant in the erythrocyte membrane did not trigger considerable hemolysis. These data, therefore, suggest that the SDS-stable oligomeric form of the Ala425Val-VCC protein generated in the human erythrocyte cell membrane did not represent the functional pore assembly of the toxin; rather, such assembly appeared to remain trapped in the membrane as an abortive, nonfunctional oligomeric state.

Ala425Val mutation disrupted interaction of VCC with Asolectin–cholesterol liposome without affecting the basal level association of the toxin with Asolectin liposome lacking cholesterol

We compared the membrane permeabilization effects of the WT-VCC and Ala425Val-VCC proteins against the Asolectin–cholesterol liposome (Asolectin/cholesterol ratio of 1 : 1 w/w). We monitored the release of calcein from the liposome upon treatment with wild-type and mutant VCC proteins. Consistent with previous reports, WT-VCC triggered substantial calcein release from the Asolectin–cholesterol liposome. By contrast, the Ala425Val-VCC protein, at equivalent concentrations, had significantly reduced potential to induce calcein release (Fig. 4A). A nominal ~ 20% calcein release was induced by Ala425Val-VCC for the highest protein concentration tested (65 μg·mL−1). This result suggested that the Ala425Val mutation critically abrogated the membrane permeabilization effect of the VCC toxin against the simple liposome system.

image

Figure 4. Effect of Ala425Val mutation on the cholesterol-dependent interaction of the VCC toxin with the membrane lipid bilayer. (A) Comparison of the membrane permeabilization effects triggered by the Ala425Val-VCC and WT-VCC proteins on Asolectin–cholesterol liposomes. Liposomes (32.5 μg·mL−1) containing trapped calcein were incubated with the proteins, and membrane permeabilization was assayed by monitoring the % calcein release. Curve 1, 65 μg·mL−1 WT-VCC; curve 2, 6.5 μg·mL−1 WT-VCC; curve 3, 65 μg·mL−1 Ala425Val-VCC; curve 4, 6.5 μg·mL−1 Ala425Val-VCC. (B) Formation of SDS-stable oligomers by the Ala425Val-VCC and WT-VCC proteins upon incubation with Asolectin–cholesterol liposomes. Purified protein samples (65 μg·mL−1) were incubated with Asolectin–cholesterol liposomes (protein/lipid ratio of 1 : 2 w/w) in a 20-μL reaction volume for the specified time, dissolved in SDS/PAGE sample buffer for 5 min at 25 °C, and analyzed by SDS/PAGE/Coomassie staining. The SDS-stable oligomers and the monomeric forms of the proteins are indicated. Ala425Val-VCC is shown in lanes 1, 3 and 5; WT-VCC is shown in lanes 2, 4 and 6. (C) Binding of the Ala425Val-VCC (■) and WT-VCC (●) proteins with Asolectin–cholesterol liposome (left) and Asolectin liposome (right) monitored by ELISA-based binding assay. Error bars indicate the standard deviations determined from three measurements. (Insets) Liposome-binding efficacy of the two VCC variants as detected by SDS/PAGE and Coomassie staining. Briefly, proteins (65 μg·mL−1) were incubated with liposomes (protein/lipid ratio of 1 : 2 w/w) for 1 h at 25 °C, in a reaction volume of 100 μL, subjected to ultracentrifugation at 105 000 g, and the pellet fractions were resuspended in 100 μL NaCl/Pi; 15 μL samples were mixed with 5 μL of 6× SDS/PAGE sample buffer, incubated for 15 min at 100 °C, and liposome-bound proteins were visualized by SDS/PAGE/Coomassie staining. (Left inset) Binding to Asolectin–cholesterol liposome. (Right inset) Binding to Asolectin liposome. (D) Oligomerization efficacy of the membrane-bound fractions of the Ala425Val-VCC and WT-VCC proteins in Asolectin–cholesterol liposome (left) and Asolectin liposome (right). Proteins (65 μg·mL−1) were incubated with the liposomes (as described in C), subjected to ultracentrifugation at 105 000 g, total pellet fractions were dissolved in 30 μL SDS/PAGE sample buffer, and 15 μL of such samples were subjected to SDS/PAGE/Coomassie staining. Boiling in the presence of SDS/PAGE sample buffer led to dissociation of the oligomeric assembly into the monomeric form, thus showing only the total amount of the liposome-bound fractions of the proteins. Unboiled samples allowed detection of the SDS-stable oligomeric assembly of the VCC proteins formed in the liposome membranes. Band intensities were estimated using the methods described in Fig. 2B. Normalized band intensities are indicated at the bottom for the purpose of qualitative comparison. (E) FRET from tryptophan residues in the VCC proteins to dansyl-PE incorporated in the Asolectin–cholesterol liposomes. Protein (32.5 μg·mL−1) was incubated with the liposome (protein/lipid ratio of 1 : 1 w/w), and the kinetics of tryptophan-to-dansyl FRET was monitored. Curve 1, WT-VCC; Curve 2, Ala425Val-VCC.

Download figure to PowerPoint

Next, we compared interactions of the WT-VCC and Ala425Val-VCC variant with the cholesterol-containing Asolectin liposome membrane. We observed that a significantly reduced fraction of the Ala425Val-VCC mutant was converted into the SDS-stable oligomeric form compared with the WT-VCC toxin, when incubated in the presence of the Asolectin–cholesterol liposome (Fig. 4B). The extent of oligomer formation for Ala425Val-VCC was only ~ 25% that of the WT-VCC, as analyzed using the method described in Fig. 2. The reduced oligomer formation by the Ala425Val-VCC protein in the presence of Asolectin–cholesterol liposome could be because of either: (a) decreased binding of the mutant with the liposome membrane, or (b) its lesser oligomerization ability in the membrane lipid bilayer. In exploring these possibilities, we found that the Ala425Val-VCC protein showed a considerably decreased binding propensity toward the Asolectin–cholesterol liposome when compared with the WT-VCC toxin (Fig. 4C,D). An ELISA-based assay revealed a modest decrease in binding efficiency toward the Asolectin–cholesterol liposome for Ala425Val-VCC compared with WT-VCC over a range of protein concentrations (Fig. 4C). A more prominent effect was noticed when the liposome-binding efficacy was tested using a pull-down assay (Fig. 4C inset, D). In this assay, Ala425Val-VCC displayed ~ 75–80% reduced binding to Asolectin–cholesterol liposome compared with the wild-type protein. For example, comparison of the total membrane-bound fractions of the wild-type and mutant proteins (lane 4 versus lane 2 in Fig. 4D, left) revealed that Ala425Val-VCC interacted with the Asolectin–cholesterol liposome with an efficiency of only ~ 20% that of the WT-VCC protein. Even higher concentrations of cholesterol in the Asolectin–cholesterol liposome (Asolectin/cholesterol ratio of 1 : 5 w/w) could not rescue the effect of compromised binding efficacy induced by the Ala425Val mutation (data not shown). Notably, the liposome-bound fractions of both proteins showed almost equivalent propensities to convert into the oligomeric form (as evident from comparison of the oligomer and monomer band intensities in lanes 1 and 3 in Fig. 4D, left). Therefore, the Ala425Val mutation caused a significant defect in the VCC toxin in terms of its binding, not its oligomerization ability, in the Asolectin–cholesterol liposome membrane. In this context, it is important to note that the interactions with the Asolectin liposome lacking cholesterol were significantly less for both WT-VCC and Ala425Val-VCC (Fig. 4C,D). Moreover, binding to the Asolectin liposome was comparable for both the wild-type and mutant variants (Fig. 4C,D). These data, therefore, suggested that the Ala425Val mutation specifically abrogated the cholesterol-dependent mechanism of interaction of the VCC protein with the membrane lipid bilayer of the liposome, without affecting its basal, low-level association with liposome devoid of cholesterol.

To obtain more insight into the differential modes of interaction of the wild-type and mutant variants of the VCC molecules with the cholesterol-containing Asolectin liposome, we monitored the fluorescence resonance energy transfer (FRET) from the tryptophan residue(s) in the VCC proteins to the dansylated phosphatidylethanolamine (dansyl-PE) incorporated in the Asolectin–cholesterol liposomes. Because the FRET efficiency depends, among other factors, on the proximity and favorable orientation of the fluorescence energy donor and acceptor groups [30], the extent of increase in the FRET signal (from tryptophan residues to dansyl-PE) could be interpreted in terms of the intimacy of the interactions between the VCC variants and the membrane lipid bilayer [24, 31]. In our experimental setup, the WT-VCC toxin showed a steady increase in the dansyl FRET signal upon incubation with the dansyl-PE-containing Asolectin–cholesterol liposomes (Fig. 4E). By contrast, the mutant protein Ala425Val-VCC could not trigger any detectable increase in the dansyl FRET signal (Fig. 4E). This result suggested that the Ala425Val-VCC mutant, although capable of associating with the Asolecting–cholesterol liposome to some detectable extent, failed to make an efficient interaction with the membrane lipid bilayer that would be optimal to trigger an efficient tryptophan-to-dansyl FRET signal.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

In our study, we found that the single point mutation of Ala425Val in VCC critically abrogated the membrane pore-formation mechanism of the toxin. Mapping of the Ala425Val mutation onto the VCC structure revealed its location within a potential membrane-interacting loop (Tyr420–Tyr421–Val422–Val423–Gly424–Ala425; Ala425 being the last residue of this ‘longest loop’ in the VCC membrane-proximal rim domain) (Fig. 1) [11]. The mutation was found to be located in the loop in such a way that it would presumably make direct contact with the membrane, partly penetrating the outer leaflet of the membrane lipid bilayer (Fig. 1) [11]. The location of this mutation also aligned well with the potential membrane-binding region of the Staphylococcus aureus α-hemolysin [12].

We observed that the single point mutation of Ala425Val critically disrupted interaction of the VCC toxin with the membrane lipid bilayer of erythrocytes, as well as cholesterol-containing liposome. However, the mutant could still display wild-type-like basal, low-level association with the liposome membrane lacking cholesterol. Consistent with its location within a potential membrane-interaction site(s) in VCC, it is possible that the Ala425Val mutation might have perturbed the critical structural feature(s) in VCC implicated in its cholesterol-dependent membrane-interaction mechanism(s). Analysis of the structural models suggested that introduction of the Ala425Val mutation might have critically altered the distribution of the surface-exposed hydrophobic patch contouring the potential membrane-interacting loop in the VCC molecule. In the process of forming transmembrane oligomeric pore by the WT-VCC, this ‘potential membrane-interacting loop’ is predicted to penetrate the outer leaflet of the membrane lipid bilayer. Thus, an optimally balanced distribution of the surface-exposed hydrophobic patch onto such region would be critical to facilitate interaction of VCC with the membrane components. As indicated by the modeled structure of the mutant protein, the Ala425Val substitution might therefore be proposed to affect the efficient interaction of the toxin with the cholesterol-containing membrane lipid bilayer, presumably because of the altered hydrophobicity of the potential membrane-interacting loop region. It needs to be explored further whether the Ala425Val mutation directly abrogates the interaction mechanism of VCC with the membrane-anchored cholesterol molecules.

The inability of the Ala425Val-VCC mutant to confer FRET from the tryptophan residues(s) in VCC to the dansyl-PE in the Asolectin–cholesterol liposome also supported the idea that the Ala425Val mutation critically affected an intimate interaction of the protein with the cholesterol-containing membrane lipid bilayer. Considering the extent of exposure on the protein surface, and also the Trp-to-dansyl Forster distance (~ 25 Å) [30], it is possible to speculate that at least one Trp residue at position 362 would make direct contact with the dansyl group in the liposome membrane (Fig. S2). Because this Trp362 is positioned in the membrane-proximal rim domain of VCC, and remains at a site proximal to the Ala425Val mutation, it is possible that the Ala425Val mutation created as yet undefined structural constraint(s) that abrogated the participation of the Trp362 residue in the FRET process. However, such a qualitative interpretation should be considered with caution, and more detailed structural studies are required to validate such propositions. Our data confirmed that a critical interaction of the VCC molecule with the membrane lipid bilayer of the Asolectin–cholesterol liposome was abrogated because of the Ala425Val mutation.

Our data suggested that the Ala425Val mutation trapped the membrane-bound VCC molecules in an abortive nonfunctional oligomeric assembly. Upon interacting with the erythrocyte membrane, Ala425Val-VCC could form SDS-stable oligomers. A transmission electron microscopy-based study showed that such oligomers possessed typical ring-like architecture commonly observed with archetypical β-PFT molecules [32]. These data suggested that the mutant protein was not defective in terms of forming the oligomeric assembly in the erythrocyte membrane. However, formation of such an oligomeric assembly of the Ala425Val mutant in the erythrocyte membrane could not trigger any cytolytic activity. It is possible that the mutation of Ala425Val caused an, as yet undefined, structural alteration in VCC that hindered formation of the functional β-barrel pore, leading to the colloid-osmotic lysis of the target cells. Constriction of the inner diameter of the transmembrane pore might be one possible effect leading to the generation of a nonfunctional pore by the Ala425Val-VCC. Formation of membrane-bound oligomeric pores with a constricted pore diameter has been indicated previously for a truncated variant of the VCC protein lacking the C-terminal β-prism lectin-like domain [33]. As observed with Ala425Val-VCC, this truncated variant of VCC has been shown to form SDS-stable oligomers on the membrane, without triggering wild-type-like cytolytic activity against erythrocytes [13, 33]. Another probable mechanism could be that introduction of the Ala425Val mutation results in inefficient/abortive membrane insertion of the channel-forming ‘stem loop’. The membrane-bound SDS-stable oligomeric assembly formed by Ala425Val-VCC appears to be quite distinct from the SDS-labile ‘pre-pore’ assembly of the VCC toxin reported previously [34]. Pre-pore oligomeric assembly of β-PFTs, including that of VCC, has been speculated to be a transient metastable intermediate that could be trapped on the membrane surface only via artificial covalent cross-linking of the interacting monomeric subunits [34]. By contrast, oligomers of the Ala425Val mutant formed in the membrane exhibit more stable and robust assembly of the molecule that are resistant to the dissociating action of the denaturing detergent SDS. It is therefore evident that the Ala425Val mutation trapped the toxin in the membrane lipid bilayer in the form of an abortive nonfunctional oligomer that might represent a distinct intermediate. Consistent with the speculated scheme of events for the transmembrane β-barrel channel formation, i.e. (a) membrane binding, (b) oligomerization, (c) membrane insertion of the prestem loops, and (d) generation of the transmembrane β-barrel, we propose that the Ala425Val mutation could have introduced deleterious effect(s) in steps (c) and/or (d) of the above scheme. Because the Ala425Val mutation is located within the potential membrane-interacting loop, and not in the membrane-spanning stem region, we further propose that the mutation might have caused as yet unknown long-range conformational change(s) in the VCC molecule, abrogating formation of the functional β-barrel transmembrane pore. However, such propositions need to be validated via high-resolution structural studies.

In summary, the Ala425Val mutation severely affected the cholesterol-dependent mechanism of membrane interaction in the VCC protein, an alteration that also abolished formation of the functional pore by the membrane-bound toxin molecules. Earlier studies suggested that the presence of cholesterol in the membranes was a critical factor in optimal membrane binding, oligomerization, membrane insertion and pore formation by the VCC toxin [24-29]. Moreover, an optimal membrane pore-forming property of VCC was found to be tightly linked with its efficient interaction with the membrane [24-29]. Consistent with this, and based on the data presented here, we hypothesize that the Ala425Val mutation in the VCC toxin compromised its optimal cholesterol-dependent membrane interaction mechanism that, in turn, abrogated crucial mechanism(s) triggering the conformational fine-tuning in the VCC molecular structure, required for the generation of functional transmembrane pore assembly.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Wild-type and mutant form of VCC

Hemolytically active mature form of the VCC molecule (WT-VCC) was generated following a method described previously [35]. The purified form of the mutant variant of VCC (Ala425Val-VCC) was also generated using the method described for the WT-VCC protein.

Assay of hemolytic activity

Hemolytic activity of the wild-type and mutant proteins was assayed against human erythrocytes suspended in NaCl/Pi (20 mm sodium phosphate, 150 mm sodium chloride, pH 7.0) corresponding to D650 = 0.8–0.9. Hemolytic activity was assessed by measuring the release of hemoglobin at 415 nm [36].

Western blotting detection of VCC

The proteins dissolved in SDS/PAGE sample buffer (50 mm Tris/HCl, pH 6.8, containing 2% w/v SDS, 2 mm beta-mercaptoethanol, 4% v/v glycerol, 0.01% w/v bromophenol blue) were separated on SDS/PAGE, transferred to a poly(vinylidene difluoride) membrane in a complete wet transfer assembly (BioRad, Hercules, CA, USA) at 90 V for 90 min at 4 °C, the poly(vinylidene difluoride) membrane was blocked with 3% non-fat dry milk (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in NaCl/Pi containing 0.05% Tween 20 (TPBS) at 4 °C overnight. After washing for three times in TPBS, the blot was incubated with rabbit anti-VCC serum (1 : 5000 v/v) for 1 h at 25 °C, washed three times with TPBS, followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit IgG (1 : 10 000, v/v) for 1 h at 25 °C. After three washes with TPBS, the blot was developed using the ECL western blotting detection kit (GE Healthcare Life Sciences, Piscataway, NJ, USA), and exposed to X-ray film (Biomax, Kodak, Sigma-Aldrich, St. Louis, MO, USA) to detect the bands of VCC monomer and oligomer. Polyclonal anti-VCC serum was generated in rabbit using the Custom Polyclonal Antibody Service of GeNie/Merck, Bangalore, India.

Flow cytometry

An earlier study had shown that at low temperature (4 °C), the oligomerization and cytolytic activity of VCC are drastically abrogated; however, at such a low temperature binding of VCC to the erythrocyte membrane is not affected [37]. Therefore, we tested and compared the binding efficacy of the VCC variants to human erythrocytes at 4 °C using a flow cytometry-based assay following the methods described by Farrand et al. [38]. Briefly, human erythrocytes (1 × 106 cells) were treated with different concentrations of the VCC variants in a reaction volume of 100 μL for 30 min at 4 °C in NaCl/Pi. We found that at concentrations as high as 4.875 μg·mL−1, WT-VCC did not induce any lysis of the erythrocytes when incubated at 4 °C. Cells were pelleted, resuspended in 50 μL NaCl/Pi containing rabbit anti-VCC serum (1 : 100 v/v dilution and 0.1% w/v BSA; Sigma-Aldrich), and incubated for 30 min at 4 °C. Subsequently, cells were treated with fluorescein isothiocyanate (FITC)-conjugated goat anti-(rabbit IgG) (1 : 100 v/v dilution; Sigma-Aldrich) in 50 μL NaCl/Pi containing 0.1% BSA for 30 min at 4 °C. Cells were pelleted, washed twice with ice-cold NaCl/Pi, resuspended in 500 μL ice-cold NaCl/Pi and analyzed using FACSCalibur (BD Biosciences, San Jose, CA, USA) flow cytometer. FITC fluorescence was monitored using an excitation wavelength of 488 nm and emission wavelength of 530 nm. Geometric mean fluorescence (GMF) values were calculated using flowjo software (www.flowjo.com). The binding data were calculated using the equation:

  • display math

GMFcontrol = GMF for the cells that were not incubated with VCC variants, but treated with anti-VCC and anti-rabbit-FITC; GMFmaximum = GMF for the cells incubated with the highest concentration of WT-VCC used in the assay (4.875 μg·mL−1), followed by treatment with anti-VCC and anti-rabbit-FITC.

Preparation of liposomes

Cholesterol, Asolectin (from soybean) and dansyl-PE were from Sigma-Aldrich. For liposome preparation, appropriate amounts of lipids were first dissolved in chloroform in a round-bottomed flask. The solvent was evaporated at room temperature and the lipid film was dried under vacuum for 2 h, followed by resuspension in NaCl/Pi for 2 h at 37 °C. Large unilamellar vesicles were prepared by repeated extrusion of the liposome suspension through polycarbonate membranes of 0.1 μm pore size using a Mini-Extruder apparatus (Avanti Polar Lipids, Inc., Alabaster, AL, USA). Asolectin–cholesterol mixed liposome was prepared by maintaining a Asolectin/cholesterol ratio of 1 : 1 w/w, unless stated otherwise. Asolectin–cholesterol liposome containing dansyl-PE was prepared by incorporating 1% dansyl-PE (weight ratio) during preparation of the lipid film.

Asolectin–cholesterol liposome containing trapped calcein was prepared as described previously [27]. Briefly, a dried lipid film of Asolectin–cholesterol was resuspended in 20 mm Hepes buffer (pH 8.0), containing 150 mm sodium chloride and 50 mm calcein. After extrusion, free calcein molecules were removed by passing the liposome preparation through a size-exclusion chromatography column of Sephadex G50 (GE Healthcare Life Sciences, Piscataway, NJ, USA) equilibrated with 20 mm Hepes buffer (pH 8.0), containing 150 mm sodium chloride.

ELISA

An ELISA-based method was used to detect the binding of the VCC proteins with the liposome, as described previously [31, 36]. Briefly, wells of a 96-well microtiter plates (Nunc, Rochester, NY, USA) were coated in triplicate with 1 μg of liposome suspension in NaCl/Pi by incubating overnight at 4 °C. Plates were washed three times with TPBS, blocked by addition of 0.2 mL of 3% non-fat dry milk powder in NaCl/Pi for 1 h. Wild-type and mutant VCC proteins were incubated for 2 h at 25 °C to allow binding to the plate-bound liposome, washed three times with TPBS, and bound VCC proteins were detected by addition of rabbit anti-VCC serum (1 : 5000, v/v) for 90 min, followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit IgG (1 : 10 000, v/v) for 1 h at 25°C. VCC proteins bound to liposomes were detected by color development upon addition of o-phenylenediamine (1 mg·mL−1) in 0.1 m sodium citrate buffer (pH 4.5) containing H2O2 (2 μL·mL−1 of 30% v/v H2O2). The reactions were stopped by the addition of 1 m H2SO4, and absorbance was measured at 490 nm using a multiwell spectrometer (iMark; BioRad, USA).

Far-UV circular dichroism (CD)

Far-UV CD experiments were carried out with Chirascan spectropolarimeter (Applied Photophysics, Leatherhead, UK) equipped with a Peltier-based temperature controller, using 5 mm pathlength quartz cuvette as described before [35]. Protein concentrations were in the range 0.5–1 μm.

Fluorescence measurements

Intrinsic tryptophan fluorescence spectra were monitored using Fluoromax-4 (Horiba Scientific, Edison, NJ, USA) spectrofluorimeter equipped with a Peltier-based temperature controller, using a 1-cm cuvette. Tryptophan fluorescence was monitored upon excitation at 290 nm, with slit widths of 2.5 and 7.5 nm for excitation and emission, respectively. Protein concentrations were in the range 250–300 nm.

Calcein fluorescence measurements were taken on a Perkin–Elmer LS 55 spectrofluorimeter (Perkin-Elmer, Waltham, MA, USA) using a 1-cm cuvette at 25 °C. Calcein fluorescence was monitored at 520 nm upon excitation at 488 nm, with excitation and emission slit widths of 2.5 and 7.5 nm, respectively. Percentage calcein release was calculated from the increase in fluorescence intensity with respect to a control liposome sample without protein treatment. The 100% calcein release was determined from the fluorescence intensity corresponding to the sample treated with 6 mm sodium deoxycholate.

FRET from tryptophan residues in VCC to dansyl-PE incorporated in Asolectin–cholesterol liposome was monitored with a Perkin-Elmer LS 55 spectrofluorimeter using a 1-cm cuvette at 25 °C. Fluorescence was monitored at 512 nm upon excitation at 280 nm, with excitation and emission slit widths of 5 nm. Data were corrected using the control measurements taken with the dansyl-PE-containing Asolectin–cholesterol liposomes in absence of any protein treatment.

Transmission electron microscopy

Wild-type and mutant form of VCC proteins (65 μg·mL−1) were incubated with the human erythrocyte stroma (with an estimated protein content of 130 μg·mL−1) (prepared by hypotonic lysis of the human erythrocytes with 5 mm sodium phosphate buffer, pH 7.4) at room temperature, samples were subsequently washed with the same buffer, added onto the carbon-coated grid and negatively stained with 1% phosphotungstic acid. The samples were examined with Hitachi H7500 Transmission Electron Microscope at the Sophisticated Analytical Instrumentation Facility of Panjab University, Chandigarh, India.

Structural models

Structural coordinates were obtained from the Protein Data Bank (PDB) (entry 1XEZ for VCC precursor Pro-VCC; entry 3O44 for the VCC oligomer) [11, 13]. Structural coordinates of the monomeric form of the mature VCC molecule were generated from the PDB file 1XEZ using pdbset in the ccp4 suit [39]. A homology-based structural model of Ala425Val-VCC mutant was generated with the SWISS-MODEL server (http://swissmodel.expasy.org/) [40, 41] using the Pro-VCC structure (1XEZ) as template, and the modeled structure was subjected to energy minimization with the Chiron server (http://troll.med.unc.edu/chiron/index.php) [42]. Protein surface hydrophobic patches were estimated using the HOTPATCH server (http://hotpatch.mbi.ucla.edu/). All the structural models were visualized using pymol [43].

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

This work was supported by a grant from the Department of Biotechnology (DBT), India (DBT Grant No. BT/PR13350/BRB/10/751/2009). We also thank IISER Mohali for financial support. We thank Dr Arunika Mukhopadhaya of IISER Mohali for assistance with the flow cytometry. We acknowledge the Sophisticated Analytical Instrumentation Facility of Panjab University, Chandigarh, India for assistance with the transmission electron microscopy study. We also thank Mr Dinesh Sharma at the Sophisticated Analytical Instrumentation Facility of Panjab University for his assistance with the transmission electron microscopy study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
febs8809-sup-0001-FigS1-S2.zipZip archive201K

Fig. S1. Intrinsic tryptophan fluorescence emission and far-UV CD spectra of the wild-type and the mutant proteins.

Fig. S2. Location of the tryptophan residues in VCC potentially implicated in FRET to dansyl-PE in Asolectin–cholesterol liposome.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.