The effect of protein stability on protein-monoglyceride interactions.

We have previously shown that proteins such as beta-lactoglobulin and lysozyme insert into monoglyceride monolayers and are able to induce an L(beta) to coagel phase transition in monoglyceride bilayers. These studies gave a first indication that protein stability could be an important factor for these interactions. This study therefore aims at further investigating the potential role of protein stability on protein-monoglyceride interactions. To this end we studied the interaction of stable and destabilized alpha-lactalbumin with monostearoylglycerol. Our results show that protein stability is important for the insertion of proteins into a monostearoylglycerol monolayer, such that the lower the stability of the protein the better the protein inserts. In marked contrast to beta-lactoglobulin and lysozyme we found that destabilized alpha-lactalbumin does not induce the L(beta) to coagel phase transition in monoglyceride bilayers. We propose that this is due to an increased surface coverage by the protein which could result from the unfolding of the protein upon binding to the interface.

Thermotropic phase behavior of monoglyceride-dicetylphosphate dispersions and interactions with proteins: a (2)H and (31)P NMR study.

The phase behavior of a 1-[(2)H(35)]-stearoyl-rac-glycerol ([(2)H(35)]-MSG)/dicetylphosphate (DCP) mixture and its interaction with beta-lactoglobulin and lysozyme were studied by (2)H and (31)P nuclear magnetic resonance (NMR). The behavior of the lipids was monitored by using deuterium-labeled [(2)H(35)]-MSG as a selective probe for (2)H NMR and DCP for (31)P NMR. Both (2)H and (31)P NMR spectra exhibit characteristic features representative of different phases. In the lamellar phases, (31)P NMR spectra of DCP are different from the spectra of natural phospholipids, which is attributable to differences in the intramolecular motions and the orientation of the shielding tensor of DCP compared with phospholipids. The presence of the negatively charged amphiphile DCP has a large effect on the phase behavior of [(2)H(35)]-MSG. At low temperature, the presence of DCP inhibits crystallization of the gel phase into the coagel. Upon increasing the temperature, the gel phase of [(2)H(35)]-MSG transforms in the liquid-crystalline lamellar phase. In the presence of DCP, the gel phase directly transforms into an isotropic phase. The negatively charged beta-lactoglobulin and the positively charged lysozyme completely neutralize the destabilizing effect of DCP on the monoglyceride liquid-crystalline phase and they even stabilize this phase. Without DCP the proteins do not seem to interact with the monoglyceride. These results suggest that interaction is facilitated by electrostatic interactions between the negatively charged DCP and positively charged residues in the proteins. In addition, the nonbilayer-forming DCP creates insertion sites for proteins in the bilayer.

Effect of nonbilayer lipids on membrane binding and insertion of the catalytic domain of leader peptidase.

Biological membranes contain a substantial amount of "nonbilayer lipids", which have a tendency to form nonlamellar phases. In this study the hypothesis was tested that the presence of nonbilayer lipids in a membrane, due to their overall small headgroup, results in a lower packing density in the headgroup region, which might facilitate the interfacial insertion of proteins. Using the catalytic domain of leader peptidase (delta2-75) from Escherichia coli as a model protein, we studied the lipid class dependence of its insertion and binding. In both lipid monolayers and vesicles, the membrane binding of (catalytically active) delta2-75 was much higher for the nonbilayer lipid DOPE compared to the bilayer lipid DOPC. For the nonbilayer lipids DOG and MGDG a similar effect was observed as for DOPE, strongly suggesting that no specific interactions are involved but that the small headgroups create hydrophobic interfacial insertion sites. On the basis of the results of the monolayer experiments, calculations were performed to estimate the space between the lipid headgroups accessible to the protein. We estimate a maximal size of the insertion sites of 15 +/- 7 A2/lipid molecule for DOPE, relative to DOPC. The size of the insertion sites decreases with an increase in headgroup size. These results show that nonbilayer lipids stimulate the membrane insertion of delta2-75 and support the idea that such lipids create insertion sites by reducing the packing density at the membrane-water interface. It is suggested that PE in the bacterial membrane facilitates membrane insertion of the catalytic domain of leader peptidase, allowing the protein to reach the cleavage site in preproteins.

Fructans insert between the headgroups of phospholipids.

Fructans are polysaccharides consisting of one glucose unit and two or more fructose units. It was hypothesized that fructans play a role in drought tolerance in plants by interacting directly with the membrane. In this paper we investigated this hypothesis by studying fructan-membrane interactions in hydrated mono- and bilayer systems. It was found that fructans inserted between the headgroups of different kinds of phospholipids with some preference for phosphatidylethanolamine. Insertion occurred even under conditions of very tight lipid packing. The presence of a surface associated layer of fructan was observed in both model systems. This layer was able to reduce the ability of a surface-active protein to interact with the lipids. Fructans showed a much stronger effect on the different lipid systems than other (poly)saccharides, which appears to be related to their hydrophobic properties. Fructans were able to stabilize the liquid-crystalline lamellar phase, which is consistent with a drought protecting role in plants.

The specificity of monoglyceride-protein interactions and mechanism of the protein induced L(beta) to coagel phase transition.

This study aims at gaining insight into the specificity and molecular mechanism of monoglyceride-protein interactions. We used beta-lactoglobulin (beta-LG) and lysozyme as model proteins and both monostearoylglycerol and monopalmitoylglycerol as defined gel phase monoglycerides. The monoglycerides were used in different combinations with the two negatively charged amphiphiles dicetylphosphate and distearylphosphate. The interactions were characterized using the monolayer technique, isothermal titration calorimetry, (2)H-nuclear magnetic resonance (NMR) using deuterium labelled monoglycerides and freeze fracture electron microscopy (EM). Our results show that lysozyme inserts efficiently into all monolayers tested, including pure monoglyceride layers. The insertion of beta-LG depends on the lipid composition of the monolayer and is promoted when the acylchains of the negatively charged amphiphile are shorter than that of the monoglyceride. The binding parameters found for the interaction of beta-LG and lysozyme with monoglyceride bilayers were generally similar. Moreover, in all cases a large exothermic binding enthalpy was observed which was found to depend on the nature of the monoglycerides but not of the proteins. (2)H-NMR and freeze fracture EM showed that this large enthalpy results from a protein mediated catalysis of the monoglyceride L(beta) to coagel phase transition. The mechanism of this phase transition consists of two steps, an initial protein mediated vesicle aggregation step which is followed by stacking and probably fusion of the bilayers.

Membrane activity of the peptide antibiotic clavanin and the importance of its glycine residues.

The peptide antibiotic clavanin A (VFQFLGKIIHHVGNFVHGFSHVF-NH(2)) is rich in histidine and glycine residues. In this study the antimicrobial activity and membrane activity of wild-type clavanin A and seven Gly --> Ala mutants thereof were investigated. Clavanin A effectively killed the test microorganism Micrococcus flavus and permeabilized its cytoplasmic membrane in the micromolar concentration range, suggesting that the membrane is the target for this molecule. Consistent with this suggestion, it was observed that clavanin A efficiently inserted into different phospholipid monolayers mainly via hydrophobic interactions. Bilayer permeabilization was observed for both low and high molecular mass fluorophores enclosed in unilamellar vesicles and occurred at the same concentration as the antimicrobial activity. It is therefore suggested that the loss of barrier function does not involve specific receptors in the target membrane. Circular dichroism spectroscopy indicated that under membrane mimicking conditions a random coil --> helical transition was induced for all clavanin derivatives tested. Observed differences in peptide-membrane interaction and biological activity between the various clavanin derivatives demonstrated the functional importance of Gly at the positions 6 and 13. These two glycines may act as flexible hinges that facilitate the hydrophobic N-terminal end of clavanin to deeply insert into the bilayer. On the contrary, no such role is evident for Gly 18, as its substitution by Ala actually stimulated membrane interaction and biological activity. This study suggests that the combined hydrophobicity, overall state of charge, and conformational flexibility of the peptide determine the (membrane) activity of clavanin A and its Gly --> Ala mutants.

Spontaneous insertion of gene 9 minor coat protein of bacteriophage M13 in model membranes.

Gene 9 minor coat protein from bacteriophage M13 is known to be located in the inner membrane after phage infection of Escherichia coli. The way of insertion of this small protein (32 amino acids) into membranes is still unknown. Here we show that the protein is able to insert in monolayers. The limiting surface pressure of 35 mN/m for 1,2-dioleoyl-sn-glycero-3-phosphocholine and 1,2-dioleoyl-sn-glycero-3-phosphoglycerol lipid systems indicates that this spontaneous insertion can also occur in vivo. By carboxyfluorescein leakage experiments of vesicles it is demonstrated that protein monomers, or at least small aggregates, are more effective in releasing carboxyfluorescein than highly aggregated protein. The final orientation of the protein in the bilayer after insertion was addressed by proteinase K digestion, thereby making use of the unique C-terminal location of the antigenic binding site. After insertion the C-terminus is still available for the enzymatic digestion, while the N-terminus is not. This leads to the overall conclusion that the protein is able to insert spontaneously into membranes without the need of any machinery or transmembrane gradient, with the positively charged C-terminus remaining on the outside. The orientation after insertion of gene 9 protein is in agreement with the 'positive inside rule'.

Lipid organization and dynamics of the monostearoylglycerol-water system. A 2H NMR study.

Deuterium labeled monostearoylglycerols with fully ([2H(35)]-MSG) and selectively ([11-(2)H(2)]-MSG) deuterated chains have been synthesized and used as a probe for 2H NMR. At low temperature monoglyceride-water systems form the coagel or crystalline phase, which transforms with increasing temperature subsequently into the gel, liquid crystalline and cubic phase. The 2H NMR spectra exhibit characteristic features representative of these phases. The gel phase is metastable and gradually transforms into the coagel at temperatures below 40 degrees C. The undercooled cubic phase transforms into the liquid crystalline phase during days. In the liquid crystalline phase, the chain order profile indicates an increase of the chain flexibility towards the methyl group. In the liquid crystalline phase, bilayers spontaneously align in a magnetic field with their normal perpendicular to the field. The results demonstrate that 2H NMR can serve as a convenient tool to study both structure and dynamics of different monoglyceride-water phases.

Dynamin is membrane-active: lipid insertion is induced by phosphoinositides and phosphatidic acid.

Dynamin is a large GTPase involved in the regulation of membrane constriction and fission during receptor-mediated endocytosis. Dynamin contains a pleckstrin-homology domain which is essential for endocytosis and which binds to anionic phospholipids. Here, we show for the first time that dynamin is a membrane-active molecule capable of penetrating into the acyl chain region of membrane lipids. Lipid penetration is strongly stimulated by phosphatidic acid (PA), phosphatidylinositol 4-phosphate, and phosphatidylinositol 4, 5-bisphosphate. Though binding is more efficient in the presence of the phosphoinositides, a much larger part of the dynamin molecule penetrates into PA-containing mixed-lipid systems. Thus, local lipid metabolism will dramatically influence dynamin-lipid interactions, and dynamin-lipid interactions are likely to play an important role in dynamin-dependent endocytosis. Our data suggest that dynamin is directly involved in membrane destabilization, a prerequisite to membrane fission.

Visualization of highly ordered striated domains induced by transmembrane peptides in supported phosphatidylcholine bilayers.

We used atomic force microscopy (AFM) to study the lateral organization of transmembrane TmAW(2)(LA)(n)W(2)Etn peptides (WALP peptides) incorporated in phospholipid bilayers. These well-studied model peptides consist of a hydrophobic alanine-leucine stretch of variable length, flanked on each side by two tryptophans. They were incorporated in saturated phosphatidylcholine (PC) vesicles, which were deposited on a solid substrate via the vesicle fusion method, yielding hydrated gel-state supported bilayers. At low concentrations (1 mol %) WALP peptides induced primarily line-type depressions in the bilayer. In addition, striated lateral domains were observed, which increased in amount and size (from 25 nm up to 10 microm) upon increasing peptide concentration. At high peptide concentration (10 mol %), the bilayer consisted mainly of striated domains. The striated domains consist of line-type depressions and elevations with a repeat distance of 8 nm, which form an extremely ordered, predominantly hexagonal pattern. Overall, this pattern was independent of the length of the peptides (19-27 amino acids) and the length of the lipid acyl chains (16-18 carbon atoms). The striated domains could be pushed down reversibly by the AFM tip and are thermodynamically stable. This is the first direct visualization of alpha-helical transmembrane peptide-lipid domains in a bilayer. We propose that these striated domains consist of arrays of WALP peptides and fluidlike PC molecules, which appear as low lines. The presence of the peptides perturbs the bilayer organization, resulting in a decrease in the tilt of the lipids between the peptide arrays. These lipids therefore appear as high lines.

Blistering of langmuir-blodgett bilayers containing anionic phospholipids as observed by atomic force microscopy.

Asymmetric bilayers of different phospholipid compositions have been prepared by the Langmuir-Blodgett (L-B) method, and imaged by atomic force microscopy (AFM). Such bilayers can function as a model for biological membranes. The first leaflet consisted of zwitterionic phospholipids phosphatidylcholine (PC) or phosphatidylethanolamine (PE). The second leaflet consisted of the anionic phospholipid phosphatidylglycerol (PG), in either the condensed or liquid phase or, for comparison, of PC. Different bilayers showed different morphology. In all bilayers defects in the form of holes were present. In some bilayers with a first leaflet consisting of PC, polygonal line-shaped defects were observed, whereas when the first leaflet consisted of PE, mainly round defects were seen. Not only the shape, but also the amount of defects varied, depending on the condition and the composition of the second leaflet. In most of the PG-containing systems the defects were surrounded by elevations, which reversibly disappeared in the presence of divalent cations. This is the first time that such elevations have been observed on phospholipid bilayers. We propose that they are induced by phospholipid exchange between the two leaflets around the defects, leading to the presence of negatively charged phospholipids in the first leaflet. Because the substrate is also negatively charged, the bilayer around the edges is repelled and lifted up. Since it was found that the elevations are indeed detached from the substrate, we refer to this effect as bilayer blistering.

Interaction mode specific reorganization of gel phase monoglyceride bilayers by beta-lactoglobulin.

The interaction between beta-lactoglobulin and sonicated aqueous dispersions of the gel phase forming monoglyceride monostearoylglycerol were studied using isothermal titration calorimetry, direct binding experiments, differential scanning calorimetry, leakage of a fluorescent dye and solid-state (31)P- and (2)H-NMR. In the absence of a charged amphiphile, monostearoylglycerol forms a precipitate. Under these conditions, no interaction with beta-lactoglobulin was observed. In the presence of the negatively charged amphiphile dicetylphosphate, the gel phase monostearoylglycerol formed stable and closed, probably unilamellar, vesicles with an average diameter of 465 nm. beta-Lactoglobulin interacts with these bilayer structures at pH 4, where the protein is positively charged, as well as at pH 7 where the protein is negatively charged. Under both conditions of pH, the binding affinity of beta-lactoglobulin is in the micromolar range as observed with ITC and the direct binding assay. At pH 4, two binding modes were found, one of which is determined with ITC while the direct binding assay determines the net result of both. The first binding mode is observed with ITC and is characterized by a large binding enthalpy, a decreased enthalpy of the MSG L(beta) to L(alpha) phase transition and leakage of a fluorescent dye. These characteristics are explained by a beta-lactoglobulin induced partial L(beta) to coagel phase transition that results from a specific electrostatic interaction between the protein and the charged amphiphile. This explanation is confirmed by solid-state (2)H-NMR using 1-monostearoylglycerol with a fully deuterated acyl chain. Upon interaction with beta-lactoglobulin, the isotropic signal in the (2)H-NMR spectrum of the monostearoylglycerol-dicetylphosphate mixture partially transforms into a broad anisotropic signal which could be assigned to coagel formation. The second binding mode probably results from an aspecific electrostatic attraction between the negatively charged bilayer and the positively charged protein and causes the precipitation of the dispersion. At pH 7, only the first binding mode is observed.

Pore formation by nisin involves translocation of its C-terminal part across the membrane.

Nisin is an amphiphilic peptide with a strong antimicrobial activity against various Gram-positive bacteria. Its activity results from permeabilization of bacterial membranes, causing efflux of cytoplasmic compounds. To get information on the molecular mechanism of membrane permeabilization, a mutant of nisin Z containing the C-terminal extension Asp-(His)6 was produced. The biological and anionic lipid-dependent membrane activity of this peptide was very similar to that of nisin Z. Analysis of the pH dependence of model membrane interactions with the elongated peptide indicated the importance of electrostatic interactions of the C-terminus with the target membrane for membrane permeabilization. Most importantly, the membrane topology of the C-terminus of the molecule could be determined by trypsin digestion experiments, in which trypsin was encapsulated in the lumen of large unilamellar vesicles. The results show that the C-terminal part of the peptide translocates across model membranes. The pH and anionic lipid dependence of translocation closely paralleled the results of membrane permeabilization studies. Binding of nickel ions to the histidines blocked translocation of the C-terminus and concomitantly resulted in a 4-fold reduced capacity to induce K+ leakage. The results demonstrate for the first time that pore formation of nisin involves translocation of the C-terminal region of the molecule across the membrane.

Fructans interact strongly with model membranes.

Bacterial fructans with a high degree of polymerisation cause a very large increase in surface pressure of lipid monolayers at the air-water interface with a broad range of lipids, including phosphatidylethanolamine and several types of phosphatidylcholines. The surface active effect of fructans contrasts strongly with the maximal effects observed for trehalose, sucrose and glucose under comparable conditions (20 and 0.6 mN/m for fructans and the other sugars, respectively). The results demonstrate a profound and specific membrane interaction of the fructans which is probably very different from the effect of the smaller carbohydrates. The fructan concentrations used in this study are within the physiological range observed in fructan-accumulating plants. The suggested water-stress protective effect of fructans may be induced by membrane-fructan interaction which prevent lipid condensation and phase transitions to take place.

Phosphatidylethanolamine mediates insertion of the catalytic domain of leader peptidase in membranes.

Leader peptidase is an integral membrane protein of E. coli and it catalyses the removal of most signal peptides from translocated precursor proteins. In this study it is shown that when the transmembrane anchors are removed in vivo, the remaining catalytic domain can bind to inner and outer membranes of E. coli. Furthermore, the purified catalytic domain binds to inner membrane vesicles and vesicles composed of purified inner membrane lipids with comparable efficiency. It is shown that the interaction is caused by penetration of a part of the catalytic domain between the lipids. Penetration is mediated by phosphatidylethanolamine, the most abundant lipid in E. coli, and does not seem to depend on electrostatic interactions. A hydrophobic segment around the catalytically important residue serine 90 is required for the interaction with membranes.

The orientation of nisin in membranes.

Nisin is a 34 residue long peptide belonging to the group A lantibiotics with antimicrobial activity against Gram-positive bacteria. The antimicrobial activity is based on pore formation in the cytoplasmic membrane of target organisms. The mechanism which leads to pore formation remains to be clarified. We studied the orientation of nisin via site-directed tryptophan fluorescence spectroscopy. Therefore, we engineered three nisin Z variants with unique tryptophan residues at positions 1, 17, and 32, respectively. The activity of the tryptophan mutants against Gram-positive bacteria and in model membrane systems composed of DOPC or DOPG was established to be similar to that of wild type nisin Z. The tryptophan fluorescence emission maximum showed an increasing blue-shift upon interaction with vesicles containing increased amounts of DOPG, with the largest effect for the 1W peptide. Studies with the aqueous quencher acrylamide showed that all tryptophans became inaccessible from the aqueous phase in the presence of negatively charged lipids in the vesicles. From these results it is concluded that anionic lipids mediate insertion of the tryptophan residues in at least three positions of the molecule into the lipid bilayer. The depth of insertion of the tryptophan residues was determined via quenching of the tryptophan fluorescence by spin-labeled lipids. The results showed that the depth of insertion was dependent on the amount of negatively charged lipids. In membranes containing 50% DOPG, the distances from the bilayer center were determined to be 15.7, 15.0, and 18.4 A for the tryptophan at position 1, 17, and 32, respectively. In membranes containing 90% DOPG, these distances were calculated to be 10.8, 11.5, and 13.1 A, respectively. These results suggest an overall parallel average orientation of nisin in the membrane, with respect to the membrane surface, with the N-terminus more deeply inserted than the C-terminus. These data were used to model the orientation of nisin in the membrane.

Lipids in total extracts from Acholeplasma laidlawii A pack more closely than the individual lipids. Monolayers studied at the air-water interface.

Pressure-area curves were obtained at 25, 35 and 45 degrees C for total lipid extracts and four individual glucolipids isolated from Acholeplasma laidlawii strain A-EF22. The glucolipids are 1,2-diacyl-3-0-(alpha-D-glucopyranosyl)-sn-glycerol (MGlcDAG), 1,2 -diacyl-3-0-[alpha-D-glucopyranosyl-(1-->2)-0-alpha-D-glucopyranosyl] -sn-glycerol (DGlcDAG), 1,2-diacyl-3-0-[alpha-D-glucopyranosyl-(1-->2)-0-(6-0-acyl-alpha-D-gluco pyranosyl)]-sn-glycerol (MADGlcDAG), and 1,2-diacyl-3-0-[glycerophosphoryl-6-0-(alpha-D-glucopyranosyl-(1-- >)-0-alpha-D-glucopyranosyl)]-sn-glycerol (GPDGlcDAG). The total lipid extracts were obtained from A. laidlawii, grown at 37 degrees C with fatty acids of varying degrees of unsaturation and chain length. The mean surface area per molecule was obtained from these pressure-area curves at surface pressures equal to 10, 20, 30 and 40 mN/m. It was found that the interfacial area of the lipids increases with increasing degree of unsaturation, but is nearly independent of the acyl chain length at constant unsaturation. The surface charge density varied between 4.7 x 10(-3) e-/angstrom(2) and 9.4 x 10(-3) e-/angstrum(2) for the total lipid extracts studied, but did not exhibit any consistent dependence on variations in degree of unsaturation or acyl chain length. The mean area per molecule was found to be smaller for the total lipid extracts than for the individual lipids. It is concluded that the bacterium strives to regulate its lipid composition in such a way that the packing of the lipids in the membrane is appropriately tight, and/or to keep a slight negative spontaneous curvature of the lipid bilayer of the cell membrane ("optimal packing"). This is in accordance with the physico-chemical model for the regulation of the lipid composition in the membrane of A. laidlaiwii previously presented by us (see e.g. Andersson, A.-S., Riffors, L., Bergqvist, M., Persson, S. and Lindblom, G. (1996) Biochemistry 35, 11119-11130).

Charge-dependent insertion of beta-lactoglobulin into monoglyceride monolayers.

The interactions between beta-lactoglobulin and 1-monostearoyl-glycerol were studied in order to gain insight into protein-gel-phase monoglyceride interactions. Using a monomolecular layer at the air-water interface, we determined the insertion of beta-lactoglobulin into the monoglycerides under different conditions of protein and surface charge by varying the pH and/or incorporating charged amphiphiles into the monolayer, respectively, and using subphases with either a low or high ionic strength. The interactions were quantified by determining the binding of 14C-labeled beta-lactoglobulin to the monolayer. Our results show the importance of electrostatics for binding of beta-lactoglobulin to condensed monoglycerides. Moreover, electrostatic interactions were found to be important for specific insertion of beta-lactoglobulin into the monolayer. A negatively charged surface in particular allowed positively charged beta-lactoglobulin to insert in a surface charge density-dependent manner, even at surface pressures as high as 36 mN/m, whereas under other conditions, the limiting insertion pressure was 32 mN/m. The rheological properties of the monolayer were not affected by the interactions with beta-lactoglobulin.

Influence of charge differences in the C-terminal part of nisin on antimicrobial activity and signaling capacity.

Three mutants of the antibiotic nisin Z, in which the Val32 residue was replaced by a Glu, Lys or Trp residue, were produced and characterized for the purpose of establishing the role of charge differences in the C-terminal part of nisin on antimicrobial activity and signaling properties. 1H-NMR analyses showed that all three mutants harbor an unmodified serine residue at position 33, instead of the usual dehydroalanine. Apparently, the nature of the residue preceding the serine to be dehydrated, strongly affects the efficiency of modification. Cleavage of [Glu32,Ser33]nisin Z by endoproteinase Glu-C yielded [Glu32]nisin Z(1-32)-peptide, which has a net charge difference of -2 relative to wild-type nisin Z. The activity of [Lys32,Ser33]nisin Z against Micrococcus flavus was similar to that of wild-type nisin, while [Trp32,Ser33]nisin Z, [Glu32,Ser33]nisin Z and [Glu32]nisin Z(1-32)-peptide exhibited 3-5-fold reduced activity, indicating that negative charges in the C-terminal part of nisin Z are detrimental for activity. All variants showed significant loss of activity against Streptococcus thermophilus. The potency of the nisin variants to act as signaling molecules for auto-induction of biosynthesis was significantly reduced. To obtain mutant production, extracellular addition of (mutant) nisin Z to the lactococcal expression strains was essential.

The C-terminal region of nisin is responsible for the initial interaction of nisin with the target membrane.

The interaction of nisin Z and a nisin Z mutant carrying a negative charge in the C-terminus ([Glu-32]-nisin Z) with anionic lipids was characterized in model membrane systems, and bacterial membrane systems. We focused on three possible steps in the mode of action of nisin, i.e., binding, insertion, and pore formation of nisin Z. Increasing amounts of anionic lipids in both model and natural membranes were found to strongly enhance the interaction of nisin Z with the membranes at all stages. The results reveal a good correlation between the anionic lipid dependency of the three stages of interaction, of which the increased binding is probably the major determinant for antimicrobial activity. Maximal nisin Z activity could be observed for negatively charged lipid concentrations exceeding 50-60%, both in model membrane systems as well as in bacterial membrane systems. We propose that the amount of negatively charged lipids of the bacterial target membrane is a major determinant for the sensitivity of the organism for nisin. Nisin Z induced leakage of the anionic carboxyfluorescein was more efficient as compared to the leakage of the potassium cation. This lead to the conclusion that an anion-selective pore is formed. In contrast to the results obtained for nisin Z, the binding of [Glu-32]-nisin Z to vesicles remained low even in the presence of high amounts of negatively charged lipids. The insertion and pore-forming ability of [Glu-32]-nisin Z were also decreased. These results demonstrate that the C-terminus of nisin is responsible for the initial interaction of nisin, i.e., binding to the target membrane.