Competitive interactions of amphipathic polycationic peptides and cationic fluorescent probes with lipid membrane: Experimental approaches and computational model

Archives of Biochemistry and Biophysics 545 (2014) 167–178

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi

Competitive interactions of amphipathic polycationic peptides and cationic fluorescent probes with lipid membrane: Experimental approaches and computational model Victor V. Lemeshko ⇑ Escuela de Física, Facultad de Ciencias, Universidad Nacional de Colombia, Sede Medellín, Calle 59A, No 63-20, Medellín, Colombia

a r t i c l e

i n f o

Article history: Received 27 December 2013 and in revised form 16 January 2014 Available online 2 February 2014 Keywords: Polycationic peptides Fluorescent probes Mitochondria Liposomes Peptide–membrane interaction Computational model

a b s t r a c t The electrostatic interaction of polycationic peptides with negatively charged biomembranes has been recognized as the first and very important step of their selective binding to many bacteria and transformed cells. In this work we demonstrated the phenomenon of competition of some earlier designed polycationic peptides and fluorescent probes for their binding to the negatively charged inner membrane of mitochondria and to the PC/PG (9:1) liposomes. Rat liver mitochondria swelling induced by the antimicrobial polycationic peptide BTM-P1 (VAPIAKYLATALAKWALKQGFAKLKS) and by the retro-BTM-P1 was significantly diminished in the presence of 10 lM fluorescent probe safranin O. In experiments with liposomes, the polycationic peptides BTM-P1 and P7-5 (IYLATALAKWALKQGF-GG-RRRRRRR) at the concentrations of 2–3 lM completely displaced the membrane-bound fluorescent probe DiSC3(5) in a low ionic strength medium. The developed computational model allowed a mathematical description of such interactions, predicting membrane surface concentrations of bound peptides as the function of the membrane surface charge and lipid quantity in the sample, the peptide charge, hydrophobicity and concentration, the ionic strength of incubation medium and of the presence of a charged fluorescent probe used for monitoring the membrane surface potential under real-time peptide–membrane interactions. Ó 2014 Elsevier Inc. All rights reserved.

Introduction The antimicrobial peptides represent high clinical interest due to their low toxicity to mammalian cells and inherent difficulty to acquire bacterial resistance to them [1–4]. Most of the natural antimicrobial peptides are polycationic due to a presence of lysine and/or arginine residues [5–8]. In addition, they consist of a large proportion of hydrophobic amino acid residues [2,7–9]. The amphipathic character of the polycationic antimicrobial peptides allows their insertion into the lipid bilayer of biomembranes that has been considered a definitive feature of their antimicrobial activity even when the membrane permeabilization does not occur [10,11]. The polycationic peptide insertion into the lipid bilayer and the membrane permeabilization has been shown to be strongly potentiated by high values of the trans-membrane potential of bacteria [12] and of mitochondria [13–15]. The permeabilization of the plasma membrane of red blood cells by such peptides has also been potentiated by a relatively high plasma membrane

⇑ Fax: +57 4 4309327. E-mail address: [email protected] http://dx.doi.org/10.1016/j.abb.2014.01.024 0003-9861/Ó 2014 Elsevier Inc. All rights reserved.

potential (minus inside) generated in the presence of valinomycin [15–17], or even by the application of external electrical pulses to the cell suspension [17]. The selectivity of action of the polycationic antimicrobial peptides is mainly attributed to a high negative charge of the membrane surface of prokaryotic cells. Cancerous cells have also been reported to possess higher negative surface charges and higher trans-membrane potentials in comparison with normal eukaryotic cells [18–20]. That is why many of the polycationic antimicrobial peptides have also been found to reveal anticancer properties [7,21–23]. The design of new peptides with antibacterial and anticancer activities and the study of the mechanism of their action require development of new experimental approaches to evaluate the peptide interactions with biological membranes, particularly with the charged lipid bilayer. In this respect, the potential-dependent cyanine fluorescent probes, normally used as indicators of the membrane permeabilization by various factors, have also been shown to sense electrical properties of the membrane surface [24]. The charged amphipathic substances interacting with biomembranes might displace membrane-bound dyes. This kind of competitive interaction has been observed, for example, as a displacement of

168

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

the cationic fluorescent probe dancyl-PMBN from the liposomal membrane by various polycationic peptides [25]. Several mathematical models have been developed to describe the peptide–membrane interactions [26–30]. Some of these models have been based on the Gouy–Chapman theory of the electrical double layer to estimate electrostatic interaction of polycationic peptides with the membrane, considering the uniformly distributed fixed charge density of the membrane surface [29,30]. On the other hand, the membrane surface charge and the surface potential might be affected by the peptide adsorption, hence inhibiting further peptide binding to the membrane, like that described for 1-anilino-8-naphthalenesulfonate, the anionic and hydrophobic fluorescent dye [31]. The combined electrostatic and hydrophobic interactions of peptides, dyes and other substances with lipid membrane might result in their competition or in a synergistic increase of the peptide adsorption on the membrane. In the present work, we first experimentally observed the influence of the widely used potential-sensitive cationic probe safranin O on mitochondria permeabilization by some polycationic peptides, and vice versa, the displacement of the cationic cyanine probe DiSC3(5) from the negatively charged liposomal membrane by the peptides. The experimental data directly suggested the competitive character of the peptide and the fluorescent probe interactions with the membrane. The computational model was developed to explain the observed phenomena in their dependence on the probe and lipid membrane quantities, and on the concentration, net charge and hydrophobicity of a peptide. An effective surface charge density was included in the Gouy–Chapman equation instead of a fixed surface charge. The calculated data demonstrated that the fluorescence intensity of the aqueous phase DiSC3(5) is the most sensitive parameter reflecting peptide interactions with the lipid membranes mimicking eukaryotic or prokaryotic cell surfaces. The model allows the prediction of the optimal experimental conditions to study the membrane–peptide interactions using charged fluorescent probes. Materials and methods Materials The polycationic peptides were designed in our laboratory and synthesized by the GenScript Corporation (NJ, USA). The used peptides were BTM-P1 (95.4% purity), with the sequence of VAPIAKYLATALAKWALKQGFAKLKS, derived from the Cry11Bb protoxin [13–15], its retro analog, retro-BTM-P1 (94.2% purity) with the sequence of SKLKAFGQKLAWKALATALYKAIPAV [15], and the peptide P7-5 (95% purity) described earlier [17], with the sequence of IYLATALAKWALKQGF-GG-RRRRRRR, showing anticancer properties. The phospholipids were purchased from Avanti Polar Lipids, Inc. Other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO, USA), Isolation of rat liver mitochondria Rat liver mitochondria were isolated by the method of differential centrifugation as described earlier [17], following the principles outlined in the Guide for the Care and Use of Laboratory Animals published by the USA National Institutes of Health (NIH Publication No. 85-23, revised 1996) and approved by the Local Ethics Committee of the National University of Colombia, Medellin Branch. The 10% liver homogenate (5 g of liver) was prepared in medium containing 210 mM mannitol, 70 mM sucrose, 2.5 mM MgCl2, 1 mM EGTA–KOH, 0.3 mg/ml bovine serum albumin (BSA, free fatty acid fraction V), 10 mM HEPES–KOH, pH 7.2, at 0–4 °C. Mitochondria were washed twice in medium containing 210 mM

mannitol, 70 mM sucrose, 50 lM EGTA–KOH, 0.3 mg/ml BSA and 10 mM HEPES–KOH, pH 7.2. Finally, the mitochondrial pellet was resuspended in 1 ml of the same medium without BSA. Monitoring of the inner membrane potential of mitochondria The inner membrane potential of mitochondria was monitored with the potential-sensitive fluorescent probe safranin O, as described in [32]. The real time fluorescence intensity (580 nm emission, 520 nm excitation) was measured using the Aminco-Bowman Series 2 spectrofluorimeter. Mitochondria, at the final concentration of 0.5 mg protein/ml, were added to the incubation medium composed of 100 mM sucrose, 75 mM KCl, 10 mM potassium phosphate, 5 mM HEPES, 50 lM EGTA, pH 7.2 (SKPH-medium), supplemented with 10 lM safranin O. To energize mitochondria, succinate was added to the final concentration of 2.5 mM. The peptides were added to the energized mitochondria at the final concentrations of 0.5 lM (BTM-P1) or 1.0 lM (retro-BTM-P1), observing subsequent increase in the safranin O fluorescence as a result of the inner membrane permeabilization. Monitoring of the redox state of mitochondrial pyridine nucleotides The level of reduced forms of mitochondrial pyridine nucleotides, NAD(P)H, was monitored using the Aminco-Bowman Series 2 Luminescence Spectrometer, as described earlier [33]. Mitochondria, at the final concentration of 0.5 mg protein/ml, were added to the SKPH medium. Where indicated, 2.5 mM succinate was added to energize the mitochondria. After that, the added peptides (0.5 lM BTM-P1 or 1.0 lM retro-BTM-P1) caused a decrease in the NAD(P)H fluorescence due to the inner membrane potential decrease, as explained in [33]. Monitoring of mitochondrial swelling Mitochondrial swelling was monitored simultaneously with the safranin O or NAD(P)H fluorescence measuring in standard quartz cuvette using a modified cuvette holder for the Aminco-Bowman Series 2 Spectrometer described in [34]. The light, emitted by an infrared light-emitting diode (920 nm) and dispersed at 90° in mitochondrial suspension, was detected by additionally mounted photodiode and amplifier. The amplifier output signal was registered by the data acquisition system of the spectrometer using one of the two auxiliary channels [34]. The samples were constantly stirred with the magnetic stirrer and maintained at the temperature of 30 °C. Liposome preparation Small unilamellar lipid vesicles (1SUV) were prepared from egg yolk PC:PG (9:1 weight ratio). The lipids dissolved in chloroform were dried under N2 to form a lipid film on the glass tube wall. After that, the sample was desiccated under a vacuum for 2 h to further remove the solvent. The obtained lipid film, composed of 18 mg PC and 2 mg of PG, was rehydrated with 2 ml of 0.1 M potassium phosphate buffer, 0.1 mM EGTA, pH 7.2 (KPB medium), to the final lipid concentration of 10 mg/ml. After that, the sample was vortexed to obtain multilamellar liposomes. To prepare SUV, the sample was sonicated (tip diameter 3 mM; at 20% power of the Cole-Parmer Ultrasonic Processor CP-500, 500 watts, 20 kHz), pulses ‘‘15 s on/ 45 s off’’ during 10 min of total sonication time at 4 °C, until the suspension was transparent. Titanium particles were removed by 1 Abbreviations used: SUV, small unilamellar lipid vesicles; PC, phosphatidilcholine; FCCP, carbonyl-cyanide-p-trifluoromethoxy phenylhydrazone; PG, phosphatidylglycerol.

169

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

centrifugation at 10,000g during 10 min. The obtained SUV sample was stored at 4 °C before experiments were conducted. Monitoring of the DiSC3(5) interaction with liposomal membrane The liposomes (5 ll of prepared SUV sample) were added to 1 ml of incubation medium after 1 min preincubation with 0.2 lM DiSC3(5). The fluorescence intensity was measured at 668 nm (648 nm excitation) using the Aminco-Bowman Series 2 Spectrometer. The used incubation media were KPB-medium (described above); 0.1 M sodium phosphate buffer, pH 7.2 (NaPBmedium); 125 mM NaCl, 5 mM HEPES, pH 7.2 (NaH-medium); 250 mM sucrose, 5 mM HEPES, pH 7.2 (SH-medium), and NaH/ SH-medium (1:1). The peptides BTM-P1 and P7-5, and a protonophoric uncoupler carbonyl-cyanide-p-trifluoromethoxy phenylhydrazone (FCCP) were added where indicated, influencing the DiSC3(5) fluorescence intensity. Computational model of competitive interactions of the polycationic peptides and the fluorescent probe DiSC3(5) with liposomal membrane Let us assume that the SUV suspension was added at the liposomal volume VL to 1 ml of incubation medium (Vao). Knowing the average diameter of liposomes (about 50 nm), it is possible to calculate their total internal aqueous volume (Vai), as well as the internal lipid leaflet volume (Vmi) and the external lipid leaflet volume (Vmo) of the membrane. Adding the fluorescent probe DisC3(5), it will be distributed in the membrane and in the aqueous phases, thus the average DisC3(5) concentration in the sample (CD) may be expressed as:

CD ¼

C ai  V ai þ C ao  V ao þ C mi  V mi þ C mo  V mo ; V ao

ð1Þ

where Cai and Cao are the concentrations of DisC3(5) in the internal and the external aqueous phases, and Cmi, Cmo are its concentrations in the internal and the external lipid leaflets, respectively (Fig. 1). As the probe is permeable through the membrane, and the trans-membrane potential was considered to be equal to zero, we can write:

C ai ¼ C ao

ð2Þ

The DiSC3(5) probe bound to a phospholipid membrane has mainly been considered located close to the water–lipid interphases, as internal and external monolyers (Fig. 1), but not in the membrane hydrophobic core [35,36]. Due to a high value of the probe partition coefficient (Q) between the phosphatidilcholine (PC) membrane and the aqueous phase (Q = 104 or higher) [20], the surface charge and the surface membrane potentials should be generated, influencing probe binding to the membrane (Fig. 1). Thus, the effective DiSC3(5) partition coefficients for the inner (QDi) and the outer (QDo) leaflets of the membrane might be expressed as:

Q Di ¼

wi C mi ¼ Q  e 25 C ai

ð3Þ

and

Q Do ¼

wo C mo ¼ Q  e 25 ; C ao

ð4Þ

respectively, were wi and wo are the internal and the external surface potentials expressed in mV (Fig. 1). The membrane might also contain charged lipids, as phosphatidylglycerol (PG) for example. It will additionally change the surface charge and the surface potential values influencing the interaction of the DiSC3(5) probe with the membrane. The membrane may be considered theoretically as a perfect plane surface

Fig. 1. Schematic representation of the electrostatic and hydrophobic interactions of a polycationic peptide and of the fluorescent probe DiSC3(5) with liposomes. A – the negative surface potentials are generated due to the presence in the membrane of negatively charged phospholipids like phosphatidylglycerol (PG), in addition to electrically neutral phospholipids like phosphatidylcholine (green space). B – the surface potentials influenced by the adsorption of DiSC3(5) on the internal leaflet, and by the adsorption of both DiSC3(5) and polycationic peptide on the external leaflet of the membrane. The external leaflet concentration of DiSC3(5) (Cmo) is decreased due to the electrostatic competition with a polycationic peptide for the electronegative external membrane leaflet. Cmi – the DiSC3(5) concentration in the internal membrane leaflet. Cai and Cao – the DiSC3(5) concentrations inside the liposome and in the external medium, respectively, Cai = Cao. QDi and QDo – effective DiSC3(5) partition coefficients for the inner and the outer membrane leaflets, respectively. Cpm and Cpa – peptide concentrations in the outer membrane leaflet and in the external medium, respectively. Kpo – the effective peptide partition coefficient. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

with a uniformly smeared charge over it, following the approximation used in [38]. The resulting surface charge density (in electronic charges per Å2) on the internal membrane leaflet can be presented as:

ri ¼ C mi  NA  25 Å 

0:01 2

 PG

ð5Þ

70 Å

assuming that DiSC3(5) is distributed as monolyer at the water– lipid interphase. The membrane area corresponding to each phospholipid molecule was taken as 70 Å2 ([38] and references therein). Here, Cmi is in molar concentration (mol/l, or mol/ 1027 Å3), 25 Å is the lipid leaflet thickness, and PG is the concentration of phosphatidylglycerol (%) in the membrane lipids. The surface potential, wi, on the Helmholtz’s layer of the inner membrane leaflet depends on the surface charge density ri and can be calculated using Graham equation ([50] and references therein):

(

X C i;in  ½expðzi ui F=RTÞ  1

ri ¼  2RT e

)12 ð6Þ

i

where R is the universal gas constant, T is the absolute temperature, F is Faraday’s constant, e is the absolute dielectric constant for water, Ci,in is the concentration of the ion i in the ‘‘infinity’’ of the

170

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

internal aqueous phase of liposomes, which can be taken as the concentration of a one–one-valent electrolyte in the bulk phase, and zi is the ion valence. An amphipathic polycationic peptide, with net charge q, can interact with the membrane by at least two type of interactions, by non-electrostatic (hydrophobic and other type of interactions) and electrostatic. The non-electrostatic interactions might be presented trough a partition coefficient Kp, which for antimicrobial peptides is known to vary in a range of several orders of magnitude [26,39–41]. Thus the electrostatic interactions, as an additional force, will influence the final effective partition coefficient of the peptide (Kpo) between the external membrane leaflet and the external aqueous phase as:

K po ¼

qwo C pm ¼ K p  e 25 ; C pa

ð7Þ

where Cpm and Cpa are the peptide concentrations in the external lipid leaflet and the external aqueous phase, respectively (Fig. 1). The membrane bound polycationic peptide will influence the surface charge density on the outer membrane leaflet, in addition to the effects caused by the negatively charged phospholipid PG and by the cationic probe DiSC3(5):

ro ¼ ðC mo þ q  C pm Þ  NA  25 Å 

0:01 2

 PG:

ð8Þ

70 Å

Here, Cpm and Cmo are the peptide and DiSC3(5) concentrations, respectively, in the external membrane leaflet (Fig. 1). In this case, the Grahame equation will be:

(

ro

X ¼  2RT e C i;out  ½expðzi uo F=RTÞ  1

)12 ð9Þ

;

i

were ro and wo are the charge density and the electrical potential, respectively, on the external membrane surface, Ci,out is the concentration of the ion i in the ‘‘infinity’’ of the external aqueous phase that can be taken as the concentration of a one–one-valent electrolyte in the bulk phase, as KCl or NaCl for example, and zi is the valence of the ion i. The average concentration of the added peptide (Cp) in the liposomal suspension can be expressed as:

Cp ¼

C pm  V mo þ C pa  V ao ; V ao

ð10Þ

where Cpm and Cpa are the concentrations of the peptide in the external membrane leaflet and the external medium, respectively (Fig. 1). To measure relative changes in the total quantities of the aqueous phase DiSC3(5) and of the membrane-bound probe, the fluorescence technique can be used, because the fluorescence maximum for the aqueous phase DiSC3(5) monomers is 668 nm, and for the membrane bound DiSC3(5) monomers is 690 nm, with a 7% higher quantum yield in the membrane, according to the data presented in [42]. On the other hand, the membrane-bound probe contributes to the 668 nm fluorescence with a weight of 0.42, and vice versa, the aqueous phase probe contributes to the 690 nm fluorescence with a weight of 0.36, both determined according to the corresponding emission spectra of DiSC3(5) [42]. Thus, the fluorescence intensity of the aqueous phase probe monomers (at 668 nm) can be expressed in arbitrary units as:

F a ¼ ðC ai  V ai þ C ao  V ao Þ þ 0:42  ðC mi  V mi þ C mo  V mo Þ;

ð11Þ

and the fluorescence intensity of the membrane bound probe monomers (at 690 nm) as:

F m ¼ 1:07  ½ðC mi  V mi þ C mo  V mo Þ þ 0:36  ðC ai  V ai þ C ao  V ao Þ ð12Þ

Having a suspension of ultrasound prepared liposomes (SUV) with a membrane thickness of 5 nm and the external diameter of approximately 50 nm, of a known total liposomal volume, VL, their total internal aqueous volume can be estimated as Vai = 0.5 VL, the internal lipid leaflet volume as Vmi = 0.2 VL, and the external lipid leaflet volume as Vmo = 0.3 VL. Here, Vmi and Vmo were considered proportional to the internal and the external membrane areas of SUV, respectively. If we add 5 ll of SUV with the lipid concentration of 10 mg/ml to 1 ml of incubation medium, the corresponding volumes will be equal to: Vao = 1.005 ml, VL = 0.1 ll, Vai = 0.05 ll, Vmi = 0.02 ll, Vmo = 0.03 ll, assuming that the lipid density is approximately equal to 1 g/ml [43]. For the computational modeling, we considered SUV composed of 100% PC, or SUV composed of 90% PC and 10% PG. The liposomal internal ionic strength was always taken as 0.15 M, and the ionic strength of the incubation medium was changed in the range of 0.002–0.150 M. The partition coefficient for DiSC3(5) in a PC membrane was taken as Q = 104, close to the reported value in [37] and the effective partition coefficients in the electronegative membranes are expressed by the Eqs. (3) and (4). For amphipathic cationic peptides, the partition coefficient, Kp, was varied in the range of 101–105 for interaction with a neutral PC membrane, and the corresponding effective partition coefficients, Kpo, for the peptide interaction with a charged membrane were determined according to Eq. (7). The peptide net charge q was varied in the range of 0–7. In this work, we do not consider the possibility of the formation of dye dimmers, because the concentrations of DiSC3(5) used in the experiments, or considered in the model, were not higher than CD = 0.4 lM. Mathcad 2001i professional software was used to describe the behavior of the computational model expressed by the Eqs. (1)–(12) under various conditions.

Results The competitive interactions of cationic peptides and a positively charged fluorescent probe with the lipid membrane were first observed when the peptide-induced mitochondrial swelling was monitored simultaneously with endogenous NAD(P)H fluorescence, depending on the inner membrane potential, or simultaneously with the fluorescence of safranin O, cationic potential-dependent probe (Fig. 2). As shown in Fig. 2B, the slope in the mitochondrial swelling induced by the addition of 0.5 lM BTM-P1 was decreased to the level of 58 ± 4% in the presence of 10 lM safranin O, comparing with that determined in the absence of the probe (Fig. 2A). Similarly, the slope in the mitochondrial swelling induced by 1.0 lM retro-BTM-P1 was decreased by safranin O to the level of 35 ± 3% (Fig. 2D and C), respectively. The relative slope in an increase of safranin O fluorescence after the addition of BTMP1 was also significantly less (Fig. 2B) than the relative slope in a decrease of NAD(P)H fluorescence (Fig. 2A), both resulting of the inner membrane potential drop. Similar results for decreased fluorescence slopes were obtained in the experiments with retro-BTM-P1 (Fig. 2D and C, respectively, for the presence and absence of safranin O). Other experimental evidences of the competitive interactions of polycationic peptides and a cationic fluorescent probe were obtained using liposomes having a negative membrane surface charge. According to the normalized DiSC3(5) spectra shown in Fig. 3A, the aqueous phase probe in the SH-medium had an emission maximum at 668 nm (Fig. 3A, a), which was shifted to 687 nm (Fig. 3A, b) after the addition of 5 ll of SUV composed of PC/PG (9:1) to the final lipid concentration of 0.05 lg/ml. The addition of SUV resulted in a significant decrease of the fluorescence intensity at 668 nm (Fig. 3B). Subsequent two additions of 1.0 lM BTMP1 allowed an almost complete recovery of the initial level of the DiSC3(5) fluorescence (Fig. 3B). No further increase in the

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

171

Fig. 2. Polycationic peptides BTM-P1 (P1) and retro-BTM-P1 (RP1) induce mitochondrial swelling monitored by light dispersion simultaneously with the fluorescence of endogenous NAD(P)H (A, C) or of added safranin O (B, D) as indicators of the inner membrane potential. Incubation medium SKPH without (A, C) or with 10 lM safranin O (B, D); Mc – mitochondria, 0.5 mg/ml; Suc – 2.5 mM succinate. The slopes in mitochondrial swelling induced by the peptide were significantly decreased in the presence of safranin O.

fluorescence intensity was observed after the third addition of 1.0 lM peptide. The emission maximum of the recovered fluorescence was shifted again towards the blue region, like the spectrum a in Fig. 3A (data not shown), presumable due to the probe displacement from the membrane by the added peptide. In contrast to the effect caused by the cationic peptide BTM-P1, the addition of 1.0 lM FCCP, a known acidic protonophoric uncoupler, caused a further remarkable decrease in the fluorescence intensity. Interestingly, in this case, the subsequent addition of only 1.0 lM BTM-P1 caused a complete recovery of the DiSC3(5) fluorescence, without an essential increase in the fluorescence intensity after the second addition of 1.0 lM peptide (Fig. 3C). The effect of the complete recovery of the aqueous phase DiSC3(5) fluorescence was also observed with 2.5 lM polycationic peptide P7-5 [17] in the SH-medium (Fig. 4A), very similar to that caused by 2–3 lM BTM-P1 (Fig. 3B). At a higher ionic strength, in the NaH/SH-medium (1:1), the DiSC3(5) fluorescence response to the addition of the same quantity of SUV was significantly less than in the SH-medium (Fig. 4B). Moreover, only 36% of that was recovered after the addition of 2.5 lM P7-5. At the physiological ionic strength, KPH-medium (Fig. 4C) or NaPH-medium (Fig. 4D), the DiSC3(5) fluorescence response to the addition of liposomes was comparable with that observed in the NaH/SH-medium (Fig. 4B), but the response to the subsequent addition of 2.5 lM P7-5 was strongly decreased (Fig. 4C and D). The obtained experimental data motivated us to develop the computational model shown in Fig. 1 and described by the mathematical Eqs. (1)–(12). The computational analysis was per-

formed for liposomes composed of phospholipids that have no net charge (as 100% PC for example) and for liposomes composed of PC/PG (9:1). The internal ionic strength of liposomes was assumed to be 150 mM, as it would be 150 mM KCl, for example. Initially, the calculations were performed for liposomal suspensions in high and low ionic strength media, 150 mM or 2 mM, respectively. The quantity of liposomes added to 1 ml of an incubation medium taken for calculations was the same as that used in the experiments described above. The calculations performed for liposomes composed of 100% PC in the presence of 0.4 lM DiSC3(5) showed that an arbitrary peptide with the net charge of +5 and the partition coefficient Kp = 103 almost did not influence the probe fluorescence in the aqueous or membrane phases (Fig. 5A, Fa and of Fm, respectively). The positive surface potential was developed on the external lipid leaflets of liposomes in a low ionic strength medium due to the DiSC3(5) adsorption, and it was further increased by the added peptide (Fig. 5B, a). The external surface potentials, developed in a high ionic strength due to the DiSC3(5) and peptide adsorption on the membrane, were significantly lower (Fig. 5B, b) than in a low ionic strength medium (Fig. 5B, a). Very low surface potential was maintained in the internal membrane leaflet, exposed to a high ionic strength internal medium, independently on the ionic strength of incubation media (Fig. 5B, c and d). The calculations also demonstrated a relatively low affinity of the considered arbitrary polycationic peptide to the electrically neutral lipid membrane of SUV, showing linear increase in the peptide concentrations in both aqueous (Fig. 5C, a and b) and

172

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

Fig. 3. The aqueous phase DiSC3(5) fluorescence in the SH-medium decreased by the addition of liposomes (SUV) (B, C) and FCCP (C), and recovered by the peptide BTM-P1 (P1). A – fluorescence spectra of 0.2 lM DiSC3(5) before (a) and after the addition of 5 ll SUV (b).

membrane phases (Fig. 5C, c and d) for low and high ionic strength media, respectively, with an increase in the peptide concentration (Fig. 5C). It was also noted that the affinity of the peptide to the hypothetical PC membrane was lower in a low than in a high ionic strength medium (Fig. 5C, c and d, respectively). The calculations were also performed for liposomes composed of PC/PG (9:1) (Fig. 6A). The obtained results clearly showed that after the addition of liposomes, the DiSC3(5) fluorescence in the aqueous phases should decrease more significantly in a low ionic strength medium (Fig. 6A, a, at Cp = 0) than in a high ionic strength medium (Fig. 6A, a, at Cp = 0). In contrast, the membrane phase probe fluorescence was prognosticated to be higher in a low ionic strength than in a high ionic strength medium (Fig. 6A, c and d, respectively, at Cp = 0). In a low ionic strength medium, the relatively high changes in the DiSC3(5) fluorescence in both the aqueous and membrane phases were calculated in response to the addition of the same polycationic peptide (with the net charge +5 and the partition coefficient Kp = 103) at the final concentration of 1 lM or larger (Fig. 6A, a and c for the membrane and the aqueous phase fluorescence, respectively). For a high ionic strength medium, the calculations showed very small responses to the peptide additions (Fig. 6A, b and d, for the membrane and the aqueous phase fluorescence, respectively). The calculations also showed that the negative surface potential of the internal membrane leaflet of the PC/PG liposomes maintained at an almost constant level when the ionic strength of incubation medium or the concentration of added polycationic peptide was changed (Fig. 6B, c and d for a low and high ionic strength, respectively). On the other hand, the external surface membrane potential was changed from near 100 mV in a low ionic strength medium to almost 24 mV in a high ionic strength medium even in the absence of a peptide (Fig. 6B, a and b, respectively, at Cp = 0). The addition of the peptide described above up to the con-

Fig. 4. The aqueous phase DiSC3(5) fluorescence decreased by the addition of liposomes (5 ll SUV/ml), recovered, partially or completely, by subsequent additions of the peptide P7-5 in the SH-medium (A), NaH/SH (1:1)-medium (B), KPBmedium (C) and NaPB-medium (D) (see Section 2.7). The media contained 0.2 lM DiSC3(5).

centration of 1 lM or higher, greatly decreased the calculated values of the external negative surface potential of liposomes in a low ionic strength medium. A significantly lower effect of the peptide on the membrane surface potential resulted for a high ionic strength medium (Fig. 6B, a and b, respectively). The peptide concentrations in the external membrane leaflet of the PC/PG liposomes (Cpm) and in the aqueous phase (Cpa) significantly depended on the peptide concentration in the system and on the ionic strength of the incubation medium, as shown in Fig. 7A. In a low ionic strength medium, the phenomenon of saturation of the concentration of membrane bound peptide was clearly observed, beginning from an approximately 0.5 lM peptide concentration in the SUV suspension (Fig. 7A, c), after which an almost linear increase in the aqueous phase peptide concentration was revealed by the calculations (Fig. 7A, a). For a high ionic strength medium, a significantly weaker effect of the saturation of the membrane-bound peptide was determined in the same range of the peptide concentrations in the sample (Fig. 7A, d and c for the membrane and the aqueous phase peptide concentrations, respectively). The computational analysis of the model (Fig. 1) also showed that the presence of the above mentioned arbitrary peptide in a low ionic strength liposomal suspension significantly affects the DiSC3(5) concentrations in both the external (Cmo) and the internal (Cmi) membrane leaflets (Fig. 7B, a and c, respectively). This is because of the displacement of DiSC3(5) from the external membrane leaflet by the amphipathic polycationic peptide (Fig. 7B, a) finally leading to an increase in the probe binding to the internal membrane leaflet (Fig. 7B, c). Relatively slight peptide effects on Cmo and Cmi were revealed in the high ionic strength medium

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

173

Fig. 6. The influence of an arbitrary peptide on the aqueous (Fa) and the membrane bound (Fm) DiSC3(5) fluorescence (panel A), and on the membrane surface potentials (panel B) calculated for PC/PG (9:1) liposomal suspensions in low (continued lines) and high ionic strength (dotted lines) media, at 0.4 lM DiSC3(5) and 5 ll PC/PG SUV/ml (0.05 lg of lipid/ml). Panel B: a, b – external leaflet; c, d – internal leaflet.

Fig. 5. The influence of an arbitrary amphipathic polycationic peptide (with the net charge of +5 and the partition coefficient Kp = 103) on the aqueous (Fa) and the membrane bound (Fm) DiSC3(5) fluorescence (panel A), on the membrane surface potentials (panel B) and the concentrations of the peptide (panel C) in the aqueous phase (Cpa) and the external membrane leaflet (Cpm) calculated for PC liposomal suspensions in low (continued lines) and high ionic strength (dotted lines) media, at 0.4 lM DiSC3(5) and 5 ll of PC SUV/ml (0.05 lg of lipid/ml). Panels A and C: a, c – at low ionic strength of incubation medium; b, d – at high ionic strength of incubation medium. Panel B: a, b – external leaflet at low and high ionic strength of incubation medium, respectively; c, d – internal leaflet at low and high ionic strength of incubation medium, respectively.

(Fig. 7B, b and d, respectively). The corresponding calculations for the DiSC3(5) fluorescence intensity in the aqueous (Fa) and the membrane (Fm) phases as the functions of the ionic strength of incubation medium and of the peptide concentration in the SUV suspension are shown in Fig. 8, indicating that higher values of both parameters lead to remarkable increase in Fa (Fig. 8A) and to a decrease in Fm (Fig. 8B). The calculations shown in Fig. 8 were made for an arbitrary peptide described above, with a net charge of +5 and Kp = 103. We also calculated Fa for other values of Kp, in the range of 101–105 and for various positive net charges of the peptide, q (Fig. 9). The obtained results for PC/PG liposomal suspension in high and low ionic strength incubation media showed that the peptides with a higher value of Kp displaced DiSC3(5) from the membrane most effectively, thus increasing the probe quantity in the aqueous phase and resulting in an increase of Fa (Fig. 9). This effect of competitive interactions was revealed more efficiently in a low ionic strength incubation medium (Fig. 9, left column) in comparison with a high ionic strength medium (Fig. 9, right column). The computational analysis of the model also showed that an increase in hydrophobicity (Kp) of peptides with low positive net charge (0–2) significantly increased the concentrations of peptide bound to the negatively charged membrane. This effect is greatly revealed in a high ionic strength medium (Fig. 10, right column),

Fig. 7. The influence of an arbitrary peptide on the concentrations of the peptide (panel A) in the aqueous phase (Cpa) and the external membrane leaflet (Cpm), and on the DiSC3(5) concentrations (panel B) in the external (a, b) and the internal (c, d) membrane leaflets calculated for PC/PG (9:1) liposomal suspensions in low (continued lines) and high ionic strength (dotted lines) media, at 0.4 lM DiSC3(5) and 5 ll PC/PG SUV/ml (0.05 lg lipids/ml).

although it is also remarkable at a low ionic strength (Fig. 10, left column). In a high ionic strength medium, the concentration of membrane-bound peptide increases with an increase in the positive net charge for low hydrophobic peptides (Fig. 10D) and significantly decreased for highly hydrophobic peptides (Fig. 10F). Hence, the model shows that the peptide hydrophobicity relatively slightly influences the saturated membrane concentrations of highly charged peptides adsorbed on the negative lipid membrane, beginning from the submicromolar range of peptide concentra-

174

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

Fig. 8. The dependence of the DiSC3(5) fluorescence in the aqueous (Fa) and the membrane (Fm) phases on the concentration of an arbitrary polycationic peptide calculated for the suspension of PC/PG (9:1) liposomes (5 ll PC/PG SUV/ml, or 0.05 lg lipids/ml) at 0.4 lM DiSC3(5). Cp – peptide concentration in the medium; Co – ionic strength of incubation medium.

tions in the medium, but it strongly increases the membrane binding for peptides with a low charge (Fig. 10). The calculations shown in Figs. 9 and 10 were performed at 0.4 lM concentration of DiSC3(5) in the suspension of PC/PG liposomes. Similar calculations were also performed at 0.1 lM DiSC3(5) in the suspensions of both PC/PG and PC liposomes. As calculated for a highly hydrophobic peptide (Kp = 105), the behavior of the aqueous phase DiSC3(5) fluorescence at 0.1 lM DiSC3(5) in the suspension of PC/PG liposomes (PG = 10%) in low and high ionic strength media (Fig. 11B and F, respectively) was very similar to that demonstrated for 0.4 lM DiSC3(5) (Fig. 9C and F, respectively), but with a lower sensitivity scale of approximately 4 times (Fig. 11). Very expressive changes of the DiSC3(5) fluorescence were also revealed for the presence of a less hydrophobic peptide (Kp = 104) in a low ionic strength medium, but not in a high ionic strength medium with PC/PG liposomes (Fig. 11A and E, respectively). Under the same conditions, but for the suspension of PC liposomes (PG = 0%), the aqueous DiSC3(5) fluorescence was only slightly changed by a highly hydrophobic cationic peptide (Kp = 105) in both low and high ionic strength media (Fig. 11D and H, respectively). The fluorescence changes were less expressive for a peptide with Kp = 104 (Fig. 11C and G, respectively) and were almost not revealed for a peptide with Kp = 103 or less (data not shown). The concentrations of a highly hydrophobic peptide (Kp = 105) bound to the negatively charged PC/PG liposomes in low and high ionic strength media at 0.1 lM DiSC3(5) (Fig. 12B and F, respectively) were very similar to those calculated at 0.4 lM

Fig. 9. The dependence of DiSC3(5) fluorescence in the aqueous phase (Fa) on the charge (q) and the concentration (Cp) of amphipathic polycationic peptides in the suspension of PC/PG (9:1) liposomes. The peptide partition coefficients: Kp = 101 (A, D), Kp = 103 (B, E) and Kp = 105 (C, F). Calculations were performed for the suspension of PC/PG (9:1) liposomes (5 ll PC/PG SUV/ml, or 0.05 lg lipids/ml) in low (A–C) and high (D–F) ionic strength media at 0.4 lM DiSC3(5).

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

175

Fig. 10. The dependence of the calculated peptide concentration in the external membrane leaflet of PC/PG (9:1) liposomes on the concentration (Cp) and the charge (q) of amphipathic polycationic peptides with the partition coefficients: Kp = 101 (A, D), Kp = 103 (B, E) and Kp = 105 (C, F). Calculations were performed for the suspension of PC/PG (9:1) liposomes (5 ll PC/PG SUV/ml, or 0.05 lg lipids/ml) in low (A–C) and high (D–F) ionic strength media at 0.4 lM DiSC3(5).

DiSC3(5) (Fig. 10C and F, respectively). The concentrations of membrane bound peptide with Kp = 104 were substantially decreased (Fig. 12A and E, respectively). Interestingly, relatively high concentrations of a peptide with Kp = 105 bound to the PC liposomes in a low ionic strength medium were only revealed for very low or zero net charges of the peptide (Fig. 12D). In a high ionic strength medium, the concentrations of the membrane bound peptides (Kp = 105) with low net charges were substantially increased (Fig. 12H) in comparison with low ionic strength medium (Fig. 12D). A decrease of the peptide hydrophobicity, from Kp = 105 to even Kp = 104, greatly decreased the peptide binding to PC liposomes in both low and high ionic strength media (Fig. 12C and G, respectively). Thus, the computational analysis clearly demonstrated the importance of the electronegative lipid membrane for the binding of amphipathic polycationic peptides at high concentrations on the membrane surface. Discussion The electrostatic attraction of polycationic antimicrobial peptides to the negatively charged membrane surfaces of bacteria and transformed cells has been recognized as the first step of the peptide–membrane interactions leading to the selective adsorption of the peptide on the membrane up to threshold concentrations [2,7,11,27,44]. During subsequent steps, the membrane can be permeabilized due to various membrane-disrupting processes, particularly by detergent-type lysis of the membrane [45] and mechanisms explained by carpet [11], barrel-stave [46] and toroidal-pore [47] models, and others [44,48,49]. The electrostatic peptide–membrane interaction can be affected by the competitive displacement of the peptide by other amphipathic compounds, like that reported for the competitive displacement of dansyl-polymyxin B from lipopolysaccharides of Gram-negative bacteria [2]

and liposomes [25] by some antimicrobial peptides. In our experiments, we observed that 10 lM safranin O, the cationic fluorescent probe normally used for monitoring the mitochondrial inner membrane potential, significantly decreased the rate of swelling of isolated mitochondria induced by the antimicrobial polycationic peptide BTM-P1 and by its retro-analog, retro-BTM-P1 (Fig. 2). The obtained experimental data allowed the assumption that safranin O is capable of displacing the membrane-bound peptides, thus decreasing their concentrations on the membrane surface. To confirm this hypothesis, we next developed the experimental liposomal model with a certain percentage of negatively charged lipids to mimic a prokaryotic cell membrane with the aim to evaluate the competitive interactions of polycationic peptides and the cationic fluorescent probe DiSC3(5). It was clearly observed that the peptides BTM-P1 (Fig. 3A) and P7-5 (Fig 4A) at the final peptide concentration of 2–3 lM completely displaced DiSC3(5) from the liposomal membrane. Under these conditions, the added FCCP, hydrophobic weak acid protonophore, induced an increased capture of the cationic probe DiSC3(5) from the aqueous phase. The affinity of the membrane to the peptide BTM-P1 was clearly increased in the presence of FCCP, because in this case only 1 lM peptide completely displaced DiSC3(5) from the membrane (Fig. 3C) seemingly due to an increase in the membrane’s negative surface charge resulting from the binding of an anionic form of FCCP. As the displacement effect significantly increased by a decrease in the ionic strength of the incubation medium (Fig. 4B–D), it might indicate the possibility of the electrostatic repulsion of DiSC3(5) from the external membrane leaflet by the adsorbed peptide. Hence, the interaction of the antimicrobial peptide BTM-P1, of the anticancer peptide P7-5 or of any other amphipathic cationic peptides with the lipid membrane might be monitored by the fluorescence intensity of the hydrophobic cationic probe DiSC3(5) or similar.

176

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

Fig. 11. The dependence of DiSC3(5) fluorescence in the aqueous phase (Fa) on the charge (q) and the concentration (Cp) of amphipathic polycationic peptides in the suspensions of PC/PG (PG = 10%) liposomes (A, E, B, F) and PC (PG = 0%) liposomes (C, G, D, H). The peptide partition coefficients: Kp = 104 (A, E, C, G) and Kp = 105 (B, F, D, H). Calculations were performed for the suspensions of liposomes (0.05 lg lipids/ml) in low (A–D) and high (E–H) ionic strength media at 0.1 lM DiSC3(5).

To describe the mentioned effects of electrostatic competition in more general form, we further developed the computational model allowing the prediction of the peptide interaction with the external leaflet of the liposomal membrane as the functions of the membrane surface charge, peptide charge and hydrophobicity, and of the ionic strength of incubation medium. The model was based on the Gouy–Chapman theory of the electrical double layer. Several models of this type have been recently suggested to describe the peptide–membrane interactions. However, the surface charge of the lipid membrane has been considered to be fixed, calculated from the mol fraction of anionic phospholipids [29,30]. On the other hand, in a model not related to the peptide–membrane interactions, but to the interaction of 1-anilino-8-naphthalenesulfonate, an anionic fluorescent probe, with the membrane, the varied value of the surface potential has been considered as an effective membrane surface charge resulting from the binding of inorganic ions to the membrane [31]. In our model, we considered the effective surface charge as the result of the algebraic sum of the charges of phospholipids, adsorbed peptide, and fluorescent probe.

The model can be extended to consider the influence of other charged compounds, as the FCCP anion for example (Fig. 3B). The developed model was focused on the demonstration that the fluorescent properties of the cyanine cationic probe DiSC3(5), or similar, can be used to study peptide–membrane interactions. The computational analysis of the model (Fig. 1) demonstrated that the membrane negative surface potential is crucial for adsorbing polycationic peptides on the membrane surface up to high, threshold concentrations to induce membrane permeabilization or damage [2,7,11,27,44]. The first calculations were performed for SUV suspensions in the presence of an arbitrary polycationic peptide with the net charge of +5 and the partition coefficient Kp = 103. The calculations showed that in the peptide concentration in the external membrane leaflet of the PC/PG (9:1) liposomes in a high ionic strength medium were more than 2 orders of magnitude higher that of the liposomes formed of electrically neutral phospholipids (Fig. 7A, d and Fig. 5C, d, respectively). For a low ionic strength medium, this difference between liposomes increases from 2 to 3 orders of magnitude (Fig. 7A, c and Fig. 5C,

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

177

Fig. 12. The dependence of the calculated peptide concentration in the external membrane leaflet of PC/PG (PG = 10%) liposomes (A, E, B, F) and of the PC (PG = 0%) liposomes (C, G, D, H) on the concentration (Cp) and the charge (q) of amphipathic polycationic peptides with the partition coefficients: Kp = 104 (A, E, C, G) and Kp = 105 (B, F, D, H). Calculations were performed for the suspension of PC/PG (PG = 10%) (A, E, B, F) and PC (PG = 0%) (C, G, D, H) liposomes (0.05 lg lipids/ml) in low (A–D) and high (E–H) ionic strength media at 0.1 lM DiSC3(5).

c, respectively). In a low ionic strength medium, the displacement of DiSC3(5) from the membrane by an adsorbed peptide might be monitored by its fluorescence in the aqueous and membrane phases (Fig. 6A, a and c, respectively, and Fig. 8). This displacement effect was due to the significant neutralization of the negative surface charge of the external membrane leaflet by the bound polycationic peptide (Fig. 6). The computational model behavior is consistent with our experimental data shown in Figs. 2–4. Experimental evidence of the neutralization of a negative surface charge of the Gram-negative bacteria E. coli by some novel antimicrobial peptides have been recently obtained by the zeta potential studies of these cells [50]. The authors found a correlation between the minimal inhibitory concentration of each peptide and membrane surface charge neutralization, thus confirming the hypothesis that surface charge neutralization occurs close to the peptide concentrations at which the bacterial surface disruptions take place [50,51]. Interestingly, many uniformly cationic peptides have been found to interact with lipid monolayers composed of anionic PG and to induce calcein release from liposomes regardless of peptide conformation, including alfa-helical, beta-sheet, extended and cyclic motifs [52]. Hence, charge affinity is an important means conferring membrane selectivity to antimicrobial peptides [2]. A number of cationic antimicrobial peptides have been shown to be inactive against human red blood cells and Candida albicans cells in a high ionic strength medium, but displayed strong activity against these cells in a low ionic strength glucose medium [53]. On the other hand, many pathogens are characterized by a highly elec-

tronegative cell membrane that has been assumed to provide an important determinant by which antimicrobial peptides target microbial versus host membranes known to have a generally neutral net charge [49], in good concordance with the model predictions (Figs. 5–7). Many antimicrobial peptides are moderately hydrophobic, in addition to displaying a positive net charge ranging from +2 to +9 [2]. The computational analysis of both this aspects, of the net charge q, varied from 0 to +7, and of the hydrophobic partition coefficient Kp, varied in the range of 101–105, clearly demonstrated the importance of moderate hydrophobicity to conserve the preferential affinity of polycationic peptides to the electronegative lipid membrane (Fig. 10). In the case of a low ionic strength medium, the model predicts the highest affinities of the PC/PG (9:1) liposomes for low hydrophobic peptides (Kp = 101) with the positive net charge equal to +2 or higher (Fig. 10A). The optimum net charge was shifted to +1 for highly hydrophobic peptides, Kp = 105 (Fig. 10C). In the case of a high ionic strength medium, significant increase in the positive charge, up to +7 or higher, was needed to essentially increase membrane binding of low hydrophobic peptides (Fig. 10C). Again, for highly hydrophobic peptides, the optimum net charge was shifted to +1 to achieve a maximum adsorption effect, which strongly decreased with an increase in the positive net charge of the peptide (Fig. 10F). Thus, the electrostatic component of the interaction of highly hydrophobic peptides with the electronegative lipid membrane is

178

V.V. Lemeshko / Archives of Biochemistry and Biophysics 545 (2014) 167–178

significantly decreased in comparison with the hydrophobic component that might explain the loss of antimicrobial specificity of such amphipathic peptides [2]. Even for highly hydrophobic polycationic peptides, the computational analysis clearly demonstrated (Fig. 12) the importance of the electronegative lipid membrane for the peptide binding at high concentrations on the membrane surface that might be monitored by the fluorescence intensity of some cationic dyes. Conclusion The described computational model allows the optimization of a net charge of a peptide with given hydrophobicity, and vice versa, to prognosticate the highest binding of the peptide to the electronegative lipid membrane without the essential loss of antimicrobial specificity. In addition, the model allows to determine the most optimal experimental conditions for monitoring the peptide–membrane interactions using fluorescent cationic probes like DiSC3(5). We also demonstrated that at least some cationic fluorescent probes, like safranin O, at the concentrations normally used for monitoring membrane potentials, might significantly decrease membrane effects caused by studied amphipathic polycationic peptides. Acknowledgments

[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]

Financial support for this work was provided by the Colciencias (Colombia) research grants #111852128625 and #5201-545-3156 (362-2011) and by the National University of Colombia, Medellin Branch.

[38] [39]

References

[42] [43] [44] [45] [46] [47] [48]

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]

M. Zasloff, Nature 415 (2002) 389–395. M.R. Yeaman, N.Y. Yount, Pharmacol. Rev. 55 (2003) 27–55. G.G. Perron, M. Zasloff, G. Bell, Proc. Biol. Sci. 271 (2006) 251–256. N.J. Afacan, A.T. Yeung, O.M. Pena, R.E. Hancock, Curr. Pharm. Des. 18 (2012) 807–819. X. Han, W. Kang, Bioinformatics 20 (2004) 970–973. K.L. Brown, R.E. Hancock, Curr. Opin. Immunol. 18 (2006) 24–30. D.W. Hoskin, A. Ramamoorthy, Biochim. Biophys. Acta 1778 (2008) 357–375. Z. Jiang, A.I. Vasil, J.D. Hale, R.E. Hancock, M.L. Vasil, R.S. Hodges, Biopolymers 90 (2008) 369–383. Ø. Rekdal, B.E. Haug, M. Kalaaji, H.N. Hunter, I. Lindin, I. Israelsson, T. Solstad, N. Yang, M. Brandl, D. Mantzilas, H.J. Vogel, J. Biol. Chem. 287 (2012) 233–244. R.E. Hancock, A. Rozek, FEMS Microbiol. Lett. 206 (2002) 143–149. Y. Shai, Biochim. Biophys. Acta 1462 (1999) 55–70. T.J. Falla, D.N. Karunaratne, R.E. Hancock, J. Biol. Chem. 271 (1996) 19298– 19303.

[40] [41]

[49] [50]

[51] [52] [53]

V.V. Lemeshko, M. Arias, S. Orduz, J. Biol. Chem. 280 (2005) 15579–15586. V.V. Lemeshko, Arch. Biochem. Biophys. 493 (2010) 213–220. V.V. Lemeshko, Peptides 32 (2011) 2010–2020. M. Arias, S. Orduz, V.V. Lemeshko, Biochim. Biophys. Acta 1788 (2009) 532– 537. V.V. Lemeshko, Biochim. Biophys. Acta 2013 (1828) 1047–1056. C. Wang, L.L. Tian, S. Li, H.B. Li, Y. Zhou, H. Wang, Q.Z. Yang, L.J. Ma, D.J. Shang, PLoS One 8 (2013) e60462. T. Iwasaki, J. Ishibashi, H. Tanaka, M. Sato, A. Asaoka, D. Taylor, M. Yamakawa, Peptides 30 (2009) 660–668. L.B. Chen, Annu. Rev. Cell Biol. 4 (1988) 155–181. N. Papo, Y. Shai, Cell. Mol. Life Sci. 62 (2005) 784–790. S. Riedl, D. Zweytick, K. Lohner, Chem. Phys. Lipids 164 (2011) 766–781. S.J. Kang, H.Y. Ji, B.J. Lee, Arch. Pharm. Res. 35 (2012) 791–799. S. Krasne, Biophys. J. 30 (1980) 441–462. R. Halevy, A. Rozek, S. Kolusheva, R.E. Hancock, R. Jelinek, Peptides 24 (2003) 1753–1761. M.N. Melo, M.A. Castanho, Biochim. Biophys. Acta 1768 (2007) 1277–1290. M.N. Melo, R. Ferre, L. Feliu, E. Bardají, M. Planas, M.A. Castanho, PLoS One 6 (2011) e28549. L. Chen, N. Jia, L. Gao, W. Fang, L. Golubovic, Int. J. Mol. Sci. 14 (2013) 7932– 7958. T. Lazaridis, Proteins 58 (2005) 518–527. D. Shental-Bechor, T. Haliloglu, N. Ben-Tal, Biophys. J. 93 (2007) 1858–1871. D.H. Haynes, J. Membr. Biol. 17 (1974) 341–366. M.R. Wieckowski, L. Wojtczak, FEBS Lett. 423 (1998) 339–342. V.V. Lemeshko, S. Orduz, Biosci. Rep. 33 (2013) 125–136. V.V. Lemeshko, V. Haridas, J.C. Quijano Perez, J.U. Gutterman, Arch. Biochem. Biophys. 454 (2006) 114–122. M.N. Ivkova, V.A. Pechatnikov, V.G. Ivkov, Gen. Physiol. Biophys. 3 (1984) 97– 117. G. Cabrini, A.S. Verkman, J. Membr. Biol. 92 (1986) 171–182. M.N. Ivkova, V.A. Pechatnikov, O.A. Gracheva, E.V. Pechatnikova, V.G. Ivkov, Gen. Physiol. Biophys. 6 (1987) 45–55. C. Xu, L.M. Loew, Biophys. J. 84 (2003) 2768–2780. R. Ferre, M.N. Melo, A.D. Correia, L. Feliu, E. Bardají, M. Planas, M. Castanho, Biophys. J. 96 (2009) 1815–1827. M. Rodrigues, A. Santos, B.G. de la Torre, G. Rádis-Baptista, D. Andreu, N.C. Santos, Biochim. Biophys. Acta 2012 (1818) 2707–2717. J.P. Powers, A. Tan, A. Ramamoorthy, R.E. Hancock, Biochemistry 44 (2005) 15504–15513. G. Cabrini, A.S. Verkman, J. Membr. Biol. 90 (1986) 163–175. C. Huang, J.P. Charlton, J. Biol. Chem. 246 (1971) 2555–2560. L.T. Nguyen, E.F. Haney, H.J. Vogel, Trends Biotechnol. 29 (2011) 464–472. B. Bechinger, K. Lohner, Biochim. Biophys. Acta 1758 (2006) 1529–1539. Z. Oren, Y. Shai, Biopolymers 47 (1998) 451–463. K.A. Brogden, Nat. Rev. Microbiol. 3 (2005) 238–250. E.J. Paredes-Gamero, M.N. Martins, F.A. Cappabianco, J.S. Ide, A. Miranda, Biochim. Biophys. Acta 2012 (1820) 1062–1072. J.M. Sanderson, Org. Biomol. Chem. 3 (2005) 201–212. C.S. Alves, M.N. Melo, H.G. Franquelim, R. Ferre, M. Planas, L. Feliu, E. Bardají, W. Kowalczyk, D. Andreu, N.C. Santos, M.X. Fernandes, M.A. Castanho, J. Biol. Chem. 285 (2010) 27536–27544. M.N. Melo, M.A. Castanho, Front. Immunol. 3 (2012) 236. L. Zhang, A. Rozek, R.E. Hancock, J. Biol. Chem. 276 (2001) 35714–35722. E.J. Helmerhorst, I.M. Reijnders, W. van ’t Hof, E.C. Veerman, A.V. Nieuw Amerongen, FEBS Lett. 449 (1999) 105–110.