A Proline-Hinge Alters the Characteristics of the Amphipathic α-helical AMPs

HP (2–20) is a 19-aa, amphipathic, α-helical peptide with antimicrobial properties that was derived from the N-terminus of Helicobacter pylori ribosomal protein L1. We previously showed that increasing the net hydrophobicity of HP (2–20) by substituting Trp for Gln17 and Asp19 (Anal 3) increased the peptide's antimicrobial activity. In hydrophobic medium, Anal 3 forms an amphipathic structure consisting of an N-terminal random coil region (residues 2–5) and an extended helical region (residues 6–20). To investigate the structure-activity relationship of Anal 3, we substituted Pro for Glu9 (Anal 3-Pro) and then examined the new peptide's three-dimensional structure, antimicrobial activity and mechanism of action. Anal 3-Pro had an α-helical structure in the presence of trifluoroethanol (TFE) and sodium dodecyl sulfate (SDS). NMR spectroscopic analysis of Anal 3-Pro's tertiary structure in SDS micelles confirmed that the kink potential introduced by Pro10 was responsible for the helix distortion. We also found that Anal 3-Pro exhibited about 4 times greater antimicrobial activity than Anal 3. Fluorescence activated flow cytometry and confocal fluorescence microscopy showed that incorporating a Pro-hinge into Anal 3 markedly reduced its membrane permeability so that it accumulated in the cytoplasm without remaining in the cell membrane. To investigate the translocation mechanism, we assessed its ability to release of FITC-dextran. The result showed Anal 3-Pro created a pore <1.8 nm in diameter, which is similar to buforin II. Notably, scanning electron microscopic observation of Candida albicans revealed that Anal 3-Pro and buforin II exert similar effects on cell membranes, whereas magainin 2 exerts a different, more damaging, effect. In addition, Anal 3-Pro assumed a helix-hinge-helix structure in the presence of biological membranes and formed micropores in both bacterial and fungal membranes, through which it entered the cytoplasm and tightly bound to DNA. These results indicate that the bending region of Anal 3- Pro peptide is prerequisite for effective antibiotic activity and may facilitate easy penetration of the lipid bilayers of the cell membrane.


Introduction
Antimicrobial peptides have been classified into four major structural groups: a-helices, b-sheets, extended helices and loops. Of those, amphipathic a-helical structures are the most widely distributed among naturally occurring antimicrobial peptides [1]. The mechanisms of action for several, including defensin [2], cecropin [3], magainin [4], protegrin [5], melittin [6] and buforins [7], have been investigated, but the precise mechanism for their broad spectrum antimicrobial activity is still not fully understood. Evidence suggests, however, that they target the outer and inner bacterial membranes, ultimately either disrupting the cell membrane [2] or causing cooperative permeabilization [8].
To better understand the structure-activity relationship of Anal 3, in the present study, we constructed an additional analogue by substituting the kink amino acid Pro for Glu 9 , thereby creating a helix-hinge-helix peptide (Anal 3-Pro). By analyzing circular dichroism (CD) spectra, we then determined the conformations of Anal 3-Pro in membrane-mimicking environments, after which NMR spectroscopy was used to determine the tertiary structure of the peptide. We then assayed the antimicrobial efficacy of these peptides against several bacteria and pathogenic fungal cells, and the cytotoxic activity against human red blood cells (RBCs). In addition, the mechanisms of action of Anal 3 and Anal 3-Pro were compared in C. albicans.

Peptide Preparation
The peptides were synthesized using the solid phase method with Fmoc(9-fluorenyl-methoxycarbonyl)-chemistry [11]. Peptide purification was then carried out using preparative HPLC on a C 18 reverse-phase column. The amino acid compositions of the purified peptides were confirmed using an amino acid analyzer (HITACHI 8500A, Japan). The molecular weights of the synthetic peptides were determined using a matrix-assisted laser desorption ionization (MALDI) mass spectrometer.

CD
To determine the secondary structure of the peptides in membrane-mimetic environments, CD experiments were carried out using a J720 spectropolarimeter (Tokyo, Japan) with a cell having a 1-mm path length. The peptide CD spectra in water, 50% TFE/water solution and 100 mM SDS micelles were recorded at 25uC at wavelengths ranging from 190 to 250 nm obtained at 0.1 nm intervals. The peptide concentration was 100 mM. For each spectrum, the data from four scans were averaged and smoothed using the J720/98 system program (Version 120C). CD data were expressed as mean residue ellipticity [h] [12].

NMR and structure calculation
Perdeuterated sodium dodecylsulfate (SDS-d 25 ) was purchased from Cambridge Isotope Laboratories Inc, after which the conformation of the peptides in a membrane-mimicking environment (SDS micelles) was determined using NMR. The sample concentration was 1.0 mM in 0.45 ml of H 2 O/D 2 O (9:1,v/v; pH 4.0) containing 150 mM SDS micelles. All phase-sensitive two-dimensional experiments (i.e., DQF-COSY, TOCSY and NOESY) were carried out using the time-proportional phase incrementation method [13,14]. Structural calculations were carried out using X-PLOR version 3.851 [15] with the topology and parameter sets topallhdg and parallhdg, respectively. Standard pseudoatom corrections were applied to the non-stereospecifically assigned restraints [16], and an additional 0.5 Å was added to the upper bounds for NOEs involving methyl protons [17]. A hybrid distance geometry-dynamical simulated annealing protocol [18] was employed to generate the structures.

Antimicrobial assays
The bacteria were grown to the mid-logarithmic phase in medium comprised of (in g/l) 10 bactotryptone, 5 yeast extract and 10 NaCl (pH 7.0). To assay antibacterial activity, each peptide was diluted stepwise (concentrations: 100, 50, 25, 12.5, 6.25, 3.125, 1.56, 0.78, 0.39, 0.195, 0.097 mM) in 1% bactopeptone medium. The tested organism (final bacterial suspension: 5610 3 colony formation units (CFU)/ml) was suspended in 100 ml in growth medium, mixed with 100 ml of test peptide solution, and seeded into the wells of a microtiter plate. Bacterial growth was then measured as an increase in optical density at 620 nm after incubating 10 h at 37uC.

Ethics Statement
This study was approved by the institutional ethics committee, and all healthy donors provided written informed consent before treatment. We processed to the Ethical standards of the Institutional Ethics Committee of Chosun University and to the checklist for ethical consideration of cytotoxicity studies (https:// www.cre.or.kr/article/policy/1382313). Human red blood cells (RBCs) were obtained from blood freshly collected from healthy donors at the Chosun University Hospital in Kwangju (Republic of Korea). Moreover, this study received ethics approval from the Institutional Ethics Committee of Chosun University. The authors of this article were blinded to all personal data from the donors, and all blood donors remained anonymous. All procedures were carried out according to rules provided by the Institutional Ethics Committee of Chosun University. Samples of blood were obtained from 5 healthy donors. The samples were immediately stored at 4uC until needed.

Preparation of human red blood cells (RBCs)
Human RBCs were collected by centrifugation and washed three times with PBS (pH 7.0). The RBCs (100 ml) were suspended at 8% (v/v) in PBS and were plated into a 96-well plate. Peptide solution (100 ml) was added to each well, after which the plates were incubated for 1 h at 37uC and then centrifuged at 1506g for 5 min.

Hemolytic activity assay
The hemolytic activities of the peptides were evaluated by determining the release of hemoglobin from an 8% suspension of fresh human RBCs. A 100-ml aliquot of the 8% RBC suspension was added to the wells of a 96-well plate, after which 100 ml of peptide solution in PBS (pH 7.0) was added to a final peptide concentration of 100 mM. Hemolysis was then measured based on absorbance at 414 nm using an Emax plate reader. As controls, no and complete hemolysis were determined in PBS and 0.1% Triton-X 100, respectively. The degree of hemolysis was calculated using the following equation: Effect of the peptides on RBC morphology RBCs were incubated for 1 h at 37uC with 60% of the minimal inhibitory concentration (MIC) of Anal 3-Pro or melittin. Negative controls were run without peptides. The RBCs were then fixed in 0.05 M cacodylate buffer (pH 7.2) containing equal volumes of 4% glutaraldehyde and 1% paraformaldehyde. After lyophilization and gold coating, the samples were examined on a HITACHI S-2400 (Tokyo, Japan).

Flow cytometry
Membrane integrity following peptide treatment was evaluated using flow cytometry. Anal 3, Anal 3-Pro or melittin (control) was added to C. albicans cells (2610 5 cells in YPD media) to the MIC, after which cell membrane permeabilization was detected by incubating the treated cells in 50 mg/ml propidium iodide (PI) for 30 min at 4uC. After removing the unbound dye by extensively washing with PBS, the cells were subjected to flow cytometric analysis using a FACScalibur flow cytometer (Becton Dickinson, San Jose, CA). PI fluorescence was monitored in the FL2-H channel.

Confocal laser-scanning microscopy
Intracellular localization of fluorescein isothiocyanate (FITC)conjugated Anal 3, Anal 3-Pro or magainin II in C. albicans was determined by confocal laser scanning microscopy. The cells were inoculated into 3 ml of yeast medium and then incubated for 12 h at 28uC. The FITC-labeled peptides were added to 100 ml of the cell suspension to the MIC, after which they cells were incubated for 15 min at 28uC. Intracellular localization of FITC-labeled peptides was determined using an Olympus IX 70 upright microscope (Olympus, Japan) equipped with a Leica TCS 4D laser-scanning system.

Scanning electron microscopy (SEM)
Midlog phase C. albicans were resuspended at 10 8 CFU/ml in Na-phosphate buffer (pH 7.4) supplemented with 100 mM NaCl (buffer A) and then incubated at 37uC with Anal 3-pro or magainin II. Controls were run in the absence of peptide solvent. After 30 min, the cells were fixed for 2 h at 4uC in 0.2 M Nacacodylate buffer (pH 7.4) containing an equal volume of 5% glutaraldehyde. After fixation, the samples were filtered on Isopore filters (0.2 mm pore size, Millipore, Bedford, MA, USA) and extensively washed with 0.1 M Na-cacodylate buffer (pH 7.4). The filters were then treated with 1% osmium tetroxide, washed with 5% sucrose in cacodylate buffer and dehydrated in a graded ethanol series. After lyophilization and gold coating, the samples were examined on a HITHACHI S-2400 instrument (HITHACHI, Japan).

Preparation of small unilamellar vesicles (SUVs)
SUVs were prepared by drying of PC/cholesterol (10:1 w/w) under nitrogen, suspending the film by vortex mixing in buffer (50 mM phosphate buffer containing 100 mM, pH 7.5) and sonicating the suspension using a tip ultrasonicator. A drop of vesicles was deposited on a carbon-coated grid and negatively stained with 2% uranyl acetate. Specimens were examined in the TECNAI 12 at an accelerating voltage of 120 kV.

DNA gel retardation
Plasmid pBluescript SK(+) (Stratagene) was purified using a plasmid extraction kit (Exprep TM Quick, GeneAll Biotechnology Co., Seoul, Korea). The plasmid DNA (200 ng) was then mixed with increasing amounts of peptides in 10 mM Tris buffer (pH 8.0) containing 1 mM EDTA, 5% glycerol, 20 mM KCl and 50 mg/ml BSA. The mixtures were incubated for 10 min at 37uC and then electrophoresed on a 0.5% agarose gel in the TBE buffer, after which the gels were stained with ethidium bromide [20]. Gel retardation was visualized under UV illumination using a Bio-Rad Gel Documentation system. . FD-entrapping liposomes (unilamellar vesicles) with different lipid compositions were prepared using the reverse-phase evaporation method [21], after which the concentrations of the FD-entrapping vesicles were determined in triplicate using a phosphorus assay [22]. To prepare FDentrapping liposomes, a buffer solution (buffer II: PBS) containing 2 mg/ml of the FD was sonicated for 30 min with a lipid solution in chloroform (20 mg/ml) on ice. The chloroform was then gradually removed using a rotary vacuum evaporator at 25uC, resulting first in the formation of a viscous gel and then a liposome suspension. Buffer (2 ml) was added, and the suspension was evaporated further to remove the remaining solvent. The liposome suspensions were then centrifuged and washed for several cycles at 22,0006g av for 30 min in order to remove unentrapped-FD. The washed liposomes were extruded 30 times through polycarbonate filters (two stacked 0.4-mm pore size filters) using an Avanti Mini-Extruder (Avanti Polar Lipids inc., Alabaster, AL) to obtain liposomes of homogeneous size (,400 nm). Aliquots of the peptide solutions at appropriate concentrations were incubated with a suspension of FD-loaded liposomes (100 mM) for 1-20 min at 25uC, and were then centrifuged for 30 min at 22,0006g. The supernatants were recovered and the leakage was recorded by monitoring the FITC fluorescence intensity (excitation wavelength: 494 nm, emission wavelength: 520 nm). 100% leakage was achieved by addition of Triton X-100 to a final concentration of 1 mM. The percent leakage value was then plotted.
We selected Glu 9 RPro substitution because the resultant peptide is negatively charged and does not favour electrostatic interaction between itself and the negatively charged bacterial membrane. Moreover, we wanted to introduce a hinge region into the middle of the peptide, and with the exception of Glu, all of the amino acids in that region of the peptide are essential for its antimicrobial activity. This mutation changed the peptide's conformation because proline is a potent a-helix breaker that has been found in the transmembrane helices of a variety of integral membrane proteins [24] and in a number of a-helical antimicrobial peptides that promote ion channel activity, including melittin [25], cecropin A [26], paradoxin [27], brevinins [28], gaegurin [29] and buforin II [8]. The primary structures of these cell-selective antimicrobial peptides consist of a hinge (a Pro residue) conecting N-and C-terminal a-helical segments, and twodimensional NMR studies have shown that these peptides do indeed assume a helix-hinge-helix structure in membrane-mimicking environments [30]. The importance of the Pro residue was confirmed by the findings that Pro 14 RAla substitution in the active analogue derived from gaegurin induced a significant reduction in antibacterial activity [31], while elimination of Pro 7 from pardaxin increased hemolytic activity [27].
The effect of incorporating a Pro-hinge into amphipathic ahelical peptides that lack cell selectivity had not been investigated until now, however. We therefore introduced a Glu 9 RPro substitution into HP (2-20) Anal 3 with the idea of creating a helix-hinge-helix peptide with no cell selectively (Anal 3-Pro). We then investigated the effect of the Pro incorporation on the ahelical structure of Anal 3 using CD and NMR studies, and compared the antimicrobial and hemolytic activities of Anal 3-Pro with those of the previously described HP (2-20) Anals 1-5 [10].

CD Spectra
We used SDS micelles to examine the behaviour of the AMPs in a membrane-mimicking environment, and we used TFE as a cosolvent in CD spectroscopy studies of peptide folding. This solvent is often used to increase the propensity of AMPs to form secondary structure. We first investigated the secondary structures of HP (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20), Anal 3 and Anal 3-Pro by examining their CD spectra in aqueous buffer, 50% TFE/water solution and SDS micelles. We found that the CD spectra for all three peptides showed a random coil structure in water ( Figure S1) but an a-helical structure in a membrane-mimicking environment, such as 50% TFE/water solution or 150 mM SDS micelles. Moreover, the data indicate that Anal 3 has a more a-helical structure than Anal 3-Pro, suggesting the proline residue in the central region caused distortion of the helix.

Resonance Assignment
Sequence-specific resonance assignment was carried out using DQF-COSY, TOCSY and NOESY data. Figure S2 and S3 show the NOESY spectra with sequential assignments of Anal 3-Pro in the HN-HN region. NOESY and TOCSY experiments carried out at 313K, 318K and 323K enabled complete assignment of the overlapping peaks. Chemical shifts within Anal 3-Pro in SDS micelles at 313K and pH 4.0 (relative to DSS) are listed in Table  S1. Anal 3-Pro has a short helix extending from Lys 3 to Leu 8 and a long helix extending from Pro 10 to Trp 19 and a hinge region between the helices. The observed values of the 3 J HNa coupling constant for the helical region in the C-terminus of the peptide were generally less then 6 Hz. The values of the amide proton temperature coefficient were used to predict hydrogen bond donors; values more positive than 24.5 ppb/K were taken as indicating that the amide proton is involved in intramolecular hydrogen bonding.

Tertiary Structures
The tertiary structures of Anal 3-Pro were determined using sequential (|i2j| = 1) and medium-range (1,|i2j|#5) as well as restraints on intraresidual distances and torsion angles (Table S2). From among the structures, which were accepted with small deviations from the idealized covalent geometry and experimental restraints (#0.05 Å for bonds, #5u for angles, #5u for chirality, #0.3 Å for NOE restraints and #3u for torsion angle restraints), the 20 output structures with the lowest energy were analyzed. The statistics for these 20 simulated annealing (SA) structures of Anal 3-Pro are given in Table S2. All displayed good covalent geometries and small NMR constraint violations. When we superimposed the 20 structures on the backbone atoms of the residues extending from Pro 10 to Trp 19 , the rmsd from the mean structure was 0.77 Å for the backbone atoms (N, Ca, C9, O) and 1.46 Å for all heavy atoms. However, the rmsd from the mean structure for all residues was much larger for all peptides because of the kink in the peptide backbone if the region in the middle of the molecule. Figure 1A shows a ribbon diagram of the lowest energy Anal 3-Pro structure, while in Figure 2B all heavy atoms in the region extending from Pro 10 to Trp 19 were superimposed with respect to the restraint-minimized average structure for all peptides. Using PROCHECK analysis, we determined that Anal 3-Pro contains a short a-helix, extending from Lys 4 to Leu 9 in the N-terminal region, and a long a-helix, extending from Pro 10 to Trp 19 in the Cterminal region, which are connected via the Pro 10 residue.

Antimicrobial and Hemolytic Activities
The antimicrobial activities of the peptides were evaluated in vitro against a representative set of bacterial strains, including three Gram-positive and four Gram-negative species, using the broth microdilution method. The respective MICs are summarized in Table 1. The data indicated that Anal 3-Pro had 4 times greater antibacterial activity than Anal 3 against all bacterial strains tested. In addition, when the MICs for the antifungal activity of the peptides were determined in MTT assays, we found that Anal 3-Pro had approximately twice the antifungal activity of Anal 3 against the microorganisms tested (Table 1). In particular, replacing Glu with Pro in the middle of the peptide sequence introduces a kink into the helix structure, which apparently increased the antimicrobial activity of Anal 3-Pro, as compared to Anal 3.
We also evaluated the hemolytic activity of the peptides. Notably, none of the synthetic peptides tested showed any hemolytic activity. Anal 3-Pro, for example, exhibited no hemolytic activity at concentrations as high as ,100 mM, despite its strong antimicrobial activity ( Figure S4). Scanning electron microscopic examination of RBCs confirmed that, even at 100 mM, Anal 3-Pro elicited no change in RBC morphology and caused no damage to the cell membrane (compare Figure S4A and B). By contrast, at a concentration inducing 60% hemolysis, melittin, an antimicrobial peptide known to target the cell membrane [6], had a significant effect on RBC morphology, frequently inducing formation of pores in the membrane ( Figure  S4C). Thus, Anal 3-Pro shows remarkable antibacterial and antifungal activity with no hemolytic activity. Its further study may therefore not only increase our understanding of the mechanism of action of Anal 3-Pro and the mechanism underlying its cell-  selectivity, but also facilitate the design of novel antibiotic peptide with improved antimicrobial activity and no cytotoxicity.

Flow Cytometric Analysis of Cell Membrane Permeability
The precise mechanism of action of helix-hinge-helix antimicrobial peptides is not yet fully understood, this finding is consistent with the idea that disruption of the cell structure through pore formation [32] and/or ion channel generation [33] is the most likely mechanism. To investigate the extent to which Anal 3-Pro acts by damaging the plasma membrane or by affecting cell physiology, C. albicans were incubated with propidium iodide (PI) [32] plus Anal 3, Anal 3-Pro or melittin, after which intracellular PI fluorescence within individual cells was analyzed by FACScalibur flow cytometry. Figure 2 shows a plot of forward light scatter (x-axis) against PI fluorescence (y-axis) in which each dot represents an individual cell. The results showed that whereas most cells left untreated (Figure 2A) or treated with Anal 3-Pro ( Figure 2C) showed little or no PI signal, cells treated for 30 min on ice with 12.5 mM Anal 3 or melittin showed a dramatic rightward shift of the peak, indicating a large influx of PI (Figure 2, B and D). We found that when the MICs of peptides (Ana 3 and melittin) were applied to the pathogenic fungus C. albicans, they exhibited a strong capacity to permeabilize the cell membrane. By contrast, Anal 3-Pro caused only a very small amount of dye influx, suggesting a different mechanism of antimicrobial activity.

Confocal Laser-Scanning Microscopy
The observed differences in the effects of Anal 3, Anal 3-Pro, buforin II and TAT on membrane permeability prompted us to investigate the site of action of Anal 3-Pro. To that end, C. albicans or T. begellii were incubated with FITC-conjugated Anal 3; Anal 3-Pro; magainin II, another antimicrobial peptide that targets the cell membrane; buforin II, an antimicrobial peptides that targets the cytoplasm after permeating a membrane micropore; or TAT, another antimicrobial peptide that targets the cytoplasm, after which the cells were visualized using confocal laser-scanning microscopy. We found that FITC-Anal 3 remained associated with the cell membrane (Figure 3, Aa, Ab, Ba and Bb), as did FITCmagainin II (Figure 3, Ae and Af). By contrast, Anal 3-Pro (Figure 3

Non-morphological Changes Induced by Anal 3-Pro
Scanning electron microscopic examination of C. albicans treated with the respective peptides revealed that untreated cells had a normal, smooth surface ( Figure S5A), that cells treated with Anal 3-Pro or buforin II had similarly smooth membranes ( Figure S5B and D) and that cells treated with magainin II had a rough cell surface ( Figure S5C). This suggests that substituting Pro for Glu significantly reduced the degree of peptide-induced membrane disruption, as compared to Anal 3 [34].   To confirm the ability of Anal 3-Pro to disrupt microbial cells, we carried out a set of experiments using liposomes. Artificial small unilamellar vesicles (SUVs) (phosphatidylcholine (PC)/cholesterol (CH); 10:1, w/w) and neutral PC vesicles were used as model membrane systems. The antimicrobial activity of Anal 3 is based on the formation of transmembrane channels. SUVs were disrupted after treatment with Anal 3 ( Figure S6B), suggesting the peptides perturb the lipid components of the plasma membranes. But, SUVs treated with Anal 3-pro were not disrupted and retained their vesicluar form ( Figure S6C). This clearly indicates that the change in overall length caused by the helix distortion likely prevents favourable interactions across the membrane, such as those seen with the extended Anal 3 helix.

Anal 3-Pro Recognizes Chitin and Peptidoglycan
The pattern motifs of the microbial cell wall components from gram-negative bacteria, gram-positive bacteria and yeast can initiate innate immune signaling. To examine the specificity of  Anal 3-Pro binding, we incubated the cell wall components with the indicated peptide as described in the ''Materials and Methods.'' Using these binding assays, we specifically detected Anal 3-Pro in extracts containing chitin and peptidoglycan, whereas no binding activity was observed with chitosan, cellulose or b-1,3-glucan ( Figure 4). Thus, Anal 3-Pro appears to recognize chitin and peptidoglycan and to permeabilize the cellular target site. By contrast, buforin II recognized only chitin.

DNA Gel Retardation
To confirm the cellular target site of Anal 3-Pro, we also assessed retardation of DNA band migration. Given the concentration-dependent translocation of Anal 3-Pro into the cytoplasm of bacterial cells shown in Figure S7, we examined the binding to intracellular DNA. Anal 3-Pro was mixed with a fixed amount of plasmid DNA (pDNA, 100 ng), after which the complexes were electrophoresed on agarose gels. Retardation of the pDNA band was detected for a 600 ng peptide, indicating aggregation of the pDNA with Anal 3-Pro ( Figure S7B). Additional in vitro experiments will be needed to determine how the peptide-DNA interaction is mediated. On the other hand, no band retardation was seen when pDNA was mixed with Anal 3 ( Figure S7A), indicating pDNA and Anal 3 do not interact. In addition, buforin II ( Figure S7C) and TAT ( Figure S7D) were shown to induce retartation at different peptide concentrations, 1 mg or 400 ng, respectively.

Sizes of the Different Pores Formed between Peptides
To gather additional clues as to the type and extent of membrane damage induced by Anal 3-Pro, TAT and buforin II, we assessed the peptide-induced release of fluorescently labeled dextran molecules of various sizes -i.e., FD4 (3.9 kDa, 1.8 nm radius), FD10 (9.9 kDa), FD20 (19.8 kDa, 3.3 nm radius), FD40 (40.5 kDa, 4.8 nm radius), FD70 (71.6 kDa, 5 nm radius) and FD500 (530 kDa). Using FD-containing liposomes incubated with 10 mM peptide (peptide/lipid ratio = 1/10) [35,36], we determined that the evoked release of FD varied inversely with its molecular mass/size ( Figure 4A and B). Anal 3-Pro released barely 20% of FD4 from phosphatidylethanolamine (PE)/phosphatidyglycerol (PG) liposomes, while TAT and buforin II released only 5% of the FD4, and were unable to release any of the larger markers. For instance, Anal 3-Pro released 39.63% of the FD4 from PC/cholesterol (CH) liposomes, but only 16.65% of the FD10. TAT and buforin II respectively released only 4.7% and 3.6% of the FD4 from PC/CH vesicles, and were unable to release any of the larger markers. This indicates that the pore created by Anal 3-Pro has an estimated radius less than 1.8 nm, which is similar to pores induced by buforin II or TAT at 1/10 (peptide/ lipid) molar ratio concentration ( Figure 5A). When we then compared the release from PE/PG liposomes with that from PC/ CH liposomes ( Figure 5B), we determined that release from the former was peptide-specific, whereas release from the latter did not vary among the peptides.
In summary, to investigate the structure and mechanism of action of Pro-kink incorporation, we designed Anal 3-Pro in which the Glu at position 9 of Anal 3 was substituted with Pro introduced kink into a-helices. Anal 3-Pro exhibited an increase in antibiotic activity without a hemolytic effect. It appears that incorporating Pro into Anal 3 induces a more selective cytotoxicity against microorganisms with a change in the mode of action of the amphipathic a-helical peptide. The modes of action of Anal 3 and Anal 3-Pro are summarized in Figure 6. Anal 3 exerts its bactericidal effects by disrupting the membrane and inducing pore formation. By contrast, Anal 3-Pro acts by inducing formation of a short-lived micropore, permeating into the cytoplasm and then binding to the DNA. Our results clearly demonstrate that Anal 3-Pro constitutes a new class of antimicrobial peptide that targets intracellular components, most likely DNA, after permeating short-lived micropores in the cell membrane, and that the proline hinge is a key structural factor for the cell penetration. Thus insertion of a single amino acid to form a proline-hinge region can change a membrane-targeted peptide to a cell-penetrating one. This finding may provide an important clue to the design of future therapeutic antibiotic drugs, given its efficacy toward bacterial and fungal cells and its lack of toxicity toward eukaryotic cells such as human erythrocytes.