Cytotoxic activity of a tumor protease-activated pore-forming toxin.

Equinatoxin II is a pore forming toxin produced by the sea anemone Actinia equina. It is able to kill very unspecifically most cell types by the membrane-perturbing action of an amphiphilic alpha-helix located at its N-terminal. A normally active N-terminal mutant, containing one single cys in the amphiphilic alpha-helix, becomes totally inactive when it is bound to avidin via a biotinylated linker. By choosing, as a linker, a peptide containing a tumor protease cleavage site, we were able to construct an enzymatically activable conjugate which should be selective for tumor cells. The introduced cleavage site was designed in order to be digested by both cathepsin B and matrix metalloproteases (MMPs). We confirmed that this conjugate could be activated in vitro by cathepsin B and MMPs. After having measured the enzymatic activity of fibrosarcoma and breast carcinoma cells, we analyzed the cytotoxic effect of the conjugate on the same lines and on human red blood cells (HRBC) as controls. We found that the conjugate was activated, at least in part, by the tumor cell lines used, whereas it was inactive on HRBC. That the activation process was dependent on the enzymatic action of cathepsin B and MMPs, was indicated by three lines of evidence: (1) binding occurred normally on all type of cells including HRBC which however were insensitive being devoid of enzymes; (2) the cytotoxic effect correlated with the amount of cathepsin B activity expressed by the cells; (3) conjugate activation was reduced by specific inhibitors of cathepsin B and MMPs. These results demonstrate the possibility of tumor cell killing by a pore-forming toxin conjugate specifically activated by tumor proteases.

Ion channels and bacterial infection: the case of beta-barrel pore-forming protein toxins of Staphylococcus aureus.

Staphylococcus aureus strains causing human pathologies produce several toxins, including a pore-forming protein family formed by the single-component alpha-hemolysin and the bicomponent leukocidins and gamma-hemolysins. The last comprise two protein elements, S and F, that co-operatively form the active toxin. alpha-Hemolysin is always expressed by S. aureus strains, whereas bicomponent leukotoxins are more specifically involved in a few diseases. X-ray crystallography of the alpha-hemolysin pore has shown it is a mushroom-shaped, hollow heptamer, almost entirely consisting of beta-structure. Monomeric F subunits have a very similar core structure, except for the transmembrane stem domain which has to refold during pore formation. Large deletions in this domain abolished activity, whereas shorter deletions sometimes improved it, possibly by removing some of the interactions stabilizing the folded structure. Even before stem extension is completed, the formation of an oligomeric pre-pore can trigger Ca(2+)-mediated activation of some white cells, initiating an inflammatory response. Within the bicomponent toxins, gamma-hemolysins define three proteins (HlgA, HlgB, HlgC) that can generate two toxins: HlgA+HlgB and HlgC+HlgB. Like alpha-hemolysin they form pores in planar bilayers with similar conductance, but opposite selectivity (cation instead of anion) for the presence of negative charges in the ion pathway. gamma-Hemolysin pores seem to be organized as alpha-hemolysin, but should contain an even number of each component, alternating in a 1:1 stoichiometry.

The phytotoxic lipodepsipeptide syringopeptin 25A from Pseudomonas syringae pv syringae forms ion channels in sugar beet vacuoles.

Syringopeptin 25A (SP(25)A) belongs to a family of cyclic lipodepsipeptides (LDPs) produced by the gram-negative bacterium Pseudomonas syringae, a phytopathogenic organism that affects several plants of agronomic interest. LDPs increase the permeability of plasma and, possibly, intracellular membranes in plant cells. Consistently, SP(25)A forms ion channels in planar lipid bilayers and other model membranes. Here we used sugar beet tonoplasts as a new biological model system to study toxin action. When applied to the vacuoles by a fast perfusion procedure, SP(25)A increases membrane permeability by forming discrete ion channels even at low applied potentials. The SP(25)A channel displays anion selectivity (with a Cl-/K+ permeability ratio of 6.7 +/- 1.3) and has intrinsic rectification properties that derive from a different channel conductance at negative and positive voltages, presumably owing to an asymmetric distribution of fixed charges on the pore. Substitution of chloride with different anions reveals the following selectivity sequence NO3- approximately Cl-> F- > gluconate-, suggesting that the permeation pore is filled with water. The properties of the SP(25)A channels in vacuolar membranes are similar to those observed in planar lipid membranes prepared with asolectin. This work provides a direct demonstration of toxin effects on a native plant membrane, extending to a biological system previous results obtained on artificial planar lipid membranes.

Retrieving biological activity from LukF-PV mutants combined with different S components implies compatibility between the stem domains of these staphylococcal bicomponent leucotoxins.

Bicomponent leucotoxins, such as Panton-Valentine leucocidin, are composed of two classes of proteins, a class S protein such as LukS-PV, which bears the cell membrane binding function, and a class F protein such as LukF-PV, which interacts to form a bipartite hexameric pore. These leucotoxins induce cell activation, linked to a Ca(2+) influx, and pore formation as two consecutive and independently inhibitable events. Knowledge of the LukF-PV monomer structure has indicated that the stem domain is folded into three antiparallel beta-strands in the water-soluble form and has to refold into a transmembrane beta-hairpin during pore formation. To investigate the requirements for the cooperative assembly of the stems of the S and F components to produce biological activity, we introduced multiple deletions or single point mutations into the stem domains of LukF-PV and HlgB. While the binding of the mutated proteins was weakly dependent on these changes, Ca(2+) influx and pore formation were affected differently, confirming that they are independent events. Ca(2+) entry into human polymorphonuclear cells requires oligomerization and may follow the formation of a prepore. The activity of some of the LukF-PV mutants, carrying the shorter deletions, was actually improved. This demonstrated that a crucial event in the action of these toxins is the transition of the prefolded stem into the extended beta-hairpins and that this step may be facilitated by small deletions that remove some of the interactions stabilizing the folded structure.

Permeabilization of model lipid membranes by Bacillus sphaericus mosquitocidal binary toxin and its individual components.

The high larvicidal effect of Bacillus sphaericus (Bs), a mosquito control agent, originates from the presence of a binary toxin (Bs Bin) composed of two proteins (BinA and BinB) that work together to lyse gut cells of susceptible larvae. We demonstrate for the first time that the binary toxin and its individual components permeabilize receptor-free large unilamellar phospholipid vesicles (LUVs) and planar lipid bilayers (PLBs) by a mechanism of pore formation. Calcein-release experiments showed that LUV permeabilization was optimally achieved at alkaline pH and in the presence of acidic lipids. BinA was more efficient than BinB, BinB facilitated the BinA effect, and their stoichiometric mixture was more effective than the full Bin toxin. In PLBs, BinA formed voltage-dependent channels of approximately 100-200 pS with long open times and a high open probability. Larger channels (> or =400 pS) were also observed. BinB, which inserted less easily, formed smaller channels (< or =100 pS) with shorter mean open times. Channels observed after sequential addition of the two components, or formed by their 1:1 mixture (w/w), displayed BinA-like activity. Bs Bin toxin was less efficient at forming channels than the BinA/BinB mixture, with channels displaying the BinA channel behavior. Our data support the concept of BinA being principally responsible for pore formation in lipid membranes with BinB, the binding component of the toxin, playing a role in promoting channel activity.

Mode of action of beta-barrel pore-forming toxins of the staphylococcal alpha-hemolysin family.

Staphylococcal alpha-hemolysin is the prototype of a family of bacterial exotoxins with membrane-damaging function, which share sequence and structure homology. These toxins are secreted in a soluble form which finally converts into a transmembrane pore by assembling an oligomeric beta-barrel, with hydrophobic residues facing the lipids and hydrophilic residues facing the lumen of the channel. Besides alpha-hemolysin the family includes other single chain toxins forming homo-oligomers, e.g. beta-toxin of Clostridium perfringens, hemolysin II and cytotoxin K of Bacillus cereus, but also the staphylococcal bi-component toxins, like gamma-hemolysins and leucocidins, which are only active as the combination of two similar proteins which form hetero-oligomers. The molecular basis of membrane insertion has become clearer after the determination of the crystal structure of both the oligomeric pore and the soluble monomer. Studies on this family of beta-barrel pore-forming toxins are important for many aspects: (i) they are involved in serious pathologies of humans and farmed animals, (ii) they are a good model system to investigate protein-membrane interaction and (iii) they are the basic elements for the construction of nanopores with biotechnological applications in various fields.

Sizing the radius of the pore formed in erythrocytes and lipid vesicles by the toxin sticholysin I from the sea anemone Stichodactyla helianthus.

The radius of the pore formed by sticholysin I and II (StI, StII) in erythrocytes and sticholysin I in lipid vesicles was investigated. The rate of colloid osmotic lysis of human erythrocytes, exposed to one of the toxins in the presence of sugars of different size, was measured. The relative permeability of each sugar was derived and the pore radius estimated with the Renkin equation. The radius was similar for sticholysin I and II and was independent of the reference sugar chosen and of the toxin concentration applied. It was also the same when erythrocytes were pretreated with different toxin doses in the presence of a polyethylene glycol (PEG) large enough to prevent lysis and thereafter transferred to solutions containing oligosaccharides of different size where they did lyse at different rates. The osmometric behavior of large unilamellar vesicles (LUV) was thereafter used to estimate the toxin lesion radius in a model system. LUV transferred to a hyperosmotic solution with a certain sugar immediately shrank and then re-swelled at a rate dependent on the bilayer permeability to water and sugar. When LUV were previously permeabilized with StI, only a fraction of them, namely those not carrying pores, continued to behave as osmometers. By increasing the size of the added sugar and approaching the pore radius, the fraction of osmometric LUV increased. Relative permeabilities were derived and used to estimate a channel radius around 1.2 nm, both for sugars and for PEGs. In conclusion the sticholysin pore has a constant size independent of toxin concentration and similar in natural and artificial membranes, suggesting it has a fixed predominant structure.

Effects of lipid composition on membrane permeabilization by sticholysin I and II, two cytolysins of the sea anemone Stichodactyla helianthus.

Sticholysin I and II (St I and St II), two basic cytolysins purified from the Caribbean sea anemone Stichodactyla helianthus, efficiently permeabilize lipid vesicles by forming pores in their membranes. A general characteristic of these toxins is their preference for membranes containing sphingomyelin (SM). As a consequence, vesicles formed by equimolar mixtures of SM with phosphatidylcholine (PC) are very good targets for St I and II. To better characterize the lipid dependence of the cytolysin-membrane interaction, we have now evaluated the effect of including different lipids in the composition of the vesicles. We observed that at low doses of either St I or St II vesicles composed of SM and phosphatidic acid (PA) were permeabilized faster and to a higher extent than vesicles of PC and SM. As in the case of PC/SM mixtures, permeabilization was optimal when the molar ratio of PA/SM was ~1. The preference for membranes containing PA was confirmed by inhibition experiments in which the hemolytic activity of St I was diminished by pre-incubation with vesicles of different composition. The inclusion of even small proportions of PA into PC/SM LUVs led to a marked increase in calcein release caused by both St I and St II, reaching maximal effect at ~5 mol % of PA. Inclusion of other negatively charged lipids (phosphatidylserine (PS), phosphatidylglycerol (PG), phosphatidylinositol (PI), or cardiolipin (CL)), all at 5 mol %, also elicited an increase in calcein release, the potency being in the order CL approximately PA >> PG approximately PI approximately PS. However, some boosting effect was also obtained, including the zwitterionic lipid phosphatidylethanolamine (PE) or even, albeit to a lesser extent, the positively charged lipid stearylamine (SA). This indicated that the effect was not mediated by electrostatic interactions between the cytolysin and the negative surface of the vesicles. In fact, increasing the ionic strength of the medium had only a small inhibitory effect on the interaction, but this was actually larger with uncharged vesicles than with negatively charged vesicles. A study of the fluidity of the different vesicles, probed by the environment-sensitive fluorescent dye diphenylhexatriene (DPH), showed that toxin activity was also not correlated to the average membrane fluidity. It is suggested that the insertion of the toxin channel could imply the formation in the bilayer of a nonlamellar structure, a toroidal lipid pore. In this case, the presence of lipids favoring a nonlamellar phase, in particular PA and CL, strong inducers of negative curvature in the bilayer, could help in the formation of the pore. This possibility is confirmed by the fact that the formation of toxin pores strongly promotes the rate of transbilayer movement of lipid molecules, which indicates local disruption of the lamellar structure.

Genetic grafting of membrane-acting peptides to the cytotoxin dianthin augments its ability to de-stabilize lipid bilayers and enhances its cytotoxic potential as the component of transferrin-toxin conjugates.

Three chimeric proteins were obtained by fusing together the dianthin gene and DNA fragments encoding for the following membrane-acting peptides: the N-terminus of protein G of the vesicular stomatitis virus (KFT25), the N terminus of the HA2 hemagglutinin of influenza virus (pHA2), and a membrane-acting peptide (pJVE). Chimeric dianthins (KFT25DIA, pHA2DIA and pJVEDIA) retained full enzymatic activity in cell-free assays and showed increased ability to induce pH-dependent calcein release from large unilamellar vesicles (LUVs). pHA2DIA and pJVEDIA also showed faster kinetics of interaction with LUVs, while KFT25DIA and pHA2DIA displayed a reduced cytotoxicity as compared to wild-type dianthin. Conjugates made by chemically cross-linking KFT25DIA or pJVEDIA and human transferrin (Tfn) showed greater cell-killing efficiency than conjugates of Tfn and wild-type dianthin. As a consequence, by fusion of membrane-acting peptides to the dianthin sequence the specificity factor (i.e., the ratio between non-specific and specific toxicity) of Tfn-KFT25DIA, Tfn-pHA2DIA and Tfn-pJVEDIA was increased with respect to that of Tfn-based conjugates made with wild-type dianthin. Taken together, our results suggest that genetic fusion of membrane-acting peptides to enzymatic cytotoxins results in the acquisition of new physico-chemical properties exploitable for designing new recombinant cytotoxins and to tackle cell-intoxication mechanisms.

Anionic phospholipids are involved in membrane association of FtsY and stimulate its GTPase activity.

FtsY, the Escherichia coli homologue of the eukaryotic signal recognition particle (SRP) receptor alpha-subunit, is located in both the cytoplasm and inner membrane. It has been proposed that FtsY has a direct targeting function, but the mechanism of its association with the membrane is unclear. FtsY is composed of two hydrophilic domains: a highly charged N-terminal domain (the A-domain) and a C-terminal GTP-binding domain (the NG-domain). FtsY does not contain any hydrophobic sequence that might explain its affinity for the inner membrane, and a membrane-anchoring protein has not been detected. In this study, we provide evidence that FtsY interacts directly with E.coli phospholipids, with a preference for anionic phospholipids. The interaction involves at least two lipid-binding sites, one of which is present in the NG-domain. Lipid association induced a conformational change in FtsY and greatly enhanced its GTPase activity. We propose that lipid binding of FtsY is important for the regulation of SRP-mediated protein targeting.

Structure-function studies of tryptophan mutants of equinatoxin II, a sea anemone pore-forming protein.

Equinatoxin II (EqtII) is a eukaryotic cytolytic toxin that avidly creates pores in natural and model lipid membranes. It contains five tryptophan residues in three different regions of the molecule. In order to study its interaction with the lipid membranes, three tryptophan mutants, EqtII Trp(45), EqtII Trp(116/117) and EqtII Trp(149), were prepared in an Escherichia coli expression system [here, the tryptophan mutants are classified according to the position of the remaining tryptophan residue(s) in each mutated protein]. They all possess a single intrinsic fluorescent centre. All mutants were less haemolytically active than the wild-type, although the mechanism of erythrocyte damage was the same. EqtII Trp(116/117) resembles the wild-type in terms of its secondary structure content, as determined from Fourier-transform infrared (FTIR) spectra and its fluorescent properties. Tryptophans at these two positions are buried within the hydrophobic interior of the protein, and are transferred to the lipid phase during the interaction with the lipid membrane. The secondary structure of the other two mutants, EqtII Trp(45) and EqtII Trp(149), was altered to a certain extent. EqtII Trp(149) was the most dissimilar from the wild-type, displaying a higher content of random-coil structure. It also retained the lowest number of nitrogen-bound protons after exchange with (2)H(2)O, which might indicate a reduced compactness of the molecule. Tryptophans in EqtII Trp(45) and EqtII Trp(149) were more exposed to water, and also remained as such in the membrane-bound form.

Lysine 77 is a key residue in aggregation of equinatoxin II, a pore-forming toxin from sea anemone Actinia equina.

Among eighteen point mutants of equinatoxin II produced in E. coli, containing a single cystein substitution at variable position, EqtIIK77C was chosen for its peculiar properties. It was almost 100 times less hemolytic than the wild-type, but its hemolytic activity could be restored by chemical modification of the thiol group, provided that a positive charge was reintroduced. This indicates that a positive charge at this position is necessary for toxin activity. The mutant formed larger pores as compared to the wild type, but displayed the same cation selectivity. The pores reverted to normal size upon reintroduction of the positive charge. The conformation of EqtIIK77C and its binding to lipid membranes, either vesicles or red blood cells, was almost normal. However the kinetics of calcein release from lipid vesicles was substantially slower than that of the wild-type. Taken together with the different size of the pore formed, this is an indication that mutation of Lys77 --> Cys influences the normal development of the aggregate which is required for assembling the functional pore.

Effects of calcium and protons on the secondary structure of the nodulation protein NodO from Rhizobium leguminosarum biovar viciae.

NodO, a 30-kDa nodulation protein secreted by Rhizobium leguminosarum biovar viciae, belongs to a family of proteins produced by Gram-negative bacteria containing a variable number of glycine/aspartates nonapeptides. In some instances, these are organized into a parallel beta-roll structure and bind Ca(2+) (one ion per repeat). To gain insight into NodO's secondary and tertiary structures, and their dependence upon Ca(2+) binding, we performed fluorescence experiments and FTIR spectroscopy. We found that calcium binds to the protein, promoting about a 10% increase in beta-structure mainly to the expense of random-coil. Protons can also induce a reversible change in NodO structure, as indicated by quenching of intrinsic tryptophan fluorescence and binding of ANS, albeit probably via a different mechanism. Tb(3+), a trivalent lanthanide, can compete with Ca(2+) for the same binding sites, but with higher affinity. The number of Ca(2+) binding sites, estimated by FTIR spectroscopy, was found to be consistent with the number of predicted repeats.

Cysteine-scanning mutagenesis of an eukaryotic pore-forming toxin from sea anemone: topology in lipid membranes.

Equinatoxin II is a cysteineless pore-forming protein from the sea anemone Actinia equina. It readily creates pores in membranes containing sphingomyelin. Its topology when bound in lipid membranes has been studied using cysteine-scanning mutagenesis. At approximately every tenth residue, a cysteine was introduced. Nineteen single cysteine mutants were produced in Escherichia coli and purified. The accessibility of the thiol groups in lipid-embedded cysteine mutants was studied by reaction with biotin maleimide. Most of the mutants were modified, except those with cysteines at positions 105 and 114. Mutants R144C and S160C were modified only at high concentrations of the probe. Similar results were obtained if membrane-bound biotinylated mutants were tested for avidin binding, but in this case three more mutants gave a negative result: S1C, S13C and K43C. Furthermore, mutants S1C, S13C, K20C, K43C and S95C reacted with biotin only after insertion into the lipid, suggesting that they were involved in major conformational changes occurring upon membrane binding. These results were further confirmed by labeling the mutants with acrylodan, a polarity-sensitive fluorescent probe. When labeled mutants were combined with vesicles, the following mutants exhibited blue-shifts, indicating the transfer of acrylodan into a hydrophobic environment: S13C, K20C, S105C, S114C, R120C, R144C and S160C. The overall results suggest that at least two regions are embedded within the lipid membrane: the N-terminal 13-20 region, probably forming an amphiphilic helix, and the tryptophan-rich 105-120 region. Arg144, Ser160 and residues nearby could be involved in making contacts with lipid headgroups. The association with the membrane appears to be unique and different from that of bacterial pore-forming proteins and therefore equinatoxin II may serve as a model for eukaryotic channel-forming toxins.

The interaction of lipodepsipeptide toxins from Pseudomonas syringae pv. syringae with biological and model membranes: a comparison of syringotoxin, syringomycin, and two syringopeptins.

Pseudomonas syringae pv. syringae produces two groups of cyclic lipodepsipeptides (LDPs): the nona-peptides syringomycins, syringostatins, and syringotoxin (ST), and the more complex syringopeptins composed of either 22 or 25 amino acid residues (SP22 and SP25). Both classes of peptides significantly contribute to bacterial pathogenesis and their primary target of action seems to be the plasma membrane. We studied and compared the activity of some members of these two classes of LDPs on red blood cells and on model membranes (monolayers and unilamellar vesicles). All peptides induced red blood cell hemolysis. The mechanism was apparently that of a colloid-osmotic shock caused by the formation of pores, as it could be prevented by osmoticants of adequate size. Application of the Renkin equation indicated a radius of approximately 1 nm for the lesions formed by syringopeptins SP22A and SP25A, whereas those formed by syringomycin E (SRE) had a variable, dose-dependent size ranging from 0.7 up to 1.7 nm. All tested LDPs displayed surface activity, forming peptide monolayers with average molecular areas of 1.2 nm2 (SRE), 1.5 nm2 (SP22A), and 1.3 nm2 (SP25A). They also partitioned into preformed lipid monolayers occupying molecular areas that ranged from 0.6 to 1.7 nm2 depending on the peptide and the lipid composition of the film. These LDPs formed channels in lipid vesicles as indicated by the release of an entrapped fluorescent dye (calcein). The extent of permeabilization was dependent on the concentration of the peptide and the composition of the lipid vesicles, with a preference for those containing a sterol. From the dose dependence of the permeabilization it was inferred that LDPs increased membrane permeability by forming oligomeric channels containing from four to seven monomers. On average, syringopeptin oligomers were smaller than SRE and ST oligomers.

Conductive properties and gating of channels formed by syringopeptin 25A, a bioactive lipodepsipeptide from Pseudomonas syringae pv. syringae, in planar lipid membranes.

Syringopeptin 25A, a pseudomonad lipodepsipeptide, can form ion channels in planar lipid membranes. Pore conductance is around 40 pS in 0.1 M NaCl. Channel opening is strongly voltage dependent and requires a negative potential on the same side of the membrane where the toxin was added. These pores open and close with a lifetime of several seconds. At negative voltages, an additional pore state of around 10 pS and a lifetime of around 30 ms is also present. The voltage dependence of the rates of opening and closing of the stable pores is exponential. This allows estimation of the equivalent charge that is moved across the membrane during the process of opening at about 2.6 elementary charges. When NaCl is present, the pore is roughly 3 times more permeant for anions than for cations. The current voltage characteristic of the pore is nonlinear, i.e., pore conductance is larger at negative than at positive voltages. The maximal conductance of the pore depends on the concentration of the salt present, in a way that varies almost linearly with the conductivity of the solution. From this, an estimate of a minimal pore radius of 0.4 nm was derived.

Antiparasite activity of sea-anemone cytolysins on Giardia duodenalis and specific targeting with anti-Giardia antibodies.

The killing activity of sea-anemone cytolysins on Giardia duodenalis was investigated. Three different toxins, sticholysin I and II from Stichodactyla helianthus (St I and St II) and equinatoxin II from Actinia equina (EqtII) were all found to be active in an acute test, with a C50 in the nanomolar range (St I, 0.5 nM; St II, 1.6 nM; and EqtII, 0.8 nM). A method to target the cytolysin activity more specifically towards the parasite cells by using anti-Giardia antibodies was then investigated. Parasite cells were sensitised with a primary murine monoclonal or polyclonal antibody followed by a biotinylated secondary anti-mouse-IgG monoclonal antibody. Subsequently, avidin and a biotinylated EqtII mutant were added, either in two separate steps or as a pre-formed conjugate. When the monoclonal antibody was used, the C50 of biotinylated EqtII was 1.3 nM with sensitised cells and 5 nM with non-sensitised cells, indicating a four-fold enhancement of activity with the cell treatment. Treatment with the polyclonal antibody was somehow more effective than with the monoclonal antibody in an acute test. This indicates that sea-anemone cytolysins can efficiently kill Giardia cells, and that it is possible to improve, to a certain extent, the anti-parasite specificity of these toxins with anti-Giardia antibodies. However, the feasibility of this approach "in vivo" remains to be demonstrated.