Modeling peptide binding to anionic membrane pores

Peptide‐induced pore formation in membranes can be dissected into two steps: pore formation and peptide binding to the pore. A computational method is proposed to study the second step in anionic membranes. The electrostatic potential is obtained from numerical solutions to the Poisson–Boltzmann equation and is then used in conjunction with IMM1 (implicit membrane model 1). A double charge layer model is used to incorporate the effects of the membrane dipole potential. Inhomogeneity of the charge density in the pore, characterized by explicit membrane simulations of toroidal pores, is included in the model. This approach was applied to two extensively studied peptides, magainin and melittin. In agreement with previous work, binding to toroidal pores is more favorable than binding to the flat membrane. The dependence of binding energy on anionic content exhibits different patterns for the two peptides, in correlation with the different lipid selectivity that has been observed experimentally. © 2013 Wiley Periodicals, Inc.     

Novel Antimicrobial Peptides from a Cecropin-Like Region of Heteroscorpine-1 from Heterometrus laoticus Venom with Membrane Disruption Activity

The increasing antimicrobial-resistant prevalence has become a severe health problem. It has led to the invention of a new antimicrobial agent such as antimicrobial peptides. Heteroscorpine-1 is an antimicrobial peptide that has the ability to kill many bacterial strains. It consists of 76 amino acid residues with a cecropin-like region in N-terminal and a defensin-like region in the C-terminal. The cecropin-like region from heteroscorpine-1 (CeHS-1) is similar to cecropin B, but it lost its glycine-proline hinge region. The bioinformatics prediction was used to help the designing of mutant peptides. The addition of glycine-proline hinge and positively charged amino acids, the deletion of negatively charged amino acids, and the optimization of the hydrophobicity of the peptide resulted in two mutant peptides, namely, CeHS-1 GP and CeHS-1 GPK. The new mutant peptide showed higher antimicrobial activity than the native peptide without increasing toxicity. The interaction of the peptides with the membrane showed that the peptides were capable of disrupting both the inner and outer bacterial cell membrane. Furthermore, the SEM analysis showed that the peptides created the pore in the bacterial cell membrane resulted in cell membrane disruption. In conclusion, the mutants of CeHS-1 had the potential to develop as novel antimicrobial peptides.     

Synergistic Biophysical Techniques Reveal Structural Mechanisms of Engineered Cationic Antimicrobial Peptides in Lipid Model Membranes

In the quest for new antibiotics, two novel engineered cationic antimicrobial peptides (eCAPs) have been rationally designed. WLBU2 and D8 (all 8 valines are the d -enantiomer) efficiently kill both Gram-negative and -positive bacteria, but WLBU2 is toxic and D8 nontoxic to eukaryotic cells. We explore protein secondary structure, location of peptides in six lipid model membranes, changes in membrane structure and pore evidence. We suggest that protein secondary structure is not a critical determinant of bactericidal activity, but that membrane thinning and dual location of WLBU2 and D8 in the membrane headgroup and hydrocarbon region may be important. While neither peptide thins the Gram-negative lipopolysaccharide outer membrane model, both locate deep into its hydrocarbon region where they are primed for self-promoted uptake into the periplasm. The partially α-helical secondary structure of WLBU2 in a red blood cell (RBC) membrane model containing 50 % cholesterol, could play a role in destabilizing this RBC membrane model causing pore formation that is not observed with the D8 random coil, which correlates with RBC hemolysis caused by WLBU2 but not by D8.     

Influence of the pore size of reversed phase materials on peptide purification processes

The influence of the pore size of a chromatographic reversed phase material on the adsorption equilibria and diffusion of two industrially relevant peptides (i.e. a small synthetic peptide and insulin) has been studied using seven different reversed phase HPLC materials having pore sizes ranging from 90 Å to 300 Å. The stationary phase pore size distribution was obtained by inverse size exclusion measurement (iSEC). The effect of the pore size on the mass transfer properties of the materials was evaluated from Van Deemter experiments. It has been shown that the lumped mass transfer coefficient increases linearly with the average pore size. The Henry coefficient and the impurity selectivity were determined in diluted conditions. The saturation capacity of the main peptides was determined in overloaded conditions using the inverse method (i.e. peak fitting). It was shown that the adsorption equilibria of the peptides on the seven materials is well described by a surface-specific adsorption isotherm. Based on this a lumped kinetic model has been developed to model the elution profile of the two peptides in overloaded conditions and to simulate the purification of the peptide from its crude mixture. It has been found that the separation of insulin from its main impurity (i.e. desamido-insulin) was not affected by the pore size. On the other hand, in the case of the synthetic peptide, it was found that the adsorption of the most significant impurity decreases with the pore size. This decrease is probably due to an increase in silanol activity with decreasing pore size.     

Membrane orientation of MSI-78 measured by sum frequency generation vibrational spectroscopy

Antimicrobial peptides (AMPs) selectively disrupt bacterial cell membranes to kill bacteria whereas they either do not or weakly interact with mammalian cells. The orientations of AMPs in lipid bilayers mimicking bacterial and mammalian cell membranes are related to their antimicrobial activity and selectivity. To understand the role of AMP-lipid interactions in the functional properties of AMPs better, we determined the membrane orientation of an AMP (MSI-78 or pexiganan) in various model membranes using sum frequency generation (SFG) vibrational spectroscopy. A solid-supported single 1,2-dipalmitoyl-an-glycero-3-[phospho-rac-(1-glycerol)] (DPPG) bilayer or 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (POPG) bilayer was used as a model bacterial cell membrane. A supported 1,2-dipalmitoyl-an-glycero-3-phosphocholine (DPPC) bilayer or a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) bilayer was used as a model mammalian cell membrane. Our SFG results indicate that the helical MSI-78 molecules are associated with the bilayer surface with ～70° deviation from the bilayer normal in the negatively charged gel-phase DPPG bilayer at 400 nM peptide concentration. However, when the concentration was increased to 600 nM, MSI-78 molecules changed their orientation to make a 25° tilt from the lipid bilayer normal whereas multiple orientations were observed for an even higher peptide concentration in agreement with toroidal-type pore formation as reported in a previous solid-state NMR study. In contrary, no interaction between MSI-78 and a zwitterionic DPPC bilayer was observed even at a much higher peptide concentration (～12,000 nM). These results demonstrate that SFG can provide insights into the antibacterial activity and selectivity of MSI-78. Interestingly, the peptide exhibits a concentration-dependent membrane orientation in the lamellar-phase POPG bilayer and was also found to induce toroidal-type pore formation. The deduced lipid flip-flop from SFG signals observed from lipids also supports MSI-78-induced toroidal-type pore formation.     

Mechanism of supported membrane disruption by antimicrobial peptide protegrin-1

While pore formation has been suggested as an important step in the membrane disruption process induced by antimicrobial peptides, membrane pore formation has never been directly visualized. We report on the dynamics of membrane disruption by antimicrobial peptide protegrin-1 (PG-1) on dimyristoyl-sn-glycero-phosphocholine-supported bilayer patches obtained via atomic force microscopy. The action of PG-1 is found to be concentration-dependent. At low PG-1 concentrations (1 < [PG-1] < 4 microg/mL), the peptide destabilizes the edge of the membrane to form fingerlike structures. At higher concentrations, PG-1 induces the formation of a sievelike nanoporous structure in the membrane. The highest degree of disruption is attained at concentrations >or=20 microg/mL, at which PG-1 disrupts the entire membrane, transforming it into stripelike structures with a well-defined and uniform stripe width. This first direct visualization of these membrane structural transformations helps elucidate the PG-1-induced membrane disruption mechanism.     

Biomolecular engineering by combinatorial design and high-throughput screening: small, soluble peptides that permeabilize membranes

Rational design and engineering of membrane-active peptides remains a largely unsatisfied goal. We have hypothesized that this is due, in part, to the fact that some membrane activities, such as permeabilization, are not dependent on specific amino acid sequences or specific three-dimensional peptide structures. Instead they depend on interfacial activity: the ability of a molecule to partition into the membrane-water interface and to alter the packing and organization of lipids. Here we test that idea by taking a nonclassical approach to biomolecular engineering and design of membrane-active peptides. A 16,384-member rational combinatorial peptide library, containing peptides of 9-15 amino acids in length, was screened for soluble members that permeabilize phospholipid membranes. A stringent, two-phase, high-throughput screen was used to identify 10 unique peptides that had potent membrane-permeabilizing activity but were also water soluble. These rare and uniquely active peptides do not share any particular sequence motif, peptide length, or net charge, but instead they share common compositional features, secondary structure, and core hydrophobicity. We show that they function by a common mechanism that depends mostly on interfacial activity and leads to transient pore formation. We demonstrate here that composition-space peptide libraries coupled with function-based high-throughput screens can lead to the discovery of diverse, soluble, and highly potent membrane-permeabilizing peptides.     

Haloduracin α binds the peptidoglycan precursor lipid II with 2:1 stoichiometry

The two-peptide lantibiotic haloduracin is composed of two post-translationally modified polycyclic peptides that synergistically act on gram-positive bacteria. We show here that Halα inhibits the transglycosylation reaction catalyzed by PBP1b by binding in a 2:1 stoichiometry to its substrate lipid II. Halβ and the mutant Halα-E22Q were not able to inhibit this step in peptidoglycan biosynthesis, but Halα with its leader peptide still attached was a potent inhibitor. Combined with previous findings, the data support a model in which a 1:2:2 lipid II:Halα:Halβ complex inhibits cell wall biosynthesis and mediates pore formation, resulting in loss of membrane potential and potassium efflux.     

Molecular design and synthesis of artificial ion channels based on cyclic peptides containing unnatural amino acids

A series of novel cyclic peptides composed of 3 to 5 dipeptide units with alternating natural-unnatural amino acid units, have been designed and synthesized, employing 5-(N-alkanoylamino)-3-aminobenzoic acid with a long alkanoyl chain as the unnatural amino acid. All cyclic peptides with systematically varying pore size, shape, and lipophilicity are found to form ion channels with a conductance of ca. 9 pS in aqueous KCl (500 mM) upon examination by the voltage clamp method. These peptide channels are cation selective with the permeability ratio P(Cl(-))/P(K(+)) of around 0.17. The ion channels formed by the neutral, cationic, and anionic cyclic peptides containing L-alanine, L-lysine, and L-aspartate, respectively, show the monovalent cation selectivity with the permeability ratio P(Na(+))/P(K(+)) of ca. 0.39. On the basis of structural information provided by voltage-dependent blockade of the single channel current of all the tested peptides by Ca(2+), we inferred that each channel is formed from a dimer of the peptide with its peptide ring constructing the channel entrance and its alkanoyl chains lining across the membrane to build up the channel pore. The experimental results are consistent with an idea that the rate of ion conduction is determined by the nature of the hydrophobic alkanoyl chain region, which is common to all the channels.     

Biomimetic Peptide‐Gated Nanoporous Membrane for On‐Demand Molecule Transport

The controllable molecule transport is crucial to realize many highly valuable applications both in vivo and in vitro. Nanoporous membranes, with nanoscopic pores, high porosity, uniform pore dimensions, and controllable surface chemical properties, hold tremendous potential to achieve this function. Herein, we report a nano‐gating system for on‐demand molecule transport based on a peptide‐gated nanoporous membrane. Acting as gatekeeper, the peptides introduced to the nanoporous membrane provide an opportunity to realize on‐demand on–off states via reversible conformational switching of the peptides. This nano‐gating system offers sustained release and can be used as a sophisticated molecule transport platform for localized drug delivery with a feedback function.     

Cell-Penetrating Peptides: Correlation between Peptide-Lipid Interaction and Penetration Efficiency

Cell-penetrating peptides are used in the delivery of peptides and biologics, with some cell-penetrating peptides found to be more efficient than others. The exact mechanism of how they interact with the cell membrane and penetrate it, however, remains unclear. This study attempts to investigate the difference in free energy profiles of three cell-penetrating peptides (TAT, CPP1 and CPP9) with a model lipid bilayer (DOPC) using molecular dynamics pulling simulations with umbrella sampling. Potential mean force (PMF) and free energy barrier between the peptides and DOPC are determined using WHAM analysis and MM-PBSA analysis, respectively. CPP9 is found to have the smallest PMF value, followed by CPP1 and TAT, consistent with the experimental data. YDEGE peptide, however, does not give the highest PMF value, although it is a non-cell-permeable peptide. YDEGE is also found to form water pores, alongside with TAT and CPP9, suggesting that it is difficult to distinguish true water pore formation from artefacts arising from pulling simulations. On the contrary, free energy analysis of the peptide-DOPC complex at the lipid-water interface with MM-PBSA provides results consistent with experimental data with CPP9 having the least interaction with DOPC and lowest free energy barrier, followed by CPP1, TAT and YDEGE. These findings suggest that peptide-lipid interaction at the lipid-water interface has a direct correlation with the penetration efficiency of peptides across the lipid bilayer.     

The interaction of a peptide with a scrambled hydrophobic/hydrophilic sequence (Pro-Asp-Ala-Asp-Ala-His-Ala-His-Ala-His-Ala-Ala-Ala-His-Gly) (PADH) with DPPC model membranes: a DSC study

Depending on their hydrophobicity, peptides can interact differently with lipid membranes inducing dramatic modifications into their host systems. In the present paper, the interaction of a synthetic peptide with a scrambled hydrophobic/hydrophilic sequence (Pro-Asp-Ala-Asp-Ala-His-Ala-His-Ala-His-Ala-Ala-Ala-His-Gly) (PADH) with 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) model membranes has been investigated by differential scanning calorimetry (DSC), adopting three different experimental approaches. In the first, the peptide is forced to be included into the hydrocarbon region of the lipid bilayer, by codissolving it with the lipid giving rise to mixed multilamellar vesicles–peptide systems; in the second, this system is passed through an extruder, thus producing large unilamellar vesicles–peptide systems; in the third, it is allowed to interact with the external surface of the membrane.

The whole of the DSC results obtained have shown that the incorporation of the peptide into the lipid bilayer by means of the first method induces a decrease in the enthalpy of the gel–liquid crystal transition of the membrane and a shift of the transition to the lower temperatures, thus resembling, in spite of its prevalently hydrophilic nature, the behavior of transbilayer hydrophobic peptides. The extrusion of these systems creates unilamellar vesicles free of peptides but of smaller size as evidenced by the decreased cooperativity of the transition. The peptide, added externally to the DPPC model membrane, has no effect on the phase behavior of the bilayer.

These findings suggest that the effect of the interaction of scrambled hydrophobic/hydrophilic peptides into lipid bilayers strongly affects the thermotropic behavior of the host membrane depending on the preparation method of the lipid/peptide systems. The whole of the results obtained in the present paper can be useful in approaching studies of bioactive peptides/lipids systems.     

Membrane binding and structure of de novo designed alpha-helical cationic coiled-coil-forming peptides.

We introduce a de novo designed peptide model system that enables the systematic study of 1) the role of a membrane environment in coiled-coil peptide folding, 2) the impact of different domains of an alpha-helical coiled-coil heptad repeat on the interaction with membranes, and 3) the dynamics of coiled-coil peptide-membrane interactions depending on environmental conditions. Starting from an ideal alpha-helical coiled-coil peptide sequence, several positively charged analogues were designed that exhibit a high propensity toward negatively charged lipid membranes. Furthermore, these peptides differ in their ability to form a stable alpha-helical coiled-coil structure. The influence of a membrane environment on peptide folding is studied. All positively charged peptides show strong interactions with negatively charged membranes. This interaction induces an alpha-helical structure of the former random-coil peptides, as revealed by circular dichroism measurements. Furthermore, vesicle aggregation is induced by a coiled-coil interaction of vesicle-bound peptides. Dynamic light scattering experiments show that the strength of vesicle aggregation increases with the peptide's intrinsic ability to form a stable alpha-helical coiled coil. Thus, the peptide variant equipped with the strongest inter- and intra-helical coiled-coil interactions shows the strongest effect on vesicle aggregation. The secondary structure of this peptide in the membrane-bound state was studied as well as its effect on the phospholipids. Peptide conformation within the peptide-lipid aggregates was analyzed by (13)C cross-polarization magic-angle spinning NMR experiments. A uniformly (13)C- and (15)N-labeled Leu residue was introduced at position 12 of the peptide chain. The (13)C chemical shift and torsion angle measurements support the finding of an alpha-helical structure of the peptide in its membrane-bound state. Neither membrane leakage nor fusion was observed upon peptide binding, which is unusual for amphiphatic peptide structures. Our results lay the foundation for a systematic study of the influence of the alpha-helical coiled-coil folding motif in membrane-active events on a molecular level.     

Molecular mechanisms of membrane perturbation by antimicrobial peptides and the use of biophysical studies in the design of novel peptide antibiotics

Antibiotic resistant bacterial strains represent a global health problem with a strong social and economic impact. Thus, there is an urgent need for the development of antibiotics with novel mechanisms of action. There is currently an extensive effort to understand the mode of action of antimicrobial peptides which are considered as one alternative to classical antibiotics. The main advantage of this class of substances, when considering bacterial resistance, is that they rapidly, within minutes, kill bacteria. Antimicrobial peptides can be found in every organism and display a wide spectrum of activity. Hence, the goal is to engineer peptides with an improved therapeutic index, i.e. high efficacy and target specificity. For the rational design of such novel antibiotics it is essential to elucidate the molecular mechanism of action. Biophysical studies have been performed using to a large extent membrane model systems demonstrating that there are distinctive different mechanisms of bacterial killing by antimicrobial peptides. One can distinguish between peptides that permeabilize and/or disrupt the bacterial cell membrane and peptides that translocate through the cell membrane and interact with a cytosolic target. Lantibiotics exhibit specific mechanisms, e.g. binding to lipid II, a precursor of the peptidoglycan layer, either resulting in membrane rupture by pore formation or preventing cell wall biosynthesis. The classical models of membrane perturbation, pore formation and carpet mechanism, are discussed and related to other mechanisms that may lead to membrane dysfunction such as formation of lipid-peptide domains or membrane disruption by formation of non-lamellar phases. Emphasis is on the role of membrane lipid composition in these processes and in the translocation of antimicrobial peptides.     

Heavy‐Atom Labeled Transmembrane β‐Peptides: Synthesis,CD‐Spectroscopy,and X‐ray Diffraction Studies in Model Lipid Multilayer

Transmembrane β‐peptides are promising candidates for the design of well‐controlled membrane anchors in lipid membranes. Here, we present the synthesis of transmembrane β‐peptides with and without tryptophan anchors, as well as a novel iodine‐labeled d ‐β³‐amino acid. By using one or more of the heavy‐atom labeled amino acids as markers, the orientation of the helical peptide was inferred based on the electron‐density profile determined by X‐ray reflectivity. The β‐peptides were synthesized through manual Fmoc‐based solid‐phase peptide synthesis (SPPS) and reconstituted in unilamellar vesicles forming a right‐handed 3₁₄‐helix secondary structure, as shown by circular dichroism spectroscopy. We then integrated the β‐peptide into solid‐supported membrane stacks and carried out X‐ray reflectivity and grazing incidence small‐angle X‐ray scattering to determine the β‐peptide orientation and its effect on the membrane bilayers. These β‐peptides adopt a well‐ordered transmembrane motif in the solid‐supported model membrane, maintaining the basic structure of the original bilayer with some distinct alterations. Notably, the helical tilt angle, which accommodates the positive hydrophobic mismatch, induces a tilt of the acyl chains. The tilted chains, in turn, lead to a membrane thinning effect.     

Inverted micelle formation of cell-penetrating peptide studied by coarse-grained simulation: importance of attractive force between cell-penetrating peptides and lipid head group

Arginine-rich peptide and Antennapedia are cell-penetrating peptides (CPPs) which have the ability to permeate plasma membrane. Deformation of the plasma membrane with CPPs is the key to understand permeation mechanism. We investigate the dynamics of CPP and the lipid bilayer membrane by coarse-grained simulation. We found that the peptide makes inverted micelle in the lipid bilayer membrane, when the attractive potential between the peptide and lipid heads is strong. The inverted micelle is formed to minimize potential energy of the peptide. For vesicle membrane, the peptide moves from the outer vesicle to the inner vesicle through the membrane. The translocation of the peptide suggests inverted micelle model as a possible mechanism of CPPs.     

Membrane interactions of host-defense peptides studied in model systems

Host-defense, antibiotic peptides are believed to generate their cytolytic effects by interacting with the membranes of bacterial cells. Direct analyses of peptide interactions with real cellular membranes are difficult, however, due to the high complexity of physiological membranes. This review summarizes experimental work aiming to understand peptide-membrane interactions and their relationships with the peptides' biological actions using specific model systems. Varied model assemblies have been constructed that generally aim to mimic the fundamental lipid bilayer organization of the membrane. The model systems we will describe include multilamellar and unilamellar vesicles, planar lipid bilayers, lipid monolayers and micelles, and colorimetric biomimetic membranes. The different artificial models have facilitated examination of specific biological or chemical parameters affecting peptide action, for example the effect of membrane lipid composition on peptide affinities and membrane penetration, the relationship between membrane fluidity and peptide interactions, the conformations of active peptides, and other factors. We evaluate the strengths and limitations of the various approaches, and point to future directions in the field.     

Phase behavior and the partitioning of caveolin-1 scaffolding domain peptides in model lipid bilayers

The membrane binding and model lipid raft interaction of synthetic peptides derived from the caveolin scaffolding domain (CSD) of the protein caveolin-1 have been investigated. CSD peptides bind preferentially to liquid-disordered domains in model lipid bilayers composed of cholesterol and an equimolar ratio of dioleoylphosphatidylcholine (DOPC) and brain sphingomyelin. Three caveolin-1 peptides were studied: the scaffolding domain (residues 83-101), a water-insoluble construct containing residues 89-101, and a water-soluble construct containing residues 89-101. Confocal and fluorescence microscopy investigation shows that the caveolin-1 peptides bind to the more fluid cholesterol-poor phase. The binding of the water-soluble peptide to lipid bilayers was measured using fluorescence correlation spectroscopy (FCS). We measured molar partition coefficients of 10(4) M(-1) between the soluble peptide and phase-separated lipid bilayers and 10(3) M(-1) between the soluble peptide and bilayers with a single liquid phase. Partial phase diagrams for our phase-separating lipid mixture with added caveolin-1 peptides were measured using fluorescence microscopy. The water-soluble peptide did not change the phase morphology or the miscibility transition in giant unilamellar vesicles (GUVs); however, the water-insoluble and full-length CSD peptides lowered the liquid-liquid melting temperature.     

Phosphate-mediated arginine insertion into lipid membranes and pore formation by a cationic membrane peptide from solid-state NMR

The insertion of charged amino acid residues into the hydrophobic part of lipid bilayers is energetically unfavorable yet found in many cationic membrane peptides and protein domains. To understand the mechanism of this translocation, we measured the (13)C-(31)P distances for an Arg-rich beta-hairpin antimicrobial peptide, PG-1, in the lipid membrane using solid-state NMR. Four residues, including two Arg's, scattered through the peptide were chosen for the distance measurements. Surprisingly, all residues show short distances to the lipid (31)P: 4.0-6.5 A in anionic POPE/POPG membranes and 6.5-8.0 A in zwitterionic POPC membranes. The shortest distance of 4.0 A, found for a guanidinium Czeta at the beta-turn, suggests N-H...O-P hydrogen bond formation. Torsion angle measurements of the two Arg's quantitatively confirm that the peptide adopts a beta-hairpin conformation in the lipid bilayer, and gel-phase 1H spin diffusion from water to the peptide indicates that PG-1 remains transmembrane in the gel phase of the membrane. For this transmembrane beta-hairpin peptide to have short (13)C-(31)P distances for multiple residues in the molecule, some phosphate groups must be embedded in the hydrophobic part of the membrane, with the local (31)P plane parallel to the beta-strand. This provides direct evidence for toroidal pores, where some lipid molecules change their orientation to merge the two monolayers. We propose that the driving force for this toroidal pore formation is guanidinium-phosphate complexation, where the cationic Arg residues drag the anionic phosphate groups along as they insert into the hydrophobic part of the membrane. This phosphate-mediated translocation of guanidinium ions may underlie the activity of other Arg-rich antimocrobial peptides and may be common among cationic membrane proteins.     

Molecular view of the role of fusion peptides in promoting positive membrane curvature

Fusion peptides are moderately hydrophobic segments of viral and nonviral membrane fusion proteins that enable these proteins to fuse two closely apposed biological membranes. In vitro assays furthermore show that even isolated fusion peptides alone can support membrane fusion in model systems. In addition, the fusion peptides have a distinct effect on the phase diagram of lipid mixtures. Here, we present molecular dynamics simulations investigating the effect of a particular fusion peptide, the influenza hemagglutinin fusion peptide and some of its mutants, on the lipid phase diagram. We detect a systematic shift toward phases with more positive mean curvature in the presence of the peptides, as well as an occurrence of bicontinuous cubic phases, which indicates a stabilization of Gaussian curvature. The wild-type fusion peptide has a stronger effect on the phase behavior as compared to the mutants, which we relate to its boomerang shape. Our results point to a different role of fusion peptides than hitherto assumed, the stabilization of pores rather than stalks along the fusion pathway.