| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 14, 14315-14322, April 2, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶
From the
Department of Microbiology and Immunology, The University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma 73190, the Biota Structural Biology Laboratory, St Vincent's Institute of Medical Research, Fitzroy, Victoria 3065, Australia, ||The Protein Crystallography Unit, Department of Biochemistry and Molecular Biology, Monash University, School of Biomedical Sciences, Melbourne, Victoria 3168, Australia, and the Department of Biochemistry and Microbiology, University of Victoria, Victoria, British Columbia V8W 3P6, Canada
Received for publication, December 16, 2003 , and in revised form, January 9, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-hairpin of AT has been identified. Spectrofluorimetric analysis of cysteine-substituted residues modified with an environmentally sensitive fluorescent probe via the cysteine sulfydryl showed that the side chains of residues 203-232 alternated between the aqueous milieu and the membrane core when the AT oligomer was inserted into membranes, consistent with the formation of an amphipathic transmembrane -hairpin. AT derivatives that contained deletions that removed up to 90% of the -hairpin did not form a pore but were similar to native toxin in all other aspects of the mechanism. Furthermore, a mutant of AT that contained an engineered disulfide, predicted to restrict the movement of the -hairpin, functioned similarly to native toxin except that it did not form a pore unless the disulfide bond was reduced. Together these studies revealed the location and structure of the membrane-spanning domain of AT. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The mechanisms of AT and aerolysin have been shown to be highly similar. Both toxins bind to cells via glycophosphatidylinositol-anchored protein receptors (7, 8), although aerolysin appears to bind to some receptors that are not recognized by AT and vice versa. Following cell binding both toxins are activated by the proteolytic cleavage of an amino-terminal propeptide by furin or furin-like protease (9, 10). Activation then allows the toxin monomers to oligomerize on the membrane and form a pore (11, 12). The difference seen in receptor specificity of the two toxins appears to be linked to the presence of an amino-terminal peptide of aerolysin that is not conserved in AT. The crystal structure of aerolysin, the only member of this toxin family whose crystal structure has been solved (13), shows that it is a bi-lobal protein consisting of four distinct domains (D1-D4), three of which are rich in -sheet structure (see Fig. 1). The small lobe of aerolysin is missing from the amino terminus in AT, and this implies that AT is a single-lobed structure consisting of three domains that are homologous to D2-D4 of aerolysin (see Fig. 1). The small lobe of aerolysin (D1) contains a lectin-binding domain that enables it to bind to various receptors that are not recognized by AT (14, 15). However, fusion of D1 of aerolysin to the amino terminus of AT converts it to a molecule with aerolysin-like receptor specificity and activity (15). Comparatively little is known about the cytolytic mechanism of enterolobin, but its primary structure appears to be more related to aerolysin than AT. Enterolobin displays sequence similarity with both the small and large lobes of aerolysin and appears to form a dimer in solution (6, 16).
|
-strands and 
-helices. Both motifs have been identified in toxins as the secondary structures used to cross the membrane during pore formation. For example, the pore-forming domain of the colicins and the T-domain of diphtheria toxin use a pair of amphipathic -helices to span the membrane (17, 18), whereas the protective antigen component of anthrax toxin (19), the -hemolysin from Staphylococcus aureus (20, 21), and perfringolysin O from Clostridium perfringens (22, 23) all use amphipathic -strands that contribute to the formation of a transmembrane -barrel (see Fig. 1). However, identification of the TMD has not been necessarily easy or without controversy. The TMD of the cholesterol-dependent cytolysins (CDCs) was first suggested to be comprised of a -sandwich (24), or an amphipathic -helix (25) and Gilbert et al. (26) suggested that the CDCs did not penetrate the membrane at all. The pore-forming region of the CDCs was unambiguously shown to be comprised of two amphipathic -hairpins located in D3 of the perfringolysin O crystal structure that are derived from six short -helices of the soluble monomer (22-24).
It was previously hypothesized that the D4 -sandwich of aerolysin penetrates the membrane and forms the TMD of aerolysin (13). More recently, Tsitrin et al. (27) also suggested that the D4 -sandwich forms the TMD of aerolysin based on an electron density map of what was proposed to be a soluble oligomer of aerolysin. However, to date, no experimental data have been reported that directly demonstrate the interaction of a specific region of aerolysin, or the related AT or enterolobin, with the membrane.
The aerolysin-based structural model of AT and the aerolysin crystal structure exhibit a conspicuous amphipathic loop that is found in D2 of AT and the corresponding D3 of aerolysin (Fig. 1). Using multiple fluorescent and biochemical approaches, we have confirmed that this amphipathic loop in D2 of AT, comprised of residues Lys-203 to Gln-232, forms an amphipathic -hairpin that spans the membrane and is necessary for pore formation. Similar analyses of residues in the D3 -sandwich of AT, corresponding to the D4 region of aerolysin indicates that it is only peripherally associated with the membrane and is unlikely to participate directly in the formation of the transmembrane -barrel.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Generation of Deletion and Point Mutations—All deletion mutations in AT were generated using a four-primer site-directed PCR mutagenesis procedure previously described (29) except that pBRS10 was used as the template and Pfu turbo thermostable polymerase (Stratagene, La Jolla, CA) was substituted for Taq. PCR overlap products were purified using the Quantum-Prep PCR Kleen Spin columns (Bio-Rad, Hercules, CA), digested with NcoI and XhoI, and ligated into pET-22(b)+ digested with NcoI and XhoI.
An active, cysteine-less derivative of AT, termed ATC86A, was generated using the above mutagenesis procedure, and the resulting plasmid was designated pBRS20. All cysteine substitutions were produced using pBRS20 as template, using either the four-primer PCR mutagenesis procedure above or QuikChange site-directed mutagenesis (Stratagene).
Expression and Purification of AT—The growth and harvesting of Escherichia coli BLR-DE3 expressing polyhistidine-tagged native AT and the various AT derivatives were performed according to Sellman et al. (28). The cell pellets were resuspended in 150 ml of buffer A (10 mM MES; Research Organics, Cleveland, OH), 150 mM NaCl, pH 6.5. Lysis of cells was carried out in an EmulsiFlex-C5 high pressure homogenizer (Avestin, Ottawa, Ontario) at 15,000 p.s.i. Cell debris was removed by centrifugation at 21,000 x g for 10 min. Purification of the His-tagged AT from the supernatant was accomplished using a cobalt-chelating column, and cation-exchange column was done as previously described (28). AT-containing fractions eluted from the cation-exchange column were combined in a Micro-ProDiCon System (Spectrum, Gardena, CA) with a 10,000 MWCO Micro-ProDiCon membrane for simultaneous dialysis and concentration. Samples were dialyzed against 10 mM MES, 500 mM NaCl, 1 mM EDTA, pH 6.5 (buffer B) overnight at 4 °C. For cysteine-substituted proteins, 1 mM dithiothreitol was included in the dialysis buffer. 10% glycerol was added to the concentrated toxin before storage at -80 °C. Protein concentration was determined by absorbance at 280 nm using a molar extinction coefficient of 63,000 M-1 cm-1.2
SupT1 Membrane Preparation—SupT1 cells were cultured in RPMI 1640 growth medium supplemented with 20% fetal calf serum and 100 units of penicillin and streptomycin. Four roller bottles of cells were grown to 
2 x 106 cells/ml, the cells were harvested, pelleted at 250 x g for 10 min, and resuspended to a final volume of 50 ml in Hanks' balanced salt solution (HBSS) (BioWhittaker, Walkersville, MD). Membranes were prepared by lysing the cell suspension in an EmulsiFlex-C5 high pressure homogenizer set at 15,000 p.s.i. Cell membranes were collected by centrifugation at 30,000 x g for 20 min. Supernatant was removed, and the cell membranes were washed five times by resuspending the membrane pellet in 30 ml of buffer C (50 mM HEPES, 0.5 M NaCl, pH 8.0) and centrifuging at 30,000 x g for 20 min. Following the final wash, the membranes were resuspended with buffer C to an A600 of 10.
Construction of the AT Homology Model—The homology model of AT was based on the 2.3-Å resolution crystal structure of proaerolysin.3 The sequence alignment, which included the related toxin, enterolobin, and subsequent homology model, were generated using the aerolysin structure as a template within the HOMOLOGY module of Insight II (Accelrys Inc., San Diego, CA) on a Silicon Graphics Indigo 2 Maximum Impact Work station. The model was built in two stages: (i) identification of significant regions of sequence identity between AT and aerolysin and assigning coordinates to these regions and (ii) assigning coordinates to the homology model where the sequence identity to aerolysin is less convincing using a data base of peptide fragments. The models were visually checked to remove any steric clashes within the software package `O' (30), using the lego side chain and lego main chain options. This was followed by energy minimization (500 cycles of Powell minimization with tight harmonic restraints on the C atoms, followed by 500 cycles with no harmonic restraints) using X-PLOR (version 3.851) (31). The quality of the model was assessed using PROCHECK (Laskowski et al. (42)) and the 3D-1D environment plot (32). The similar molecular mechanism, the pairwise sequence identity, the good stereochemistry, and three-dimensional profiles of the AT model, are factors indicative of a good quality homology model.
Modification of Cysteine-substituted AT with NBD or IAF—All cysteine-substituted mutants of AT were stored in 1 mM dithiothreitol (DTT). Therefore, excess DTT was removed before labeling the unique cysteines of each AT mutant with the sulfhydryl-specific derivative of the fluorescent dye N,N'-dimethyl-N-(iodoacetyl-N'-(7-nitrobenz-2-oxa-1,3,-diazoyl)ethylenediamine (NBD) or 5-iodoacetamidofluoroscein (IAF) (Molecular Probes, Eugene, OR). Approximately 1 mg of toxin was passed through a gel filtration column (1.5-cm inside diameter x 20 cm) containing Sephadex G-50 equilibrated in buffer C at room temperature. NBD or IAF was then added to the protein to a 20-fold molar excess. The reaction was incubated in the dark for 1 h at room temperature and passed again over a Sephadex G-50 column equilibrated in buffer C to remove unbound dye. The extent of labeling was determined spectroscopically using an 
478 nm of 25,000-1 M-1 cm-1 for NBD or an 492 nm of 75,000-1 M-1 cm-1 for IAF (33). Protein concentration was determined by absorbance at 280 nm using a molar extinction coefficient of 63,000 M-1 cm-1.2
SupT1 Cell Viability Assay—SupT1 cells were cultured in RPMI 1640 growth medium supplemented with 20% fetal calf serum and 100 units of penicillin and streptomycin. SupT1 cells at 500,000 cells/ml were placed into the wells of a flat-bottom microtiter plate in 100-µl aliquots for a total number of cells per well of 50,000. Toxin was diluted in a separate round-bottom microtiter plate in which the wells were prefilled with 10 µl of HBSS. Toxin (1 µg) was placed into the first well of a single row and brought up to 20 µl of final volume with HBSS. The toxin was carried through 2-fold serial dilutions in the remaining wells of the row. Next, 10 µl from each well was added to respective wells of the flat-bottom plate containing the SupT1 cells and incubated for 4 h at 37 °C. Following incubation, 10 µl of Cell Counting Kit-8 reagent (Dojindo Molecular Technologies, Gaithersburg, MD) was added to each well and incubated for 4 h at 37 °C. The plates were read at A450 nm, and the concentration of toxin that resulted in 50% cell death (tissue culture lethal dose 50%) determined. Results from cells treated with mutant toxins were expressed as a percentage of the value obtained from cells treated with wild type AT.
Steady-state Fluorescence Spectroscopy—Steady-state fluorescence was measured using an SLM 8100 photon-counting spectrofluorometer with a double monochromator in the excitation light path, a single emission monochromator, cooled PMT housings, and a 450-watt xenon lamp (Spectronic Instruments, Rochester, NY), with the band-pass set at 4 nm. The emission wavelength was 540 nm, and the excitation wavelength was 480 nm. Emission scans were taken at 1-nm intervals between 500 and 600 nm for both monomeric and oligomeric forms of each mutant. For each NBD-labeled residue tested, the net NBD-intensity was determined by subtracting the fluorescence intensity of an equivalent NBD-free sample.
For each sample, toxin was pre-activated with trypsin at a 1:3500 ratio (w/w; trypsin:toxin) for 30 min at 37 °C. The reaction was stopped with the addition of a 30-fold molar excess of the protease inhibitor TLCK. 10 µg of pre-activated toxin was incubated alone (monomer) or in the presence of 50 µl of SupT1 membranes (oligomeric) and brought up to 400 µl of final volume with buffer C. All samples were incubated overnight at 37 °C in the dark to allow binding, oligomerization, and insertion to go to completion. For intensity measurements, reactions were placed into quartz cuvettes (1 x 1 cm) filled with 1.6 ml of buffer C at 37 °C and placed into the spectrofluorometer set at a constant temperature of 37 °C.
Quenching of NBD Emission by Membrane-restricted Spin-labeled Phospholipids—To determine if a particular NBD dye was present within the nonpolar core of the bilayer, we incorporated nitroxides into the SupT1 membrane suspensions as follows. 5-DOXYL-stearic acid and 16-DOXYL-stearic acid (DSA) (Sigma, St. Louis, MO) were suspended together to a final concentration (each) of 50 mg/ml in methanol. SupT1 membrane suspensions were made 4% in the DSA solution by directly adding the DSA mixture to the membranes. The membranes were vortexed immediately for 30 s and rocked at room temperature for 1 h.
For each experiment, 10 µg of pre-activated toxin was incubated with 50 µl of DSA-containing membranes in a total volume of 400 µl (brought up to volume with buffer C) overnight at 37 °C in the dark. The samples were placed into quartz cuvettes (1 x 1 cm) prefilled with 1.6 ml of buffer C at 37 °C and placed into the spectrofluorometer sample holder set at a constant temperature of 37 °C. Emission scans were taken at 1-nm intervals between 500 and 600 nm with an integration time of 1 s. Due to the fact that the DSA is dissolved in methanol, samples containing membranes made 4% in methanol alone were used as controls for net NBD intensity on membranes. Emission scans of NBD-free samples were taken and subtracted from the equivalent NBD-labeled samples to determine the net NBD intensity.
Saturable Binding of ATT224C-IAF and Kd Determination—The affinity of ATT224C-IAF for SupT1 cells was determined using a liquid phase binding assay. Increasing amounts of ATT224C-IAF (10-500 nM) were incubated with 1 x 106 SupT1 cells, brought up to 100-µl final volume in ice-cold HBSS and incubated for 30 min at 4 °C. Following incubation, cells were pelleted and washed once with 200 µl of ice-cold HBSS to remove unbound toxin and brought up to a final volume of 400 µl with ice-cold HBSS. The geometric mean fluorescence of the cells was determined by analysis on a FACSCalibur flow cytometer (Flow Cytometry and Confocal Microscopy Laboratory, Williams Medical Research Institute, Oklahoma University Health Sciences Center, Oklahoma City, OK) equipped with a 488-nm laser.
Each experiment was carried out in duplicate, and all data points represent the mean of four independent experiments. Nonspecific binding was determined by adding 25 µM unlabeled toxin to each amount of labeled toxin tested, and this value was subtracted from that obtained with labeled toxin alone to calculate specific binding. GraphPad Prism software, version 3.0, was used for nonlinear regression, curve-fitting analysis to determine an approximate Kd.
Competitive Binding Assay and Ki Determination—The ability of wild type toxin or deletion mutants to compete for binding with ATT224C-IAF to SupT1 cells was determined using a modified version of the above liquid phase binding assay. ATT224C-IAF at 208 nM was mixed with increasing concentrations of either unlabeled wild type or mutant toxin (0.5 nM to 10 µM) and brought up to a final volume of 100 µl in ice-cold HBSS before adding to 1 x 106 SupT1 cells. Cells incubated with ATT224C-IAF alone were considered maximum binding, and the mean fluorescence value was set at 100%.
Each ligand concentration was analyzed in duplicate. To determine the inhibition constants (Ki) for each competing toxin, the geometric mean fluorescence of the cells was plotted versus the concentration of competitor toxin. GraphPad Prism, version 3.0, was used for nonlinear regression, curve-fitting analysis for determination of Ki values for each competitor with a Kd for ATT224C-IAF set at 92 nM.
Activation and Oligomerization of TMD Deletion Mutants and ATS220C-S269C on SupT1 Membranes—Activation and oligomerization of toxin in the presence of membranes was carried out as follows. Wild type and mutant protoxin (ATPro) (5 µg) was activated using trypsin at a ratio of 1:1000 trypsin:toxin (w/w). The mixture was incubated at 37 °C for 30 min, and the trypsin was inhibited by the addition of a 30-fold molar excess of the protease inhibitor TLCK. SupT1 membranes (10 µl) were added to the activated toxin, and the volume was adjusted to 30 µl with buffer C and incubated at 37 °C for 2 h. SDS sample buffer (8 µl 6x) and 7 µl of 10% SDS were added to the samples, boiled at 90 °C for 2 min, and separated on a 4-15% gradient gel. The proteins were transferred to nitrocellulose paper, and the blot was incubated with affinity-purified anti-AT antibody. After 1 h, unbound primary antibody was removed by washing the blot three times in buffer D (10 mM Tris-HCl, 150 mM NaCl, 0.05% Tween 20, pH 8.0), and then secondary antibody, conjugated to horseradish peroxidase, was added to the blot, and the mixture was incubated for an additional 45 min. The blot was again washed three times in buffer D to remove unbound antibody. Colorimetric development of the bands recognized by the antibody was accomplished by developing the blot with the color development solution 4-chloro-1-naphthol according to manufacturer's instructions (Bio-Rad, Hercules, CA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-helices and -strands are defined by the periodicity of the interaction of their side chains with the membrane and aqueous environments; approximately every 3.5 residues of an amphipathic -helix interacts with the membrane (18), whereas alternating residues of an amphipathic -strand interact with the membrane (22). Residues Lys-203 to Gln-232 of AT exhibit characteristics of an amphipathic transmembrane -hairpin. The fluorescence-based method of Shepard et al. (22) was used to determine if this loop penetrates the membrane and then to examine the periodicity, and therefore its secondary structure, if it is inserted into the membrane. This technique takes advantage of the environmentally sensitive properties of the sulfhydryl-specific fluorescent dye iodoacetamide-NBD (NBD). The fluorescence intensity (FI) of NBD is quenched when it is in an aqueous environment, such as the channel of the pore, but increases when it is in a nonpolar environment, such as the core of the bilayer.
Residues Lys-203 to Gln-232 were individually mutated to cysteines, and the proteins were purified and labeled with NBD. The FI of each NBD-labeled mutant was examined in the absence and presence of SupT1 cell-derived membranes that contain one or more receptors for AT (7). Fig. 2 shows example emission scans from adjacent residues, Glu-227 and Phe-228, in the monomeric and membrane-bound forms. In an amphipathic -hairpin the side chain of Phe-228 is predicted to interact with the bilayer during pore formation, whereas the side chain of Glu-227 is predicted to reside in the hydrophilic channel of the pore. Consistent with this prediction, no significant increase in FI was seen for ATE227C-NBD following its incubation with membranes (Fig. 2A), whereas the FI of ATF228C-NBD increased 3-fold under the same conditions (Fig. 2B).
|
Residues of AT Domain 2 Span the Membrane—Although an increase in the FI of NBD indicates movement of the NBD into a nonpolar environment such as the bilayer, it does not necessarily confirm the location of the labeled residue in the bilayer. The membrane location of those residues shown to enter a nonpolar environment was confirmed by collisional quenching analysis. Nitroxides are efficient quenching agents of NBD as previously shown by Shepard et al. (22), and, when attached to the fatty acyl chain of stearic acid, nitroxides can be introduced into natural membranes and will only quench fluorophores exposed to the bilayer core (22, 23).
As seen in Fig. 3, the pattern of quenching by DOXYL-stearic acid (DSA) was consistent with the alternating periodicity previously seen for the FI in Fig. 2. The greatest level of quenching was seen between residues Gly-205 and Glu-231. Val-212, which did not show an increase in FI on membranes, was quenched significantly by the membrane-restricted nitroxide. This observation suggests that the Val-212 side chain initially resides in a nonpolar environment in the soluble monomer of AT, so the FI of the NBD-modified cysteine at this position does not change significantly as the side chain moves from its nonpolar location in the soluble monomer to its position in the membrane.
|
-strands may exist within the span of residues from 215 to 218, and, as a result, these residues may be only partially exposed to the membrane.
The cytolytic activity of each of the NBD-labeled mutants was tested on SupT1 cells. Of the 30 residues mutated and labeled in D2, 15 retained 25% or greater activity, 12 retained 5-25% activity, and 3 retained 
1% activity (data not shown). Those mutants that lost 99% or greater activity are indicated in Figs. 2 and 3, and each are predicted to face the membrane when inserted (Ile-204, Phe-210, and Phe-226). Due to the severe loss of activity by these mutants, they were not characterized further.
Deletion of Residues between Lys-203 and Gln-232 Abolishes Pore Formation—Based on the molecular model of AT (Fig. 1), it appears that deletion of the amphipathic loop would not significantly affect folding of AT, because it does not appear to be a core domain with extensive contacts. We therefore hypothesized that deletion of these residues would only disrupt pore formation if the residues are in the membrane-spanning region. Three mutants within this region were constructed, deleting increasing lengths of the loop region: AT
212-222, AT208-226, and AT204-230. The largest deletion (AT204-230) removed 90% of the predicted TMD. All deletion mutants were found to lack cytolytic activity on SupT1 cells (data not shown).
The three deletion mutants did not appear to be misfolded, because receptor binding, activation, and oligomerization were not affected. Analysis of the binding data showed that ATT224C-IAF bound to a single class of receptors these cells with a Kd of 92 nM (Fig. 4A). Competition experiments showed that the deletion mutants exhibited similar Ki values to that for wild type AT (Fig. 4B). Thus, the deletions did not significantly affect receptor binding. Furthermore, the deletion mutants were activated in vitro by trypsin similar to wild type and each oligomerized on membranes as well as or better than wild type AT, suggesting their overall structure had not been compromised by the loop deletions (Fig. 5A). In vitro activation by trypsin also did not yield peptides that were not found after activation of the native toxin in solution (data not shown). Hence, the general structure of these mutants was apparently not altered to an extent that resulted in the exposure of new trypsin sites.
|
|
-Sheet Backbone of D2 Prevents Pore Formation—Previous studies on the analogous loop region in D3 of aerolysin showed that it could be tethered to the backbone of the molecule by means of a disulfide bond between the loop and the backbone of the molecule (34). Once formed, the disulfide prevented oligomerization and pore formation until it was reduced with dithiothreitol (DTT). From these studies, it was concluded that the loop region must move away from the D3 -sheet in order for aerolysin to form a stable heptamer and generate pores in membranes. A similar experiment was performed with AT, in which residues analogous to those mutated to cysteine in aerolysin (based on the AT structural model in Fig. 1) were replaced with cysteines. Purified ATS220C-S269C was found to be inactive in its oxidized form, but exhibited >80% of the activity of native toxin when reduced (data not shown). When activated by trypsin and incubated with SupT1 membranes, ATS220C-S269C bound and oligomerized on SupT1 cells under both reducing and nonreducing conditions in a manner similar to native AT (Fig. 5B). Thus movement of the loop region in D2 of AT is only necessary for pore formation and not binding or oligomerization.
Domain 3 Residues of AT Do Not Appear to Span the Bilayer—The TMD of aerolysin was originally proposed to be comprised of the D4 -sandwich, which formed a hydrophobic dagger that penetrated the membrane (13). This idea has been reiterated in a recent study of an aerolysin mutant that forms a water-soluble heptamer, in which an aromatic belt was said to distinguish the upper boundary of a transmembrane region that was proposed to be present behind the propeptide (27). Because D4 of aerolysin is homologous to D3 of AT, we examined residues within this region of AT to determine if they interact with the membrane during pore formation.
Two hydrophobic residues in D4 of aerolysin, Leu-277 and Leu-196, were suggested to interact with the membrane when pores are formed. These residues border three aromatic residues on different -strands within D4 (Fig. 6A), which were proposed (27) to make up an aromatic belt as is found in porins (35). Leu-277 and Leu-196 reside on two different -strands in aerolysin (Fig. 6A) and are not conserved in AT. For Leu-277 of aerolysin, the equivalent residue in AT is a glutamic acid at position 246 (Fig. 6B), whereas for Leu-196, the homologous residue is a proline at position 162 (Fig. 6B). Because Pro-162 represented a more conservative change, we mutated it to a cysteine and labeled the mutant with NBD. When labeled with NBD and incubated in the presence of membranes, ATP162C-NBD displayed a slight increase in FI that was not quenched upon addition of DSA to the membranes (Fig. 6C), suggesting that it does not interact with the membrane during pore formation.
|
-strands extend between D2 and D3 of AT (Fig. 6B). Pro-162 lies on one of these -strands (-strand 1, Fig. 6B) near the predicted D2-D3 junction. Seven residues downstream of Pro-162 were examined on the same -strand, which spans the length of the D3 -sandwich of AT behind the propeptide, to determine if any of the residues within this region interact with the membrane. If D3 entered the membrane then these residues, which are on the same face as the predicted location of the propeptide, should interact with the membrane after toxin activation. These residues were analyzed in the same fashion as for those in Figs. 2 and 3. As seen in Fig. 6C, all of the residues tested on this -strand, except ATL163C-NBD, display at least a 2-fold increase in FI on membranes compared with the soluble monomer. When incubated with membranes containing the membrane restricted collisional quenching agent DSA, most of the residues exhibited little or no quenching indicating that these residues do not interact with the membrane to a significant extent.
As described above, Glu-246 of AT, corresponding to Leu-277 of aerolysin, was also predicted to interact with the membrane (27). This residue is on a different -strand than Pro-162 (-strand 2, Fig. 6B). We examined two consecutive residues downstream of Glu-246 to determine if this strand interacts with the membrane during pore formation. Using the same techniques as above, both residues showed an increase in FI upon incubation with SupT1 membranes (Fig. 6C), but displayed little or no quenching when DSA was present in the bilayer. Leu-264, which resides on -strand 3 in the middle of D3 of AT (Fig. 6B), lies directly behind the propeptide. Residues on this strand are predicted to be masked by the propeptide until its dissociation and are immediately below one of the residues of the putative aromatic belt. ATL264C-NBD showed a very slight increase in FI that was not quenched on membranes containing DSA (Fig. 6C). Therefore, the residues of the D4 -strands do not appear to have significant interactions with the membrane.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-hemolysin by Song et al. (20). The structure of the -hemolysin membrane pore showed how the amphipathic -hairpins of individual monomers of the toxin formed an amphipathic -barrel structure in the membrane. Since that time, only anthrax toxin (19) and perfringolysin O (22, 23) have been shown to utilize transmembrane -hairpins to form a -barrel. In the present study we have shown that residues 203-232 of AT also participate in the formation of a membrane-spanning -hairpin (Fig. 7). As with S. aureus -hemolysin, the hairpins of the individual monomers of the AT oligomer likely combine and insert into the membrane to generate the pore-forming -barrel. Consistent with this scenario is the alternating pattern in the FI of the NBD attached to cysteine-substituted residues; those residues that were predicted to face the membrane generally exhibited a greater increase in FI than the residues predicted to face the channel of the bilayer. Furthermore, collisional quenching analysis with a membrane-restricted quencher confirmed that the same residues identified by their increased FI were indeed interacting with the nonpolar environment of the membrane. The size of the AT hairpin, 30 residues, is also similar in length to that for the transmembrane hairpins of S. aureus -hemolysin (20), the anthrax protective antigen (19), and the two transmembrane hairpins of perfringolysin O (22, 23).
|
To insert into the membrane this loop apparently moves away from the D2 backbone -sheet. Rossjohn et al. (34) had previously shown that the introduction of a disulfide into aerolysin by the substitution of residues Thr-253 and Ala-300 with cysteines prevented pore formation. This disulfide locked the D3 loop region in aerolysin to the D3 -sheet. They found that this disulfide prevented oligomerization of aerolysin and proposed that this loop must move in order for aerolysin to become insertion competent. Upon reduction of the disulfide, the disulfide-trapped aerolysin regained its cytolytic activity. Based on the molecular model of AT, the analogous residues of AT were mutated to cysteines. As expected, the structurally analogous disulfide in AT prevented pore formation when oxidized. However, like the deletion mutants, it retained the ability to bind membranes, undergo proteolytic activation, and to form SDS-resistant oligomers. Also, like the aerolysin disulfide-trapped mutant, ATS220C-S269C regained nearly full cytolytic activity when the disulfide was reduced. Therefore, it also appears that this D2 transmembrane loop must move away from the backbone of AT to allow the -hairpin to insert into the membrane.
One important difference between the aerolysin and AT disulfide mutants should be noted. We found that ATS220C-S269C can oligomerize when the engineered disulfide is oxidized, whereas the disulfide mutant of aerolysin does not appear to oligomerize (34). This difference may reflect the fact that aerolysin exists as a head-to-tail dimer in solution (13), whereas AT exists as a monomer,4 and so the aerolysin dimer must dissociate prior to oligomerization. Aerolysin dimer formation is, in part, mediated via its small lobe (13, 37), which is missing in AT. When the small lobe is removed from aerolysin-like AT, it exists as a monomer in solution (37). Therefore, it appears that the movement of the D3 loop of aerolysin may also be necessary for the disruption of the aerolysin dimer after receptor binding (38).
Is the amphipathic structure of the AT TMD conserved in aerolysin and enterolobin? Both aerolysin and enterolobin maintain an amphipathic structure that is located in their primary structures and is positionally conserved with that of AT (Fig. 7). Although enterolobin contains an arginine in the middle of the predicted hairpin, this residue may reside near the predicted turn and face the aqueous milieu. However, it could also penetrate the membrane with the aliphatic part of the side chain and snorkel the charged amine to the surface, similar to that found in the perfringolysin O TMDs (22). Hence, it appears reasonable that, like AT, aerolysin and enterolobin may also use an amphipathic -hairpin structure to form their transmembrane pore.
It has been previously suggested (13, 27) that the D4 -sandwich of aerolysin forms a hydrophobic dagger that penetrates and forms the membrane pore. D3 and D4 of AT and aerolysin, respectively (Fig. 1), contain the propeptides of these two toxins, and the proteolytic cleavage of the propeptide has been shown to be essential for toxin oligomerization, presumably by exposing sites protected by the propeptide (12, 39, 40). Although we cannot rule out a superficial interaction of some residues within D3 of AT (D4 of aerolysin) with the membrane, the data herein do not support a role for D3 as the TMD of AT. Also, the fact that we can delete the D2 loop of AT and maintain its ability to bind and oligomerize without pore formation is also consistent with the previously established role of D3 in the oligomerization of AT (39) and D4 in the oligomerization of aerolysin (12, 13, 39, 40).
These studies provide the first experimental evidence for the location and structure of the transmembrane domain of a member of the aerolysin family of cytolytic toxins, C. septicum alpha toxin. Similar amphipathic regions also exist within the structures of the related aerolysin and enterolobin toxins suggesting that they may also utilize an amphipathic -hairpin to span the bilayer. Therefore, we propose that this family of toxins belongs to the -pore-forming class of toxins (41) that utilize an amphipathic -hairpin to form a membrane-spanning -barrel.
|
FOOTNOTES |
|---|
¶ A National Health Medical Research Council Senior Principal Research Fellow.
** A Wellcome Trust Senior Research Fellow.
To whom correspondence should be addressed: Dept. of Microbiology and Immunology, 940 Stanton L. Young Blvd, The University of Oklahoma Health Sciences Center, Oklahoma City, OK 73190. Tel.: 405-271-2133; Fax: 405-271-3117; E-mail: Rod-Tweten{at}ouhsc.edu.
1 The abbreviations used are: AT, alpha toxin; TMD, transmembrane domain; CDCs, cholesterol-dependent cytolysins; NBD, N,N'-dimethyl-N-(iodoacetyl-N'-(7-nitrobenz-2-oxa-1,3-diazoyl)ethylendiamine; IAF, 5-iodoacetamidofluoroscein; FI, fluorescence intensity; D1-D4, domains 1-4; DOXYL, 2-(3-carboxypropyl)-4,4-dimethyl-2-tridecyl-3-oxazolidinyloxy; DSA, DOXYL-stearic acid; TLCK, 1-chloro-3-tosylamido-7-amino-2-heptanone; MES, 4-morpholineethanesulfonic acid; HBSS, Hanks' balanced salt solution; DTT, dithiothreitol.
2 R. K. Tweten, unpublished data.
3 S. C. Feil, J. Rossjohn, J. T. Buckley, and M. W. Parker, unpublished results.
4 J. A. Melton and R. K. Tweten, unpublished data.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Osusky, L. Teschke, X. Wang, K. Wong, and J. T. Buckley A Chimera of Interleukin 2 and a Binding Variant of Aerolysin Is Selectively Toxic to Cells Displaying the Interleukin 2 Receptor J. Biol. Chem., January 18, 2008; 283(3): 1572 - 1579. [Abstract] [Full Text] [PDF]
|
|
|
|
