To understand the structure–activity relationship of chensinin-1, a anti-microbial peptide (AMP) with an unusual structure, and to develop novel AMPs as therapeutic agents.
To understand the structure–activity relationship of chensinin-1, a anti-microbial peptide (AMP) with an unusual structure, and to develop novel AMPs as therapeutic agents.
A series of chensinin-1 analogues were designed and synthesized by one to three replacement of glycines with leucines at the hydrophilic face of chensinin-1 or rearrangement of some of the residues in its sequence. Circular dichroism spectroscopy showed that the analogues adopted α-helical-type conformations in 50% trifluoroethanol/water but adopted β-strand-type conformations in 30 mmol l−1 sodium dodecyl sulphate. The anti-microbial activities of the peptides against Gram-positive bacteria increased 5- to 30-fold, and these increases paralleled the increases in the peptides' hydrophobicities. Their haemolytic activities also increased. Amphipathicities had little influence on the bactericidal activity of chensinin-1. All peptides caused leakage of calcein entrapped in negatively charged liposomes although with different efficiencies. The peptides did not induce leakage of calcein from uncharged liposomes.
Peptide adopted an aperiodic structure can improve the anti-microbial potency by increasing peptide hydrophobicity. Its target is bacteria plasma membrane.
Chensinin-1 can act as a new lead molecule for the study of AMPs with atypical structures.
The widespread use and misuse of antibiotics has resulted in the emergence of many multidrug-resistant micro-organisms. Therefore, antibiotic resistance is an increasingly serious problem, and the development of new types of antibiotics is urgently needed. Anti-microbial peptides (AMPs) have become important candidates as potential therapeutic agents. Cationic AMPs are especially important because they have many desirable features and would form a new class of antibiotics. AMPs increase bacterial membrane permeability, which causes leakage of cytoplasmic components and, consequently, micro-organism death (Sansom 1991; Matsuzaki et al. 1996; Shai 1999). Evolution of pathogen resistance to AMPs may therefore be a slow process caused by a fundamental change in membrane composition. The cationicity, hydrophobicity, amphipathicity (the segregation of hydrophilic and hydrophobic residues in an amphipathic α-helical AMP) and the secondary structure of many cationic AMPs are believed to be important for anti-microbial activity (Lequin et al. 2006; Chou et al. 2008; Jiang et al. 2011). The anti-microbial activity of AMPs seems to stem from their electrostatic interactions with the charged head groups of membrane lipids, whereas their hydrophobicities are apparently important for insertion into bacterial membranes (Yeaman and Yount 2003). However, AMP activities may also be influenced by other structural properties, including the positions of charged residues and the relative sizes of hydrophobic and hydrophilic regions (Leptihn et al. 2010). Many studies have used only linear AMPs with amphipathic α-helical conformations or amphipathic β-sheet peptides stabilized by multiple disulphides and compared them with derivatives having different net positive charges and hydrophobicities (Lequin et al. 2006; Chen et al. 2007a; Chi et al. 2007; Chou et al. 2008). Furthermore, only a few amphipathic β-sheeted AMPs containing a single disulphide have been used in structure–activity studies (Rozek et al. 2000; Szabo et al. 2010).
Chensinin-1 (SAVGRHGRRFGLRKHRKH) is a naturally occurring 18-residue peptide from the skin secretions of Rana chensinensis (Shang et al. 2009). It contains no negatively charged residues and +7 or +10 positively charged residues under physiological or acidic conditions, respectively. Unlike other known amphibian AMPs, chensinin-1 has a prevalence of three histidine residues that may lead to a dramatic change in anti-microbial activity at different pH values (Kan et al. 2003; Li et al. 2010). Furthermore, the N-terminal sequence (S-A-V) of chensinin-1 differs from that of other relatively short AMPs (20–24 residues) from Ranidae (Conlon et al. 2004). Thus, chensinin-1 is distinctly different from other known AMPs, including the ‘Rana box’-containing brevinin family of peptides. Notably, chensinin-1 mainly adopts an aperiodic structure in a membrane-mimetic environment and may be a novel lead molecule for the design of AMPs with possible therapeutic applications. Importantly, chensinin-1 has no haemolysis and is hopeful to be used as a novel antibiotics. In the study reported herein, we systematically designed and synthesized a series of chensinin-1 analogues with increased amphipathicity and/or hydrophobicity, while attempting to maintain the low haemolytic activity of chensinin-1, and compared their activities to determine which of their properties are responsible for the bactericidal activity, low haemolytic activity and selectivity toward Gram-positive bacterial membranes found for chensinin-1.
The bacterial strains Escherichia coli (AS 1.349), Pseudomonas aeruginosa (CGMCC 1.860), Enterobacter aerogenes (CGMCC 1.876), Enterobacter cloacae (CGMCC 1.58), Klebsiella pneumoniae subsp. Pneumoniae (CGMCC 1.176), Staphylococcus aureus (AS 1.72), Bacillus cereus (AS 1.126), Streptococcus lactis (AS 1.1690), Enterococcus faecalis (CGMCC 1.595) and Enterococcus faecium (CGMCC 1.2334) were acquired from the China General Microbiological Culture Collection Centre. The bacterial strains Proteus mirabilis (CICC22931) and Acinetobacter baumannii (CICC22934) were acquired from the China Center of Industrial Culture Collection.
Peptides were synthesized in crude form using standard Fmoc solid-phase peptide synthesis protocols (GL Biochem (Shanghai) Ltd, Shanghai, China) and were purified to near homogeneity (>95%) using reverse-phase (RP) HPLC through a Vydac 218TP1022 C-18 column (2·2 cm × 25 cm; Separations Group, Hesperia, CA, USA) equilibrated with acetonitrile/water/trifluoroacetic acid. The relative masses of the peptides were obtained using HPLC/enhanced product ion scan mass spectrometry and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Shimadzu, Japan).
The effective hydrophobicity of each peptide was assessed by RP-HPLC (Chen et al. 2007a; Conlon et al. 2009). Each peptide was chromatographed through an RP-HPLC Vydac 218TP54 C-18 column (0·46 cm × 25 cm) equilibrated with acetonitrile/water/trifluoroacetic acid (10·0/89·9/0·1, v/v/v). The concentration of acetonitrile was linearly increased to 56·0% (v/v) over 60 min at 25°C. The flow rate was 0·7 ml min−1. The A214 nm was measured for each fraction. The mean hydrophobicity value (H) for each peptide was calculated using the following values: Trp, 32·4; Phe, 29·1; Leu, 23·3; Ile, 21·4; Met, 15·7; Tyr, 14·7; Val, 13·4; Pro, 9·0; Cys, 7·6; Lys, 2·8; Glu, 2·8; Ala, 2·8; Thr, 2·3; Asp, 1·6; Arg, 0·6; Gln, 0·6; His, 0; Ser, 0; Gly, 0; Asn, −0·6 (Mant et al. 2009). The amphipathicity value (MH) for each peptide was calculated using its hydrophobic moment (Eisenberg et al. 1982) and Jemboss version 1.5 (Carver and Bleasby 2003).
Circular dichroism (CD) spectrum of each peptide was recorded using a J-810 spectropolarimeter (JASCO, Victoria, BC, Canada) between 190 and 250 nm at 0·1-nm intervals and 25°C with a scan rate of 20 nm min−1. Each peptide (0·3 mg ml−1) was dissolved in water, or 30 mmol l−1 aqueous sodium dodecyl sulphate (SDS), or 50% (v/v) trifluoroethanol (TFE)/water. A quartz cuvette with a 1-mm path length was used. Three consecutive scans were averaged, and the CD spectrum of the appropriate solvent was subtracted from the corresponding peptide spectrum. The program CDNN was used to calculate the percentage of secondary structural features.
The minimum inhibitor concentration (MIC) of each peptide was measured using a standard micro-dilution method. Briefly, serially diluted solutions (50 μl) of each peptide were individually added into a well of a 96-well microtitre plate. Then, 50 μl of an inoculum (106 CFU ml−1) of a log-phase bacterial culture was added. The A600 nm was recorded using a microtitre plate reader after the cultures had been incubated for 18–24 h at 37°C. The MIC was defined as the smallest peptide concentration that completely inhibited bacterial growth. To assess the precision of the assay, parallel incubations in the presence of gentamycin sulphate were performed.
The haemolytic activity of each peptide was determined as reported in (Lequin et al. 2006). Briefly, 2 × 107 erythrocytes from the blood of healthy humans were washed three times with 0·9% (w/v) NaCl and then incubated with serial dilutions of each peptide in 0·9% (w/v) NaCl for 3 h at 37°C. After centrifugation, the A545 nm of each resuspended pellet solution was measured. The positive control was an erythrocyte suspension incubated in water (100% haemolysis), and the negative control was an erythrocyte suspension incubated in 0·9% (w/v) NaCl (0% haemolysis). The HC50 was defined as the mean peptide concentration from three independent experiments that caused haemolysis of 50% of the erythrocytes. We estimated the HC50 value for a peptide by linear extrapolation of the slope between 400 and 500 μmol l−1 to the x-axis intersect of the dose–response curve when the per cent lysis for a peptide was <50% at 500 μmol l−1.
Calcein-loaded liposomes were prepared according to (Ishibashi et al. 1999; Coccia et al. 2011; Epand et al. 2006) with modifications. Phosphatidylcholine (PC), phosphatidylglycerol (PG), phosphatidylethanolamine (PE), cholesterol and cardiolipin (CL) were purchased from Sigma (Shanghai, China). Three types of liposomes were prepared: the Staph. aureus type contained a PG/CL molar ratio of 3 : 1, the E. coli type contained a PG/CL/PE molar ratio of 2 : 1 : 7, and the human erythrocyte type contained a PC/cholesterol molar ratio of 10 : 1. The phospholipids were dissolved in chloroform at each of the aforementioned ratios. After vacuum evaporation and drying overnight, a dye solution (60 mmol l−1 calcein, 50 mmol l−1 Tris-HCl, 100 mmol l−1 NaCl, pH 7·4) was added to each dried sample. Each mixture was ultrasonicated and subjected to 10 cycles of freezing and thawing in liquid nitrogen until homogeneous, and then extruded 11 times through a 200-nm polycarbonate membrane. Unencapsulated calcein was removed by gel filtration through a Sephadex G-50 column and eluted using phosphate. The eluted calcein-loaded liposomes were diluted to a final lipid concentration of 70 μmol l−1. Calcein release from the liposomes after addition of chensinin-1 or an analogue (final concentration, 10 μmol l−1) was measured by monitoring the fluorescence intensity at an excitation wavelength of 490 nm and emission wavelength of 520 nm. To quantify 100% dye release, 10% (v/v) Triton X-100 in 20 μl Tris-HCl (pH 7·4) was added to dissolve the vesicles. The percentage of leaked dye was calculated as:
where F is the fluorescence intensity of a peptide liposome/calcein solution, and F0 and Ft are fluorescence intensities of the solvent without or with Triton X-100, respectively.
The H value for chensinin-1 is 4·29, meaning that its hydrophobicity is relatively low. To enhance the hydrophobicity of chensinin-1, three peptides were synthesized with Leu substituted for Gly11, or Gly11 and Gly7, or Gly4, Gly7, and Gly11. ANTHEPROT 4.3 (Institute of Biology and Chemistry of Proteins, Lyon, France) predicted that chensinin-1 can assume an α-helical conformation without its nonpolar and polar side chains being segregated [Fig. 1(1)]. To change the amphipathicity of chensinin-1, its sequence was rearranged. The resulting peptide chensinin-1a had considerable amphipathic character, with its hydrophilic/polar residues Arg, Lys, His and Ser on one face and its hydrophobic/nonpolar residues Leu, Val, Phe, Ala and Gly on the opposite face [Fig. 1(5)]. To enhance the hydrophobicity of chensinin-1a, the three glycines, which reside on the hydrophobic face, were substituted with leucines in the same manner as they were substituted in chensinin-1. ANTHEPROT 4.3 predicted that chensinin-1a and its analogues would have amphipathic α-helical conformations [Fig. 1(5–8)]. Table 1 lists the sequences of the peptides, their theoretical molecular weights, net charges, retention times, mean H values and amphipathicity values. All analogues are strongly cationic, with a net positive charge of +7 or +10 under physiological or acidic conditions, respectively. The MH values of the peptides varied from 0·809 (chensinin-1) to 1·031 (1a-G4LG7G11L).
|Peptides sequence||Molecular weight||Net charge||Mean hydrophobicity||Amphipathicity||RP-HPLC retention time|
|pH 7·4||pH 5·0|
The experimental hydrophobicities of the peptides were determined by measuring their RP-HPLC retention times, which varied between 14·9 and 24·0 min and reflected the calculated H values. The chensinin-1 and chensinin-1a analogues containing three leucines, 1-G4LG7LG11L and 1a-G4LG7LG11L, were the most hydrophobic of the peptides. Notably, chensinin-1 and chensinin-1a (and each corresponding pair of chensinin-1 and chensinin-1a analogues, i.e., those with the same number of leucine replacements) had the same H value because they have the same amino acid composition; however, the observed retention times differed for each chensinin-1 analogue and its corresponding chensinin-1a analogue. For each chensinin-1/chensinin-1a analogue pair, the retention time for the chensinin-1a analogue was less than that of its partner.
The secondary structures of the peptides in water, or in 30 mmol l−1 SDS, or in 50% (v/v) TFE/water, were assessed by CD spectroscopy. All peptides had an aperiodic structure in water (Fig. 2a) and had a slightly ordered structure in membrane-mimetic environments such as 50% TFE/water and in 30 mmol l−1 SDS (Fig. 2b or c). The calculated percentage of secondary structure elements for chensinin-1 and its analogues indicated that the peptides had different types of secondary structure when in 50% TFE/water and 30 mmol l−1 SDS. All analogues in TFE/water had a spectrum with double maxima at 208 and 222 nm, indicating that the peptides were partially α-helical. In 30 mmol l−1 SDS, each peptide appeared to contain a mixture of a β-sheet-type structure and an aperiodic-type structure. 1a-G4LG7LG11L and 1-G4LG7LG11L had the most defined structures among the peptides, being 66·3 and 61·0% α-helical in 50% TFE. 1a-G7L was the least structured peptide (15·3% α-helical). The α-helix content decreased in the order: 1-G4LG7LG11L, 1a-G4LG7LG11L, 1-G7LG11L, 1a-G7LG11L, 1-G7L, 1a-G7L, chensinin-1 and chensinin-1a. Notably, in 30 mmol l−1 SDS, all peptides adopted a β-sheet-like structure to the extent of ~46–60%, with the exception of 1a-G7L for which a β-turn-like structure (60·7%) was determined. The order for the percentage of β-sheet-like structures for the peptides in 30 mmol l−1 SDS was same as that for the helical content of the peptides in 50% TFE/water.
The MICs of the peptides for the bacteria are listed in Table 2. All peptides, except chensinin-1a, had substantially improved anti-microbial activity against the Gram-positive bacteria in comparison with chensinin-1, but they had no activity against the Gram-negative bacteria. Chensinin-1a exhibited slightly poorer activities against B. cereus and S. latics than did chensinin-1, possibly because of its lesser experimentally determined hydrophobicity. In comparison with chesinin-1 and chensinin-1a, their analogues displayed 7- to 27-fold and 5- to 30-fold greater anti-bacterial activities, respectively, against the Gram-positive bacteria. The substitution of Gly with Leu increases the hydrophobicity of a peptide. The anti-microbial activities of the peptides increased as their hydrophobicities increased. The triple mutants, 1-G4LG7LG11L and 1a-G4LG7LG11L, exhibited the greatest activity against the Gram-positive bacteria (the geometric means for the MICs for the Gram-positive bacteria were 2·81 and 2·66 μmol l−1, respectively). However, the amphipathicity had little influence on the anti-microbial activities of chensinin-1 and its analogues.
|Microorganism||MIC (μmol l−1)|
|HC50 (μmol l−1)||>500||452·55||452·37||448·97||>500||500·00||493·92||460·78|
The effects of peptide concentration on erythrocyte haemolysis are shown in Fig. 3 and Table 2. The haemolytic activities of the chensinin-1 analogues significantly increased (HC50 ranged from >500 to 450 μmol l−1) with increasing hydrophobicity and amphipathicity, although the concentration of each peptide that caused 50% haemolysis was always much greater than the concentration that completely inhibited bacterial viability. With the exception of chensinin-1a, the HC50 values of the peptides did not differ significantly.
The therapeutic index of an AMP is the ratio of the average HC50 to the MIC, and therefore, larger values indicate greater potency. Among the peptides, 1a-G4LG7LG11L and 1-G4LG7LG11L exhibited the greatest anti-microbial potencies against the Gram-positive bacteria—therapeutic index: 173 and 159, respectively (Table 2).
To investigate the membrane-permeabilizing abilities of chensinin-1 and its analogues, we measured the release of calcein from liposomes of different composition. The composition of the uncharged liposomes (10 : 1 (w/w) PC/cholesterol) mimicked that of the human erythrocyte membrane. The negatively charged liposomes (3 : 1 (w/w) PG/CL or 2 : 1 : 7 (w/w/w) PG/CL/PE) mimicked those of Gram-positive (Staph. aureus) and Gram-negative (E. coli) bacteria, respectively (Ishibashi et al. 1999; Chen et al. 2007b). Entrapped calcein leaked from both types of negatively charged liposomes when they were exposed to the peptides, although relatively large differences in efficiency were observed. Conversely, the peptides did not cause the uncharged liposomes to leak calcein (Fig. 4). The maximum percentage of calcein leakage was achieved within 10 min of exposure to the peptides; 1a-G4LG7LG11L induced 92% leakage from the Staph. aureus–type liposomes. The leucine-containing peptides caused the most leakage (relative extent of calcein leakage, 1-G4LG7LG11L > 1a-G7LG11L > 1-G7LG11L > 1-G7L > 1a-G7L), whereas chensinin-1 induced only 60% leakage (Fig. 4a). Quite a different pattern was found for calcein leakage from the negatively charged, E. coli-type liposomes; all liposomes leaked to the same extent (~50%). The addition of the membrane-disrupting agent Triton X-100 led to an additional 50% increase in calcein leakage (Fig. 4b).
With the number and type of multidrug-resistant bacteria increasing, there is an urgent need to develop new and more potent classes of antibiotics. AMPs may be valuable alternatives to classic antibiotics; in particular, cationic AMPs have many features that would make them a novel class of antibiotics. The type of AMP for this study should be able to assume an amphipathic helical conformation in bacterial membranes, such that their polar/charged residues are segregated on one side of the helix and their hydrophobic residues on the other side (Shai 1999; Lequin et al. 2006; Chou et al. 2008; Jiang et al. 2011). Chensinin-1 is a short, linear AMP and contains seven positively charged arginine residues and three positively charged histidine residues. Chensinin-1 adopts a structure, according to deconvolution of its far-UV CD spectrum, that is composed of 20% β-strand, 8% α-helix and 72% aperiodic conformation in membrane-mimetic environments such as 50% TFE/water or 30 mmol l−1 SDS solution. However, when the glycines in chensinin-1 are replaced by leucines, the CD spectra of the chensinin-1 analogues in 50% TFE/water are those typical for α-helices. Additionally, the α-helix content increases as the number of substituted leucines increases. 1a-G4LG7LG11L has the most organized structure with 66·3% α-helical content in 50% TFE/water, probably because it could adopt an uninterrupted hydrophobic surface that could stabilize its helical structure (Jiang et al. 2007). Notably, the conformations of the chensinin-1 analogues in 30 mmol l−1 SDS were determined to contain ~46–60% β-sheet conformations, with aperiodic conformations accounting for the remaining ~40–54%. 1a-G7L was the exception as its conformation was determined to be 60·7% β-turn and 39·3% aperiodic.
Hydrophobicity is an important factor that affects the anti-microbial activities of amphipathic α-helical AMPs. Increasing the hydrophobicity of the nonpolar face of amphipathic α-helical AMPs increases their anti-microbial activities (Chen et al. 2005, 2007a; Jiang et al. 2011). Notably, studies on magainin-2 analogues showed than an increase in the hydrophobicity had relatively little influence on their activities against Gram-negative bacteria, although it enhanced their activities against Gram-positive bacteria and increased haemolysis (Dathe et al. 1997; Wieprecht et al. 1997). Similarly, increasing the hydrophobicity of chensinin-1 resulted in a 5- to 30-fold increase in killing activity against Gram-positive bacteria, had no effect on activity against Gram-negative bacteria, and increased haemolytic activity. Notably, only at 500 μmol l−1, a concentration much greater than that needed to kill 100% of the bacteria, did the chensinin-1 analogues cause ~50% haemolysis. Several studies have also demonstrated that increasing amphipathicity has a relatively modest effect on anti-microbial and haemolytic activity in α-helical peptides (Matsuzaki et al. 1996; Giangaspero et al. 2001; Pàl et al. 2005). We rearranged the chensinin-1 sequence to create an amphipathic conformation, while retaining its charge and amino acid composition. But the observation that chensinin-1a and its analogues had little changes in their anti-microbial activities comparable with chensinin-1 and its analogues suggests that amphipathicity is not an important factor for bactericidal activity of chensinin-1.
The phospholipid composition of cell membranes differs greatly between Gram-negative bacteria, Gram-positive bacteria, and erythrocytes. Human erythrocyte membranes contain the most PC; for E. coli (Gram negative) membranes, the major lipid is PE, and for Staph. aureus (Gram positive) membranes, the anionic PG and CL are both present in substantial amounts (Chen et al. 2007b). One property of α-helical AMPs is their ability to permeabilize targeted plasma membranes (Rasul et al. 2010). Our calcein leakage study suggests that chensinin-1 and its analogues interact with negatively charged membranes, although large differences in the relative efficiencies were observed. Conversely, the peptides did not affect liposomes that mimicked the zwitterionic phospholipid membrane of erythrocytes. The relative abilities of these peptides to permeabilize liposomes are consistent with their biological and haemolytic activities. The Staph. aureus membrane-type liposomes were more sensitive to the peptides than were the E. coli membrane-type liposomes. Although chensinin-1 and its analogues had no killing activity against the Gram-negative bacteria, they induced the release of calcein from E. coli-type liposomes. The inability of chensinin-1 and its analogues to permeabilize Gram-negative bacteria may be related to the presence of specific cell-wall components found in those bacteria, for example, lipopolysaccharides. Although lipopolysaccharides are negatively charged and considered by some to be the first target of cationic AMPs, they may also be a barrier against a variety of host defence factors (Papo and Shai 2005). Our current results are consistent with our previous observations using chensinin-1 (Shang et al. 2012). In conclusion, the chensinin-1 analogues, despite their obvious sequence dissimilarities, basically share the same bactericidal mechanism among themselves, albeit with some small differences, for example, variable affinity for negatively charged lipids.
This work was supported by the National Natural Science Foundation of China (grant no. 30970352) and a Science Grant (2011E12SF031) from the Dalian Science and Technology Bureau. The authors thank Prof. Xuefang Zheng, Department of Chemistry, Dalian University, for help with CD spectroscopy.