N-terminal amphipathic helix as a trigger of hemolytic activity in antimicrobial peptides: A case study in latarcins

FEBS Letters 583 (2009) 2425–2428

journal homepage: www.FEBSLetters.org

N-terminal amphipathic helix as a trigger of hemolytic activity in antimicrobial peptides: A case study in latarcins Anton A. Polyansky *,1, Alexander A. Vassilevski 1, Pavel E. Volynsky, Olga V. Vorontsova, Olga V. Samsonova, Natalya S. Egorova, Nicolay A. Krylov, Alexei V. Feofanov, Alexander S. Arseniev, Eugene V. Grishin, Roman G. Efremov M.M. Shemyakin and Yu.A. Ovchinnikov Institute of Bioorganic Chemistry, Russsian Academy of Sciences, Ul. Miklukho-Maklaya, 16/10, 117997 GSP, Moscow V-437, Russia

a r t i c l e

i n f o

Article history: Received 30 March 2009 Revised 27 May 2009 Accepted 24 June 2009 Available online 27 June 2009 Edited by Peter Brzezinski Keywords: Antimicrobial peptide Cytolytic peptide Hemolytic activity Molecular modeling Molecular hydrophobicity potential Protein engineering

a b s t r a c t In silico structural analyses of sets of a-helical antimicrobial peptides (AMPs) are performed. Differences between hemolytic and non-hemolytic AMPs are revealed in organization of their N-terminal region. A parameter related to hydrophobicity of the N-terminal part is proposed as a measure of the peptide propensity to exhibit hemolytic and other unwanted cytotoxic activities. Based on the information acquired, a rational approach for selective removal of these properties in AMPs is suggested. A proof of concept is gained through engineering specific mutations that resulted in elimination of the hemolytic activity of AMPs (latarcins) while leaving the beneficial antimicrobial effect intact. Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction To protect themselves from invading noxious microorganisms, multicellular animals and plants produce an array of antimicrobial peptides (AMPs) as a vital part of the so-called innate immune system [1]. To date, several hundred AMPs have been described and deposited into specialized databases [2]. A simplified classification envisages three major structural groups of AMPs: (1) linear, with a strong propensity for a-helix formation; (2) disulfide-containing; (3) linear, non-a-helical, usually proline-, tryptophane-, histidineor glycine-rich. Amphipilic a-helical (class 1) peptides represent one of the major subgroups of AMPs and are rather convenient in terms of structure–function investigations and production both via chemical synthesis or gene expression. Most AMPs target bacterial cell membranes and some have attracted considerable interest as a possible new generation of anti-

Abbreviations: AMPs, antimicrobial peptides; EC50, effective peptide concentration inducing a 50% cell death; MHP, molecular hydrophobicity potential; MIC, minimal inhibitory concentration * Corresponding author. Fax: +7 495 336 2000. E-mail address: [email protected] (A.A. Polyansky). 1 These two authors contributed equally to the work.

biotics [1,3]. One of the major hindrances to the clinical use of many effective AMPs, however, is their ability to damage mammalian cell membranes that leads to high hemolytic and cytotoxic activity dangerous to the host organism. For some peptides, certain amino acid substitutions strongly decrease the unwanted hemolytic and cytotoxic activity while leaving the beneficial antimicrobial effect intact [4,5]. Several best studied examples include the cyclic peptide antibiotic gramicidin S [6], disulfide-containing bhairpin (class 2) protegrins [7] and proline/tryptophane-rich (class 3) indolicidin [8]. Structure–function investigations of these peptides suggested specific amino acid replacements that ensured selectivity of their antimicrobial activity. In case of a-helical (class 1) AMPs, impressive efforts have been made to study the effects of different biophysical parameters including the net charge, chain length, helicity, hydrophobicity and hydrophobic moment on activity [3,9–12]. Combinatorial chemistry approaches suggested that peptide functional properties may not be related to specific sequences or spatial structures [13,14], while careful examination of naturally occurring AMPs identified some sequence patterns associated with antimicrobial activity [15]. In this paper, we propose a novel approach to rational design of AMPs devoid of unwanted hemolytic and cytotoxic activity and thus enhancement of their ability to discriminate between bacterial and eukaryotic cell membranes. The proposed procedure is

0014-5793/$36.00 Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2009.06.044

2426

A.A. Polyansky et al. / FEBS Letters 583 (2009) 2425–2428

based on structural differences in hydrophobic/hydrophilic organization noted between N-terminal regions in naturally occurring hemolytic and non-hemolytic peptides and is believed to work well for most of a-helical AMPs. We use the molecular hydrophobicity potential (MHP) approach [16] for bulk analysis of a-helical AMP sets derived from specialized databases as well as for prediction of single point mutations to eliminate the hemolytic activity of a selected peptide. We test our assumptions on two recently discovered highly active a-helical AMPs latarcin 2a (Ltc2a) and latarcin 5 (Ltc5) from Lachesana tarabaevi spider venom [17]. For Ltc2a, the three-dimensional structure in detergent micelles is known and represents the common helix–hinge–helix structural motif [18]. As for Ltc5, its consideration as a mostly helical peptide is corroborated by secondary structure predictions and circular dichroism spectra measurements [17].

2. Materials and methods 2.1. Analysis of hydrophobic/hydrophilic properties of AMP sequences Hydrophobic properties of a-helices were calculated and visualized using the molecular hydrophobicity potential (MHP) approach as described elsewhere [16]. Sequences of a-helical AMPs with low degree of similarity in their N-terminal regions were selected for analysis of their hydrophobic/hydrophilic properties from the APD database [2] and divided into two groups: prominent hemolytic peptides (HEM+, 14 sequences); non-hemolytic peptides (HEM, 46 sequences). Another AMP database of a-helical peptides [19] was filtered and taken as a reference set (hAMP, 107 sequences). For each sequence, the following procedure was applied: (1) building of the peptide 3D-structure in an ideal helix conformation; (2) calculation of MHP in each point of the peptide surface and further interpolation of these data on the plane (2D-MHP (a,Z) map, where a and Z are rotation angle around the helix axis and rise along it, respectively); (3) estimation of average values of hydrophobic (MHP > 0.09) areas (Aphob) and MHP (hmhpi) in the N-terminal region of the 2D-MHP map (Z < 20 Å); and (4) calculation of scoring value F according to the following equation:

F ¼ ðhmhpi  Aphob Þ1=2 The F-values were related to specific AMP sequences and their distribution along the AMPs set was schematically presented in histograms. A similar procedure was carried out for Ltc2a, Ltc5, their several single-point mutants and peptides with deletions of the first three N-terminal residues. 2D-MHP maps for AMP sequences were produced using the PLATINUM program developed in our laboratory (http://model.nmr.ru/platinum/).

incubation at 37 °C for 24 h, growth inhibition was determined by measuring absorbance at 595 nm. 2.4. Cytotoxicity assays The hemolytic activity as well as toxicity of the peptides for human leukocytes and erythroleukemia K562 cells was determined as described in [20]. Briefly, hemolytic activity was assayed using fresh capillary human blood. Red blood cells (107 cells/ml) were incubated with peptides at various concentrations for 3 h at 37 °C. Hemoglobin release was monitored by measuring the absorbance of the supernatant at 414 nm. Leukocyte-enriched fraction was obtained as a buffy coat during blood sedimentation. The K562 cells or leukocytes (106 cells/ml) were incubated with serially diluted peptides for 3 h (5% CO2, 37 °C) and stained with Hoechst 33342 and propidium iodide. The stained cells were examined with an Axiovert 200 M fluorescence microscope (Carl Zeiss) and classified as dead (Hoechst 33342 (+), propidium iodide (+)) or viable (Hoechst 33342 (+), propidium iodide ()). Cytotoxic activity of the peptides was characterized by the effective concentration inducing a 50% cell death (EC50). 3. Results and discussion 3.1. Analysis of N-terminal hydrophobic properties along AMP sets Normalized histograms of F-values for HEM+, HEM, and hAMP sets (see Section 2) are given in Fig. 1. It is seen that in contrast to non-hemolytic (HEM) and merely a-helical AMPs, the distribution for hemolytic (HEM+) peptides is narrow and its maximum is shifted toward the region of large F-values. This indicates that hemolytic activity of a-helical AMPs might be correlated with hydrophobic properties of their N-terminal helices (approximately 13 residues). Namely, a peptide with a prominent hydrophobic Nterminus (large Aphob and hmhpi values for this region), most probably, would possess hemolytic activity. Taking into account that majority of membrane-active peptides prefer interacting with membranes primarily by their N-terminal regions [21–23] and that mammalian cell membranes are less polar than bacterial ones [24,25], this conclusion does not appear too much surprising. Thus, slight modifications of hydrophobicity of a-helical AMPs in their N-terminal parts may produce profound effects on their hemolytic activity. To test this assumption, we selected two highly active and hemolytic AMPs, Ltc2a and Ltc5, and predicted several single-point

2.2. Peptide synthesis Peptides were synthesized using Fmoc/t-butyl chemistry as described in [17]. Ninety-nine percent purity of synthetic peptides was achieved, verified by high-performance liquid chromatography and matrix-assisted laser desorption/ionization mass spectrometry. 2.3. Antimicrobial assays Determination of minimal inhibitory concentration (MIC) values for the peptides was performed using a twofold microtiter broth dilution assay as described in [17]. Briefly, bacterial (Bacillus subtilis VKM B-501, Escherichia coli C600, Staphylococcus aureus 209-P) mid-log phase cultures were diluted in low-salt Luria– Bertani broth to 105 colony-forming units/ml; 90 ll were mixed with 10 ll of peptide serial dilutions in 96-well sterile plates. After

Fig. 1. Hydrophobic properties of N-terminal region of AMPs. Normalized distribution of AMP sequences over their F-values. Data for different sets are colored according to the scale given in the inset. Reference non-hemolytic (magainin 2) and hemolytic (melittin) AMPs along with Ltc2a and its several mutants are shown with respect to their F-values.

2427

A.A. Polyansky et al. / FEBS Letters 583 (2009) 2425–2428

Fig. 2. Hydrophobic organization of the N-terminal region (20 angstroms in length) of peptide helix for Ltc2a and its single-point mutants. The molecular hydrophobicity potential (MHP) on the surface of the helices is given in octanol–water partition coefficient log P units. The value of the rotation angle about the axis helix and the rise along it are plotted on the x and y axes, respectively. Only the hydrophobic areas with MHP > 0.09 are shown. Contour intervals are 0.015. The positions of residue Ca-atoms are indicated by letters and numbers. Regions of mutations are schematically shown with circles.

3.2. Mutants of Ltc2a andLtc5 with corrected biological activity

Table 1 Amino acid sequences of latarcins 2a and 5 mutants. Peptide

Amino acid sequence

F-value

Ltc2a (native) Ltc2a_I7Q Ltc2a_F10K Ltc2a_G11L Ltc2a_N-trim Ltc5 (native) Ltc5_M6R Ltc5_N-trim

1

GLFGKLIKKFGRKAISYAVKKARGKH26-OH GLFGKLQKKFGRKAISYAVKKARGKH-OH GLFGKLIKKKGRKAISYAVKKARGKH-OH GLFGKLIKKFLRKAISYAVKKARGKH-OH 1 GKLIKKFGRKAISYAVKKARGKH23-OH 1 GFFGKMKEYFKKFGASFKRRFANLKKRL28-NH2 GFFGKRKEYFKKFGASFKRRFANLKKRL-NH2 1 GKMKEYFKKFGASFKRRFANLKKRL25-NH2

343.7 278.5 261.9 405.1 244.5 327.3 264.6 241.5

mutations along with deletion of several residues in their N-terminal region with further examination of mutant activity profiles in a series of biological tests.

According to the 2D-MHP map, Ltc2a possesses a prominent hydrophobic patch in its N-terminal part (residues 1–10, Fig. 2). Introduction of polar instead of hydrophobic residues in this region allows modification of size and shape of the patch and thus would presumably affect the peptide’s activity. We designed two mutants (Ltc2a_I7Q, Ltc2a_F10K) with proposed lack of hemolytic activity and also a mutant peptide (Ltc2a_G11L) with an opposite amino acid replacement, which should be more active on erythrocytes than the wild-type AMP (Table 1 and Fig. 2). The F-values for Ltc2a and its mutants as well as for the reference hemolytic (melittin) and non-hemolytic (magainin 2) peptides are depicted in Fig. 1. It is seen that the F-value for Ltc2a coincides with the maximum of the hemolytic distribution (gray bars, Fig. 1), while ‘‘non-hemolytic” and ‘‘hemolytic” mutants of Ltc2a locate in the corresponding

Table 2 Biological activity of latarcins 2a and 5 mutants. Peptide

Target cells E. coli

B. subtilis

S. aureus

MIC, lM Ltc2a (native) Ltc2a_I7Q Ltc2a_F10K Ltc2a_G11L Ltc2a_N-trim Ltc5 (native) Ltc5_M6R Ltc5_N-trim a

n.t., not tested.

3 6 6 3 6 0.7 1.5 1.5

Erythrocytes

Leukocytes

K562 cells

18 >36 >36 3 n.t.a n.t. n.t. n.t.

3 18 18 0.7 >36 12 >36 >36

EC50, lM 0.7 1.5 0.7 0.7 0.7 0.7 0.7 0.7

1.5 36 18 1.5 36 0.7 6 3

6 >36 >36 3 >36 12 >36 >36

2428

A.A. Polyansky et al. / FEBS Letters 583 (2009) 2425–2428

regions of F. Design of Ltc5 ‘‘non-hemolytic” mutants was performed in a similar manner. Introducing a polar residue in the N-terminal region of Ltc5 (Ltc5_M6R) decreases the F-value compared to the wild-type peptide (Table 1). To further test the role of the N-terminal hydrophobic residues, we designed two peptides with deletion of the first three N-terminal residues (Ltc2a_N-trim, Ltc5_N-trim). In both cases, this removal leads to a significant decrease of the F-values (Table 1). The designed mutants were synthesized and their biological activity was compared to that of the wild-type peptides (Table 2). In Ltc2a_I7Q, Ltc2a_F10K and Ltc5_M6R single-point mutants, we have seen minor if any decrease in activity against B. subtilis and E. coli, while the unwanted hemolytic activity and, furthermore, the cytotoxic activity to leukocytes and, to a lesser extent, erythroleukemia K562 cells, practically disappeared. The effect observed on mammalian cells correlated with a decrease in activity towards S. aureus known for its ability to modify its cell membrane into one of a zwitterionic type that resembles eukaryotic membranes [26]. Truncation of three N-terminal residues (Ltc2a_Ntrim, Ltc5_N-trim) produced similar effects on the peptides’ activity profile (Table 2). This result further emphasizes a significant impact of the hydrophobic N-termini of the peptides on their cytotoxic activity. As predicted, the Ltc2a_G11L mutant exhibited higher hemolytic and cytotoxic activities, while its antimicrobial properties were unaffected. 3.3. Conclusion To summarize, several issues should be outlined. In this study, we have developed an easy-to-follow approach to fine-tuning AMPs’ selectivity. We have uncovered differences in organization of the N-terminal region of a-helical AMPs: those peptides exhibiting high hemolytic activity have higher F-values that relate to the distribution of hydrophobic properties along the molecule (Fig. 1). We therefore believe that this region serves a trigger of unwanted hemolytic and other cytotoxic activities of AMPs. Proof of the proposed concept was gained through engineering novel peptides virtually devoid of hemolytic activity out of highly hemolytic prototypes. The resulting changes of the peptide’s F-value strictly correlated with its biological activity (Tables 1 and 2). The reported amino acid substitutions for Ltc2a and Ltc5 follow the more or less general trend of introducing polar residues into the hydrophobic side of AMPs’ helices to reduce their hemolytic activity, taken up after the extensive structure–function studies referenced, but our approach represents a new valuable tool of quantification. We believe that it may be applied to other a-helical AMPs to adjust their properties. Acknowledgements This work was supported by the Russian Foundation for Basic Research (Grants 07-0401166, 08-04-00454, and 07-04-01514), by the Russian Federation Federal Agency for Science and Innovations (Grants SS-4728.2006.4, MK-69.2008.4), and by the RAS Programmes (MCB and ‘‘Basic fundamental research of nanotechnologies and nanomaterials”). References [1] Lehrer, R.E. and Hancock, R. (1998) Cationic peptides: a new source of antibiotics. Trends Biotechnol. 16, 82–88. [2] Wang, Z. and Wang, G. (2004) APD: the Antimicrobial Peptide Database. Nucleic Acids Res. 32, D590-D592; .

[3] Yount, M.R. and Yeaman, N.Y. (2003) Mechanisms of antimicrobial peptide action and resistance. Pharmacol. Rev. 55, 27–55. [4] Asthana, N., Ghosh, S.P. and Yadav, J.K. (2004) Dissection of antibacterial and toxic activity of melittin: a leucine zipper motif plays a crucial role in determining its hemolytic activity but not antibacterial activity. J. Biol. Chem. 279, 55042–55050. [5] Chen, Y., Mant, C.T., Farmer, S.W., Hancock, R.E., Hodges, M.L. and Vasil, R.S. (2005) Rational design of alpha-helical antimicrobial peptides with enhanced activities and specificity/therapeutic index. J. Biol. Chem. 280, 12316–12329. [6] Prenner, E.J., Kiricsi, M., Jelokhani-Niaraki, M., Lewis, R.N., Hodges, R.S. and McElhaney, R.N. (2005) Structure–activity relationships of diastereomeric lysine ring size analogs of the antimicrobial peptide gramicidin S: mechanism of action and discrimination between bacterial and animal cell membranes. J. Biol. Chem. 280, 2002–2011. [7] Langham, A.A., Khandelia, H., Schuster, B., Waring, A.J., Lehrer, R.I. and Kaznessis, Y.N. (2008) Correlation between simulated physicochemical properties and hemolycity of protegrin-like antimicrobial peptides: predicting experimental toxicity. Peptides 29, 1085–1093. [8] Kim, S.M., Kim, J.M., Joshi, B.P., Cho, H. and Lee, K.H. (2009) Indolicidin-derived antimicrobial peptide analogs with greater bacterial selectivity and requirements for antibacterial and hemolytic activities. Biochim. Biophys. Acta 1794, 185–192. [9] Shai, Y. and Oren, Z. (2001) From ‘‘carpet” mechanism to de-novo designed diastereomeric cell-selective antimicrobial peptides. Peptides 22, 1629–1641. [10] Dathe, M., Meyer, J., Beyermann, M., Maul, B., Bienert, C. and Hoischen, M. (2002) General aspects of peptide selectivity towards lipid bilayers and cell membranes studied by variation of the structural parameters of amphipathic helical model peptides. Biochim. Biophys. Acta 1558, 171–186. [11] Chen, Y., Guarnieri, M.T., Vasil, A.I., Vasil, M.L., Mant, C.T. and Hodges, R.S. (2008) Role of peptide hydrophobicity in the mechanism of action of alphahelical antimicrobial peptides. Antimicrob. Agents Chemother. 51, 1398–1406. [12] Jiang, Z., Vasil, A.I., Hale, J.D., Hancock, R.E., Vasil, M.L. and Hodges, R.S. (2008) Effects of net charge and the number of positively charged residues on the biological activity of amphipathic alpha-helical cationic antimicrobial peptides. Biopolymers 90, 369–383. [13] Hilpert, K., Elliott, M.R., Volkmer-Engert, R., Henklein, P., Donini, O., Zhou, Q., Winkler, D.F. and Hancock, R.E. (2006) Sequence requirements and an optimization strategy for short antimicrobial peptides. Chem. Biol. 13, 1101–1107. [14] Rathinakumar, R. and Wimley, W.C. (2008) Biomolecular engineering by combinatorial design and high-throughput screening: small, soluble peptides that permeabilize membranes. J. Am. Chem. Soc. 130, 9849–9858. [15] Zelezetsky, I. and Tossi, A. (2006) Alpha-helical antimicrobial peptides – using a sequence template to guide structure–activity relationship studies. Biochim. Biophys. Acta 1758, 1436–1449. [16] Efremov, R.G., Chugunov, A.O., Pyrkov, T.V., Priestle, J.P., Jacoby, A.S. and Arseniev, E. (2007) Molecular lipophilicity in protein modeling and drug design. Curr. Med. Chem. 14, 393–415. [17] Kozlov, S.A., Vassilevski, A.A., Feofanov, A.V., Surovoy, A.Y., Grishin, D.V. and Karpunin, E.V. (2006) Latarcins, antimicrobial and cytolytic peptides from the venom of the spider Lachesana tarabaevi (Zodariidae) that exemplify biomolecular diversity. J. Biol. Chem. 281, 20983–20992. [18] Dubovskii, P.V., Volynsky, P.E., Polyansky, A.A., Chupin, V.V., Efremov, R.G. and Arseniev, A.S. (2006) Spatial structure and activity mechanism of a novel spider antimicrobial peptide. Biochemistry 45, 10759–10767. [19] Dennison, S.R., Wallace, J., Phoenix, F. and Harris, D.A. (2005) Amphiphilic alpha-helical antimicrobial peptides and their structure/function relationships. Protein Pept. Lett. 12, 31–39. [20] Vassilevski, A.A., Kozlov, S.A., Samsonova, O.V., Egorova, N.S., Karpunin, D.V., Pluzhnikov, K.A., Grishin, A.V. and Feofanov, E.V. (2008) Cyto-insectotoxins, a novel class of cytolytic and insecticidal peptides from spider venom. Biochem. J. 411, 687–696. [21] Volynsky, P.E., Polyansky, A.A., Simakov, N.A., Efremov, A.S. and Arseniev, R.G. (2005) Effect of lipid composition on the ‘‘membrane response” induced by a fusion peptide. Biochemistry 44, 14626–14637. [22] Dennison, S.R., Phoenix, F. and Harris, D.A. (2005) Are oblique orientated alpha-helices used by antimicrobial peptides for membrane invasion? Protein Pept. Lett. 12, 27–29. [23] Lins, L., Decaffmeyer, M., Brasseur, A. and Thomas, R. (2008) Relationships between the orientation and the structural properties of peptides and their membrane interactions. Biochim. Biophys. Acta 1778, 1537–1544. [24] Glukhov, E., Stark, M., Deber, L.L. and Burrows, C.M. (2005) Basis for selectivity of cationic antimicrobial peptides for bacterial versus mammalian membranes. J. Biol. Chem. 280, 33960–33967. [25] Epand, R.M. and Epand, R.F. (2009) Lipid domains in bacterial membranes and the action of antimicrobial agents. Biochim. Biophys. Acta 1788, 289–294. [26] Sahl, A. and Peschel, H.G. (2006) The co-evolution of host cationic antimicrobial peptides and microbial resistance. Nat. Rev. Microbiol. 4, 529– 536.