Báo cáo khoa học: Salt-resistant homodimeric bactenecin, a cathelicidin-derived antimicrobial peptide potx

Salt-resistant homodimeric bactenecin, a cathelicidin-derived antimicrobial peptide Ju Y. Lee 1 , Sung-Tae Yang 1,3 , Seung K. Lee 1 , Hyun H. Jung 1 , Song Y. Shin 2 , Kyung-Soo Hahm 2 and Jae I. Kim 1 1 Department of Life Science, BioImaging Research Center, Gwangju Institute of Science and Technology, Korea 2 Department of Bio-Materials, Graduate School and Research Center for Proteineous Materials, Chosun University, Gwangju, Korea 3 Section on Membrane Biology, Laboratory of Cellular and Molecular Biophysics, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD, USA Over the course of evolution, endogenous antimicro- bial peptides have assumed the role of providing a first line of defense against pathogenic infections in both mammalian and nonmammalian species [1]. Among these host defense peptides, the cathelicidins are char- acterized by conserved cathelin-like domains (the proregion) and highly variable C-terminal antimicro- bial domains [2] that enable them to be classified into three structural classes: amphipathic a-helical peptides, b-hairpin peptides stabilized by disulfide bridges, and linear Trp-rich, Pro-rich peptides [3,4]. Bactenecin, a cathelicidin purified from the granules of bovine neutrophils, is a b-hairpin monomer with one intra- molecular disulfide bond, and has been shown to have Keywords antimicrobial peptides; bactenecin; dimerization; peptide–membrane interaction; salt resistance Correspondence J. I. Kim, Department of Life Science, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea Fax: +82 62 970 2484 Tel: +82 62 970 2494 E-mail: jikim@gist.ac.kr (Received 19 July 2007, revised 28 April 2008, accepted 4 June 2008) doi:10.1111/j.1742-4658.2008.06536.x The cathelicidin antimicrobial peptide bactenecin is a b-hairpin molecule with a single disulfide bond and broad antimicrobial activity. The proform of bactenecin exists as a dimer, however, and it has been proposed that bactenecin is released as a dimer in vivo, although there has been little study of the dimeric form of bactenecin. To investigate the effect of bacten- ecin dimerization on its biological activity, we characterized the dimer’s effect on phospholipid membranes, the kinetics of its bactericidal activity, and its salt sensitivity. We initially synthesized two bactenecin dimers (anti- parallel and parallel) and two monomers (b-hairpin and linear). Under oxi- dative folding conditions, reduced linear bactenecin preferentially folded into a dimer forming a ladder-like structure via intermolecular disulfide bonding. As compared to the monomer, the dimer had a greater ability to induce lysis of lipid bilayers and was more rapidly bactericidal. Interest- ingly, the dimer retained antimicrobial activity at physiological salt concen- trations (150 mm NaCl), although the monomer was inactivated. This salt resistance was also seen with bactenecin dimer containing one intermole- cular disulfide bond, and the bactenecin dimer appears to undergo multi- meric oligomerization at high salt concentrations. Overall, dimeric bactenecin shows potent and rapid antimicrobial activity, and resists salt- induced inactivation under physiological conditions through condensation and oligomerization. These characteristics shed light on the features that a peptide would need to serve as an effective therapeutic agent. Abbreviations ABD, antiparallel dimer bactenecin; Acm, acetamidomethyl; CDB, C-terminal dimeric bactenecin; CFU, colony-forming unit; hRBC, human red blood cell; KCTC, Korean Collection for Type Cultures; MIC, minimal inhibitory concentration; MTB, monomeric turn bactenecin; NDB, N-terminal dimeric bactenecin; PDB, parallel dimer bactenecin; POPC, 1-palmitoyl-2-oleoyl-phosphatidylcholine; POPG, 1-palmitoyl-2-oleoyl- phosphatidylglycerol; SLB, Ser-substituted linear bactenecin. FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS 3911 antibacterial activity against both Gram-negative and certain Gram-positive bacteria [5]. In addition, two lin- ear variants of bactenecin, Bac2S and Bac2A, show similar activities against Gram-negative bacteria and stronger activities against Gram-positive bacteria [6], and Bac2A also acts as a potent chemoattractant, inducing chemotaxis in undifferentiated THP-1 cells [7]. However, although bactenecin has largely been studied as a monomeric molecule, its proform report- edly exists as a dimer formed by intermolecular disul- fide bridges in the C-terminal antimicrobial domain. Moreover, it is known that the synthetic cyclic peptide is mainly active against Gram-negative bacteria [6,8], whereas the isolated native peptide showed activity against both Escherichia coli and Staphylococcus aureus [5]. This suggests that it may be necessary to recon- sider the structure of the mature native bactenecin in vivo [9]. Although b-hairpin bactenecin and its analogs have been the subjects of numerous studies, little is known about the antimicrobial activity of the dimeric form, or the way in which it interacts with the bacterial membrane. That said, earlier studies suggest that dimerization of antimicrobial peptides leads to the appearance of a more diverse spectrum of antimicro- bial activity than is exhibited by monomers. For instance, Tencza et al. reported that dimeric LLP1, which is a Cys-containing peptide derived from a lenti- virus envelope protein that spontaneously forms disulfide-linked dimers, possesses much greater antimi- crobial activity against S. aureus than monomeric LLP1 [10]. In addition, disulfide-dimerized magainin 2 [(mag-N22C) 2 ] induces membrane permeabilization at lower concentrations than the monomeric form [11]. In the case of the channel-forming peptide alamethicin, channels formed by covalent dimers displayed lifetimes at a particular conductance that were up to 170-fold longer than those of monomers [12]. Consistent with that finding, in many cases dimerization was closely connected to enhanced antimicrobial activity mediated by the formation of pores or channels in the lipid membrane [13,14]. For effective use in clinical pharmacotherapy, anti- microbial peptides need to remain active in the pres- ence of physiological levels of salt (120–150 mm NaCl), and structural constraints such as dimerization or Cys-knot formation are also related to the salt sen- sitivity of antimicrobial peptides. For instance, defen- sins, a group of b-form antimicrobial peptides, are generally degraded under high-salt conditions [15], but oxidized b-defensin (Defr-1), which contains five Cys residues that associate to produce dimers through for- mation of various intramolecular and intermolecular disulfide bridges, exhibits potent and broad-spectrum antimicrobial activity that is not suppressed at high salt concentrations [16]. In addition, the study of protegrin-1 and rhesus theta defensin-1, which have b-strand and cyclic structures, respectively, has shown that structural rigidity resulting from Cys-stabilization enables the peptides to retain activity against most bacteria in high-salt environments [17]. It was previously reported that bactenecin is too small to disrupt the bacterial membrane unless a multi- mer is involved in forming pores or channels [8], and that the native peptide may occur in both monomeric and dimeric forms [9,18]. To test that idea, in the pres- ent study we chemically synthesized two dimers that adopt parallel and antiparallel conformations and two monomers that adopt b-hairpin and linear confor- mations, and investigated their biological activities. Results and Discussion Peptide folding and its characterization To investigate the effect of dimerization on the antimi- crobial activity of bactenecin, we designed four bacten- ecin derivatives with differing chemical ⁄ physical properties reflecting the interactions among their Cys residues (Fig. 1). Under most oxidative folding condi- tions, reduced linear bactenecin folded into a specific form (yield, 70–80%) that trypsin digestion experi- ments revealed to be an antiparallel dimer [antiparallel dimer bactenecin (ADB)] (supplementary Figs S1 and S2). Because the majority of reduced linear bactenecin spontaneously dimerizes, even at very low oxidative folding concentrations (e.g. 10 lm), we attempted to synthesize monomeric turn bactenecin (MTB) by utiliz- ing an iodine oxidation strategy often used for oxida- tive cyclization of Cys-containing peptides having a free Cys residue and to remove protective S-acetami- domethyl (Acm) groups, although in this case there was no S-Acm group [19]. Under these conditions, dimerization was completely blocked, and MTB was obtained with a yield of about 90%. Interestingly, we failed to produce any parallel dimer bactenecin (PDB) when the oxidative folding condition was applied to unprotected ADB or MTB peptide, suggesting that ADB is thermodynamically more favorable than PDB in an air oxidative folding pathway. As ADB and PDB differ only in the orientations of their two strands with respect to one another, we suggest that mainly unfavorable terminal charge repulsion inhibits PDB formation. By adding one protective S-Acm to reduced linear bactenecin (Fig. 1), we were able to utilize an iodine oxidation strategy to synthesize PDB Salt-resistant homodimeric bactenecin J. Y. Lee et al. 3912 FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS and successfully harvest the dimer after a sequential two-step reaction leading to disulfide formation (yield, 80%). The position of the Cys residue carrying the S-Acm group was alternated because the amino acid compositions near the two Cys residues were similar to one another. Finally, we synthesized Ser-substituted linear bactenecin (SLB) to investigate the structural and ⁄ or functional role of the disulfide bond. Conformational studies We used CD spectroscopy to estimate the secondary structure of the bactenecin derivatives in buffer and in a membrane-mimicking environment achieved with the addition of SDS (Fig. 2). Consistent with previously reported CD spectra [8], those for MTB showed a typi- cal type I b-turn structure with a negative band in the vicinity of 205 nm in both environments [20]. For the two dimers, ADB and PDB, a spectrum exhibiting a negative band at 210 nm was observed in buffer, whereas an ordered b-strand structure with a maximum near 200 nm and a minimum at 220 nm was obtained in the presence of SDS micelles. This was well fitted to typical b-strand globular proteins, which show a strong positive band near 200 nm and a nega- tive band below 220 nm [21,22]. A more ordered struc- ture indicated by the red-shift from 210 to 220 nm, as well as the presence of a positive band at 200 nm, may be caused not only by the interaction of b-strands within a given dimer, but also by the interaction of b-strands between dimers. Interestingly, SLB showed a disordered structure in buffer but, upon interaction with SDS micelles, the CD spectrum changed to one similar to those of the dimers. The spectral behavior observed for SLB suggests that linear bactenecin has a strong propensity to form a b-structure in a membrane environment, and may be indicative of the importance of the dimeric structure for specific interactions with the bacterial membrane. Antimicrobial and hemolytic activities The peptides’ antimicrobial activities against selected Gram-positive and Gram-negative bacteria, as well as their hemolytic activities, are summarized in Table 1. As previously reported, MTB was more potent against Gram-negative bacteria [minimal inhibitory concen- tration (MIC) = 2–4 lm] than against Gram-positive bacteria (MIC = 4–8 lm), and was without hemolytic activity. ADB and PDB displayed activity similar to that of MTB against Gram-negative bacteria, with some hemolytic activity (10–20% hemolysis at 100 lm), but exhibited about four times greater potency (MIC = 1–2 lm) against Gram-positive bacteria. These results are consistent with those obtained with the isolated native peptide, which displayed broad-spectrum antimi- crobial activity against Gram-positive and Gram-nega- tive bacteria, and suggests to us that it is probable that native dimeric forms are also active in vivo. It was previously reported that Bac2A, in which a Cys residue was substituted with Ala, had somewhat better activity against Gram-positive bacteria than MTB [6]. Like Bac2A, SLB also showed slightly better antimicrobial activity against Gram-positive bacteria than MTB, with no hemolytic activity. Taken together with the results of the CD analysis, these findings suggest that in a membrane environment, ADB, PDB and SLB take on a common b-structure that enables better interaction with Gram-positive bacteria. Dye leakage from liposomes It is well known that dimerization can cause a signifi- cant change in a peptide’s interaction with the bacte- MTB SLB H 2 N -RLCRIVVIRVCR- CO 2 H RLCRIVVIRVCR H 2 N-RLCRIVVIRVCR-CO 2 H H 2 N - -CO 2 H Acm SH PDBADB H 2 N -RLSRIVVIRVSR -CO 2 H H 2 N-RLCRIVVIRVCR-CO 2 H H 2 N-RLCRIVVIRVCR-CO 2 H Acm SH Acm H 2 N-RLCRIVVIRVCR-CO 2 H H 2 N-RLCRIVVIRVCR-CO 2 H H 2 N-RLCRIVVIRVCR-CO 2 H Acm HO 2 C -RCVRIVVIRCLR- NH 2 H 2 N -RLCRIVVIRVCR- CO 2 H H 2 N-RLCRIVVIRVCR-CO 2 H Fig. 1. Scheme employed for the synthesis of bactenecin and its derivatives through formation of disulfide bridges. (A) ADB was folded in 2 M acetic acid ⁄ H 2 O ⁄ dimethylsulfoxide (1 : 2 : 1, v ⁄ v ⁄ v) solution for 24 h at room temperature. (B) MTB was oxidized in acetic acid ⁄ H 2 O (4 : 1, v ⁄ v) solution, after which iodine was added (10 equivalents to the number of disulfide bonds). (C) PDB was prepared in two steps: air oxidation in distilled water at 47 °C was carried out for 5 days, after which the partially oxidized peptides were dissolved in acetic acid ⁄ H 2 O (4 : 1, v ⁄ v) solution, and iodine was added (10 equivalents to the number of disulfide bonds) for 2h. J. Y. Lee et al. Salt-resistant homodimeric bactenecin FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS 3913 rial membrane, whether or not it enhances the antimi- crobial activity of the peptide [10–14]. To assess the effect of dimerization on peptide-induced membrane disruption leading to microbial cell death, we examined the capacity of peptides to release calcein from liposomes composed of 1-palmitoyl-2-oleoyl-phosphat- idylglycerol (POPG) ⁄ 1-palmitoyl-2-oleoyl-phosphati- dylcholine (POPC) (1 : 1), which served as a model of the bacterial membrane (Fig. 3). At a molar pep- tide ⁄ liposome ratio of 1 : 10, PDB and ADB induced leakage in about 90% and 70% of liposomes, respec- tively. By contrast, both MTB and SLB displayed only weak membrane lytic activity, with about 20% of lipo- somes showing leakage. In terms of the structure– activity relationships, it is noteworthy that ADB, PDB and SLB all assume a common b-structure in a mem- brane environment, despite the significant differences in their membrane lytic activities. In that regard, a two-step mechanism for membrane disruption leading to leakage has been suggested [23]. The peptide first 10 60 MTB SLB –10 0 0 20 40 MTB SLB –30 –20 –40 –20 0 m 2 · dmol –1 ) –40 –60 10 15 PDB 15 20 ADB 0 3 (deg·c m –10 –5 0 5 0 5 10 [ ] X1 –25 –20 –15 –20 –15 –10 –5 190 200 210 220 230 240 250 190 200 210 220 230 240 250 Wavelength (nm) Fig. 2. CD spectra for MTB, SLB, ADB and PDB. Spectra were recorded at 25 °Cin 10 m M sodium phosphate buffer (pH 7.4) (d)orin30m M SDS micelles ( ). Each peptide was used at a concentration of 25 l M. Table 1. MIC (lM) values and hemolytic activities of the peptides. Results indicate the ranges of three independent experiments, each performed in triplicate. The hemolytic activity was determined using 100 l M peptide, and the results represent the means of duplicate measurements from three independent assays. MIC (l M) MTB SLB ADB PDB Bacterial strain S. aureus 4–8 2–4 1–2 1–2 S. epidermidis 2–4 2 1–2 2 B. subtilis 4 2 1–2 2 E. coli 1–2 2–4 1–2 2–4 P. aeruginosa 2–4 4–8 2–4 4 Sa. typhimurium 4–8 4–8 2–4 4 % Hemolysis 0 0 10 20 100 60 80 20 40 0.001 0.01 0.1 1 10 Calcein release (%) 0 [Peptide]/[Lipid] Fig. 3. Calcein release from liposomes was measured as a function of the molar peptide ⁄ lipid ratio. Peptide concentrations were 5 l M for POPC ⁄ POPG (1 : 1) liposomes. Fluorescence from liposomes lysed with Triton X-100 was used as an indicator of 100% leakage. s, MTB; d, SLB; ,, ADB; ., PDB. Results represent the means of three independent experiments. Salt-resistant homodimeric bactenecin J. Y. Lee et al. 3914 FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS binds to the membrane (the membrane affinity of the peptide), after which it elicits membrane disruption (the membrane-perturbing activity). It is thus likely that even though all three peptides exhibit a similar structural transition upon binding to membrane sur- faces, only the dimers show enhanced membrane lytic activity, perhaps due to induction of oligomerization of the dimeric peptides by the hydrophobic membrane environment. Kinetics of the bactericidal activity To further study the antibacterial activity of ADB and PDB, the kinetics of their bactericidal activity against both Gram-positive (S. aureus; Fig. 4A) and Gram- negative (E. coli; Fig. 4B) bacteria were investigated, with magainin 2 serving as a control. The time needed for PDB and ADB to induce 100% cell death was as little as 5 min for Gram-positive bacteria, and both peptides showed the same kinetics. About 30 min or more were needed to kill 100% of Gram-negative bac- teria, with PDB acting more rapidly than ADB. The kinetics of MTB’s bactericidal activity were similar to those of ADB for Gram-negative bacteria, but were very slow for Gram-positive bacteria, with about 20% of cells remaining viable even after exposure for 60 min. SLB acted almost as rapidly as ADB or PDB against Gram-positive bacteria, but acted more slowly than the other three peptides against Gram-negative bacteria. Although, overall, PDB and ADB showed only slightly greater antimicrobial activity than MTB and SLB, we suggest that the capacity of the dimers to kill bacteria quickly enough to prevent replication gives them a greater ability to control bacterial expan- sion, thereby reducing the likelihood that resistance will develop [24]. In other words, the rapidity with which bacteria are killed may be an important factor when evaluating the activity of antimicrobial peptides in vivo and when assessing their potential for clinical use [25]. Effect of salt Studies of cationic antimicrobial peptides have shown that the salt concentration can affect their activity, even at less than physiological levels [26]. To determine whether dimerization affects salt sensitivity, S. aureus and E. coli were exposed to 8 lm peptide in the pres- ence or absence of 150 mm NaCl. In the absence of salt, all four peptides killed 100% of the bacteria. In its presence, the two dimers exhibited generally unal- tered activity against both Gram-positive (S. aureus; Fig. 5A) and Gram-negative (E. coli; Fig. 5B) bacteria. By contrast, the antimicrobial activity of MTB against S. aureus was completely lost in the presence of 150 mm NaCl, and the activity against E. coli was reduced by > 60%. Although in a membrane environ- ment SLB showed a potency and CD pattern that were similar to those of the dimers, in the presence of 150 mm NaCl, it killed only about 75% of S. aureus and was completely inactive against E. coli. It thus appears that Cys-derived dimerization enables the peptides to retain potent bactericidal activity in the presence of physiological levels of salt. Similarly, it was previously reported that the antimi- crobial activity of the guinea pig 11 kDa polypeptide, which is a homodimer joined by intermolecular disul- fide bonds, was unaffected by the presence of NaCl, whereas the activities of the guinea pig 5 kDa peptide and various defensins, which all contain intramolecular disulfide bonds, were inactivated by NaCl [27]. Together, these results strongly suggest that intermo- lecular disulfide connections contribute greatly to retention of a peptide’s antibacterial activity at high salt concentrations. Peptide oligomerization A high ionic strength may reduce the electrostatic interaction between cationic peptides and anionic lipid head groups through counterion screening, thereby 80 100 A B 80 100 40 60 40 60 CFU·mL –1 (%) 0 20 0 20 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Time (min) Fig. 4. Kinetics of the bactericidal activity of the four bactenecin derivatives against S. aureus (A) and E. coli (B). Bacteria treated with the respective peptides (8 l M) were diluted at the indicated times and then pla- ted on LB agar. The CFUs were then counted after 24 h of incubation at 37 °C. s, MTB; d, SLB; ,, ADB; ., PDB; , mag- ainin 2. Results represent the means of two independent experiments. J. Y. Lee et al. Salt-resistant homodimeric bactenecin FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS 3915 reinforcing the membrane surface region [23,28]. In the case of defensin, it was reported that the number of positive charges in the molecule was directly propor- tional to its ability to retain antimicrobial activity at higher salt concentrations [28]. Consistent with that relationship, the bactenecin dimers used in the present study have twice as many positively charged residues as bactenecin monomer. On the other hand, the salt tolerance of dimeric bac- tenecins may reflect, to some degree, the structural rigidity afforded by the two intermolecular disulfide bonds. To test that idea, we substituted the Cys resi- due at position 11 or 3 of bactenecin with a Ser and synthesized two PDB derivatives containing a single disulfide bond: N-terminal dimeric bactenecin (NDB) and C-terminal dimeric bactenecin (CDB) (Fig. 6A). In the presence of SDS micelles, the CD spectra of both NDB and CDB showed b-structure patterns, similar to those of ADB and PDB, but in buffer solution NDB showed a b-structure with reduced molar ellipticity, whereas CDB showed a random-like conformation (Fig. 6B). In addition, both NDB and CDB exhibited unexpectedly lower antimicrobial potency in the absence of salt (MIC = 32 and 16 lm for S. aureus and E. coli, respectively), and just 30–40% of the activ- ity of MTB. In the presence of 150 mm NaCl, how- ever, NDB and CDB almost completely killed the tested Gram-positive and Gram-negative bacteria, and showed a potency similar to that of ADB and PDB, which have two disulfide bonds (Fig. 6C). It is note- worthy that, in the absence of salt, both NDB and CDB exhibited much less antibacterial activity than MTB, SLB, ADB or PDB, and they displayed very different potencies in the presence or absence of salt. Finally, we compared the multimeric state of bactenecin derivatives by carrying out an electrophore- sis experiment on Tricine–acrylamide gel. Four deriva- tives, SLB (linear bactenecin), MTB (bactenecin having one intramolecular disulfide bond), NDB (bactenecin having one intermolecular disulfide bond) and PDB (bactenecin having two intermolecular disulfide bonds) were selected and exposed to an environment con- taining a high concentration of salt (300 mm NaCl). As shown in Fig. 7, the two monomers (SLB and MTB) migrated with apparent molecular masses of  1.5 kDa, whereas the two dimers (NDB and PDB) migrated with apparent molecular masses of  6.5 kDa or more in both the presence and the absence of salt. In addition, both SLB and MTB showed somewhat fainter bands in the presence of salt, but PDB exhibited a strong band whether salt was present or not. This suggests that both SLB and MTB are monomers and that PDB is a dimer in both the presence and the absence of salt. Interestingly, NDB showed a weak band at around  6.5 kDa in the absence of salt, but a strong band at around  14.2 kDa at the presence of salt, implying that NDB undergoes multimeric oligomerization in the presence of 300 mm NaCl. Thus, although the same amount of each peptide was loaded, these peptides exhibited significantly different band densities and mobilities on a Tricine–acrylamide gel, which provides a clue as to why bactenecin dimers retain their potent antibacterial activity at high salt concentrations. In conclusion, our findings are noteworthy in part because they confirm the potential importance of dimeric forms of antimicrobial peptides in vivo, and because the ladder-like structure of homodimeric antimicrobial peptides makes them relatively easy to 100 100 A B 80 80 40 60 40 60 0 20 % Bacteria killed % Bacteria killed 0 20 MTB S LB ADB PDB MTB SLB ADB PDB Fig. 5. Salt sensitivity of the antimicrobial activity of the four bac- tenecin derivatives against S. aureus (A) and E. coli (B). To deter- mine the effect of salt on the antimicrobial activity of the peptides, each peptide (8 l M) was incubated with bacteria for 3 h in the absence (gray bars) or presence (black bars) of 150 m M NaCl, after which 50 lL aliquots of the suspension were plated on LB agar for colony counts. Results represent the means of two independent experiments. Salt-resistant homodimeric bactenecin J. Y. Lee et al. 3916 FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS synthesize. Although the two dimers studied, ADB and PDB, had similar activities, synthesis of PDB was com- plex. By contrast, ADB is easily folded under most folding conditions. Interestingly, bactenecin dimers undergo multimeric oligomerization at high salt concentrations. Further studies on the structural changes in PDB and NDB that occur at the mem- brane are in progress so as to better understand the mechanism by which each dimer interacts with the membrane. Experimental procedures Peptide synthesis, disulfide formation and characterization All peptides were synthesized using the solid-phase peptide synthesis method performed manually with Fmoc chemis- try. The peptides were cleaved from the resin using trifluo- H 2 N-RLCRIVVIRVSR-CO 2 H H 2 N-RLCRIVVIRVSR-CO 2 H H 2 N-RLCRIVVIRVSR-CO 2 H NDB B A C H 2 N-RLSRIVVIRVCR-CO 2 H H 2 N-RLSRIVVIRVCR-CO 2 H H 2 N-RLSRIVVIRVCR-CO 2 H CDB 30 NDB CDB 10 20 m 2 ·dmol –1 ) NDB CDB -10 0 0 3 (deg·cm 30 -20 [ ] X 10 190 200 210 220 230 240 250 190 200 210 220 230 240 25 0 -30 30 10 20 -10 0 -20 -30 Wavelength (nm) lled 80 100 80 100 S. aureus E. coli acteria ki 40 60 40 60 % Ba 0 20 0 20 MTB NDB CDB MTB NDB C DB 0 Fig. 6. Synthesis, secondary structure and salt sensitivity of NDB and CDB, two bactenecin derivatives containing a single disulfide bond. (A) NDB and CDB were completely folded in 2 M acetic acid ⁄ H 2 O ⁄ dimethylsulfoxide (1 : 2 : 1, v ⁄ v ⁄ v) solution for 36 h with gentle stirring at room temperature. (B) CD spectra were recorded at 25 °Cin10m M sodium phos- phate buffer (pH 7.4) (d) and in 30 m M SDS micelles ( ). (C) Each peptide (8 lM) was incubated with bacteria for 3 h in the absence (gray bars) or presence (black bars) of 150 m M NaCl, after which 50 lL aliquots of the suspension were plated on LB agar for colony counts. Results represent the means of two independent experiments. No salt 300 mM salt MK NDB SLB PDB MTB NDB SLB PDB MTB 26.6 17 14.2 6.5 3.5 Fig. 7. Coomassie-stained 15% Tricine gel of bactenecin and its derivatives without salt and with 300 m M NaCl. Fifteen micrograms of each peptide were loaded. Mass markers in kDa are shown on the left. J. Y. Lee et al. Salt-resistant homodimeric bactenecin FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS 3917 roacetic acid containing various scavengers and purified by preparative RP-HPLC (Shimadzu, Tokyo, Japan). The pur- ity of peptides was verified by analytical RP-HPLC, and correct peptide masses were confirmed by MALDI- TOF MS (Shimadzu). Dissolving reduced linear bactenecin to a concentration of 1 mm in buffer solution containing 2 m acetic acid ⁄ H 2 O ⁄ dimethylsulfoxide (1 : 2 : 1) at room temperature for 24 h with gentle stirring effectively yielded ADB. MTB exhibiting a b-hairpin conformation was oxidized in acetic acid ⁄ H 2 O (4 : 1), and this was followed by addition of iodine (10 equivalents to the number of disulfide bonds). A two-step method for disulfide bond formation was used to prepare PDB. Briefly, partially protected peptides were joined using Fmoc solid-phase chemistry on Wang resin. The free thiol groups of the peptides were bonded by air oxidation in distilled water at 47 °C, while the course of the reaction was monitored using HPLC. Peptides linked by single disulfide bonds were obtained after 5 days at a yield of > 90%. The second procedure was initiated by dissolv- ing the peptide in acetic acid ⁄ H 2 O (4 : 1) and adding iodine (10 equivalents to the number of disulfide bonds), after which stirring was continued for an additional 2 h to effect removal of the Acm groups and conversion to PDB with a yield of 80%. ADB and PDB were confirmed by enzymatic digestion with trypsin (supplementary Figs S1 and S2). A linear peptide SLB, in which Cys was substituted with Ser, was also synthesized. Trypsin digestion A trypsin digestion was carried out to distinguish between PDB and ADB (supplementary Fig. S1). Samples of PDB (100 lg) and ADB (100 lg) were dissolved in 0.2 mL of 50 mm Tris ⁄ HCl buffer (pH 8), after which modified tryp- sin (5 lg) was added to a final protease ⁄ protein ratio of 1 : 20 (w ⁄ w), and the mixture was incubated at 37 °C for 6 h. Analytical RP-HPLC analysis of the reaction mixture was then carried out (supplementary Fig. S2), and MALDI-TOF MS was used to analyze the mass of each peptide. CD analysis The CD spectra of the peptides were recorded using a Jasco J-710 CD spectrophotometer (Jasco, Tokyo, Japan) with a 1 mm path-length cell. Wavelengths were measured from 190 to 250 nm (bandwidth, 1 nm; step resolution, 0.1 nm; speed, 50 nmÆmin )1 ; response time, 0.5 s). The col- lected CD spectra for the peptides were averaged over 16 scans in 0.5 mm POPC ⁄ POPG (1 : 1) liposomes and over four scans in 10 mm sodium phosphate buffer (pH 7.4) or 30 mm SDS micelles at 25 °C. The spectra are expressed as molar ellipticity [h] versus wavelength. Antibacterial activity Antimicrobial activities of each peptide against six selected organisms, including three Gram-positive and three Gram- negative bacteria, were determined using broth microdilu- tion assays [29]. Six organisms obtained from the Korean Collection for Type Cultures (KCTC), Korea Research Institute of Bioscience and Biotechnology (Taejon, Korea) were used for the assays. The Gram-negative bacteria were E. coli KCTC 1682, Salmonella typhimurium KCTC 1926, and Pseudomonas aeruginosa KCTC 1637. The three Gram-positive bacteria were Bacillus subtilis KCTC 3068, Staphylococcus epidermidis KCTC 1917, and S. aureus KCTC 1621. Briefly, single colonies of bacteria were inocu- lated into medium (LB broth) and cultured overnight at 37 °C. An aliquot of the culture was then transferred to 10 mL of fresh medium and incubated for an additional 3–5 h at 37 °C until mid-logarithmic phase. A two-fold dilution series of peptides in 1% peptone was prepared, after which serial dilutions (100 lL) were added to 100 lL of cells [2 · 10 5 colony-forming units (CFU)ÆmL )1 ]in 96-well microtiter plates (F96 microtiter plates; Nunc, Odense, Denmark) and incubated at 37 °C for 16 h. The low- est concentration of peptide that completely inhibited growth was defined as the MIC. MIC values were acquired as aver- age or triplicate measurements in three independent assays. Hemolytic activity The hemolytic activities of the peptides were determined using human red blood cells (hRBCs). After washing of fresh hRBCs three times with NaCl ⁄ P i (35 mm phosphate buffer, 150 mm NaCl, pH 7.4), 100 lL of a 4% (v ⁄ v) hRBC suspension in NaCl ⁄ P i was dispensed into sterilized 96-well plates along with 100 lL of peptide solution. The plates were then incubated for 1 h at 37 °C and centrifuged for 5 min at 1000 g. Aliquots (100 lL) of supernatant were transferred to 96-well plates, and hemoglobin release was monitored on the basis of the absorbance at 414 nm using an ELISA plate reader (Molecular Devices, Sunnyvale, CA, USA). Percentage hemolysis was calculated using the following formula: hemolysis (%) = [(A 405 nm sample ) A 405 nm zero lysis) ⁄ (A 405 nm 100% lysis ) A 405 nm zero lysis)] · 100. Zero and 100% hemolysis were determined in NaCl ⁄ P i and 0.1% Triton X-100, respectively. The recorded hemolysis (%) was the average of duplicate measurements in three independent assays. Preparation of liposomes Large unilamellar vesicles (average diameter, 100 nm) con- taining the fluorescent probe calcein were prepared by extrusion [30]. Briefly, phospholipids composed of POPG ⁄ POPC (1 : 1) were dissolved in chloroform and then dried Salt-resistant homodimeric bactenecin J. Y. Lee et al. 3918 FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS overnight under vacuum to make a thin lipid film. The dried film was then hydrated with Tris ⁄ HCl buffer (10 mm Tris, 150 mm NaCl, 1 mm EDTA, pH 7.4) containing 70 mm calcein (pH adjusted to 7.4 with NaOH) and vortex- mixed. The suspensions were subjected to five freeze–thaw cycles and then pressure-extruded through polycarbonate filters (LiposoFast, 0.1 lm pore size, 20 times). Vesicles containing entrapped calcein were separated from free calcein by gel filtration on Sephadex G-50 columns (Phar- macia, Uppsala, Sweden) equilibrated with Tris ⁄ HCl buffer. To prepare the small unilamellar vesicles used for CD spectroscopy, dried lipid film was hydrated with Tris ⁄ HCl buffer and then sonicated in an ice bath for 30 min using a titanium-tipped sonicator. The lipid concentration was 0.5 mm. Calcein leakage studies As mentioned above, the fluorescent probe calcein was encapsulated in large unilamellar vesicles at a self-quench- ing concentration of 70 mm. For leakage experiments, the indicated amounts of peptide were added to 3 mL of buffer containing calcein-loaded liposomes. The fluorescence inten- sity of the calcein released from the liposomes, which was measured with mixing after the addition of a peptide, was monitored at 520 nm (excited at 490 nm) in a Shima- dzu RF-5301 spectrofluorometer. Fluorescence from liposomes lysed with Triton X-100 (20% in Tris buffer) was used as an indicator of 100% leakage. Kinetics of bactericidal activity and salt sensitivity The kinetics of the peptides’ bactericidal activity was assessed using E. coli KCTC 1682 and S. aureus KCTC 1621 at a peptide concentration of 8 lm, which was the highest MIC for any bactenecin derivative against the strains used. The initial density of the cultures was approxi- mately 2 · 10 5 CFUÆmL )1 . After 0, 5, 10, 30 or 60 min of exposure to the peptides at 37 °C, 50 lL aliquots of serial 10-fold dilutions (up to 10 )3 ) of the cultures were plated onto LB agar plates to obtain viability counts. Colonies were counted after incubation for 24 h at 37 °C. To determine the salt sensitivity of the antimicrobial activity, peptides were incubated at 37 ° Cin100lLof1% peptone solution also containing 2 · 10 5 CFUÆmL )1 bacte- ria and 0 or 150 mm NaCl. After incubation for 3 h at 37 °C, 50 lL of the suspension was plated on LB agar for colony counts. Tricine gel electrophoresis Electrophoresis was performed with 15 lg samples of each bactenecin derivative dissolved in 2· sample buffer (125 mm Tris ⁄ HCl, pH 6.8, 20% glycerol, 2% mercaptoeth- anol, 0.04% bromophenol blue, and 4% SDS). The entire sample was loaded onto a 15% Tricine gel, after which the gel was fixed and stained with Coomassie dye. Acknowledgements This study was supported by the SRC ⁄ ERC program of MOST ⁄ KOSEF (R11-2000-083-00000-0) and the Brain Research Center of the 21st Century Frontier Research Program (M103KV010005-06K2201-00510). References 1 Hoffmann JA, Kafatos FC, Janeway CA & Ezekowitz RA (1999) Phylogenetic perspectives in innate immu- nity. Science 284, 1313–1318. 2 Zelezetsky I, Pontillo A, Puzzi L, Antcheva N, Segat L, Pacor S, Crovella S & Tossi A (2006) Evolution of the primate cathelicidin. Correlation between structural variations and antimicrobial activity. J Biol Chem 281, 19861–19871. 3 Zanetti M (2004) Cathelicidins, multifunctional peptides of the innate immunity. J Leukoc Biol 75, 39–48. 4 Bals R & Wilson JM (2003) Cathelicidins – a family of multifunctional antimicrobial peptides. Cell Mol Life Sci 60, 711–720. 5 Romeo D, Skerlavaj B, Bolognesi M & Gennaro R (1988) Structure and bactericidal activity of an antibi- otic dodecapeptide purified from bovine neutrophils. J Biol Chem 263, 9573–9575. 6 Wu M & Hancock REW (1999) Improved derivatives of bactenecin, a cyclic dodecameric antimicrobial cationic peptide. Antimicrob Agents Chemother 43, 1274–1276. 7 Bowdish DM, Davidson DJ, Scott MG & Hancock REW (2005) Immunomodulatory activities of small host defense peptides. Antimicrob Agents Chemother 49, 1727–1732. 8 Wu M & Hancock REW (1999) Interaction of the cyclic antimicrobial cationic peptide bactenecin with the outer and cytoplasmic membrane. J Biol Chem 274, 29–35. 9 Storici P, Tossi A, Lenarcic B & Romeo D (1996) Puri- fication and structural characterization of bovine cath- elicidins, precursors of antimicrobial peptides. Eur J Biochem 238, 769–776. 10 Tencza SB, Creighton DJ, Yuan T, Vogel HJ, Montel- aro RC & Mietzner TA (1999) Lentivirus-derived anti- microbial peptides: increased potency by sequence engineering and dimerization. J Antimicrob Chemother 44, 33–41. 11 Dempsey CE, Ueno S & Avison MB (2003) Enhanced membrane permeabilization and antibacterial activity of a disulfide-dimerized magainin analogue. Biochemistry 42, 402–409. J. Y. Lee et al. Salt-resistant homodimeric bactenecin FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS 3919 12 You S, Peng S, Lien L, Breed J, Sansom MS & Wool- ley GA (1996) Engineering stabilized ion channels: covalent dimers of alamethicin. Biochemistry 35, 6225– 6232. 13 Hara T, Kodama H, Kondo M, Wakamatsu K, Takeda A, Tachi T & Matsuzaki K (2001) Effects of peptide dimerization on pore formation: antiparallel disulfide- dimerized magainin 2 analogue. Biopolymers 58, 437– 446. 14 Takei J, Remenyi A, Clarke AR & Dempsey CE (1998) Self-association of disulfide-dimerized melittin ana- logues. Biochemistry 37, 5699–5708. 15 Singh PK, Jia HP, Wiles K, Hesselberth J, Liu L, Con- way BA, Greenberg EP, Valore EV, Welsh MJ, Ganz T et al. (1998) Production of beta-defensins by human airway epithelia. Proc Natl Acad Sci USA 95, 14961– 14966. 16 Campopiano DJ, Clarke DJ, Polfer NC, Barran PE, Langley RJ, Govan JR, Maxwell A & Dorin JR (2004) Structure–activity relationships in defensin dimers: a novel link between beta-defensin tertiary structure and antimicrobial activity. J Biol Chem 279, 48671– 48679. 17 Tam JP, Wu C & Yang JL. (2000) Membranolytic selectivity of cystine-stabilized cyclic protegrins. Eur J Biochem 267, 3289–3300. 18 Scocchi M, Zelezetsky I, Benincasa M, Gennaro R, Mazzoli A & Tossi A (2005) Structural aspects and bio- logical properties of the cathelicidin PMAP-36. FEBS J 272, 4398–4406. 19 Basak A & Lotfipour F (2005) Modulating furin activ- ity with designed mini-PDX peptides: synthesis and in vitro kinetic evaluation. FEBS Lett 579, 4813–4821. 20 Perczel A & Hollosi M (1996) Circular Dichroism and the Conformational Analysis of Biomolecules. Plenum Press, New York. 21 Chang CT, Wu CS & Yang JT (1978) Circular dichroic analysis of protein conformation: inclusion of the beta- turns. Anal Biochem 91, 13–31. 22 Woody RW. (1995) Circular dichroism. Methods Enzymol 246, 34–71. 23 Matsuzaki K, Harada M, Funakoshi S, Fujii N & Miyajima K (1991) Physicochemical determinants for the interactions of magainins 1 and 2 with acidic lipid bilayers. Biochim Biophys Acta 1063, 162–170. 24 Hornef MW, Putsep K, Karlsson J, Refai E & Anders- son M (2004) Increased diversity of intestinal antimicro- bial peptides by covalent dimer formation. Nat Immunol 5, 836–843. 25 Travis SM, Anderson NN, Forsyth WR, Espiritu C, Conway BD, Greenberg EP, McCray PB, Lehrer RI, Welsh MJ & Tack BF (2000) Bactericidal activity of mammalian cathelicidin-derived peptides. Infect Immun 68, 2748–2755. 26 Goldman MJ, Anderson GM, Stolzenberg ED, Kari UP, Zasloff M & Wilson JM (1997) Human beta-defen- sin-1 is a salt-sensitive antibiotic in lung that is inacti- vated in cystic fibrosis. Cell 88, 553–560. 27 Yomogida S, Nagaoka I & Yamashita T (1996) Purifi- cation of the 11- and 5-kDa antibacterial polypeptides from guinea pig neutrophils. Arch Biochem Biophys 328, 219–226. 28 Wu Z, Hoover DM, Yang D, Boule ` gue C, Santamaria F, Oppenheim JJ, Lubkowski J & Lu W (2003) Engi- neering disulfide bridges to dissect antimicrobial and chemotactic activities of human beta-defensin 3. Proc Natl Acad Sci USA 100, 8880. 29 Yang ST, Lee JY, Kim HJ, Eu YJ, Shin SY, Hahm KS & Kim JI (2006) Contribution of a central proline in model amphipathic alpha-helical peptides to self-associ- ation, interaction with phospholipids, and antimicrobial mode of action. FEBS J 273, 4040–4054. 30 Yang ST, Jeon JH, Kim Y, Shin SY, Hahm KS & Kim JI (2006) Possible role of a PXXP central hinge in the antibacterial activity and membrane interaction of PMAP-23, a member of cathelicidin family. Biochemis- try 45, 1775–1784. Supplementary material The following supplementary material is available online: Fig. S1. Trypsin cleavage sites and mass values of each peptide. Fig. S2. HPLC profiles of the peptide fragments after trypsin digestion. This material is available as part of the online article from http://www.blackwell-synergy.com Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corre- sponding author for the article. Salt-resistant homodimeric bactenecin J. Y. Lee et al. 3920 FEBS Journal 275 (2008) 3911–3920 ª 2008 The Authors Journal compilation ª 2008 FEBS . bacteria [5]. In addition, two lin- ear variants of bactenecin, Bac2S and Bac 2A, show similar activities against Gram-negative bacteria and stronger activities. that adopt parallel and antiparallel conformations and two monomers that adopt b-hairpin and linear confor- mations, and investigated their biological activities. Results