Diphtheria Toxin Forms Pores of Different Sizes Depending on Its Concentration in Membranes: Probable Relationship to Oligomerization

Diphtheria toxin forms pores in biological and model membranes upon exposure to low pH. These pores may play a critical role in the translocation of the A chain of the toxin into the cytoplasm. The effect of protein concentration on diphtheria toxin pore formation in model membrane systems was assayed by using a new fluorescence quenching method. In this method, the movement of Cascade Blue labeled dextrans of various sizes across membranes is detected by antibodies which quench Cascade Blue fluorescence. It was found that at low pH the toxin makes pores in phosphatidylcholine/phosphatidylglycerol vesicles with a size that depends on protein concentration. At the lowest toxin concentrations only the entrapped free fluorophore (MW 538) could be released from model membranes. At intermediate toxin concentrations, a 3 kD dextran could be released. At the highest toxin concentration, a 10 kD dextran could be released, but not a 70 kD dextran. Similar pore properties were found using vesicles lacking phosphatidylglycerol or containing 30% cholesterol. However, larger pores formed at lower protein concentrations in the presence of cholesterol. The dependence of pore size on toxin concentration suggests that toxin oligomerization regulates pore size. This behavior may explain some of the conflicting data on the size of the pores formed by diphtheria toxin. The formation of oligomers by membrane-inserted toxin is consistent with the results of chemical crosslinking and measurements of the self-quenching of rhodamine-labeled toxin. Based on these experiments we propose diphtheria toxin forms oligomers with a variable stoichiometry, and that pore size depends on the oligomerization state. Reasons why oligomerization could assist proper membrane insertion of the toxin and other proteins that convert from soluble to membrane-inserted states are discussed. Received: 10 March 1999/Revised: 22 June 1999     

Electrokinetic Properties of the Sarcoplasmic Reticulum Membrane Obtained from Reconstitution Studies

Electrophoretic mobility data of SR vesicles reconstituted with uncharged and two mixtures of charged and uncharged lipids (Brethes, D., Dulon, D., Johannin, G., Arrio, B., Gulik-Krzywicki, T., Chevallier, J. 1986. Study of the electrokinetic properties of reconstituted sarcoplasmic reticulum vesicles. Arch. Biochem. Biophys. 246:355–356) were analyzed in terms of four models of the membrane-water interface: (I) a smooth, negatively charged surface; (II) a negatively charged surface of lipid bilayer covered with an electrically neutral surface frictional layer; (III) an electrically neutral lipid bilayer covered with a neutral frictional layer containing a sheet of negative charge at some distance above the surface of the bilayer; (IV) an electrically neutral lipid bilayer covered with a homogeneously charged frictional layer. The electrophoretic mobility was predicted from the numerical integration of Poisson-Boltzmann and Navier-Stokes equations. Experimental results were consistent only with predictions based on Model-III with charged sheet about 4 nm above the bilayer and frictional layer about 10 nm thick. Assuming that the charge of the SR membrane is solely due to that on Ca⁺⁺-ATPase pumps, the dominant SR protein, the mobility data of SR and reconstituted SR vesicles are consistent with 12 electron charges/ATPase. This value compares well to the net charge of the cytoplasmic portion of ATPase estimated from the amino acid sequence (-11e). The position of the charged sheet suggests that the charge on the ATPase is concentrated in the middle of the cytoplasmic portion. The frictional layer of SR can be also assigned to the cytoplasmic portion of Ca⁺⁺-ATPase. The layer has been characterized with hydrodynamic shielding length of 1.1 nm. Its thickness is comparable to the height of the cytoplasmic portion of Ca⁺⁺-ATPase. Received: 15 June 1998/Revised: 8 October 1998     

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 ≈100–200 pS with long open times and a high open probability. Larger channels (≥400 pS) were also observed. BinB, which inserted less easily, formed smaller channels (≤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. Received: 29 March 2001/Revised: 20 July 2001     

Permeability Increase Induced by Escherichia coli Hemolysin A in Human Macrophages is Due to the Formation of Ionic Pores: A Patch Clamp Characterization

Escherichia coli hemolysin is known to cause hemolysis of red blood cells by forming hydrophilic pores in their cell membrane. Hemolysin-induced pores have been directly visualized in model systems such as planar lipid membranes and unilamellar vesicles. However this hemolysin, like all the members of a related family of toxins called Repeat Toxins, is a potent leukotoxin. To investigate whether the formation of channels is involved also in its leukotoxic activity, we used patch-clamped human macrophages as targets. Indeed, when exposed to the hemolysin, these cells developed additional pores into their membrane. Such exogenous pores had properties very different from the endogenous channels already present in the cell membrane (primarily K⁺ channels), but very similar to the pores formed by the toxin in purely lipidic model membranes. Observed properties were: large single channel conductance, cation over anion selectivity but weak discrimination among different cations, quasilinear current-voltage characteristic and the existence of a flickering pre-open state of small conductance. The selectivity properties of the toxin channels appearing in phospholipid vesicles were also investigated, using a specially adapted polarization/depolarization assay, and were found to be completely consistent with that of the current fluctuations observed in excised macrophage patches. Received: 14 August 1995/Revised: 2 October 1995     

Influence of Cys-130 S. aureus Alpha-toxin on Planar Lipid Bilayer and Erythrocyte Membranes

Replacement of an amino acid residue at position 130 -Gly by Cys- in the primary structure of Staphylococcus aureus alpha-toxin decreases the single-channel conductance induced by the toxin in planar lipid bilayers. Concomitantly, the pH value at which the channel becomes unable to discriminate between Cl⁻ and K⁺ ions is also decreased. By contrast, the pH dependence of the efficiency of the mutant toxin to form ion channels in lipid bilayers was unchanged (maximum efficiency at pH 5.5–6.0). The asymmetry and nonlinearity of the current-voltage characteristics of the channel were increased by the point mutation but the diameter of the water pore induced by the mutant toxin, evaluated in lipid bilayers and in erythrocyte membranes, was found to be indistinguishable from that formed by wild-type toxin and equal to 2.4–2.6 nm. Alterations at the ``trans mouth' were found to be responsible for all observed changes of the channel properties. This mouth is situated close to the surface of the second leaflet of a bilayer lipid membrane. The data obtained allows us to propose that the region around residue 130 in fact determines the main features of the ST-channel and takes part in the formation of the trans entrance of the channel. Received: 8 September 1995/Revised: 20 November 1996     

The Effects of Inhibitors Upon Pore Formation by Diphtheria Toxin and Diphtheria Toxin T Domain

The formation of pores by membrane-inserted diphtheria toxin is closely linked to the translocation of its catalytic chain across membranes. In this report a number of aromatic polyanionic molecules were identified that inhibit toxin-induced leakage of molecules from model membrane vesicles. One inhibitor, Cibacron blue, totally blocked pore formation. Aniline blue and Fast Green decreased the size of the molecule released by a given concentration of toxin. Amaranth appeared to reduce the maximal amount of leakage, without greatly affecting the size of the molecule released at a given toxin concentration. Finally, Ponceau S and Cibacron brilliant red appeared to exhibit a mixture of these various types of inhibition. The inhibitors neither prevented the conformational transition of the toxin to form a hydrophobic state at low pH, nor (with the exception of Cibacron Brilliant Red) appeared to strongly inhibit toxin binding to model membranes. Additional experiments showed release of trapped materials from model membranes by isolated T domain of the toxin was similar to that by whole toxin. The effects of inhibitors on T domain induced release was also similar to that they have on whole toxin. Therefore, it is likely that the inhibition of pore formation by whole toxin involves inhibitor interaction with the T domain. The inhibitors identified in this study may be helpful for development of agents that interfere with toxin action in vivo. Received: 10 March 1999/Revised: 22 June 1999     

The Role of Membrane Lateral Tension in Calcium-Induced Membrane Fusion

Calcium-induced fusion of liposomes was studied with a view to understand the role of membrane tension in this process. Lipid mixing due to fusion was monitored by following fluorescence of rhodamine-phosphatidyl-ethanolamine incorporated into liposomal membrane at a self-quenching concentration. The extent of lipid mixing was found to depend on the rate of calcium addition: at slow rates it was significantly lower than when calcium was injected instantly. The vesicle inner volume was then made accessible to external calcium by adding calcium ionophore A23187. No effect on fusion was observed at high rates of calcium addition while at slow rates lipid mixing was eliminated. Fusion of labeled vesicles with a planar phospholipid membrane (BLM) was studied using fluorescence microscopy. Above a threshold concentration specific for each ion, Ca²⁺, Mg²⁺, Cd²⁺ and La³⁺ induce fusion of both charged and neutral membranes. The threshold calcium concentration required for fusion was found to be dependent on the vesicle charge, but not on the BLM charge. Pretreatment of vesicles with ionophore and calcium inhibited vesicle fusion with BLM. This effect was reversible: chelation of calcium prior to the application of vesicle to BLM completely restored their ability to fuse. These results support the hypothesis that tension in the outer monolayer of lipid vesicle is a primary reason for membrane destabilization promoting membrane fusion. How this may be a common mechanism for both purely lipidic and protein-mediated membrane fusion is discussed. Received: 27 September 1999/Revised: 22 March 2000     

The Role of MIP in Lens Fiber Cell Membrane Transport

MIP has been hypothesized to be a gap junction protein, a membrane ion channel, a membrane water channel and a facilitator of glycerol transport and metabolism. These possible roles have been indirectly suggested by the localization of MIP in lens gap junctional plaques and the properties of MIP when reconstituted into artificial membranes or exogenously expressed in oocytes. We have examined lens fiber cells to see if these functions are present and whether they are affected by a mutation of MIP found in Cat ^Fr mouse lens. Of these five hypothesized functions, only one, the role of water channel, appears to be true of fiber cells in situ. Based on the rate of volume change of vesicles placed in a hypertonic solution, fiber cell membrane lipids have a low water permeability (p _H2O ) on the order of 1 μm/sec whereas normal fiber cell membrane p _H2O was 17 μm/sec frog, 32 μm/sec rabbit and 43 μm/sec mouse. Cat ^Fr mouse lens fiber cell p _H2O was reduced by 13 μm/sec for heterozygous and 30 μm/sec for homozygous mutants when compared to wild type. Lastly, when expressed in oocytes, the p _H2O conferred by MIP is not sensitive to Hg²⁺ whereas that of CHIP28 (AQP1) is blocked by Hg²⁺. The fiber cell membrane p _H2O was also not sensitive to Hg²⁺ whereas lens epithelial cell p _H2O (136 μm/sec in rabbit) was blocked by Hg²⁺. With regard to the other hypothesized roles, fiber cell membrane or lipid vesicles had a glycerol permeability on the order of 1 nm/sec, an order of magnitude less than that conferred by MIP when expressed in oocytes. Impedance studies were employed to determine gap junctional coupling and fiber cell membrane conductance in wild-type and heterozygous Cat ^Fr mouse lenses. There was no detectable difference in either coupling or conductance between the wild-type and the mutant lenses. Received: 17 February 1999/Revised: 16 April 1999     

Monitoring of the Membrane Potential in Proteoliposomes with Incorporated Cytochrome-c Oxidase Using the Fluorescent Dye Indocyanine

A method has been developed to monitor changes of the membrane potential across vesicle membranes in real time. Using the potential-sensitive fluorescent dye indocyanine and on the basis of a water/lipid redistribution model, a calculation procedure has been introduced to estimate the membrane potential in vesicles with incorporated cytochrome-c oxidase. Physical parameters, such as vesicle size distribution and density of the lipid bilayer were estimated and used as calculation parameters. By extrapolation of the transient potential change to zero time, the initial rate of the potential change (dU/dt) could be calculated. It is also shown, that the initial potential change (dU/dt) may be used to study the proton/electron stoichiometry of cytochrome-c oxidase incorporated in the vesicles. Received: 28 September 1995/Revised: 6 February 1996     

Influence of a Transmembrane Protein on the Permeability of Small Molecules Across Lipid Membranes

The influence of the nonchannel conformation of the transmembrane protein gramicidin A on the permeability coefficients of neutral and ionized α-X-p-methyl-hippuric acid analogues (XMHA) (X = H, OCH₃, CN, OH, COOH, and CONH₂) across egg-lecithin membranes has been investigated in vesicle efflux experiments. Although 10 mol% gramicidin A increases lipid chain ordering, it enhances the transport of neutral XMHA analogues up to 8-fold, with more hydrophilic permeants exhibiting the greatest increase. Substituent contributions to the free energies of transfer of both neutral and anionic XMHA analogues from water into the bilayer barrier domain were calculated. Linear free-energy relationships were established between these values and those for solute partitioning from water into decadiene, chlorobutane, butyl ether, and octanol to assess barrier hydrophobicity. The barrier domain is similar for both neutral and ionized permeants and substantially more hydrophobic than octanol, thus establishing its location as being beyond the hydrated headgroup region and eliminating transient water pores as the transport pathway for these permeants, as the hydrated interface or water pores would be expected to be more hydrophilic than octanol. The addition of 10 mol% gramicidin A alters the barrier domain from a decadiene-like solvent to one possessing a greater hydrogen-bond accepting capacity. The permeability coefficients for ionized XMHAs increase with Na⁺ or K⁺ concentration, exhibiting saturability at high ion concentrations. This behavior can be quantitatively rationalized by Gouy-Chapman theory, though ion-pairing cannot be conclusively ruled out. The finding that transmembrane proteins alter barrier selectivity, favoring polar permeant transport, constitutes an important step toward understanding permeability in biomembranes. Received: 12 July 1999/Revised: 20 October 1999     

Cation Permeability and Selectivity of a Root Plasma Membrane Calcium Channel

Calcium channels in the plasma membrane of root cells fulfill both nutritional and signaling roles. The permeability of these channels to different cations determines the magnitude of their cation conductances, their effects on cell membrane potential and their contribution to cation toxicities. The selectivity of the rca channel, a Ca²⁺-permeable channel from the plasma membrane of wheat (Triticum aestivum L.) roots, was studied following its incorporation into planar lipid bilayers. The permeation of K⁺, Na⁺, Ca²⁺ and Mg²⁺ through the pore of the rca channel was modeled. It was assumed that cations permeated in single file through a pore with three energy barriers and two ion-binding sites. Differences in permeation between divalent and monovalent cations were attributed largely to the affinity of the ion binding sites. The model suggested that significant negative surface charge was present in the vestibules to the pore and that the pore could accommodate two cations simultaneously, which repelled each other strongly. The pore structure of the rca channel appeared to differ from that of L-type calcium channels from animal cell membranes since its ion binding sites had a lower affinity for divalent cations. The model adequately accounted for the diverse permeation phenomena observed for the rca channel. It described the apparent submillimolar K _m for the relationship between unitary conductance and Ca²⁺ activity, the differences in selectivity sequences obtained from measurements of conductance and permeability ratios, the changes in relative cation permeabilities with solution ionic composition, and the complex effects of Ca²⁺ on K⁺ and Na⁺ currents through the channel. Having established the adequacy of the model, it was used to predict the unitary currents that would be observed under the ionic conditions employed in patch-clamp experiments and to demonstrate the high selectivity of the rca channel for Ca²⁺ influx under physiological conditions. Received: 23 August 1999/Revised: 12 November 1999     

The Functional Surface Charge Density of a Fast K Channel in the Myelinated Axon of Xenopus laevis

The action of Mg²⁺ on the putative xKv1.1 channel in the myelinated axon of Xenopus laevis was analyzed in voltage clamp experiments. The main effect was a shift in positive direction of the open probability curve (16 mV at 20 mm Mg²⁺), calculated from measurements of the instantaneous current at Na reversal potential after 50–100 msec steps to different potentials. The shift was measured at an open probability level of 25% to separate it from shifts of other K channel populations in the nodal region. The results could be explained in terms of screening effects on fixed charges located on the surface of the channel protein. Using the Grahame equation the functional charge density was estimated to −0.45 e nm⁻². Analyzing this value, together with previously estimated values from other K channels, with reference to the charge of different extracellular loops of the channel protein, we conclude that the loop between the transmembrane S5 segment and the pore forming P segment determines the functional charge density of voltage-gated K channels. Received: 11 December 1997/Revised: 24 April 1998     

The influence of surface charges on quaternary ammonium block of shaker K+ channels

Block of K⁺ channels can be influenced by the ability of charged residues on the protein surface to accumulate cationic blocking ions to concentrations greater than those in bulk solution. We examined the ionic strength dependence of extracellular block of Shaker K⁺ channels by tetraethylammonium ions (TEA⁺) and by a trivalent quaternary ammonium ion, gallamine³⁺. Wild-type and mutant channels were expressed in Xenopus oocytes and currents recorded with the cut-open oocyte technique. Channel block by both compounds was substantially increased when the bathing electrolyte ionic strength was lowered, but with a much larger effect for trivalent gallamine. These data were quantitatively well described by a simple electrostatic model, accounting for accumulation of blocking ions near the pore of the channel by surface charges. The surface charge density of the wild-type channel consistent with the results was −0.1 e nm⁻². Shaker channels with T449Y mutations have an increased affinity for both TEA and gallamine but the ionic strength dependence of block was described with the same surface charge density as wild-type channels. Much of the increased sensitivity of Shaker K⁺ channels to gallamine may be due to a larger local accumulation of the trivalent ion. The negative charge at position 431 contributes to the sensitivity of channels to TEA (MacKinnon & Yellen, 1990). A charge reversal mutation at this location had little effect on the ionic strength dependence of quaternary ammonium ion block, suggesting that the charge on this amino acid may directly affect binding affinity but not local ion accumulation. Received: 7 December 2000/Revised: 27 April 2001     

Mutants of the Lactose Carrier of Escherichia coli Which Show Altered Sugar Recognition Plus a Severe Defect in Sugar Accumulation

Lactose and melibiose are actively accumulated by the wild-type Escherichia coli lactose carrier, which is an integral membrane protein energized by the proton motive force. Mutants of the E. coli lactose carrier were isolated by their ability to grow on minimal plates with succinate plus IPTG in the presence of the toxic lactose analog β-thio-o-nitrophenylgalactoside (TONPG). TONPG-resistant mutants were streaked on melibiose MacConkey indicator plates, and red clones were picked. These melibiose positive mutants were then streaked on lactose MacConkey plates, and white clones were picked. Transport assays indicated that the mutants had altered sugar recognition and a defect in sugar accumulation. The mutants had a poor apparent K _m for both lactose and melibiose in transport. One mutant had almost no ability to take up lactose, but melibiose downhill transport was 58% (V _max ) of normal. All of the mutants accumulated methyl-α-d-galactopyranoside (TMG) to only 8% or less of normal, and two failed to accumulate. Immunoblot analysis of the mutant lactose carrier proteins indicated that loss of sugar transport activity was not due to loss of expression in the membrane. Nucleotide sequencing of the lacY gene from the mutants revealed changes in the following amino acids of the lactose carrier: M23I, W151L, G257D, A295D and G377V. Two of the mutants (G257D and G377V) are novel in that they represent the first amino acids in periplasmic loops to be implicated with changes in sugar recognition. We conclude that the amino acids M23, W151, G257, A295 and G377 of the E. coli lactose carrier play either a direct or an indirect role in sugar recognition and accumulation. Received: 12 October 1999/Revised: 21 December 1999     

Cysteine Residues and the Structure of the Rat Renal Proximal Tubular Type II Sodium Phosphate Cotransporter (Rat NaPi IIa)

The rat renal Na/P _i cotransporter type IIa (rat NaP_i IIa) is a 637 amino acid protein containing 12 cysteine residues. We examined the effect of different cysteine modifying methanethiosulfonate (MTS)-reagents and the disulfide bond reducing agent tris(2-carboxyethyl)phosphine (TCEP) on the transport activity of wild-type and 12 single cysteine substitution mutants of rat NaPi IIa expressed in Xenopus laevis oocytes. The transport activity of the wild-type protein was resistant to three membrane impermeant MTS-reagents (MTSEA, MTSET and MTSES). In contrast, membrane permeant methyl methanethiosulfonate (MMTS) and TCEP inhibited the transport activity of both the wild-type, as well as all the single mutant proteins. This indicated the existence of more than one functionally important cysteine residue, not accessible extracellularly, and at least 2 disulfide bridges. To identify the disulfide bridges, three double mutants lacking 2 of the 3 cysteine residues predicted to be extracellular in different combinations were examined. This led to the identification of one disulfide bridge between C306 and C334; reconsideration of the topological model predictions suggested a second disulfide bridge between C225 and C520. Evaluation of a fourth double mutant indicated that at least one of two disulfide bridges (C306 and C334; C225 and C520) has to be formed to allow the surface expression of a functional cotransporter. A revised secondary structure is proposed which includes two partially repeated motifs that are connected by disulfide bridges formed between cysteine pairs C306-C334 and C225-C520. Received: 13 December 1999/Revised: 31 March 2000     

Direct Comparison of NPPB and DPC as Probes of CFTR Expressed in Xenopus Oocytes

Blockers of CFTR with well-characterized kinetics and mechanism of action will be useful as probes of pore structure. We have studied the mechanism of block of CFTR by the arylaminobenzoates NPPB and DPC. Block of macroscopic currents by NPPB and DPC exhibited similar voltage-dependence, suggestive of an overlapping binding region. Kinetic analysis of single-channel currents in the presence of NPPB indicate drug-induced closed time constants averaging 2.2 msec at −100 mV. The affinity for NPPB calculated from single-channel block, K _D = 35 μm, exceeds that for other arylaminobenzoates studied thus far. These drugs do not affect the rate of activation of wild-type (WT) channels expressed in oocytes, consistent with a simple mechanism of block by pore occlusion, and appear to have a single binding site in the pore. Block by NPPB and DPC were affected by pore-domain mutations in different ways. In contrast to its effects on block by DPC, mutation T1134F-CFTR decreased the affinity and reduced the voltage-dependence for block by NPPB. We also show that the alteration of macroscopic block by NPPB and DPC upon changes in bath pH is due to both direct effects (i.e., alteration of voltage-dependence) and indirect effects (alteration of cytoplasmic drug loading). These results indicate that both NPPB and DPC block CFTR by entering the pore from the cytoplasmic side and that the structural requirements for binding are not the same, although the binding regions within the pore are similar. The two drugs may be useful as probes for overlapping regions in the pore. Received: 14 October 1999/Revised: 18 January 2000     

Amphotericin B Kills Unicellular Leishmanias by Forming Aqueous Pores Permeable to Small Cations and Anions

The polyene antibiotic amphotericin B (AmB) is known to form two types of ionic channels across sterol-containing liposomes, depending on its concentration and time after mixing (Cohen, 1992). In the present study, it is shown that AmB only kills unicellular Leishmania promastigotes (LPs) when aqueous pores permeable to small cations and anions are formed. Changes of membrane potential across ergosterol-containing liposomes and LPs were followed by fluorescence changes of 3,3′ dipropylthiadicarbocyanine (DiSC₃(5)). In KCl-loaded liposomes suspended in an iso-osmotic sucrose solution, low AmB concentrations (≤0.1 μm) induced a polarization potential, indicating K⁺ leakage, but no movement of cations and anions was allowed until AmB concentrations greater than 0.1 μm were added. In agreement with these data, it was found that AmB altered the negative membrane potential held across LPs in a manner consistent with the differential cation/anion selectivity exhibited by the channels formed in liposomes. Thus, LPs suspended in an iso-osmotic sucrose solution did not exhibit any AmB-induced membrane depolarization effect brought about by efflux of anions until 0.1 μm or higher AmB concentrations were added. By contrast, LPs suspended in an iso-osmotic NaCl solution and exposed to 0.05 μm AmB exhibited a nearly total collapse of the negative membrane potential, indicating Na⁺ entry into the cells. The concentration dependence of the AmB-induced permeability to different salts was also measured across vesicles derived from the plasma membrane of leishmanias (LMVs), by using a rapid mixing technique. At concentrations above 0.1 μm, AmB induced the formation of aqueous pores across LMVs with a positive cooperativity, yielding Hill coefficients between 2 to 3. Measured anion selectivity across such aqueous pores followed the sequence: SCN > NO3 > Cl > I > Br > acetate (SO²⁻ ₄ being impermeable). Cell killing by AmB was followed by fluorescence changes of the DNA-binding compound ethidium bromide (EB). At low concentrations (≤0.1 μm), AmB was found to be nonlethal against LPs but, above this concentration, leishmanias were rapidly killed. The rate and extent of such an effect were found to be dependent on the type of cation and anion present in the external aqueous solution. For both NH⁺ ₄ and Na⁺ salts, the measured rank order of AmB cell killing followed the same sequence that was determined for AmB-induced salt permeation across LMVs. Further, replacement of either extracellular Na⁺ by choline or Cl⁻ by SO²⁻ ₄, or its partial substitution by sucrose, in iso-osmotic conditions, led to a complete inhibition of the killing effect exerted by otherwise lethal AmB concentrations. Finally, it was shown that tetraethylammonium (TEA⁺), an organic cation that is known to block AmB-induced salt permeation across LMVs was able to retard the time lag observed for EB incorporation across LPs, indicating that this parameter can be taken to represent the time taken for salt accumulation inside the parasites. The present results thus indicate clearly that low AmB concentrations (≤0.1 μm) were able to form across LPs, cation channels that collapsed the parasite membrane potential but are not lytic. At high concentrations (<≥0.1 μm), a salt influx via the aqueous pores formed by the antibiotic was followed by osmotic changes leading to cell lysis. This last stage is supported by electron microscopy observations of the changes of parasite morphology immediately upon addition of AmB, which indicated that the typical elongated promastigote cell forms became rounded and the flagella swells and round up. The present work is the first demonstration of the in vitro sensitivity of Leishmania promastigotes to osmotic lysis by AmB. Received: 25 September 1995/Revised: 11 March 1996     

Complement Activation by Bacterial Surface Glycolipids: A Study with Planar Bilayer Membranes

Planar asymmetric glycolipid/phospholipid bilayer membranes were used as a reconstitution model of the lipid matrix of the outer membrane of Gram-negative bacteria to study complement (C) activation by various bacterial surface glycolipids with the aim of defining the C activation pathway. As glycolipids the lipopolysaccharides of Salmonella enterica serovar Minnesota R mutant strains R595 (Re LPS) and R4 (Rd₂ LPS), pentaacyl lipid A from the LPS of the Escherichia coli Re mutant F515, and glycosphingolipid GSL-1 of Sphingomonas paucimobilis IAM 12576 were used. Methylester and carboxyl-reduced derivatives of GSL-1 were used to elucidate the role of the carboxyl group as common functional group of LPS and GSL-1 for C activation. The formation of lytic pores was monitored via the measurement of changes in membrane current. For all glycolipids we observed a considerable increase in membrane current soon after addition of whole human serum due to the formation of lytic pores in the membranes. Pore formation was dependent on the presence of C9, indicating that the observed current changes were due to C activation. We found that in our reconstitution system of the outer membrane lipid A, Re LPS, and Rd₂ LPS activated the classical pathway, the activation being independent of specific anti-LPS antibodies. In contrast, GSL-1 and the methylester derivative of GSL-1 activated the alternative pathway even at the low serum concentrations used in this study (about 0.2% v/v). Interestingly, the carboxyl reduced GSL-1 activated the classical pathway. Received: 16 July 1998/Revised: 28 October 1998     

Ion Channels Induced in Planar Lipid Bilayers by the Bacillus thuringiensis Toxin Cry1Aa in the Presence of Gypsy Moth (Lymantria dispar) Brush Border Membrane

The apical brush border membrane, the main target site of Bacillus thuringiensis toxins, was isolated from gypsy moth (Lymantria dispar) larval midguts and fused to artificial planar lipid bilayer membranes. Under asymmetrical N-methyl-d-glucamine-HCl conditions (450 mm cis/150 mm trans, pH 9.0), which significantly reduce endogenous channel activity, trypsin-activated Cry1Aa, a B. thuringiensis insecticidal protein active against the gypsy moth in vivo, induced a large increase in bilayer membrane conductance at much lower concentrations (1.1–2.15 nm) than in receptor-free bilayer membranes. At least 5 main single-channel transitions with conductances ranging from 85 to 420 pS were resolved. These Cry1Aa channels share similar ionic selectivity with P _Cl/P _NMDG permeability ratios ranging from 4 to 8. They show no evidence of current rectification. Analysis of the macroscopic current flowing through the composite bilayer suggested voltage-dependence of several channels. In comparison, the conductance of the pores formed by 100–500 nm Cry1Aa in receptor-free bilayer membranes was significantly smaller (about 8-fold) and their P _Cl/P _NMDG permeability ratios were also reduced (2- to 4-fold). This study provides a detailed demonstration that the target insect midgut brush border membrane material promotes considerably pore formation by a B. thuringiensis Cry toxin and that this interaction results in altered channel properties. Received: 23 February 2001/Revised: 15 June 2001     

ATP-Induced Shape Change of Nuclear Pores Visualized with the Atomic Force Microscope

Bidirectional transport of molecules between nucleus and cytoplasm through the nuclear pore complexes (NPCs) spanning the nuclear envelope plays a fundamental role in cell function and metabolism. Nuclear import of macromolecules is a two-step process involving initial recognition of targeting signals, docking to the pore and energy-driven translocation. ATP depletion inhibits the translocation step. The mechanism of translocation itself and the conformational changes of the NPC components that occur during macromolecular transport, are still unclear. The present study investigates the effect of ATP on nuclear pore conformation in isolated nuclear envelopes from Xenopus laevis oocytes using the atomic force microscope. All experiments were conducted in a saline solution mimicking the cytosol using unfixed nuclear envelopes. ATP (1 mm) was added during the scanning procedure and the resultant conformational changes of the NPCs were directly monitored. Images of the same nuclear pores recorded before and during ATP exposure revealed dramatic conformational changes of NPCs subsequent to the addition of ATP. The height of the pores protruding from the cytoplasmic surface of the nuclear envelope visibly increased while the diameter of the pore opening decreased. The observed changes occurred within minutes and were transient. The slow-hydrolyzing ATP analogue, ATP-γ-S, in equimolar concentrations did not exert any effects. The ATP-induced shape change could represent a nuclear pore ``contraction.' Received: 10 February 1997/Revised: 10 February 1998