The antimicrobial peptide Polybia-MP1 differentiates membranes with the hopanoid, diplopterol from those with cholesterol

Highlights

• •Membrane poration upon MP1 addition is not prevented by diplopterol (DP)
• •Relaxation times for membranes with DP and MP1 first increase and then decrease
• •Z-potential of DP-containing membranes decreases for low MP1 concentrations
• •Z-potential of DP-containing membranes increases for high MP1 concentrations

Graphical abstract

Abstract

Polybia-MP1 is an antimicrobial peptide that shows a decreased activity in membranes with cholesterol (CHO). Since it is now accepted that hopanoids act as sterol-surrogates in some sterol-lacking bacteria, we here inquire about the impact of Polybia-MP1 on membranes containing the hopanoid diplopterol (DP) in comparison to membranes with CHO. We found that, despite the properties induced on lipid membranes by DP are similar to those induced by CHO, the effect of Polybia-MP1 on membranes with CHO or DP was significantly different. DP did not prevent dye release from LUVs, nor the insertion of Polybia-MP1 into monolayers, and peptide-membrane affinity was higher for those with DP than with CHO. Zeta potentials ($\zeta$) for DP-containing LUVs showed a complex behavior at increasing peptide concentration. The effect of the peptide on membrane elasticity, investigated by nanotube retraction experiments, showed that peptide addition softened all membrane compositions, but membranes with DP got stiffer at long times. Considering this, and the $\zeta$ results, we propose that peptides accumulate at the interface adopting different arrangements, leading to a non-monotonic behavior. Possible correlations with cell membranes were inquired testing the antimicrobial activity of Polybia-MP1 against hopanoid-lacking bacteria pre-incubated with DP or CHO. The fraction of surviving cells was lower in cultures incubated with DP compared to those incubated with CHO. We propose that the higher activity of Polybia-MP1 against some bacteria compared to mammalian cells is not only related to membrane electrostatics, but also the composition of neutral lipids, particularly the hopanoids, could be important.

1. Introduction

Cell membranes are simultaneously compact and fluid. These properties, which are conserved in different species, have been proposed to be those that are compatible with life [1], and it is accepted that sterols are inducers of these particular mechanical conditions in different kingdoms [2], [3], [4], [5]. It has long been known that prokaryotes generally lack cholesterol, but their mechanical properties are similar to those found in eukaryotic membranes [6], [7]. Some bacteria produce hopanoids, a group of anaerobically evolved triterpene derivatives. Assays using hopanoid-deficient bacteria strains of species that normally synthesize these molecules suggest that they promote resistance to antibiotics, detergents, extreme pH, high temperature, oxidation, high osmolarity, and oxygen permeation [8], [9], [10], [11], [12], [13], [14], [15], [16]. It is proposed that hopanoids promote stress resistance by increasing membrane order leading to a decrease in membrane permeability. In this regard, artificial membranes and computational simulations have been employed to gather evidence that hopanoids play a role similar to that of sterols, being thus key components for regulating membrane mechanical properties [6], [8], [9], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28]. Polybia-MP1 (MP1 from now on) is an antimicrobial peptide (AMP) originally found in the venom of the Brazilian wasp Polybia paulista, with sequence IDWKKLLDAAKQIL-NH${}_2$. This AMP has a broad-spectrum bactericidal activity against both Gram-positive and Gram-negative bacteria without being hemolytic or cytotoxic [29]. The mechanism of action of this peptide has been studied using artificial membranes, and it has been shown that cholesterol (CHO) hinders membrane poration and disruption by MP1 [30], [31]. Given the similar properties of artificial membranes with hopanoids and those with sterols, and the antibiotics resistance promoted by hopanoids, we inquired whether diplopterol (DP, an abundant hopanoid) protects membranes from the action of MP1 as CHO does.

Using monolayers and bilayers as model membranes, we investigated how peptide-membrane affinity, peptide-lytic activity, and peptide effects on membrane properties are affected by the presence of DP compared to that of CHO. In order to directly answer our question, we decided to use a simple membrane composition: we studied binary mixtures with CHO or DP and a phospholipid, and pure phospholipid membranes. We have chosen POPC (2-oleoyl-1-palmitoyl-sn-glycero-3-phosphocholine) as the phospholipid because it is a widely studied lipid that forms liquid-disordered membranes, being a good model for fluid biological membranes. We are aware that the selected lipid mixtures are not the best model for bacterial membranes, but our aim is not to resemble bacteria and mammalian cells but to detect differences and similarities between DP-containing membranes compared to those with CHO. The proportion of DP and CHO was 20$\%,$ and also 40$\%$ for some experiments. This is again, not comparable to bacteria membrane proportions, but allows to test dose-dependent phenomena. In this context, it is important to recall that local membrane composition in cells can be very different from the global average that is usually determined experimentally [32].

Our results show that MP1 distinguishes DP-containing membranes from those with CHO. Changes in surface pressure upon peptide insertion were higher in monolayers with DP than in those with CHO, DP was not able to preclude membrane permeation as CHO did, peptide affinity to DP-containing vesicles was higher than to CHO-containing ones, and the peptide translocated faster bilayers with DP than those with CHO. Zeta-potential measurements indicated a decrease in the surface potential of vesicles with DP incubated with low MP1 concentration and an increase at higher peptide concentrations. The relaxation process of highly deformed vesicles with DP got slower as soon as the peptide was added but returned to the initial behavior at longer times. Both nonlinear phenomena (in the zeta potential and dynamic elasticity) were not observed in membranes containing CHO.

Given our results using artificial membranes, we wonder whether the differences observed on DP or CHO-containing membranes upon peptide addition could be detected in a cell membrane, with a lot of components and the complexity related to life. Therefore, we performed exploratory assays using Pseudomonas aeruginosa, which is a hopanoid-lacking and sterol-lacking bacteria. Previous studies performed in our lab using this bacterium have shown that the exposition of the microorganism to DP decreases the permeability and increases the resistance to some antibiotics [26]. Therefore, we incubated the bacteria with DP or CHO in the growth medium and exposed them to MP1. Cell resistance against MP1 was higher for bacteria incubated with CHO than for those incubated with DP. This suggests that the presence of DP in the cell membrane is not as effective as CHO for protecting the bacteria from the action of this peptide.

2. Methods and materials

2.1. Materials

Lipids: 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocho- line (POPC), cholesterol (CHO), 1,2-dioleoyl-sn-glycero-3-phos- phoethanolamine-N-(lissamine rhodamine B sulfonyl) (PE-Rho), and 1,2-distearoyl-sn-glycero-3-phospho- ethanolamine-N-[biotinyl(polyethylene glycol)-2000] (DSPE-PEG (2000)-Biotin) were purchased from Avanti Polar Lipids (Alabaster-Al-USA). Diplopterol (DP) was purchased from Chiron (Norway). Streptavidin-coated micro-beads (diameter of 3 µm) were purchased from Bangs Laboratories Inc. (Fishers, IN, USA). MP1 peptide and MP1 with fluorescein isothiocyanate (fMP1) attached to the N-terminal were acquired from BioSynthesis (Lewisville-TX-USA) with RP-HPLC purity level $>98$%. Chloroform HPLC grade was obtained from Merck (Darmstadt, Germany). Sodium chloride, sodium hydroxide, sucrose, HEPES, Triton X-100, carboxyfluorescein (CF), and avidin were from Sigma (St. Louis-MO-USA). Water used to prepare solutions was deionized with a resistivity of  18 M $\Omega$cm, filtered with an Osmoion system (Apema, BA, Quilmes, Argentina).

2.2. Experiments using lipid monolayers

The insertion of peptides into lipid films composed of pure POPC, and POPC mixed with CHO or DP (8:2 and 6:4 molar ratio) was assessed in experiments at constant area using a home-made circular trough (volume 1 mL, surface area 7 cm${}^2$). Lipid monolayers were prepared by spreading a lipid solution (in chloroform) onto the surface of a HEPES buffer (20 mM HEPES, 150 mM NaCl, pH 7.4) up to the desired surface pressure (${\pi }_i$). An aliquot of concentrated MP1 solution was injected in the subphase above the previously prepared monolayers, through a hole in the wall of the trough. The final peptide concentration was 0.6 µM. The adsorption kinetics of MP1 was followed by the increase in surface pressure ($\Delta \pi$) as a function of time (Figure S1-A). In order to obtain the maximum insertion pressure (MIP, surface pressure above which no more peptide molecules penetrate the lipid monolayer), $\Delta \pi$ induced by 0.6 µM of MP1 as a function of different values of ${\pi }_i$ was determined (Figure S1-B). All experiments were carried out at 25 ${}^{\circ }$C in HEPES buffer (20 mM HEPES, 150 mM NaCl, pH 7.4). The standard deviations were obtained from at least three independent measurements.

Surface pressure changes due to a molecule adsorption depend on the amount of incorporated molecules, on the extent of perturbation induced to the lipids, and on the stiffness of the host monolayer. In order to get rid of this last factor, we calculated the area change supported by the lipid monolayer due to peptide insertion. This was performed assuming that the film can be tessellated into two types of regions. The first ones are defined by regions where the peptide penetrates and disrupts the monolayer structure, forming a lipid-peptide aggregate. The second ones are formed by pure lipids. According to this model, the pure lipid regions get progressively compressed as the peptide penetrates and the peptide-lipid aggregate is formed. $\Delta$A corresponds then to the change in the area per lipid, within the pure lipid regions, which can be estimated from the mean molecular areas at the final and initial surface pressures, as read from the compression isotherms of the lipid monolayers [33], [34].

Compression isotherms of the lipid films composed of pure POPC and POPC mixed with CHO or DP (8:2 and 6:4 molar ratio) were carried out in a NIMA 102M (NIMA Technology, Coventry, UK) Langmuir balance enclosed in an acrylic box, using a Pt plate by the Wilhelmy method. Lipid monolayers were prepared by spreading the lipid solution in chloroform onto the surface of HEPES buffer (20 mM HEPES, 150 mM NaCl, pH 7.4). After 10 min to allow organic solvent evaporation, the monolayers were compressed at 5 mm/min. All measurements were performed at 25 ${}^{\circ }$C.

2.3. Experiments using large unillamelar vesicles (LUVs)

2.3.1. Preparation of LUVs

Multilamelar vesicles (MLVs) of pure POPC or containing 20 or 40$\%$ of CHO or DP were obtained from lipids dissolved in chloroform in round-bottom flasks. The solvent was dried under a stream of nitrogen and stored under vacuum for 3 h to remove the organic solvent. Afterward, the lipid films were hydrated with a carboxyfluorescein (CF)-containing buffer (25 mM CF) and subjected to intense vortexing and several freeze-thaw cycles. LUVs were formed by extrusion of MLVs suspension, 15 times through a polycarbonate membrane of 0.1 µm pore size. The free dye was removed by exclusion chromatography using a Sephadex G-25M column (Amersham Pharmacia, Uppsala, Sweden) eluting the samples with HEPES buffer (20 mM HEPES, 150 mM NaCl, pH 7.4). To avoid osmotic pressure effects, the osmolarity parity of CF and HEPES solutions was checked with an automatic micro-osmometer OM-806 (Vogel, Germany). The size distribution was confirmed by Dynamic Light Scattering (DLS) using a submicron particle sizer (Nicomp 380). The average diameter of the liposomes is shown in Figure S2. In order to determine the actual amount of phospholipid after extrusion, Bartlett’s method was used [35]. Briefly, phospholipids were digested, and inorganic phosphorus was quantified colorimetrically. Absorbance was measured with a Shimadzu UV-visible Spectrophotometer (Biospec-mini, Chiyoda-ku, Tokyo, Japan) at $\lambda =830$ nm.

2.3.2. Dye release from CF-loaded LUVs

CF release from the vesicles (100 µM total lipid concentration) was monitored detecting the dye emission at 520 nm (excited at 490 nm) on a Fluoromax-P spectrofluorometer (Horiba Jobin-Yvon). The maximum fluorescence intensity (100%) was determined by adding 10% Triton X-100 solution into a cuvette with the liposomes (complete membrane solubilization). The percentage of dye release induced by different MP1 concentrations was calculated at regular time intervals using the Eq. (1) :

$$\%C{F}_{release}=100(F_t-F_0)/(F_{100}-F_0)$$ (1)

where F${}_t$ is the observed fluorescence intensity, F${}_0$ and F${}_{100}$ correspond to the fluorescence intensity in the absence of peptide, and for 100$\%$ leakage, respectively [36]. The experiments were performed at 25 ${}^{\circ }$C. In each experiment, vesicles were used the same day of preparation. Standard deviations were obtained from three independent experiments.

2.3.3. Zeta potential measurements

Zeta potential ($\zeta$) of LUVs (40 µM total lipid concentration) incubated with the peptide at increasing peptide-to-lipid molar ratio ([P]/[L]) were determined using a Z-sizer SZ-100-Z equipment (Horiba, Ltd., Kyoto, Japan). $\zeta$ of a vesicle suspension was calculated from electrophoretic mobility ($\mu$) using Henry’s relation (Eq. (2)):

$$\zeta =\frac{3\mu {\eta }_w}{2{\epsilon }_r{\epsilon }_{0}f\left(\kappa R\right)}$$ (2)

where ${\epsilon }_r$ and ${\epsilon }_0$ are the dielectric constant of water (78.5) and the vacuum permittivity, respectively, ${\eta }_w$ is the water viscosity, R is the average vesicle radius, $\kappa$ is the inverse of Debye length and f($\kappa$R) in this case is 1.5 considering the Smoluchowsky approximation [37]. Zeta potential represents the charge at the shear plane, however, at low ionic strength (up to 0.1 M) it provides a good approximation for the surface potential [38]. For these measurements, 15 mM NaCl was used as the aqueous milieu. All experiments were performed at 25 ${}^{\circ }$C.

2.4. Experiments using giant unillamelar vesicles (GUVs)

2.4.1. Preparation of GUVs

GUVs were prepared by the electroformation technique as described by Angelova and Dimitrov [39] using a home-made wave generator and a chamber with stainless steel electrodes [40]. Briefly, 7 µL of a 0.5 mM lipid solution (in chloroform) was spread on two stainless steel electrodes and were left under vacuum for at least 3 h to remove all traces of the organic solvent. The mixtures contained the desired lipid composition plus 0.5 mole$\%$ PE-Rho and 0.0001 mole$\%$ DSPE-PEG (2000)-Biotin. The lipid films were hydrated by filling a home-made chamber of acrylic containing the electrodes with a 0.3 M sucrose solution. The electrodes were connected to the wave generator, and a sinusoidal tension of  1-2 V amplitude and 10 Hz frequency were applied for 1–3 h at 25 ${}^{\circ }$C for POPC vesicles, and 60 ${}^{\circ }$C for vesicles with CHO or DP. GUVs were then harvested and added into the corresponding chamber containing an iso-osmolar HEPES buffer (20 mM HEPES + NaCl approximately 150 mM, pH 7.4) for further experiments. To avoid osmotic pressure effects, osmolarity parity of inside and outside solutions was checked with an automatic micro-osmometer OM-806 (Vogel, Germany).

2.4.2. Peptide interaction with GUVs

GUVs were directly observed under an inverted confocal laser-scanning microscope (Olympus Fluorview FV1000, Tokyo, Japan). For this, GUVs were suspended in 250 µL of the iso-osmolar HEPES buffer in an 8-well Lab-Tek Chamber. Then, confocal time-series images taken with an oil immersion 60X objective (NA: 1.4, Olympus), were sequentially acquired with two lasers (Argon - 488 and Helium-Neon - 543 nm) before and after the addition of fluorescently-labeled MP1 (fMP1) with the help of a micropipette to yield the desired peptide concentration. Peptide surface excess on the bilayers was estimated from the fluorescence intensity at the GUV’s rim in the green channel. This intensity was determined as a function of time by drawing a line at the vesicle contour (Figure S3). This intensity was normalized by the bulk intensity outside the GUV under study.

To determine peptide penetration into the vesicle, the fluorescence intensity at the GUV interior was measured using a small circle in the center of the GUV as exemplified in Figure S3. The fluorescence intensities were measured using the NIH free ImageJ/FIJI software. The amount of GUVs that had fMP1 in their interior was determined at different times after peptide addition, and the fraction of GUVs in which peptide had entered ($f_e$) was calculated at each time. The analysis was performed using at least 25 GUVs in three independent experiments.

When an active peptide (that forms pores and permeates into the vesicle lumen) is added to the aqueous solution, it is expected to diffuse to the vesicle, and then adsorb at the membrane surface. Eventually, it penetrates into the membrane and forms dimers, multimers, pores, and/or defects. Finally, the peptide desorbs into the aqueous solution into the interior of the vesicle. In our experiments, the different peptide species in the membrane (dimers, multimers, pores) cannot be distinguished since we are following the peptide fluorescence at a micrometer-resolution, and only the total peptide concentration in the membrane ($C_M$) can be ascertained through the fluorescence intensity at the membrane rim. We will consider that initially, the reaction occurs in a unidirectional fashion: peptides adsorb (without desorption), accumulate in the membrane forming dimers, multimers, or pores, and thereafter, the peptide aggregates disassemble and desorb (without readsorption) mainly in the vesicle lumen, where their concentration ($C_i$) is initially zero. In other words, initially (for low peptide concentration in the vesicle lumen) we can neglect the backward reactions, and only consider adsorption/incorporation of peptides from outside the vesicle to the membrane, formation and disassemble of peptide aggregates, and desorption from the membrane to the vesicle lumen. At these conditions, the following holds (Eq. (3)):

$$\frac{d{C}_M}{dt}=k_a{C}_o-k_d{C}_M$$ (3)

Here, $C_M$ includes all possible peptide species in the membrane, $C_o$ is the peptide concentration outside the GUV, and $k_a$ is the rate constant for the processes of peptide adsorption from outside the vesicle and $k_d$ is the rate constant for peptide desorption into the vesicle interior. Since the local peptide concentration is proportional to the local fluorescence intensity (neglecting changes in the quantum yield of FITC with the environment), we can fit the data from fluorescence intensity at intermediate times (before $C_i$ reached a constant value, and after $C_o$ remained constant to avoid the diffusional regime) using Eq. (4):

$$\frac{d{F}_M}{dt}\frac{1}{F_M}=k_a\frac{F_o}{F_M}-k_d$$ (4)

Here, $F_M$ and $F_o$ are the fluorescence intensities of the peptide at the vesicle rim and outside the vesicle, respectively. Using Eq. (4), the fluorescence data for experiments with 0.6 µM fMP1 were fitted (see a representative plot in Figure S6-A). Higher peptide concentrations were not used due to peptide-lipid aggregation at the micrometer scale, membrane fluctuations, and vesicle rupture. The data for POPC + 40% CHO did not allow fitting due to the low fluorescence signal at the vesicle rim. From the obtained values, the ratio between the rate constants was determined $K_{ef}=\frac{k_a}{k_d}$. $K_{ef}$ gives a measure of the ability of the peptides to accumulate in the membrane.

2.4.3. Analysis of GUV’s shape

Images were collected using two channels: red, to visualize the membrane marker (PE-Rho), and green for the FITC-labeled peptide. Vesicle shape at the equator was analyzed using NIH ImageJ/FIJI software through the determination of the aspect ratio (AR, the ratio between the lengths of the major and minor axis) as a function of time, see Figure S8. Figures S9, S10, and S11 show a representative GUV’s sequence for POPC, POPC + 20% CHO, and POPC + 20% DP, respectively.

2.4.4. Membrane tether retraction assays

Nanotubes were generated by means of an optical trap. The optical tweezers setup has been described previously [41]. Briefly, a 3W power ND:YVO4 laser (Spectra-Physics, Santa Clara, CA, USA) was focused through a $100\times$ objective (water immersion, NA = 1, Zeiss). The tweezers were mounted in an Axiovert 200 inverted microscope (Zeiss) equipped with a motorized stage (Beijing Winner Optical Ins. CO., LTD, Langfang, Hebei, China), and the experiments were registered with a fast CCD video camera (Ixon EM+ model DU-897, Andor Technology, Abingdon, Oxfordshire, England).

The experiments were performed in a Teflon homemade chamber with a coverslip (Marienfeld Superior, Germany) as an optical window. The coverslip was treated with an avidin solution (0.5 mg/mL for 30 min at 10 ${}^{\circ }$C). The chamber was placed on the motorized stage of the microscope. A 3.0 µm streptavidin-coated bead solution was added to the chamber containing the iso-osmolar HEPES buffer. After incubation for 10 min, a GUV solution was added to the chamber and left to attach to the avidin-coated glass due to avidin-biotin interactions. A bead was trapped with the optical tweezers and was brought into contact with the surface of a proximal GUV. Bead-membrane adhesion was achieved by means of streptavidin-biotin interactions. Once the bead adhered to the membrane, a membrane tether was pulled at a controlled speed of 15 µm/s up to a length of 40 ($\pm$5) µm. At $t=0$ the laser was switched off and tether retraction was registered by contrast microscopy at a rate of 24 frames/s. Movie S2 shows an example. A representative image sequence is shown in Figure S12.

Tether retraction was followed through the bead motion. The equation of motion of the bead can be obtained after applying Newton’s second law (Eq. (5)):

$$F_{trap}=F_{\gamma }+F_{in}+F_{el}$$ (5)

where $F_{\gamma },$ $F_{in},$ and $F_{el}$ are the viscous, inertial, and elastic forces, respectively. $F_{\gamma }$ is the force that opposes motion, and includes the dissipation of the bead in the aqueous media and membrane dissipation. Bead dissipation can be computed through the Stokes law (Eq. (6)):

$$F_{\gamma }=\gamma \dot{L}$$ (6)

where $\dot{L}$ is the bead speed and $\gamma =6\pi \eta R_b=28nNs/m$ ($\eta$ is the solution viscosity, and $R_b,$ the radius of the bead). Membrane dissipation has the contribution of shear (${\eta }_s$) and of the friction between leaflets (${\eta }_c$), and the sum of ${\eta }_s$ and ${\eta }_c$ is in the range of 1–100 nN s/m [42], [43]. The inertial term can be written as $F_{in}=m\ddot{L},$ and can be neglected for low Reynolds numbers [44]. Finally, the driving force for retraction that corresponds to membrane elasticity for a membrane with elastic constant $k$ can be described by Eq. (7):

$$F_{el}=kL$$ (7)

When the trap is turned off, $F_{\gamma }=-F_{el},$ leading to:

$$L=L_{0}exp\left(\frac{-kt}{\gamma }\right)$$ (8)

where $L_0$ is the initial tether length (40 µm). Since we cannot split the contribution of membrane dissipation from membrane elasticity, the characteristic time $\tau =\frac{\gamma }{k}$ is obtained from the fitting process. The used fitting equation therefore is Eq. (9):

$$L=L_{0}exp(-\frac{t}{\tau })$$ (9)

2.5. In vivo assay: antimicrobial activity against Pseudomonas aeruginosa

The bacterial strain used in this study was P. aeruginosa PAO1. Bacteria were cultured on Luria-Bertani (LB) agar plates from frozen stocks, and a single colony was grown in LB liquid medium at 37 ${}^{\circ }$C under shaking at 180 rpm overnight, up to an optical density of 1--1.2 at 600 nm. Cell suspensions were prepared in 0.15 M NaCl solutions to $2\times {10}^4$ cells/mL. 14.4 µL of pure chloroform, or 6 mM solution of CHO or DP dissolved in chloroform, were added to 120 µL of the medium and left for 4 h at 37 ${}^{\circ }$C for solvent evaporation. After that, 30 µL of the media with lipids or solvent were added to 40 µL of cell suspensions and incubated for 2 h at 37 ${}^{\circ }$C without aeration. It has been reported that P. aeruginosa have the ability to incorporate exogenous fatty acids [45], and the method described here has been previously employed for the incorporation of DP and CHO [25]. To evaluate susceptibility to MP1, cells were subsequently treated with 15 or 30 µg/mL of MP1 and incubated for 3 h at 37 ${}^{\circ }$C in a 96-multiwell plate. Serial dilutions were plated on LB plates to quantify colony-forming units (CFU). The fraction of surviving cells was estimated from the ratio of colony-forming units (CFU) in the presence of peptide to that in its absence (control cells) as shown in Figure S14. The informed data correspond to the average of three independent experiments.

3. Results and discussion

DP increases the perturbation induced by MP1 on lipid monolayers

The process of insertion of MP1 into the external hemilayer of membranes was inquired with Langmuir monolayers as model. We tested peptide concentrations lower than the MIC values (5–10 µM depending on bacteria [29]) in order to study non-lethal concentrations. It is important to state that in MIC-determination assays, peptides may interact with different species in the culture, and thus, the effective peptide concentration reaching the cell membranes is unknown and probably lower than the informed MIC. Thus, we started with concentrations as low as 0.6 µM with membranes composed of POPC or POPC + 20$\%$ DP or CHO. For 0.6 µM peptide concentration, the maximum insertion pressure (MIP) was similar for all tested monolayers, although slightly higher for monolayers with 20$\%$ DP than those with the same amount of CHO (Figure S1-B). Given the small differences found in MIP, we decided to test films with higher amounts of CHO or DP at 30 mN/m, where the monolayer lipid compaction is comparable to that in bilayers [46], [47]. Fig. 1A shows the area change promoted in the lipid monolayer upon peptide penetration (see the changes in surface pressure in Figure S1-C). Our results showed that for higher CHO levels, MP1-induced perturbation on lipid-packing decreased. Contrary to this, the increase in DP levels promoted larger area changes. Molecules that incorporate into monolayers may form interfacial structures with the lipids [34], [48], [49] and/or modify the film properties, depending on the host film [50], [51], [52], [53]. Our results indicate that DP enhances the sensitivity of monolayers to AMP. To test whether this is also the case for vesicles, we next investigated MP1-induced vesicle leakage.

Fig. 1.

(A) Area change supported by lipid monolayers due to peptide insertion. Film composition: pure POPC (black), POPC + 20$\%$ (blue) or 40$\%$ (cyan) of CHO, or POPC + 20$\%$ (olive) or 40$\%$ (green) of DP. All data correspond to the average ($\pm$ SD) of at least three independent experiments. * indicates significant statistical differences determined by one-way ANOVA, Tukey’s test at $p<0.05$. (B) Lytic activity of the peptide. Fluorescence of released CF from LUVs composed of pure POPC (black), POPC + 20% CHO (blue), or POPC + 20% DP (olive). The vertical lines represent the [P]/[L]${}_{50}$ values. All data correspond to the average ($\pm$ SD) of three independent experiments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

MP1-induced dye release from liposomes is not affected by DP

The loss of vesicle content promoted by MP1 was assessed by recording the fluorescence of carboxyfluorescein (CF) released from large unilamellar vesicles (LUVs) exposed to peptide concentrations in the range 0.6–10 µM. MP1-induced leakage of POPC and POPC + 20% DP LUVs occurred without changes in the vesicle diameter, while those with 20% CHO formed larger aggregates at high peptide/lipid molar ratios ([P]/[L]) (see Figure S2).

The [P]/[L] value that induced 50% of CF release ($\left[P\right]/{\left[L\right]}_{50}$) was 3-fold higher for membranes with 20% CHO than for LUVs of pure POPC (see vertical lines in Fig. 1B). This has been previously related to a decrease in the probability of pore/defects formation by the AMP in CHO-containing membranes [54], [55]. Opposite to the CHO-induced resistance, $\left[P\right]/{\left[L\right]}_{50}$ for liposomes composed of POPC + 20% DP was similar to those of pure POPC. Interestingly, monolayer experiments indicated that MP1 promoted larger perturbations on DP-containing films than pure POPC, whereas $\left[P\right]/{\left[L\right]}_{50}$ did not show differences.

The fluorescently-labeled MP1 distinguishes membranes with DP from those with CHO

The interaction of the different membranes with the peptide was also studied by means of fluorescence microscopy, using the peptide labeled with fluorescein isothiocyanate (fMP1). It is important to note here that fMP1 and MP1 are different chemical species and thus, a direct comparison of the results from this section with the other results cannot be performed. We instead will try to correlate all the results.

Fluorescence of fMP1 was recorded during peptide binding at the rim of GUVs, and outside and inside them as shown in Figure S3. Fig. 2 shows the results for a POPC GUV, and other compositions are shown in Figures S4 and S5. Rim’s fluorescence was normalized by the fluorescence outside each GUV ($F_M$/$F_0$) to get an insight into the amount of accumulated peptide on the surface. Considering a negligible change of the emission spectra of the fluorophore in different environments, values of $F_M$/$F_0$ higher than one indicate that fMP1 concentration at the membrane is higher than outside the vesicle, thus peptide accumulates at the membrane. $F_M$/$F_0$ increased up to a maximal value at steady state, which is shown in Fig. 3A. $(F_M$/$F_0){}_{\max }$ was the smallest for CHO-containing vesicles, in agreement with the lowest $\Delta A$ values (Fig. 1A). Interestingly, two different scenarios were observed in GUVs containing 20% CHO: the peptide did not adsorb to some GUVs but accumulated on the surface of others, followed by peptide entry into the lumen (Figure S5).

Fig. 2.

(A)Kinetic curves for the fluorescence intensity signal of MP1 labeled with FITC (fMP1) at the rim (magenta), at the lumen (orange), and outside a vesicle (violet) normalized by the final fluorescence outside the vesicle. GUV composition: POPC + 0.5 mol% of PE-Rho in the presence of 0.6 µM fMP1. (B) representative confocal images at selected times after peptide addition. Upper sequence: red channel (PE-Rho), lower sequence: green channel (fMP1). Scale bar: 10 µm. For better visualization of the images, the brightness and contrast range was reduced from an original range of 0--255 to 0--132. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3.

(A) Final fluorescence intensity of fMP1 at the membrane rim normalized by the intensity outside the GUVs ($(F_M$/$F_0){}_{\max }$) (filled bars, left scale); and ratio between $k_a$ and $k_d$ ($K_{ef}$), see text for details (open bars, right scale). Data correspond to the average ($\pm$ SD) of at least 30 GUVs for $(F_M$/$F_0){}_{\max }$) or 15 for $K_{ef}$. (B) Times for achieving half the value of $(F_M$/$F_0){}_{\max }$ at the vesicle’s rim (filled bars), and times at which peptide fluorescence is detected at the lumen of 50% of the vesicles (open bars). All data correspond to the average ($\pm$ SD) of at least 45 GUVs. Different letters indicate significant differences analyzed by one-way ANOVA and Tukey’s test at $p<0.05$.

In the presence of DP, $(F_M$/$F_0){}_{\max }$ values were higher than those for membranes with the same amount of CHO but smaller than for pure POPC membranes. This result, together with those in Fig. 1A showing that monolayer perturbation increased when DP was present, may explain the lack of influence of DP on the peptide-induced leakage from LUVs: the presence of DP increases membrane perturbation induced by MP1, whereas at the same time, decreases membrane affinity for MP1.

The rate constants for the processes of peptide adsorption to the membrane from the outside ($k_a$), and for peptide desorption from the membrane into vesicle lumen ($k_d$) were estimated as detailed in Material and Methods, and the ratio between $k_a$ and $k_d$ ($K_{ef}$) was calculated. As expected, $K_{ef}$ shows a trend similar to $(F_M/F_0){}_{\max }$ (Fig. 3A). The estimation of the kinetic constants allows to get an insight of the reasons for peptide accumulation in the membrane. In the case of pure POPC, $k_a$ was higher than $k_d,$ indicating that the peptide adsorbed fast and desorbed slowly, thus remaining in the membrane, probably forming peptide aggregates. In vesicles with 20% CHO, $k_a$ and $k_d$ were both large (Figure S6-B), peptide desorbed fast, leading to a low extent of peptide in the membrane. It is important to note that for this composition, the values of $k_a$ and $k_d$ were determined only for vesicles with appreciable peptide accumulation, and not with all recorded vesicles.

The presence of 20% DP in POPC vesicles induced a decrease in both constants, while higher DP percentages increased their values, thus the trend was dose-dependent for this lipid. Since both rate constant were affected by DP similarly, the amount of peptide accumulated on the membrane was only slightly affected by the DP proportion (see $(F_M/F_0){}_{\max }$ and $K_{ef}$ values in Fig. 3A), being higher than on CHO-containing membranes and lower than on pure POPC membranes.

The rate for peptide entry into vesicles with DP was between that for pure POPC vesicles and those with CHO (Figure S7), indicating that despite MP1 induced similar permeability to the soluble CF species in membranes with or without DP (Fig. 1B), peptide entry was affected by the presence of the hopanoid. Fig. 3B depicts the times at which 50% of the GUVs had peptide in their lumen, together with the times at which $F_M$/$F_0$ reached a value equal to 50% $(F_M$/$F_0){}_{\max }$. The similarity between the values indicates that peptide accumulation at the membrane is highly correlated with peptide entry, and thus we propose that the longer times for desorption into the vesicle lumen (lower $k_d$ values) are related to the formation of peptide dimers/multimers that leads to the development of pores/defects, and peptide entry.

No visible thermal shape fluctuation was detected for all GUV compositions before or after the addition of 0.6 µM peptide (Figure S8). On the contrary, experiments with a 10-fold higher concentration of peptide revealed strong shape fluctuations (Fig. 4, S8 to S11 and Movie S1). Furthermore, visible local lipid/peptide aggregates were detected, followed by membrane disintegration in GUVs composed of POPC and POPC + 20% CHO. In the latter, fluctuations took more time to appear, as well as peptide-lipid aggregates and vesicle rupture (Figure S8).

Fig. 4.

Confocal images at selected times after addition of 6.0 µM fMP1 to POPC vesicles (upper panels) or POPC + 20% DP vesicles (lower panels). The images correspond to the merge of green (fMP1) and red (PE-Rho) channels. Brightness and contrast range: 0–150, scale bar: 10 µm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

An interesting behavior was observed for membranes containing 20% DP in the presence of 6.0 µM fMP1. The peptide induced strong membrane fluctuations, followed by a return to their initial non-fluctuating state (Fig. 4 and S10, and Movie S1). In this scenario, no peptide/lipid aggregation or vesicle rupture was detected.

According to Méléard et al. [56], from all the factors characterizing membrane mechanical properties introduced by Helfrich [57], thermal fluctuations in GUV’s shape depend only on the bending elastic modulus ${\kappa }_c,$ and the difference between the actual curvature and the spontaneous curvature $\Delta C$. Thus, at high-dose, the peptide decreased ${\kappa }_c$ and/or $\Delta C$ for membranes with POPC and POPC containing 20% CHO leading to vesicle rupture. On the other hand, the same peptide levels induced first a decrease and then an increase in ${\kappa }_c$ and/or $\Delta C,$ leading to stable vesicles that remain spherical even at 6 µM peptide. This intriguing non-linear effect was further studied at a low concentration of the unlabeled peptide using a dynamic active method.

The dynamic elasticity of membranes with different compositions is affected by MP1 in different manners

The effect of the unlabeled peptide on membrane dynamic elasticity at low peptide dose (0.6 µM) was investigated following the retraction of membrane tethers (40 $\pm$ 5 µm length) generated from single GUVs using optical tweezers (Movie S2 and Figure S12). Values of [P]/[L] = 0.2–0.3 were used in these experiments, which are higher than those from 100% leakage in LUVs.

The static force opposing tether growth depends on membrane elasticity, determined by the product of ${\kappa }_c$ and $\Delta C=(\frac{1}{2R}-C_0),$ where $C_0$ is the spontaneous curvature and $R$ is the tether radius (in the order of the nanometer) [58]. At non-static situations, tether formation requires an additional force $F_{\gamma }$ which accounts for dissipation [42]. In tether retraction, membrane elasticity is the driving force while $F_{\gamma }$ opposes the motion. The characteristic membrane relaxation time ($\tau$) was determined by fitting curves such as that shown in Fig. 5A with Eq. (9) (see Material and Methods). $\tau$ depend on both, membrane elasticity and dissipation, being $\tau =\frac{\gamma }{k},$ where $k$ accounts for the elastic terms, and thus depends on ${\kappa }_c$ and $\Delta C,$ and $\gamma$ is the dissipation coefficient. In our system, $\gamma$ is formed by Stokes dissipation of the bead in the aqueous media, membrane shear viscosity (${\eta }_s$), and interlayer dissipation that depends on the coupling between hemilayers (${\eta }_c$) [42], [43]. Therefore, the rate of tether retraction decreases (and $\tau$ increases) if ${\kappa }_c$ or $C_0$ decreases, or if ${\eta }_s$ or ${\eta }_c$ increases.

Fig. 5.

Membrane dynamic elasticity. (A) Retraction process of tethers pulled from a POPC GUV after 0, 10, 13, 17, and 25 min of peptide addition. The solid lines represent fits to the data using $L=L_0\exp (\frac{-t}{\tau }$), where $\tau$ is the characteristic time. Inset: accumulated images obtained using DIC showing the bead motion (24 frames/s) after the laser was switched off. Left: without peptide, right: after 17 min of MP1 addition, note the slower bead motion and the loss of vesicle contrast due to the increased permeability. (B) Characteristic times $\tau$ as a function of time after the addition of 0.6 µM peptide to different GUVs composed of POPC + 20% CHO (each symbol correspond to a single GUV) (C) Characteristic times $\tau$ as a function of time after the addition of 0.6 µM peptide to GUVs composed of POPC (black), POPC + 20% DP (olive), POPC + 40% CHO (cyan) or DP (green). Data correspond to the average ($\pm$ SD) of at least 5 GUVs from at least two independent experiments. * and ** indicate significant statistical differences between the values at the indicated times and at the start and end of the experiment determined by one-way ANOVA, Tukey’s test at $p<0.02$ or $p<0.002,$ respectively.

The values of $\tau$ in the absence of peptide were higher for pure POPC in comparison to POPC-CHO membranes (see Fig. 5B and C, and S13). This is expected since CHO is known to increase ${\kappa }_c$ [59], [60]. DP promoted the same changes in $\tau$ as CHO, indicating a comparable effect of this lipid on the mechanical properties of POPC bilayers.

For POPC, POPC + 40% CHO and POPC + DP (both proportions), all analyzed GUVs showed the same trend in $\tau ,$ and thus, the values of $\tau$ for different GUVs after peptide addition were averaged and plotted in Fig. 5C. On the contrary, for membranes composed of 20% CHO, each GUV showed different behavior, and thus, the $\tau$ values for individual vesicles are shown in Fig. 5B. In the absence of peptide, $\tau$ did not change in time nor due to consecutive tether formation-retraction (Figure S13).

For POPC membranes, 0.6 µM MP1 promoted a progressive increase in $\tau ,$ requiring about 5 min for half the total change. This time was similar to those shown in Fig. 3B indicating a correlation between peptide accumulation in the membrane, peptide entry and membrane softening. This increase in $\tau ,$ together with the effect on thermal fluctuations promoted by 6.0 µM fMP1 suggest that the peptide decreases the bilayer’s elasticity, as previously reported for other AMPs [60], [61], [62]. This observation has previously been related to local membrane thinning [63], [64], decrease in the lipid-lipid cohesion energy [65], peptide-peptide interactions [64], [66], [67], decoupling of the hemilayers [62], entropy effects [68], or inhomogeneous spontaneous curvature [60], [61], [65], [66], [67], [69], [70], [71], [72].

The values of $\tau$ remained constant when GUVs with 40% CHO were exposed to the peptide, in agreement with the low effect of MP1 on this membrane composition. GUVs with 20% CHO appear to represent a composition close to a critical point, being more sensitive to uncontrolled parameters such as the precise vesicle composition or the surface tension. Interestingly, the behavior in $\tau$ correlates with the stochastic adsorption and entry of fMP1 into this GUV composition (Figure S5).

Surprisingly, a non-monotonic behavior was observed in DP-containing membranes. In 20% DP-containing GUVs, a subtle increase in $\tau$ was first observed, followed by a decrease after 35 min of MP1 exposure, reaching a final value close to the initial one. This effect was more marked for membranes with 40% DP, in which MP1 induced a fast increase in $\tau$ (5--10 min after peptide addition), followed by a slow decrease. This behavior points to the presence of two opposing factors with different kinetics. Non-monotonic effects on membrane elasticity as the amount of peptide increases has been previously reported for Colistin in a three-component membrane. The behavior was explained considering dose-dependent specific peptide-lipid interactions [73]. In those experiments, vesicles were prepared with the peptide, and thus, reported data are at equilibrium conditions and bilayers are symmetric. Here, we analyzed the time-evolution of membrane mechanical response. Our membranes are two-components, and DP is not able to form stable vesicles when pure. Therefore, despite we cannot discard the above-mentioned explanation, we propose an alternative hypothesis.

We also studied membrane electrostatics in the presence of MP1 in an independent series of experiments. Zeta potential ($\zeta$) in LUVs incubated with MP1 showed a non-monotonic behavior that may shed light on the observed softening/stiffening effect of DP-containing membranes. Fig. 6 shows the $\zeta$ values for LUVs composed of POPC containing 20 mol% DP or 40% DP incubated with increasing concentrations of peptide. For both compositions, a decrease followed by an increase in $\zeta$ values was observed as $\left[P\right]/\left[L\right]$ increased. The value of $\left[P\right]/\left[L\right]$ where the minimum occurred was lower than for 40% DP, and the minimum was more marked. This particular behavior in $\zeta$ indicates that the trend in surface electrostatics changed as the amount of peptide in the membrane increased, suggesting a dose-dependent organization of the peptide at the membrane. This non-monotonic trend was not observed previously for MP1 in other membrane compositions [31], and was not observed for the other compositions studied in this work (Figure S14). The addition of MP1 to POPC LUVs led to changes in the zeta potential ($\zeta$), going from negative to positive values (Figure S14). This was expected because, even though POPC is neutral, it has been reported that preferential adsorption of counterions onto the membrane and/or preferential orientations of the zwitterionic polar headgroup led to phospholipid vesicles with negative $\zeta$ [74], [75], [76], [77], [78]. Being a cationic peptide (net charge +2), MP1-adsorption neutralized the negative charge on the vesicles. The addition of CHO leads to the same trend, though less marked (Figure S14).

Fig. 6.

Effect of MP1 on the zeta potential ($\zeta$) of vesicles composed of POPC + 20$\%$ (olive) or 40$\%$ of DP (green). All data correspond to the average ($\pm$ SD) of three independent experiments. Inset: wheel representation of MP1 in an $\alpha$-helix configuration (green represents positively charged; red, negatively charged; blue, polar uncharged and yellow, hydrophobic residues, and N and C indicate the N-terminus and C-terminus of the peptide, respectively.) (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Despite MP1 has a positive net charge, it contains two acidic residues. It has been previously shown using circular dichroism spectroscopy that this peptide adopts alpha-helix structure when interacting with membranes composed of pure PC, and PC mixed with phosphatidylglycerol (30 and 60%), or cardiolipin (30%), or cholesterol (20%) [30], [79]. Based on those reports, we here assume an alpha-helix structure on DP-containing membranes, with the aspartic acids in the positions indicated in the inset of Fig. 6. If the peptide adsorbs parallel to the interface, the hydrophilic part is expected to remain exposed to the solvent. The decrease in $\zeta$ may be a consequence of a higher exposition of the acidic residues than the cationic ones, and/or a reorganization of the ionic cloud due to peptide adsorption. As the peptide accumulates at the interface, acid/basic residues exposition change and/or counter-ions are displaced by the peptide, leading to a different trend in the change of $\zeta$ with $\left[P\right]/\left[L\right]$.

Zemel et al. analyzed the membrane perturbation free energy during peptide-membrane interaction for low-charged peptides, considering interfacially-adsorbed monomeric and dimeric peptide species, and the multi-peptide transmembrane pore state. They found that in the transmembrane pore state, the lipid perturbation energy per peptide is smaller than in the adsorbed state, and proposed that the gain in conformational freedom of the lipid chains is a central driving force for pore formation [80]. They also found a weak lipid-mediated gain in membrane perturbation free energy upon dimerization of interfacially-adsorbed peptides. Taking into account their results, we propose that at low $\left[P\right]/\left[L\right]$ values, peptides adsorbed at the membrane in a diluted regime, giving rise to a disposition of the ions that leads to a negative zeta potential. This would happen until a critical $\left[P\right]/\left[L\right]$ value was reached, which depend on membrane mechanical properties and thus, on DP percentage.

As the amount of peptide increased at the interface, the membrane perturbation free energy increased, and eventually peptide dimers or multimers would emerge. We showed previously for MP1 that the presence of peptide-peptide interactions was stabilized by inter-molecular salt-bridges involving the acidic residues [50]. Therefore, cluster formation at the membrane would imply changes in the surface electrostatics that may give rise to the increase in $\zeta$. The fact that the $\left[P\right]/\left[L\right]$ ratio for the minimal value of $\zeta$ was double for 20% in comparison to 40% DP indicates that as DP increased, the effect of monomeric peptides on the membrane perturbation free energy increased, and a lesser amount of peptide was needed for multimer formation.

Going back to the retraction experiments, despite these are kinetic assays and therefore not directly comparable to the $\zeta$ data, we propose that at short times, a low amount of peptide accumulated in the external hemilayer in a diluted regime. This first stage is followed by the formation of peptide dimers/multimers until pores/defects form, allowing vesicle leakage and the passage of the peptide to the vesicle lumen. The pores/defects may be formed by lipids and peptides (disordered toroidal pore [81]), or with one peptide at each hemilayer, since the thickness of the bilayers ($\sim$0.4 nm) [82] are twice the length of the peptide ($\sim$0.2 nm) [50].

The increase in $\tau$ occurred up to longer times for 20% than for 40% DP because in membranes with 20% DP, the amount of peptide required for the reorganization was higher. When peptides are in the outer hemilayer in a diluted regime, they may decrease ${\kappa }_c$ by decreasing the lipid chain order [65], [83], increase ${\eta }_s$ [84] or decrease the term $\frac{1}{2R}-C_0$ due to an increase in $C_0$. This last effect is expected since at this stage membranes would be asymmetric, with peptides only in the external leaflet. We discard effects due to ${\eta }_c$ because a decrease (and not an increase) in hemilayer coupling is expected during this stage [62], which would lead to a decrease in ${\eta }_c$ (and in $\tau$).

When peptides reach a threshold concentration at the interface, they would form dimers/multimers, and eventually, pores. This is accompanied by a decrease in $\tau$ that could be due to an increase in $C_0$ since membrane symmetry would be recovered. Besides, a decrease in the effective ${\eta }_s$ and an increase in the effective ${\kappa }_c$ could occur in the presence of pores. In this sense, non-monotonic changes in ${\kappa }_c$ have been proposed as the peptide moves inside the membrane [83]. Again, effects on the hemilayer coupling are discarded since, in the presence of pores, an increase and not a decrease in ${\eta }_c$ is expected.

It is important to state that despite membrane dissipation terms cannot be neglected, the elastic response of the POPC/DP membranes determined by ${\kappa }_c$ and $C_0$ likely commands the effect of the peptide on $\tau ,$ since nanotube retraction data are in agreement with the changes in membrane elasticity detected by thermal fluctuations in GUVs composed of POPC + 20% DP in the presence of 6 µM fMP1. The membrane-peptide association states proposed for POPC/DP membranes may also occur in pure POPC and POPC/CHO membranes. However, $\zeta$ and $\tau$ data indicate that membrane electrostatics and mechanical properties are affected in a different fashion in those membrane compositions.

P. aeruginosa cells exposed to DP are more sensitive to MP1 than those exposed to CHO

The results shown up to now indicate that MP1 affects differently artificial membranes containing DP or CHO, being the DP-containing membranes less protected against leakage and peptide entry, despite they recover their elastic properties, and remain spherical at high peptide doses.

Since it was previously reported that incubation of P. aeruginosa with DP or CHO leads to an decreased membrane permeability and an increased sensitivity to some antibiotics [25], we decided to investigate whether incubation of P. aeruginosa with DP or CHO affects cell resistance against MP1. These exploratory assays were performed with the aim to find possible correlations between the effect of DP in experiments with binary mixtures and DP in a membrane with a complex and not-constant composition, such as that of the bacteria. P. aeruginosa is a good model for this test, because this bacteria does not synthesize DP or CHO, and thus the molecules can be supplied externally.

Cell survival was determined in cultures preincubated with CHO or DP after exposing the bacteria to peptide concentrations higher than the MIC value (5 µM) [29]. As shown in Fig. 7, the fraction of cell survival to MP1 exposition for untreated bacteria (control 1) was similar to the survival of cultures incubated with the lipid solvent (control 2). In the presence of CHO, cells became insensitive to MP1 at 9 and 18 µM, being necessary to increase MP1 dose to 30 µM in order to observe cell death (data not shown). DP elicited a higher sensitivity to MP1 compared to CHO, as showed by the lower fraction of surviving cells. These data indicate that DP was not as effective as CHO in increasing cell resistance to MP1, and suggest that MP1 interact deferentially with cell membranes containing DP than with those containing CHO.

Fig. 7.

Activity of MP1 against P. aeruginosa previously exposed to CHO or DP. Fraction of surviving cells without treatment (control 1) or incubated in the solvent (control 2), or in a solution of CHO (blue) or DP (olive) after 3 h of exposure to 9 µM or 18 µM of MP1. All data correspond to the average ($\pm$ SD) of three independent experiments. * and ** indicate significant statistical differences determined by one-way ANOVA, Tukey’s test at $p<0.0$5 or $p<0.001,$ respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4. Conclusion

In the present work, we showed that besides electrostatic interactions, which are accepted as key factors for an AMP to distinguish between bacteria and mammalian cells, neutral lipids such as sterols and hopanoids may also be important factors in MP1 selectivity. We found that the permeability of CHO-containing membranes was affected by MP1 to a lesser extent than those with DP. Furthermore, peptide adsorption to the membranes was hampered more efficiently by CHO than by DP, and similar results were found for peptide entry into the vesicle’s lumen. An interesting effect of MP1 on the bending ability of DP-containing membranes was described, which was explained as the consequence of peptide-peptide aggregation in the membrane. For membranes with DP, the results indicate that MP1-induced poration and MP1 translocation did not lead to vesicle rupture. Bending properties were recovered after long periods in the presence of MP1 without membrane disintegration, while CHO-containing vesicles burst when exposed to high levels of the fluorescently-labeled peptide. On the other hand, survival of P. aeruginosa incubated with DP was lower than that for cells incubated with CHO. All together,the results suggest that cell death would be related with peptide entry or an enhanced membrane permeability and not with membrane disintegration.The presence of anionic lipids may enhance the sensitivity of DP-containing membranes, and this should be investigated in turn.

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

NATALIA WILKE reports equipment, drugs, or supplies was provided by FONCYT (PICT 2015- 0662). JOAO RUGGIERO NETO reports equipment, drugs, or supplies was provided by FAPESP (2015 25619-9). NATALIA WILKE reports equipment, drugs, or supplies was provided by SECyT-UNC. DAYANE S ALVARES reports financial support, equipment, drugs, or supplies, and travel were provided by FAPESP BEPE-grant (2018 14215-2).

Appendix A. Supplementary materials

Supplementary Data S1

Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/

Supplementary Data S2

Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/

Supplementary Data S3

Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/