Article (5) - 2008 - Kolusheva Et Al

peptides 29 (2008) 1620–1625

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/peptides

Short communication

Pardaxin, a fish toxin peptide interaction with a biomimetic phospholipid/polydiacetylene membrane assay Sofiya Kolusheva a, Shimon Lecht b, Yael Derazon b, Raz Jelinek a, Philip Lazarovici b,* a

Ilse Katz Center for Meso- and Nano-Scale Science and Technology, Ben Gurion University, Beer-Sheva, 84105, Israel Department of Pharmacology and Experimental Therapeutics, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem, 91120, Israel

b

article info

abstract

Article history:

Pardaxin is a fish toxin belonging to the a-helical, pore-forming peptide family, used in

Received 1 April 2008

toxicological and biophysical research to study toxin-cell and -lipid-artificial membranes

Received in revised form

interactions. We investigated the membrane interaction of two pardaxin analogues using a

12 May 2008

colorimetric phospholipid/polydiacetylene biomimetic assay. In this assay, polydiacetylene

Accepted 13 May 2008

undergoes visible, concentration dependent, blue–red transformation induced through

Published on line 23 May 2008

interactions of pardaxins with the vesicle membrane. Pardaxins P4 and P5, are composed of 33 amino acids, but differ in a single amino acid substitution at the carboxy-terminal (G31 to D31, respectively) known to decrease the pore forming activity. Addition of pardaxins in the colorimetric assay induced dose-dependent color transitions with different kinetics. The colorimetric analysis could distinguish between different pardaxins–membrane interaction profiles, suggesting bilayer surface association for P4 and vesicle membrane penetration for P5. The colorimetric assay could distinguish between pardaxins membrane interaction profiles although circular dichroism spectra of vesicle-interacting pardaxins did not indicate a significant difference in the secondary structure between these two toxin analogues. The colorimetric platform utilized in the present report represents a useful assay with general applications for studying membrane interactions of peptides in general and pore-forming toxins in particular, and may become an important tool for evaluating quantitative toxin structure–activity relationship. # 2008 Elsevier Inc. All rights reserved.

1.

Introduction

Interest in membrane-targeting peptides has increased because of the tremendous problem of antibiotic resistance. Antimicrobial, pore-forming peptides from bacterial, plant and animal origins, acting as perforators of biological membranes are suggested as alternative tools to overcome antibiotic resistance [26].

Pardaxins represent a well-known family of potent poreforming peptidic toxins secreted by Moses sole fish from the Red Sea and Pacific Ocean [13]. These single-chain, 33-residue peptides exhibit acidic and amphipatic properties [2]. Pardaxin peptides tend to interact with phospholipid bilayer membranes of different compositions [6,16,19]. Previous studies employing artificial membranes and whole-cell systems have correlated pardaxin-induced cytotoxicity with pore-formation

* Corresponding author at: Department of Pharmacology and Experimental Therapeutics, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Ein-Kerem POB 12065, Jerusalem 91120, Israel. Tel.: +972 2 6758729; fax: +972 2 6757490. E-mail address: [email protected] (P. Lazarovici). 0196-9781/$ – see front matter # 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.peptides.2008.05.012

peptides 29 (2008) 1620–1625

activity [15,20]. Molecular characteristics of pardaxins, e.g. amphipaticity and a-helical structure, are similar to other membrane-active peptides such as melittin, gramicidin, alamethicin, dermaseptin, cecropin, magainin, defensins and cathelicidins [2,3,5,6]. In this report, we focused on two pardaxin analogues: native pardaxin, denoted pardaxin P4, isolated from the secretion of the Moses sole, P. marmoratus, fish [13] and a synthetic pardaxin, denoted pardaxin P5, whose sequence differs by a single point mutation [1]. Despite having highly similar sequences the two peptides show different levels of toxicity and pore-forming activity [1,13]. The following study utilized a newly-developed biomimetic colorimetric vesicle assay [10] to characterize the interaction of pardaxins with the phospholipids organized in an artificial membrane. This biomimetic assay was previously used for the study of membrane disruption by anti-microbial peptides [8,11] and drugs [9]. The chemical basis of this assay is the production of vesicular assemblies of phospholipids and polydiacetylene (PDA) in aqueous solutions, that exhibit blue color due to the conjugated poly(ene-yne) backbone of PDA. Importantly, it has been demonstrated that the phospholipid molecules in the colorimetric vesicles assay form ‘‘microdomains’’ within the PDA matrix that resemble biological plasma membrane environment [21]. In such assemblies, the blue–red transformations of the PDA occurs following interactions between membrane active compounds and the phospholipids bilayer micro-domains. Therefore, this assay is very fast, simple and useful for rapid screening of chemicals interaction with biomimetic membrane [21]. The aim of the present study was to characterize pardaxins’ interactions with the membrane phospholipids using this new colorimetric assay and to assess whether the assay can distinguish between pardaxin analogues differing by a single point mutation. Differences in kinetics, potency and efficacy have been found between pardaxin analogues, emphasizing the applicability of this assay for quantitative structure–activity relationship of membrane-acting peptides.

2.

Materials and methods

2.1.

Materials and toxins

The diacetylene monomer 10,12-tricosadiynoic acid was purchased from GFS Chemicals (Powell, OH, USA). Dimyristoyl-phosphatidylethanolamine (DMPE) and dimyristoylphosphatidylglycerol (DMPG) were purchased from Sigma (St. Louis, MO, USA). Pardaxin P4 was isolated from the secretion of the P. marmoratus fish as previously described [13]. The pardaxin P5 analogue was synthesized on a 433A peptide synthesizer (Perkin Elmer/ABI) using standard Fmoc solid phase chemistry as previously described [1]. The identity and purity of the pardaxin analogues were confirmed by analytical HPLC and ES-MS [1].

2.2.

Vesicle preparation

Vesicle emulsion containing DMPE, DMPG and PDA (1:1:3 molar ratio) were prepared as previously described [10]. Briefly, the lipid constituents were dried together in vacuo

1621

followed by addition of de-ionized water and sonication at 70 8C. The vesicle emulsion was then cooled to room temperature and kept at 4 8C overnight, then polymerized by irradiation at 254 nm for a few seconds. The resulting emulsion exhibits a blue appearance. Vesicle samples for experiments were prepared at concentrations of 1 mM (total lipids) in 50 mM Tris-base, pH 8. The various toxins at different concentrations were mixed with the vesicles prior to addition of the Tris buffer.

2.3. Determination of the toxin-vesicle partition coefficients Analysis of the colorimetric transitions-induced by the pardaxin analogues require comparison between the effects of peptides bound to the phospholipid/PDA vesicles, rather than total concentration of the dissolved toxins in the suspension. Accordingly, we have determined the concentration of vesicle-bound peptide by a partition coefficient assay. Specifically, a calibration graph relating pardaxin concentration with UV absorbance at 220 nm was initially prepared and used to determine the concentration of the soluble, unbound toxin. Toxins at different concentrations were added to the aqueous vesicle emulsion (0.5 mM, Tris buffer 25 mM at pH 8). The emulsion was left at room temperature for a few minutes to allow equilibration and followed by ultracentrifugation at 30,000 rpm for 40 min in order to precipitate the vesicle-toxin aggregates. The concentration of soluble (unbound) pardaxin in the supernatant was determined from the calibration curve, and was subtracted from the initial concentration to yield the amount of vesicle-bound toxin. Concentrations of unbound toxin were confirmed using the Lowry method for protein quantification.

2.4.

UV–vis spectroscopy

All spectroscopic measurements were carried out at 27 8C using a Jasco V550 UV-visible spectrophotometer (Osaka, Japan), with a 1 cm optical path cell. Quantification of the extent of blue–red color transition is given by the colorimetric response (CR%), defined by [17]: CR% = ((PB0  PBI)/PB0)  100 and PB = Ablue/(Ablue + Ared), where A is the absorbance measured at either the ‘‘blue’’ component (640 nm) or the ‘‘red’’ component (500 nm) of the visible spectrum. PB0 is the red/blue ratio of the control sample (i.e., before induction of color changes by addition of peptides), while PBI is the value obtained for the vesicle emulsion after occurrence of the color change. All reported CR% values represent average between 6–7 independent measurements, carried out for different batches of vesicles emulsions.

2.5.

Circular dichroism spectroscopy

CD spectra were acquired on an Aviv 62A-DS Circular Dichroism Spectrometer (Aviv Inc., NJ). Three scans were recorded between 195 and 250 nm with 1 nm acquisition steps. A 2 mm optical path length was used. Samples were prepared by adding 0.4 mM of peptides to vesicle solutions at 0.25 and 0.5 mM total lipid concentration and 10 mM Tris-base buffer pH 8. To compensate for scattering due to the vesicles, the CD

1622

peptides 29 (2008) 1620–1625

spectrum of a vesicle solution alone was subtracted from that of the peptide in the presence of the vesicles. The measured deflections in milidegrees were converted into molar elipticities and the estimation of the protein secondary structure were made as previously described by us [14].

2.6.

Molecular modeling

The modeling work was performed on a Silicon Graphics IRIS6.5 workstation using InsightII modeling software (MSI Inc., UK) with the assistance of the biopolymer module. Pardaxin P4 model was built up from its amino acids sequence from the N terminal to the C terminal. Secondary structure of a-helix was assign to the amino acids numbered 2–10 (first helix) and 16–25 (second helix), as previously described [2]. All models were optimized by molecular mechanics using the cvff force field. Optimization of the initial models included Steepest Descent minimization, dielectric constant dependent, followed by conjugate one to retrieve an optimize structure of the pardaxin P4 [7]. In order to achieve the pardaxin P5 structure, a similar procedure was performed but G31 was changed to D31.

3.

Results

Pardaxins P4 and P5 contain 33 amino acids and differ by a single point mutation G31 ) D31 as presented in Fig. 1A. Molecular modeling of pardaxin analogues in solution (Fig. 1B) indicated a 97% homology, as compared by amino acid backbone structure. Overall, the molecular modeling analysis suggests that the mutation G31 ) D31 gives rise to conformational change as depicted in Fig. 1B. Pardaxin structure is bend-helix–bend-helix: residues 7–12 are in a right handed helix (Fig. 1B: N-terminal proximal helix), whereas amino acids 14–26 form an alpha helix (Fig. 1B: C-terminal proximal helix). A hinge center on proline 13 separates the two helixes, causes molecule bending and is essential for the toxin function [6,14,23]. The carboxy-terminal of pardaxin, including the mutated glycine at position 31, is unstructured [14]. This model is similar to the high resolution structure of pardaxin measured by solution and solid-state NMR spectroscopy [18,25]. According to these models, the substitution of the positively charged glycine with the negatively charged aspartic acid is hypothesized to reduce the overall charge of the peptide which in turn may affect the interaction of mutated pardaxin, P5, with the membrane bilayer phospholipids. Based on the molecular model of superimposed pardaxins, presented in the Fig. 1B, this amino acid substitution may induce a deformation in the neighboring C-terminal helical part as well as a change in the coil organization of the mutated (P5) compared to native (P4) pardaxin. Since the Cterminal domain of pardaxin is involved in the formation of the ion channel [15,24] this mutation may affect pardaxin channel selectivity for ions. These hypotheses deserve future investigations. CD measurements were made to examine the effect of phospholipid binding on the secondary structure of mutated pardaxin P5 in comparison to native toxin pardaxin P4 (Fig. 1C). In reasonable agreement with the predicted

Fig. 1 – (A) The primary structure of native pardaxin (P4) and single point mutated analogue (P5). (B) Superposition of molecular modeled pardaxin P4 (blue) and pardaxin P5 (red). The N-terminal and C-terminal including the G31 ) D31 mutation point are indicated. (C) CD spectra expressed as molar elipticity of pardaxins P4 and P5 (0.4 mM) alone (i) or in the presence of 0.25 mM (ii) or 0.5 mM (iii) phospholipid vesicle concentrations. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

conformation (Fig. 1B), the CD measurements of both pardaxin analogues in TFE (Fig. 1C: i spectra) indicate about 20% alpha helix and 25% beta sheet and about 55% coil similar to previous measurements [14,23]. Upon interaction of pardaxins with phospholipid vesicles, at two different concentrations (Fig. 1C – ii, iii) several spectral changes were observed. These changes mainly indicate an increase of about 50% in alpha helicity in the presence of 0.5 mM phospholipids. The changes in the CD spectrum upon lipid vesicles interactions were similar between pardaxins P4 and P5 and cannot explain the differences in the pore-forming activity of these two toxin analogues [1]. Therefore, to find a functional difference between these two analogues we investigated their interaction with phospholipid biomimetic vesicles using a novel, sensitive, colorimetric-spectroscopic assay.

peptides 29 (2008) 1620–1625

1623

Fig. 2 – (A) A schematic drawing of the structural transition of phospholipid/polydiacetylene assemblies corresponding to the blue–red color changes induced by peptide insertion (grey helix). (B) Dose–response and kinetic (0.3 mM peptide, insert) relationships between colorimetric response (CR%) and peptide concentrations (mM) in emulsions of DMPG/DMPE/PDA (1:1:3 mol ratio) vesicles. Filled squares: pardaxin P4; empty squares: pardaxin P5; filled triangles: magainin. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

The phospholipid/PDA colorimetric vesicle assay (Fig. 2A) measures peptide-membrane interactions through observation of blue–red color transitions. The degrees of blue–red color transitions can be generally correlated to the extent of bilayer incorporation by the vesicle-bound peptide [10,11]. Accordingly, the different dose–response-induced transitions by the two pardaxins analogues (Fig. 2B) suggest that perturbation of phospholipid bilayers following binding by pardaxin P4 was higher compared to pardaxin P5. Essentially, higher colorimetric response (CR%) values represent more pronounced red appearance of the vesicle emulsion. The results in Fig. 2B indicate that CR values correlate with the concentration of vesicle-bound pardaxins (after accounting for the partition coefficients as described under Materials and Methods). This increase in CR is dose-dependent between 0 and 0.4 mM for pardaxins P4 (EC50 = 0.21 mM) and between 0 and 0.6 mM for pardaxins P5 (EC50 = 0.27 mM). Although the potency of these toxins is similar, as expressed in similar EC50 values, their efficacy is different as expressed in the maximal CR values. Pardaxin P4 maximal effect is two fold higher than that of pardaxins P5 (Fig. 2B). Kinetic measurements of the colorimetric transitionsinduced by pardaxins P4, P5 and magainin, a well known pore forming peptide used as a reference peptide, were recorded using similar concentration (0.4 mM) of each peptide (Fig. 2B insert). In contrast to magainin, which induced an almost instantaneous color transition, pardaxins P4 and P5 induced gradual color changes. Pardaxin P4 induced a faster

color transition reaching a plateau after 30 s compared to pardaxin P5 that slowly reached plateau at about 200 s.

4.

Discussion

A significant difference between the dose–response curves of pardaxins P4 and P5 (Fig. 2B) is suggestive of different ways of membrane interactions. The differences in kinetics shed further light on the characteristics of membrane interactions of these peptides. The extremely rapid color transition of magainin corresponds to the strong bilayer attachment of this amphipathic peptide [4]. The slow increase of CR for pardaxins is ascribed to slower cooperative phenomena occurring within the bilayer, most likely due to pore organization processes [15]. The faster transition of P4 compared to P5 is probably related to bilayer-surface localization of the peptide (which was also apparent from the dose–response curves in Fig. 2B). In principle, a direct relationship exists between higher CR% and bilayer-surface binding, because such interactions give rise to higher mobility of the pendant side-chains of the adjacent polydiacetylene matrix [21]. Based on the differences in efficacy and kinetics of pardaxins interaction with phospholipid/PDA vesicles, it is hypothesized that pardaxin P4 experiences more pronounced membrane surface association, while pardaxin P5 exhibits a deeper insertion into the phospholipid vesicle bilayer. Indeed, perturbation of the phospholipid headgroup region by membrane-

1624

peptides 29 (2008) 1620–1625

active peptides was previously shown to induce a greater increase in CR% as a function of the quantity of bound peptide, while peptides that penetrated deeper into the hydrophobic core of the membrane bilayer produced a lower chromatic shift [9,11,22]. The pardaxin analogues penetrate slowly into the lipid bilayer, and most likely require supramolecular organization – thus giving rise to the slower CR increase. Indeed, such an organization was previously hypothesized as an aggregation process occurring both within phospholipid vesicles and in planar phospholipid bilayers [2]. Another supporting evidence for the membrane surface association of P4 compared to membrane penetration of P5 is provided by CD measurements, which indicated a molecular elipticity ratio 220 nm/209 nm higher than 1 for P4 and lower than 1 for P5 (Fig. 1C). A ratio higher than 1 is indicative of an oligomeric state of the peptide as for P4 and a ratio lower than 1 is indicative of an monomeric state of the peptide as for P5 [12]. Therefore, we assume that upon interaction with the phospholipid/PDA vesicle, P4 being oligomeric is more surface associated while P5 being monomeric tends to penetrate into the membrane. The present hypothesis may also suggest that the higher cytolytic activity on human erythrocytes of pardaxin P4 compared to P5 [1] is due to surface association rather than membrane penetration. It is evident that the phospholipid/PDA vesicle assay exhibits high sensitivity to single amino changes, thus enabling quantitative structure–activity relationship of membrane-acting peptides, in particular pore-forming peptides. Since pore-forming peptides are considered today as leading compounds for development of non-resistance inducible antibiotic drugs, the phospholipid/PDA vesicle assay could become a highly useful tool for predicting membrane interactions and bilayer penetration at early stages of drug discovery and development.

Acknowledgements The authors thank Dr. Boris Fain for the help with the molecular modeling. The generous financial support from the Pharmalogica Consortium, MAGNET Program of the Israeli Ministry of Trade and Industry is gratefully acknowledged. PL is affiliated and supported in part by David R. Bloom Center for Pharmacy, School of Pharmacy, The Hebrew University of Jerusalem. SL is supported by ‘‘Eshkol fellowship’’ from the Israeli Ministry of Science, Culture and Sport.

[4] [5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

references [19] [1] Adermann K, Raida M, Paul Y, Abu-Raya S, BlochShilderman E, Lazarovici P, et al. Isolation, characterization and synthesis of a novel paradaxin isoform. FEBS Lett 1998;435:173–7. [2] Bloch-Shilderman E, Abu-Raya S, Lazarovici P. Ionophore polypeptide toxins and signal transduction. In: Gutman Y, Lazarovici P, editors. Cellular and molecular mechanisms of toxin action: toxins and signal transduction. Amsterdam: Harwood Academic Publishers, Gmbh; 1997. p. 211–32. [3] Dhople V, Krukemeyer A, Ramamoorthy A. The human beta-defensin-3, an antibacterial peptide with multiple

[20]

[21]

[22]

biological functions. Biochim Biophys Acta 2006;1758: 1499–512. Duclohier H. Anion pores from magainins and related defensive peptides. Toxicology 1994;87:175–88. Du¨rr UH, Sudheendra US, Ramamoorthy A. LL-37, the only human member of the cathelicidin family of antimicrobial peptides. Biochim Biophys Acta 2006;1758:1408–25. Hallock KJ, Lee DK, Omnaas J, Mosberg HI, Ramamoorthy A. Membrane composition determines pardaxin’s mechanism of lipid bilayer disruption. Biophys J 2002;83:1004–13. Ittah S, Michaeli A, Goldblum A, Gat U. A model for the structure of the C-terminal domain of dragline spider silk and the role of its conserved cysteine. Biomacromolecules 2007;8:2768–73. Katz M, Tsubery H, Kolusheva S, Shames A, Fridkin M, Jelinek R. Lipid binding and membrane penetration of polymyxin b derivatives studied in a biomimetic vesicle system. Biochem J 2003;375:405–13. Katz M, Ben-Shlush I, Kolusheva S, Jelinek R. Rapid colorimetric screening of drug interaction and penetration through lipid barriers. Pharm Res 2006;23:580–8. Kolusheva S, Boyer L, Jelinek R. A colorimetric assay for rapid screening of antimicrobial peptides. Nat Biotechnol 2000;18:225–7. Kolusheva S, Shahal T, Jelinek R. Peptide–membrane interactions studied by a new phospholipid/ polydiacetylene colorimetric vesicle assay. Biochemistry 2000;39:15851–9. Lau SY, Taneja AK, Hodges RS. Synthesis of a model protein of defined secondary and quaternary structure. Effect of chain length on the stabilization and formation of twostranded alpha-helical coiled-coils. J Biol Chem 1984;259:13253–61. Lazarovici P, Primor N, Loew LM. Purification and poreforming activity of two hydrophobic polypeptides from the secretion of the red sea moses sole (pardachirus marmoratus). J Biol Chem 1986;261:16704–13. Lazarovici P, Edwards C, Raghunathan G, Guy HR. Secondary structure, permeability and molecular modeling of pardaxin pores. J Nat Toxins 1992;1:1–15. Lazarovici P. Pardaxins, pore-forming neurotoxins as pharmacological tools in dissecting neurotransmitter exocytosis and neurotoxicity. In: Menestrina G, Serra MD, Lazarovici P, editors. Pore-forming peptides and protein toxins. London: Taylor and Francis Books Ltd.; 2003. p. 178–208. Lelkes PI, Lazarovici P. Pardaxin induces aggregation but not fusion of phosphatidylserine vesicles. FEBS Lett 1988;230:131–6. Okada S, Peng S, Spevak W, Charych D. Color and chromism of polydiacetylene vesicles. Acc Chem Res 1998;31:229–39. Porcelli F, Buck B, Lee DK, Hallock KJ, Ramamoorthy A, Veglia G. Structure and orientation of pardaxin determined by NMR experiments in model membranes. J Biol Chem 2004;279:45815–23. Rapaport D, Shai Y. Interaction of fluorescently labeled pardaxin and its analogueues with lipid bilayers. J Biol Chem 1991;266:23769–75. Rapaport D, Shai Y. Aggregation and organization of pardaxin in phospholipid membranes. A fluorescence energy transfer study. J Biol Chem 1992;267:6502–9. Rozner S, Kolusheva S, Cohen Z, Dowhan W, Eichler J, Jelinek R. Detection and analysis of membrane interactions by a biomimetic colorimetric lipid/polydiacetylene assay. Anal Biochem 2003;319:96–104. Satchell DP, Sheynis T, Shirafuji Y, Kolusheva S, Ouellette AJ, Jelinek R. Interactions of mouse paneth cell alphadefensins and alpha-defensin precursors with membranes.

peptides 29 (2008) 1620–1625

Prosegment inhibition of peptide association with biomimetic membranes. J Biol Chem 2003;278: 13838–46. [23] Shai Y, Bach D, Yanovsky A. Channel formation properties of synthetic pardaxin and analogues. J Biol Chem 1990;265:20202–9. [24] Shai Y. Mechanism of the binding, insertion and destabilization of phospholipid bilayer membranes by

1625

alpha-helical antimicrobial and cell non-selective membrane-lytic peptides. Biochim Biophys Acta 1999;1462:55–70. [25] Zagorski MG, Norman DG, Barrow CJ, Iwashita T, Tachibana K, Patel DJ. Solution structure of pardaxin P-2. Biochemistry 1991;30:8009–17. [26] Zasloff M. Reconstructing one of nature’s designs. Trends Pharmacol Sci 2000;21:236–8.