High Potency and Broad-Spectrum Antimicrobial Peptides

60

Biomacromolecules 2010, 11, 60–67

High Potency and Broad-Spectrum Antimicrobial Peptides Synthesized via Ring-Opening Polymerization of r-Aminoacid-N-carboxyanhydrides Chuncai Zhou, Xiaobao Qi, Peng Li, Wei Ning Chen, Lamrani Mouad,† Matthew W. Chang, Susanna Su Jan Leong,* and Mary B. Chan-Park* School of Chemical and Biomedical Engineering, Nanyang Technological University, 62 Nanyang Drive, Singapore 637459, Singapore, and Menicon Company, 104 rue Martres, 92583 Clichy cedex, France Received August 6, 2009; Revised Manuscript Received November 12, 2009

Antimicrobial peptides (AMPs), particularly those effective against methicillin-resistant Staphylococcus aureus (S. aureus) and antibiotic-resistant Pseudomonas aeruginosa (P. aeruginosa), are important alternatives to antibiotics. Typical peptide synthesis methods involving solid-phase sequential synthesis are slow and costly, which are obstacles to their more widespread application. In this paper, we synthesize peptides via ring-opening polymerization of R-amino acid N-carboxyanhydrides (NCA) using a transition metal initiator. This method offers high potential for inexpensive synthesis of substantial quantities of AMPs. Lysine (K) was chosen as the hydrophilic amino acid and alanine (A), phenylalanine (F), and leucine (L) as the hydrophobic amino acids. We synthesized five series of AMPs (i.e., P(KA), P(KL), P(KF), P(KAL), and P(KFL)), varied the hydrophobic amino acid content from 0 to 100%, and determined minimal inhibitory concentrations (MICs) against clinically important Gramnegative and Gram-positive bacteria and fungi (i.e., Escherichia coli (E. coli), P. aeruginosa, Serratia marcescens (S. marcescens), and Candida albicans (C. albicans). We found that P(K10F7.5L7.5) and P(K10F15) show the broadest activity against all five pathogens and have the lowest MICs against these pathogens. For P(K10F7.5L7.5), the MICs against E. coli, P. aeruginosa, S. marcescens, S. aureus, and C. albicans are 31 µg/mL, 31 µg/mL, 250 µg/mL, 31 µg/mL, and 62.5 µg/mL, while for P(K10F15) the respective MICs are 31 µg/mL, 31 µg/mL, 250 µg/mL, 31 µg/mL, and 125 µg/mL. These are lower than the MICs of many naturally occurring AMPs. The membrane depolarization and SEM assays confirm that the mechanism of microbe killing by P(K10F7.5L7.5) copeptide includes membrane disruption, which is likely to inhibit rapid induction of AMP-resistance in pathogens.

1. Introduction The increasing use in recent decades of antibiotics for biomedical and agricultural purposes has led to the emergence of more resistant and virulent strains of pathogens, including methicillin-resistant S. aureus and antibiotic-resistant P. aeruginosa.1 The human cost of these developments is not small; it is estimated that annual mortality from S. aureus-associated infections is about 90000 in the U.S. alone. There is urgent need for highly effective antimicrobials against clinically significant bacteria (such as S. aureus and P. aeruginosa) and fungi (such as Candida albicans (C. albicans)). A promising approach is the use of antimicrobial peptides (AMPs) such as defensins, cathelicidins (LL-37), and magainins,2–6 which are components of the natural immune system. Further, AMPs typically show broad antimicrobial activity against bacteria, fungi, and some types of viruses7–11 and have low propensity to induce resistance in pathogens.12,13 In addition to poor proteolytic stability and toxicity to mammalian cells, cost is a major obstacle to widespread use of AMPs as antimicrobial agents.14,15 Research has shown that polymers such as polymethacrylates, polynorbornene, polypyridinium, and polyarylamide can be tuned to have impressive antimicrobial activity with high selectivity, but most of these are unsuitable for topical and systemic treatment, as they are nonbiodegradable.15–27 Instead, these polymers are typically used as disinfectant additives in * To whom correspondence should be addressed. E-mail: mbechan@ ntu.edu.sg (M.B.C.-P.); [email protected] (S.S.J.L.). † Menicon Company.

coatings, handwash, detergents, filters, and so on, rather than as therapeutic agents.26–28 On the other hand, suitable AMPs with controlled proteolysis can be highly bioavailable but degraded after systemic/topical treatment.29,30 The common techniques for making AMPs, which typically have 10-30 amino acid residues,31–33 include solid-phase synthesis and solution-coupling.34–37 Although these serial techniques offer versatility in the precise control of the peptide chain composition, they are time-consuming and expensive. Deming38,39 has developed transition metal initiators that allow living and controlled polymerization of R-amino acid Ncarboxyanhydrides (NCA) for synthesizing peptides with narrowly defined chain length and intrachain composition.40 These NCA-based peptides have been shown to interact with cell membranes.40 However, previous results did not address the question of possible antimicrobial activity of these kinds of peptides. Peptides prepared by NCA ring-opening polymerization can easily be produced in large amounts (up to kilogram quantities) and this method offers versatile control of peptide composition and length. No publication to date has demonstrated that peptides synthesized with R-amino acid NCA monomer using transition metal initiators have antimicrobial activity and their optimal composition. This paper reports the NCA ring-opening copolymerization synthesis and characterization of five series of copeptides containing cationic lysine and hydrophobic leucine, alanine, or phenylalanine residues. The syntheses were performed using Ni(COD)2 initiator (Figure 1). The five series of copeptides were (1) poly(lysine-co-alanine) (hereafter denoted as P(KA)), (2)

10.1021/bm900896h  2010 American Chemical Society Published on Web 12/03/2009

Broad-Spectrum Antimicrobial Peptides

Biomacromolecules, Vol. 11, No. 1, 2010

61

give a homogeneous solution of the corresponding NCA in 1-3 h. Briefly, 10 g of amino acid was suspended in 100 mL of THF, and the mixture was heated to 50 °C, followed by the addition of an equivalent amount of triphosgene. In the absence of a clarified solution within 1 h, 2 to 3 aliquots (0.05 equiv) of triphosgene were added to the mixture at 30 min intervals. After 3 h, the reaction mixture was poured into 300 mL of hexane, and the suspension was stored for 16-20 h at -20 °C to ensure complete crystallization. The product was recrystallized from THF/hexane to yield NCA-monomer crystals. Synthesis of Poly(Nε-benzyloxycarbonyl-L-lysine-ran-L-phenylalanine). The initiator Ni(COD)2 (71.9 mg, 0.26 mmol) was added to a 40 mL mixture of DMF/THF (10:1) in a dried flask. NCA-Lphenylalanine (0.578 g, 3.27 mmol) and NCA-Nε-benzyloxy carbonylL-lysine (1 g, 3.27 mmol) were dissolved in DMF (10 mL) and then added into the initiator solution. The mixture was stirred at room temperature in argon for 4 h. The polymers were isolated by addition of water, containing HCl (1 mmol), to the reaction mixture, causing precipitation of the polymers. The polymers were then washed with water and centrifuged several times to remove excess DMF. The polymers were dried in vacuum to give poly(Nε-benzyloxycarbonylL-lysine-ran-L-phenylalanine) copolymers as white solids (yield, >87%).

Figure 1. Schematic of (a) R-amino NCA monomer synthesis. (b) Synthesis of copeptides by the NCA method.

poly(lysine-co-leucine) (i.e., P(KL)), (3) poly(lysine-co-phenylalanine) (i.e., P(KF)), (4) poly(lysine-co-alanine-co-leucine) (i.e., P(KAL)), and (5) poly(lysine-co-phenylalanine-co-leucine) (i.e., P(KFL)). The length of the copeptides was fixed at 25 residues by fixing the ratio of NCA-monomers to initiator to 25. For the P(KF) series, the effect of length variation was also studied. The effect of composition and length of the NCA peptides on their minimum inhibitory concentrations (MICs) against clinically significant Gram-negative bacteria (i.e., E. coli ATCC8739, P. aeruginosa ATCC9027, and S. marcescens ATC13880) and Gram-positive bacteria (S. aureus ATCC6538) and fungi (C. albicans ATCC10231) were studied. Gel permeation chromatography (GPC) and proton nuclear magnetic resonance (1H NMR) were used to characterize the peptides. The mechanism for pathogen killing was also investigated using the cytoplasmic membrane potential-sensitive dye DiSC3(5) and scanning electron microscopy (SEM) to observe cell structure change after peptide exposure. The cytotoxicity of the peptides was characterized by hemolytic activity with red blood cells.

2. Experimental Details Materials. L-Alanine, L-leucine, L-phenylalanine, Nε-benzyloxycarbonyl-L-lysine, phosphatidylcholine (POPC), Ni(COD)2, bis(trichloromethyl)carbonate, hexane, diethyl ether, tetrahydrofuran (THF), trifluoroacetic acid (TFA), N,N-dimethylformamide (DMF), hydrogen bromide (30% in acetic acid), and other chemical reagents were purchased from Sigma-Aldrich and used without further purification unless otherwise specified. Drying of THF and hexane was achieved by heating in the presence of sodium strips and calcium hydride, respectively, for 48 h under reflux conditions. DMF was dried using molecular sieves. Cecropin A, Melittin, LL-37, Indolicidin, Magainin I, and Defensin (HNP-1) were purchased from Quality Controlled Biochemicals Ltd. (U.S.A.). Synthesis of NCA-Monomers. One-third equivalence of triphosgene was added to an amino acid suspension in anhydrous THF at 50 °C to

Synthesis of Protonated Poly(L-lysine · HBr-ran-L-phenylalanine). The poly(Nε- benzyloxycarbonyl-L-lysine-ran-L-phenylalanine) samples previously synthesized were each dissolved in 10 mL of TFA in a 50 mL flask. Excess HBr (30% in acetic acid) was then added and the mixtures were stirred for 5 h at room temperature. Deprotected polymers were isolated by addition of diethyl ether to the reaction mixtures, causing precipitation of the peptides. The peptides were then washed with excess diethyl ether and acetone and finally dialyzed against deionized water and dried in vacuum at 50 °C. All other copolymers were deprotected in the same manner. Spectroscopy Characterization. 1H NMR spectra were recorded on a Bruker DMX-3000 spectrometer with either DMSO or D2O as solvent. GPC was performed on an Agilent 1100 Series equipped with GPC-SEC (size exclusion chromatography) data analysis software. Peptides were dissolved in 0.1 M LiBr in DMF and the GPC eluent was kept at 60 °C. Measurement of Minimum Inhibitory Concentrations (MIC). Minimal inhibitory concentrations (MICs) were determined by dilution in the Mueller Hinton Broth (MHB) medium. Briefly, bacteria cells were grown overnight at 37 °C in MHB medium to a midlog phase and diluted to 104-105 colony forming units (CFU) mL-1. A dilution series for the peptides was made by diluting 1000 mg/mL peptide stock in MHB broth to final peptide concentrations ranging from 1000 to 8 µg/mL. A total of 100 µL of each diluted peptide from the serial dilutions was added to microtiter plates, followed by addition of 100 µL of bacterial suspension, to give a final inoculum of 5 × 105 CFU/mL. The plates were incubated at 37 °C for 18 h, and microbial growth was determined by measuring optical density absorbance at 600 nm. Two replicates were used for each pathogen, peptide, and concentration. Positive controls containing no antimicrobial peptides and negative controls containing no bacterial cells were also used. The experiments were independently repeated twice. The MIC was defined as the lowest peptide concentration that arrested bacterial growth after 18 h.41 For the fungal pathogen, C. albicans, the growth condition was the same except the (1) growth medium was RPMI1640 supplemented with 2% glucose, and (2) MIC end points were determined at 24 and 48 h and recorded as the minimum concentration with fungi growth.42,43 The MICs of other common AMPs, including Cecropin A, Melittin, LL-37, Indolicidin, Magainin I, and Defensin (HNP-1), were also measured for comparison. Hemolytic Activity. Peptide hemolytic activity was tested by determining the release of hemoglobins in erythrocyte suspensions derived from fresh human blood (5% v/v). Blood was aseptically collected and stored for less than 2 h at 4 °C. The erythrocyte suspensions were then centrifuged at 2000 × g for 5 min and washed three times with Tris buffer (10 mM Tris, 150 mM NaCl, pH7.2) before

62


being 2-fold diluted in Tris buffer. Peptides were solubilized in Tris buffer to final concentrations of 1000, 500, 250, 125, 62.5, 32, 16, 8, 4, 2, and 1 µg/mL and incubated with the erythrocyte suspension at equal volume ratios (i.e., 50:50 µL) under continuous shaking for 1 h at 37 °C. The mixtures were then centrifuged at 3500 × g for 10 min. Aliquots of 80 µL of the supernatant were transferred to a 96-well microplate and diluted with 80 µL of distilled water. Hemoglobin release was monitored by absorbance measurement of the erythrocyte samples at 540 nm using a microplate reader (Biorad). Complete hemolysis was determined in Tris buffer containing 1% Triton-X100 (positive control). Erythrocytes in Tris buffer were used as negative control. Percentage of hemolysis (H) was calculated as H (%) ) 100 × [(Op - Ob)/(OT - Ob)], where Op is the peptide density for a given peptide concentration, Ob is the Tris buffer density, and OT is the positive control (1% Triton-X100) density. HC50 is the peptide concentration resulting in 50% cell hemolysis. Circular Dichroism (CD). CD spectra were recorded at room temperature on a Chirascan CD spectrometer (Applied Photophysics Limited, U.K.). A quartz cell with a path length of 0.5 mm was used. Spectra were generated from 190 to 250 nm wavelengths at 0.1 nm intervals, 50 nm/min speed, 0.5 s response time, and 1 nm bandwidth. Peptides were dissolved to a final concentration of 1 µg/mL in 10 mM sodium phosphate buffer (pH 7.4) containing 50 µM phosphatidylcholine (POPC) small unilamellar vesicles. The spectra were plotted as mean residue ellipticity versus wavelength. Membrane Depolarization of Intact Bacterial Cells. The peptides’ ability to depolarize membranes was determined using the membrane potential-sensitive fluorescent dye, DiSC3-5 on intact S. aureus and C. albicans cells, based on the methods reported by Friedrich et al.44 S. aureus and C. albicans were grown at 37 and 28 °C, respectively, under shaking conditions to midlog phase and harvested by centrifugation. The cells were washed three times with wash buffer (20 mM glucose, 5 mM HEPES, pH 7.4) and resuspended to A600 ) 0.05 in the same buffer but also containing 0.1 M KCl. A DiSC3-5 stock solution was added to the cell suspension to a final concentration of 20 nM DiSC35, and incubated until a constant fluorescence reading was achieved, indicating the successful uptake of the dye into the bacterial membrane. Changes in fluorescence were recorded with a luminescence spectrometer (Aminco Bowman II). After the addition of peptides, membrane depolarization was determined by monitoring the change in fluorescence emission intensity of the DiSC3-5 dye at excitation and emission wavelengths of 622 and 670 nm, respectively. Complete dissipation of the membrane potential was achieved by adding gramicidin D (0.2 nM). Morphology Change of Pathogens. Pathogen morphologies were examined by SEM after 15 min treatment with peptide at MIC concentration. Bacterial cells were fixed with 4% gluteraldehyde in 0.15 M sodium phosphate buffer (pH 7.4) at equal volume ratios. The sample tube was mixed by gently inverting the tube up and down for several minutes to prevent clumping of the cells. The fixed cell suspension was transferred to a cover slide and allowed to air-dry. The slides were rinsed with 0.15 M sodium phosphate buffer (pH 7.4) and dehydrated through a graded series of ethanols (30-100%). After ethanol dehydration, the slides were oven-dried at 60 °C for 5 min. The slides were sputter-coated with gold-palladium for 80 s at 20 mA and imaged with a field emission scanning electron microscope (JEOL Field Electron Microscope, JSM-6700F, Japan).

3. Results 1

Figure 2A shows the H NMR spectrum of a typical protected peptide, poly(Nε-Benzyl oxycarbonyl-L-lysine-ran-L-phenylalanine), and the spectrum confirms that the ring-opening reaction (Figure 1b, step 1) proceeded as planned. After deprotection, copeptides were obtained and the 1H NMR spectrum (Figure 2B) confirms that the deprotected peptide, specifically protonated poly(L-lysine-ran-L-phenylalanine), has been obtained. GPC (Table 1 and Supporting Information, Figure S1) confirms that

Zhou et al.

Figure 2. NMR spectra of (A) poly(Nε-benzyloxycarbonyl-L-lysineran-L-phenylalanine) (lysine mol content is 50%) in DMSO and (B) protonated poly(L-lysine-ran-L-phenylalanine) (lysine mol content is 50%) in D2O. Table 1. GPC Results of Poly(Nε-benzyloxycarbonyl-Llysine-ran-L-phenylalanine)a ratio of NCA-monomers /initiator

Mn

Mw/Mn

15 25 35 45 100

5100 9300 13400 17600 35100

1.18 1.11 1.28 1.14 1.16

a

Lysine content is 40%.

the NCA polymerization with Ni(COD)2 initiator is controllable and the number-average molecular weight (Mn) of the peptides can be varied from around 5K to 35K. The molecular weight polydispersities of these peptides are narrow and between 1.1-1.3. The hydrophobic amino acid (HAA) content of each of the five series of copeptides was varied from 0 to 100%. The MIC results for each series are shown in Figure 3A-E and Table 2. The antimicrobial activity of the 100% HAA polymers was not measured because these polymers could not dissolve in water. Figure 3A shows that the best antibacterial activity of the P(KA) copeptides against E. coli, P. aeruginosa, and S. aureus occurs when the hydrophobic alanine residue content is in the 20-50% range; outside this range, the P(KA) peptides exhibit very limited antimicrobial activity. None of the P(KA) copeptides exhibit significant activity against S. marcescens and C. albicans, indicating that P(KA) shows relatively narrow antimicrobial activity. The antibacterial efficacy of P(KL) copeptides is shown in Figure 3B. The P(KL) copeptides exhibit antibacterial activities over a wider range of HAA content compared to P(KA)



63

Figure 3. MIC results of the peptides (A) P(KA), (B) P(KL), (C) P(KF), (D) P(KAL), (E) P(KFL), and (F) P(KF) with varying peptide length. Table 2. MIC (µg/mL) of Optimized Copolypeptides polypeptide

E. coli (Gram-)

P. aeruginosa (Gram-)

S. marcescens (Gram-)

S. aureus (Gram+)

C. albicans (fungi)

P(K10F15) P(K10F7.5L7.5) Cecropin A Melittin LL-37 Indolicidin Magainin I Defensin (HNP-1)

31 31 4 64 >256 >256 128 >128

31 31 64 64 >256 128 >256 >128

250 250 64 32 >256 >256 >256 >128

31 31 128 8 >256 128 128 >128

125 62.5 256 32 >256 128 256 >512

copeptides, and the most potent formulations of the P(KL) copeptides are distinctly more effective with lower MICs than the P(KA) copeptides: with 50% L content, the MICs against E. coli, P. aeruginosa, and S. aureus are 62.5, 125, and 62.5 µg/mL compared to 125, 250, and 62.5 µg/mL, respectively, for the most optimum composition in the P(KA) series. Like the P(KA) copeptides, the P(KL) copeptides do not have significant activity against S. marcescens and C. albicans. The P(KF) copeptides are effective against all five pathogens tested (Figure 3C) and are effective against most of them over

HC50 (µg/mL) 16 16 8 >100

a broad range of phenylalanine content. It is particularly interesting to observe that the optimal antimicrobial behavior, or the lowest MIC, of the P(KF) copeptides series occurs at the same F content for all five pathogens (i.e., 60%) and the MICs are 31, 31, 250, 31, and 125 µg/mL against E. coli, P. aeruginosa, S. marcescens, S. aureus, and C. albicans, respectively. This copeptide has properties resembling that of a broadspectrum antimicrobial agent. Biologically produced cationic antimicrobial peptides (CAPs) usually carry two or more types of HAA in their amino acid

64


Zhou et al.

Table 3. Hemolytic Activities of the Peptides (Human Red Blood Cells) peptide

P(K10F15)

P(K10F7.5L7.5)

HC50 (µg/mL)

16

16

sequences; this favors formation of energetically stable structures and an antimicrobial peptide function.12 The simple copeptides whose activities are shown in Figure 3A-C, however, lack the HAA diversity inherent in natural CAPs. As a step in the direction of exploring the effect of HAA diversity, we synthesized cationic antimicrobial peptides with two different species of hydrophobic residue. The species diversity was limited to two for simplicity and the mole ratio of the HAAs was set to unity. The two series tested were P(KAL) and P(KFL). As in the syntheses of copolymers with single HAAs, the target copolymer length was 25 residues, which was controlled through the mole ratio of NCA-monomer to initiator. Figure 3D shows the MIC results of the P(KAL) series with the total content of hydrophobic A and L varying from 0 to 100%. The P(KAL) copeptide has no antimicrobial activity against C. albicans and S. marcescens, like the P(KA) and P(KL) series (Figures 3A,B). Figure 3E shows that the P(KFL) copeptide exhibits the best antimicrobial activity among the five series with the lowest MICs across a broad spectrum of pathogens. For E. coli, P. aeruginosa, and S. aureus, the optimum range of hydrophobic F and L residues is 30-80% but with the inclusion of S. marcescens and C. albicans, the optimum concentration of F plus L is 60%. The average number of peptide repeat units is 25 (Table 1) and the average reside number is 10K, 7.5F, and 7.5L; this peptide (more precisely, this population of peptides) is hereafter denoted P(K10F7.5L7.5). It has MICs of 31, 31, 31, 250, and 62.5 µg/mL for E. coli, P. aeruginosa, S. aureus, S. marcescens, and C. albicans, respectively. Most natural antimicrobials are less than 50 amino acids in length (usually 10-30 residues long). Some synthetic short peptides (e.g., those having 9 or 11 residues45,46) have also shown excellent antibacterial activity. Of the three series of peptides with only one kind of hydrophobic amino residue, P(KF) appears to have a broader antimicrobial activity and we have therefore optimized the effect of length on antimicrobial activity using this copeptide. Figure 3F shows the MIC results versus synthesized peptide length over the range of 15 to 100 residues for P(KF) copeptides. We find that the P(KF) copeptide has broader antibacterial activity when it is 25 residues long than at shorter or longer length (it is, of course, possible that the precise minimum may lie between our sampled lengths, particularly in the case of S. marcescens). This most effective length for broad antimicrobial activity agrees well with the length range of naturally occurring cationic antibacterial peptides.12,47,48 It is possible that this length range is most ideal for conformational stability and disruption of lipid bilayers in pathogen cell walls. The hemolytic activities of two optimum copeptides (i.e., P(K10F15) and P(K10F7.5L7.5) were determined using human erythrocytes (Table 3). The HC50 of these two copeptides are both 16 µg/mL, which is slightly better than naturally derived melittin (with HC50 of 8 µg/mL), but our peptides are more hemolytic than other natural and synthetic peptides (e.g., Indolicidin with HC50 > 100 µg/mL).49 Our peptides were synthesized from two or three kinds of amino acid monomers using random copolymerization of NCA monomers, so the amino acid sequences vary with different peptide chains; some

chains will have higher than average HAA content. Peptide chains having a higher HAA content have been shown to be more hemolytic to human red blood cells.20,50 The influence of lipids which mimic bacterial membranes, on peptide secondary structure was studied. CD analysis of the P(KFL) and P(KF) peptide series in the presence of 50 µM POPC showed that only 30 and 40% HAA P(KF) peptides exhibit random coil structures as evidenced from the characteristic peaks obtained (195 nm trough and 215 nm peak), while the rest of the peptides did not show any distinct secondary structures (Figure 4). No distinct R-helix, β-sheet, or random coil structures were evident in all the KFL peptides. This result is not unexpected considering the short peptide lengths of the P(KFL) and P(KF) series and, hence, a lesser tendency to form distinct secondary structures for thermodynamic stability. Nonetheless, it is clear that the lack of a distinct secondary structure in our P(KFL) and P(KF) peptide series did not compromise the peptides’ antimicrobial activity. The cytoplasmic membrane depolarization activities of the optimum peptide, P(K10F7.5L7.5), on S. aureus (a Gram-positive bacteria) and C. albicans (a fungi) were determined by a fluorimetric method using the membrane potential-sensitive dye DiSC3(5). P(K25) (0%HAA) was used as the control peptide. The distribution of DiSC3(5) between the medium and pathogen cells is influenced by the cytoplasmic membrane potential. The cationic DiSC3(5) dye aggregates within the cytoplasmic membrane, leading to self-quenching of the fluorescence. However, upon membrane lysis, the dye is released and dissociates into the medium, resulting in fluorescence increase. Figure 5 shows the time course of fluorescence over 400s. Addition of P(K10F7.5L7.5) (10 µg/mL) to the buffer resulted in an immediate increase in fluorescence, while no significant change was observed when P(K25) peptide was added. It is clear that the peptide P(K10F7.5L7.5) is fast-acting on both bacterial and fungal membranes, achieving maximum fluorescence immediately upon addition of peptide to the buffer. This observation also suggests that the rates of peptide association with S. aureus and C. albicans membranes leading to depolarization were comparable.51 Figure 6 shows SEM images of P(K10F7.5L7.5)-treated and untreated E. coli, S. aureus, P. aeruginosa, C. albicans, and F. solani. The treated pathogens exhibit obvious morphological changes compared with the untreated controls. The surfaces of untreated bacteria and fungi appear smooth and rounded, whereas the peptide-treated bacteria and fungi exhibit wrinkled and withered surfaces.

4. Discussion We have shown for the first time that NCA ring-opening copolymerization can be used for synthesis of copeptides that exhibit broad spectrum antimicrobial activity against a wide range of clinically important pathogens, including both Gramnegative and Gram-positive bacteria and fungi. Our optimum copeptides via NCA ring-opening polymerization are P(K10F15) and P(K10F7.5L7.5) and the MICs of these two optimum copeptides were lower than those of many naturally occurring AMPs (such as LL-37, Indolicidin, Magainin I, and defensin). Of potentially great clinical importance, the P(K10F15) and P(K10F7.5L7.5) copeptides are effective, even at low MIC values (i.e., < 100 µg/mL), against antibiotic-resistant S. aureus and P. aeruginosa.1 Our data suggest that the mechanism by which our AMPs kill pathogens includes membrane disruption. Therefore, the



65

Figure 4. CD spectra of copolymers in the presence of 50 µM POPC lipid with varying hydrophobic monomer content: (a) P(KFL) and (b) P(KF), where the hydrophobic monomers are (a) F and L (1:1) and (b) F.

Figure 5. Cytoplasmic membrane depolarization of (A) S. aureus and (B) C. albicans by the copeptide (I) P(K25) (0%HAA) and (II) P(K10F7.5L7.5).

likelihood of emergence of copeptide-resistant pathogenic strains should be low as cell membranes, being a fundamental structure of the organism, are not expected to evolve under selective pressure over a short period of time. From the antimicrobial peptide database (APD),52 it has been observed that lysine is a typical hydrophilic amino acid residue of natural cationic AMPs; this is the hydrophilic amino acid used in our polymer series. However, our two optimum copeptides have higher HAA content than that of different natural cationic AMPs, which range from 41 to 49%. The ringopening polymerization process does not precisely control either

the lengths or sequences of the synthesized peptides, so that a mixture of lengths and compositions inevitably results from this method. Even the highly potent 60% HAA samples (i.e., (K10F15) and P(K10F7.5L7.5)) contain a significant number of low-activity peptides. An estimate of the variance in the number of hydrophobic residues per peptide (binomial distribution, n ) 25, p ) 0.6) gives a standard deviation of the HAA% of about 10%, with about 31% of peptides outside the range of 50-70% hydrophobic residues and only 16% of the sample having precisely 60% (i.e., 15) hydrophobic residues. The actual diversity in a synthesized sample would be somewhat greater

66


Figure 6. SEM photographs of treated and untreated bacteria and fungi with copeptide P(K10F7.5L7.5).

because there is also some variation in peptide length. The diversity can possibly explain our higher optimal HAA content of 60% compared to those reported for other AMPs (41-49%). Despite the variance, the antimicrobial activities of our two optimal peptides are still surprisingly excellent. Although the efficacies of our peptides against clinically important and resistant pathogens are high, they have the disadvantage of high hemolytic activity, which is an obstacle for their use as therapeutic agents. However, high hemolytic activity will not impact their use in other applications such as disinfectants or as antimicrobial surfaces. Others have shown that the mechanism of peptide binding to cell membranes that leads eventually to pathogen death involves a combination of electrostatic and hydrophobic interactions.53,54 The different compositions of bacterial and mammalian cell membranes allow future peptide composition fine-tuning to reduce the relative

Zhou et al.

toxicity to mammalian cells versus pathogens. The negative charge on bacterial cell surfaces is attributed to the presence of hydroxylated phospholipids such as phosphatidylglycerol, cardiolipin, and phosphatidylserine.55 On the other hand, the outer leaflet of mammalian cell membranes is rich in zwitterionic phospholipids such as phosphatidylethanolamine, phosphatidylcholine, and sphingomyelin, thus rendering mammalian cytoplasmic membranes neutral in net charge.28 The binding mechanisms of peptides with anionic bacterial lipid bilayer and zwitterionic mammalian lipid bilayer are therefore expected to be different, although other factors such as cell energetics and cell structure orientation could also readily influence peptide specificity. Increase of either the cationic charge or the hydrophobicity of an AMP can help its binding to the anionic bacterial lipid bilayers to improve the MICs.20 However, considering that mammalian cells are electrically neutral, increasing peptide hydrophobicity instead of cationic charge will increase the propensity of AMP binding to mammalian cell membranes. In this screening of NCA-polymerized copeptides, P(K10F15) and P(K10F7.5L7.5) have the optimum compositions of cationic and hydrophobic amino acids that kill bacteria effectively, but their high hydrophobicity (60%) or more hydrophobic species present may have resulted in high toxicity to mammalian red blood cells.56 Conversely, increased cationic charge and decreased hydrophobicity of peptides are expected to increase selectivity. Ring-opening copolymerization of different NCA monomers or other types of monomers afford many possibilities for controlling the composition, charge, and hydrophobicity of these peptides or peptide-polymer conjugates to generate AMPs with superior antimicrobial activities and acceptable toxicities to mammalian cells. We have demonstrated in this preliminary study that we can synthesize highly effectively broad-spectrum antimicrobial copeptides easily and inexpensively by the ring-opening of NCA method, which will make these peptides widely applicable. Naturally derived AMPs typically have approximately 30 residues and are expensive to synthesize by solid-phase sequential synthesis techniques which can cost 100000 dollars/gram.36,37 In contrast, medium/large scale production of our peptides would cost less than 10 dollars/gram and kilogram quantities can be easily produced. The production cost of these peptides is so low that their contribution to the cost of therapeutic agents, disinfectants, and antimicrobial surface coatings would be almost negligible. Future research for a simple polymerization method that could more precisely control the HAA content might further lower the MICs and toxicities of these peptides.

5. Conclusion Ring-opening polymerization of NCA monomers has been successfully applied to synthesize five series of copeptides (i.e., P(KA), P(KL), P(KF), P(KAL), and P(KFL)) with 0-100% hydrophobic amino acids. We found that the P(KF) copeptides have broader antimicrobial activity and are more effective than P(KL) and P(KA) series. Similarly, the P(KFL) series was more effective than the P(KAL) series. The optimum peptides against all five pathogens appear to be P(K10F7.5L7.5) and P(K10F15) and the MICs against E. coli, P. aeruginosa, S. marcescens, S. aureus, and C. albicans are 31, 31, 250, 31, and 62.5 µg/mL and 31, 31, 250, 31, and 125 µg/mL, respectively, which are lower than values of many naturally occurring AMPs. The membrane depolarization and SEM assays confirm the mechanism of microbe killing by P(K10F7.5L7.5) copeptide to include membrane disruption, which is unlikely to lead to microbe


resistance in clinical use. The low cost and large volume synthesis potential of the NCA synthetic route as well as the potential to produce highly effective antimicrobial polymers that can kill a number of highly resistant and clinically significant bacteria may lead to a wide future application of these peptides. Acknowledgment. This work was supported by funding from Menicon Company (Japan) and Nanyang Technological University (Singapore). Supporting Information Available. GPC spectra, 1H NMR spectrum, and composition analyses of the synthesized peptides. This material is available free of charge via the Internet at http:// pubs.acs.org.

References and Notes (1) Zhang, L.; Parente, J.; Harris, S. M.; Woods, D. E.; Hancock, R. E.; Falla, T. J. Antimicrob. Agents Chemother. 2005, 49 (7), 2921–2927. (2) Hancock, R. E.; Scott, M. G. Proc. Natl. Acad. Sci. U.S.A. 2000, 97 (16), 8856–8861. (3) Hoffmann, J. A.; Kafatos, F. C.; Janeway, C. A.; Ezekowitz, R. A. Science 1999, 284 (5418), 1313–1318. (4) Baroni, A.; Donnarumma, G.; Paoletti, I.; Longanesi-Cattani, I.; Bifulco, K.; Tufano, M. A.; Carriero, M. V. Peptides 2009, 30 (2), 267–272. (5) Schauber, J.; Gallo, R. L. J. Allergy Clin. Immunol. 2008, 122 (2), 261–266. (6) Epand, R. F.; Umezawa, N.; Porter, E. A.; Gellman, S. H.; Epand, R. M. Eur. J. Biochem. 2003, 270 (6), 1240–1248. (7) Gordon, Y. J.; Romanowski, E. G.; McDermott, A. M. Curr. Eye Res. 2005, 30 (7), 505–515. (8) Tang, Y. L.; Shi, Y. H.; Zhao, W.; Hao, G.; Le, G. W. J. Pharm. Biomed. Anal. 2008, 48 (4), 1187–1194. (9) Pistolesi, S.; Pogni, R.; Feix, J. B. Biophys. J. 2007, 93 (5), 1651– 1660. (10) Lam, K. L.; Ishitsuka, Y.; Cheng, Y.; Chien, K.; Waring, A. J.; Lehrer, R. I.; Lee, K. Y. J. Phys. Chem. B 2006, 110 (42), 21282–21286. (11) Kristiansen, P. E.; Fimland, G.; Mantzilas, D.; Nissen-Meyer, J. J. Biol. Chem. 2005, 280 (24), 22945–22950. (12) Zasloff, M. Nature 2002, 415, 389–395. (13) Hancock, R. E.; Sahl, H. G. Nat. Biotechnol. 2006, 24, 1551–1557. (14) Kuroda, K.; Caputo, G. A.; DeGrado, W. F. Chem.sEur. J. 2009, 15 (5), 1123–1133. (15) Choi, S.; Isaacs, A.; Clements, D.; Liu, D. H.; Kim, H.; Scott, R. W.; Winkler, J. D.; DeGrado, W. F. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (17), 6968–6973. (16) Kuroda, K.; DeGrado, W. F. J. Am. Chem. Soc. 2005, 127, 4128– 4129. (17) Ilker, M. F.; Nusslein, K.; Tew, G. N.; Coughlin, E. B. J. Am. Chem. Soc. 2004, 126, 15870–15875. (18) Liu, D. H.; Choi, S.; Chen, B.; Doerksen, R. J.; Clements, D. J.; Winkler, J. D.; Klein, M. L. Angew. Chem., Int. Ed. 2004, 43, 1158– 1162. (19) Sambhy, V.; Peterson, B. R.; Sen, A. Angew. Chem., Int. Ed. 2008, 47 (7), 1250–1254. (20) Al-Badri, Z. M.; Som, A.; Lyon, S.; Nelson, C. F.; Nusslein, K.; Tew, G. N. Biomacromolecules 2008, 9, 2805–2810. (21) Tew, G. N.; Liu, D.; Chen, B.; Doerksen, R. J.; Kaplan, J.; Carroll, P. J.; Klein, M. L.; DeGrado, W. F. Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (8), 5110–5114. (22) Mowery, B. P.; Lee, S. E.; Kissounko, D. A.; Epand, R. F.; Epand, R. M.; Weisblum, B.; Stahl, S. S.; Gellman, S. H. J. Am. Chem. Soc. 2007, 129, 15474–15476.


67

(23) Kawabata, N.; Nishiguchi, M. Appl. EnViron. Microbiol. 1988, 54, 2532–2535. (24) Ikeda, T.; Tazuke, S. Makromol. Chem., Rapid Commun. 1983, 4, 459–461. (25) Li, G.; Shen, J.; Zhu, Y. J. Appl. Polym. Sci. 1998, 67, 1761–1768. (26) Haldar, J.; Weight, A. K.; Klibanov, A. M. Nat. Protoc. 2007, 2 (10), 2412–2417. (27) Mukherjee, K.; Rivera, J. J.; Klibanov, A. M. Appl. Biochem. Biotechnol. 2008, 151 (1), 61–70. (28) Gabriel, G. J.; Som, A.; Madkour, A. E.; Eren, T.; Tew, G. N. Mater. Sci. Eng. R 2007, 57, 28–64. (29) Radzishevsky, I. S.; Rotem, S.; Bourdetsky, D.; Navon-Venezia, S.; Carmeli, Y.; Mor, A. Nat. Biotechnol. 2007, 25 (6), 657–659. (30) Lienkamp, K.; Madkour, A. E.; Musante, A.; Nelson, C. F.; Nusslein, K.; Tew, G. N. J. Am. Chem. Soc. 2008, 130, 9836–9843. (31) Brogden, K. A. Nat. ReV. Microbiol. 2005, 3 (3), 238–250. (32) Beisswenger, C.; Bals, R. Curr. Protein Pept. Sci. 2005, 6 (3), 255– 264. (33) Hancock, R. E.; Lehrer, R. Trends Biotechnol. 1998, 16 (2), 82–88. (34) Poon, K. W.; Liang, N.; Datta, A. Nucleosides, Nucleotides Nucleic Acids 2008, 27 (4), 389–407. (35) Arnusch, C. J.; Branderhorst, H.; de Kruijff, B.; Liskamp, R. M.; Breukink, E.; Pieters, R. J. Biochemistry 2007, 46 (46), 13437–13442. (36) Ding, Y.; Qin, C.; Guo, Z.; Niu, W.; Zhang, R.; Li, Y. Chem. BiodiVersity 2007, 4 (12), 2827–2834. (37) Haynes, S. R.; Hagins, S. D.; Juban, M. M.; Elzer, P. H.; Hammer, R. P. J. Pept. Res. 2005, 66 (6), 333–347. (38) Deming, T. J. Nature 1997, 390 (6658), 386–389. (39) Deming, T. J. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 3011– 3018. (40) Wyrsta, M. D.; Cogen, A. L.; Deming, T. J. J. Am. Chem. Soc. 2001, 123 (51), 12919–12920. (41) Clinical and Laboratory Standards; Clinical and Laboratory Standards Institute: Wayne, PA, 2005; Vol 188, p 328. (42) Lass-Florl, C.; Mayr, A.; Perkhofer, S.; Hinterberger, G.; Hausdorfer, J.; Speth, C.; File, M. Antimicrob. Agents Chemother. 2008, 52, 3637– 3641. (43) Heyn, K.; Tredup, A.; Salvetmoser, S.; Muller, F. M. Antimicrob. Agents Chemother. 2005, 49, 5157–5159. (44) Friedrich, C. L.; Moyles, D.; Beveridge, T. J.; Hancock, R. E. Antimicrob. Agents Chemother. 2000, 44 (8), 2086–2092. (45) Ferre, R.; Badosa, E.; Feliu, L.; Planas, M.; Montesinos, E.; Bardaji, E. Appl. EnViron. Microbiol. 2006, 72 (5), 3302–3308. (46) Fasman, G. D. Prediction of protein structure and the principles of protein conformation; Plenum: New York, 1989. (47) Jenssen, H.; Hamill, P.; Hancock, R. E. Clin. Microbiol. ReV. 2006, 19 (3), 491–511. (48) Brogden, K. A.; Guthmiller, J. M.; Salzet, M.; Zasloff, M. Nat. Immunol. 2005, 6 (6), 558–564. (49) Padmaja, J.; Satyanarayana, V.; Merrifield, R. B. J. Am. Chem. Soc. 1996, 118, 8989–8997. (50) Vinardell, M. P.; Infante, M. R. Comp. Biochem. Physiol., Part C: Toxicol. Pharmacol. 1999, 124, 117–120. (51) Park, K. H.; Park, Y.; Park, I. S.; Hahm, K. S.; Shin, S. Y. J. Pept. Sci. 2008, 14 (7), 876–882. (52) Wang, Z.; Wang, G. APD: the Antimicrobial Peptide Database. Nucleic Acids Res. 2004, 32 (Database issue), D590–D592. (53) Shai, Y. Biochim. Biophys. Acta, Biomembr. 1999, 1462, 55–70. (54) Brogden, K. A. Nature ReV. Microbiol. 2005, 3, 238–250. (55) Yeaman, M. R.; Yount, N. Y. Pharmacol. ReV. 2003, 55, 27–54. (56) Colak, S.; Nelson, C. F.; Nusslein, K.; Tew, G. N. Biomacromolecules 2009, 10, 353–359.

BM900896H

High Potency and Broad-Spectrum Antimicrobial Peptides

Recommend Documents