SEARCH

SEARCH BY CITATION

Keywords:

  • cationic antimicrobial peptide;
  • alanine substitution;
  • Porphyromonas gingivalis;
  • giant unilamellar vesicle;
  • membrane-active peptide

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

The antimicrobial activity of analogs obtained by substituting arginine and lysine in CL(14-25), a cationic α-helical dodecapeptide, with alanine against Porphyromonas gingivalis, a periodontal pathogen, varied significantly depending on the number and position of cationic amino acids. The alanine-substituted analogs had no hemolytic activity, even at a concentration of 1 mM. The antimicrobial activities of CL(K20A) and CL(K20A, K25A) were 3.8-fold and 9.1-fold higher, respectively, than that of CL(14-25). The antimicrobial activity of CL(R15A) was slightly lower than that of CL(14-25), suggesting that arginine at position 15 is not essential but is important for the antimicrobial activity. The experiments in which the alanine-substituted analogs bearing the replacement of arginine at position 24 and/or lysine at position 25 were used showed that arginine at position 24 was crucial for the antimicrobial activity whenever lysine at position 25 was substituted with alanine. Helical wheel projections of the alanine-substituted analogs indicate that the hydrophobicity in the vicinity of leucine at position 16 and alanines at positions 18 and/or 21 increased by substituting lysine at positions 20 and 25 with alanine, respectively. The degrees of diSC3-5 release from P. gingivalis cells and disruption of GUVs induced by the alanine-substituted analogs with different positive charges were not closely related to their antimicrobial activities. The enhanced antimicrobial activities of the alanine-substituted analogs appear to be mainly attributable to the changes in properties such as hydrophobicity and amphipathic propensity due to alanine substitution and not to their extents of positive charge (cationicity). © 2013 Wiley Periodicals, Inc. Biopolymers (Pept Sci) 102: 58–68, 2014.


INTRODUCTION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

Among the antimicrobial peptides (AMPs), cationic AMPs have received increased attention as promising antibacterial agents. A large number of cationic AMPs, including LL37 in humans, cecropins and melittin in insects, magainin and buforin 2 in amphibians, and fowlicidins in chickens, have been identified and extensively characterized.[1-5] These cationic AMPs are known to be unlikely to evoke bacterial resistance.[2-4, 6] This advantage of cationic AMPs is attributable to their ability to disrupt bacterial membranes via nonspecific electrostatic interaction with membrane lipid components. For cationic AMPs to be useful as therapeutic agents against human pathogenic bacteria, they must have both high antimicrobial potency and the ability to selectively distinguish bacterial cells from mammalian cells on the basis of different membrane lipid compositions. Thus, AMPs displaying low or no hemolytic activity would be attractive candidates for use as therapeutic agents.[3, 4, 7-9]

We recently reported that a potent dodecapeptide, derived from a region (residues 14-25) near the N-terminus of cyanate lyase (CL, EC 4.3.99.1, GenBank ID: Os10g0471300) from rice (Oryza sativa L. japonica), inhibits the growth of Porphyromonas gingivalis,[10, 11] a Gram-negative, obligate anaerobic, and asaccharolytic black pigmented bacterium. This bacterium is known to be a major etiological agent in the onset and progression of chronic periodontitis.[12-14] This dodecapeptide (CL(14-25): RRLMAAKAESRK, Table 1) contains three arginine and two lysine residues that may be important for the antimicrobial activity against P. gingivalis cells.[11] Structural analysis using PEP-HOLD, an online resource for de novo peptide structure prediction (http://bioserv.rpbs.univ-paris-diderot.fr/PEP-FOLD/), indicated that CL(14-25) adopts an α-helical structure, as reported previously.[11] However, the α-helical structure of CL(14-25) was not distinctly observed by CD spectroscopy, most probably because the peptide is too short, consisting of only 12 amino acid residues. In the previous study,[11] we investigated the antimicrobial activity and mechanism of action of the parent CL(14-25) against P. gingivalis cells and its hemolytic activity against mammalian blood cells. CL(14-25), even at a concentration of 0.8 mM, had no hemolytic activity. CL(14-25) appears to exhibit antimicrobial activity through membrane disruption because the degrees of membrane-disrupting activities induced by CL(14-25) and its truncated CL peptide analogs were closely related to their antimicrobial activities. Microscopic image analysis also suggested that CL(14-25) disrupted giant unilamellar vesicles (GUVs) in a detergent-like manner. The antimicrobial activities of CL(15-25) and CL(14–24), which are CL(14-25) analogs truncated by removing a single cationic amino acid residue at the N- and C-termini, respectively, were ∼50% of those of CL(14-25), whereas CL(16–25) and CL(14–23), which are CL(14-25) analogs truncated by deleting two cationic amino acid residues at the N- and C-termini, respectively, possessed little or no antimicrobial activity. Therefore, we confirmed that one or both arginine residues at positions 15 and 24 are critical for the antimicrobial activity against P. gingivalis cells.[11] However, to completely comprehend the contribution of the cationic amino acids in CL(14-25) to its antimicrobial activity, further investigations using CL peptide analogs bearing substitutions with uncharged amino acids are required.

Table 1. The Amino Acid Sequences and Properties Used in This Study
No.PeptideSequenceMolecular WeightNet ChargeIC50 [mM]b
CalculatedMeasureda
  1. a

    The values indicate molecular mass of each purified peptide confirmed by MALDI-TOF analysis.

  2. b

    The values of IC50 indicate 50% growth-inhibitory concentration against P. gingivalis JCM 8525.

PCL(14-25)RRLMAAKAESRK1416.751416.72+40.145 ± 0.029
1CL(R14A)ARLMAAKAESRK1331.641331.60+30.104 ± 0.011
2CL(R15A)RALMAAKAESRK1331.641331.61+30.165 ± 0.038
3CL(K20A)RRLMAAAAESRK1359.661359.63+30.038 ± 0.006
4CL(R24A)RRLMAAKAESAK1331.641331.61+30.111 ± 0.004
5CL(K25A)RRLMAAKAESRA1359.661359.63+30.051 ± 0.016
6CL(R14A, R15A)AALMAAKAESRK1246.521246.51+20.089 ± 0.011
7CL(R14A, K20A)ARLMAAAAESRK1272.541274.52+20.055 ± 0.011
8CL(K20A, K25A)RRLMAAAAESRA1302.561302.53+20.016 ± 0.007
9CL(R14A, K25A)ARLMAAKAESRA1274.541274.52+20.078 ± 0.016
10CL(R24A, K25A)RRLMAAKAESAA1274.541274.52+2>0.525

The substitution of each amino acid of a parent peptide with an alanine (alanine scan) is useful to identify the amino acids that are crucial for the antimicrobial activity.[15-17] Alanine is chosen because it is considered to be the most neutral amino acid and a side chain of alanine is the least voluminous of all the amino acids. Furthermore, it is not highly hydrophobic, has no charge, and does not fit in any pocket. Alanine substitution has been frequently used for defining the structure−antimicrobial activity relationship of many cationic AMP, including arginine and lysine residues, as well as for designing analogs with improved antimicrobial activity and/or cell selectivity.[15-19] For example, the alanine substitution of a partial sequence (residues 20–31, CFQWQRNMRKYR) from human lactoferrin showed that the antimicrobial activity against Escherichia coli of a derivative obtained by replacing an arginine residue at position 28 was significantly reduced but the antimicrobial activities of derivatives obtained by replacing arginine residues at positions 25 and 31 remained unchanged.[18] The replacement of all 18 residues of apidaecin 1b (GNNRPVTIPQPRPPHPRL) individually by alanine (alanine scan) indicated that by replacing the arginine residue at position 17, the antimicrobial activity against E. coli and Micrococcus luteus was completely abolished and the substitution of arginine residues at positions 4 and 12 partially reduced the antimicrobial activity.[19] These results suggest that the same amino acid at two different positions may not have the same crucial role in the antimicrobial activity of the peptide.

In this study, we investigated the antimicrobial activity of the alanine-substituted analogs of CL(14-25) against P. gingivalis cells and their hemolytic activity against mammalian blood cells. To elucidate the contribution of the five cationic amino acid residues (three arginine and two lysine residues) in CL(14-25) to its antimicrobial activity, we synthesized 10 types of alanine-substituted CL peptide analogs (CL(R14A), CL(R15A), CL(K20A), CL(R24A), CL(K25A), CL(R14A, R15A), CL(R14A, K20A), CL(K20A, K25A), CL(R14A, K25K), and CL(R24A, K25A), Table 1) and examined their antimicrobial activity against P. gingivalis cells. To clarify the modes of action of the alanine-substituted derivatives of CL(14-25) on the bacterial membrane, we applied the two experimental systems[11]: (1) measurement of 3,3'-dipropylthiadicarbocyanine iodide (diSC3-5, a membrane potential-sensitive fluorescent dye) release from intact P. gingivalis cells that were previously loaded with diSC3-5 and (2) microscopic observation of the morphological changes in GUVs when the alanine-substituted derivatives of CL(14-25) were added. Finally, we discuss the importance of arginine and lysine residues of CL(14-25) for the interaction with bacterial membranes and the antimicrobial activity against P. gingivalis cells.

MATERIALS AND METHODS

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

Materials

Melittin (a component of bee venom known to disrupt the cell membrane) and diSC3-5 were purchased from Sigma–Aldrich Japan (Tokyo, Japan). Sterile defibrinated sheep blood was obtained from Cosmo Bio Co., Ltd. (Tokyo, Japan). 1,2-Dioleoyl-sn-glycero-3- phosphoethanolamine (DOPE) and 1,2-dioleoyl-sn-glycerol-3-phospho-(1′-rac-glycerol) (DOPG) sodium salt were supplied by Avanti Polar Lipids, Inc. (Alabaster, AL).

Peptides Used in This Study

The amino acid sequences and properties of CL peptides used in this study are summarized in Table 1. Chemically synthesized CL peptides, including CL(14-25), CL(R14A), CL(R15A), CL(K20A), CL(R24A), CL(K25A), CL(R14A, R15A), CL(R14A, K20A), CL(K20A, K25A), CL(R14A, K25A), and CL(R24A, K25A), were obtained from Hokkaido System Science Co., Ltd. (Sapporo, Japan). Synthetic peptides were purified to more than 95% purity by reversed-phase HPLC; molecular weights of the purified CL peptides were confirmed by MALDI TOF-MS.

Determination of Antimicrobial Activity

P. ginigivalis JCM 8525 was used as a test microorganism for estimating the antimicrobial activity of the CL peptides. The bacterial cells were precultivated overnight at 37°C in test tubes containing modified Gifu anaerobic medium (GAM, Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) under anaerobic conditions. The cells obtained were further cultivated under the same culture conditions to obtain mid-logarithmic phase cells. The antimicrobial activity of the CL peptides was measured using 96 well plates under anaerobic conditions in the presence of a deoxygenating reagent (AnaeroPack A03; Mitsubishi Gas Chemical Co., Tokyo, Japan), as reported previously.[11] Using the BacTiter-Glo™ reagent (Progema, USA) according to the manufacturer's instructions, the viable cell concentration was estimated as relative light units (RLUs), which depend on the level of ATP released from intact cells. The survival rate was calculated using the following equation:

  • display math

where LP and Lc represent RLUs of the samples treated with and without peptide, respectively, and L0 is RLU of medium alone.

The 50% growth-inhibitory concentration (IC50) for each peptide was determined using a plot of the survival rate versus log peptide concentration. The results were expressed as the mean ± SD of three individual experiments.

Hemolytic Activity

The hemolytic activity of the CL peptides was measured as reported previously.[11] In brief, fresh sheep red blood cells (sRBCs) were washed three times with PBS (35 mM phosphate buffer and 150 mM NaCl, pH 7.2) by centrifugation at 2000g for 5 min and resuspended in PBS.[20] Aliquots (50 μL) of peptide solutions (serial twofold dilutions with PBS) were added to 50 μL of 4% (v/v) sRBC suspension in a 96 well plate and incubated without agitation for 1 h at 37°C. The samples were centrifuged at 2000g for 5 min, and hemoglobin release was monitored by measuring the absorbance of the supernatant at 405 nm using the plate reader (2030 ARVO™ MX, PerkinElmer, USA). The percentage of hemolysis was calculated using the following formula:

  • display math

where Ap and AT represent the absorbance of sRBC sample treated with peptide and 0.1% (w/v) Triton X-100, respectively, and A0 is the absorbance of sample without the treatment. The results were expressed as the mean ± SD of three individual experiments.

Measurement of Cell Membrane Depolarization

The membrane depolarization activity of the CL peptides was determined using intact P. gingivalis cells and a membrane potential-sensitive fluorescent dye, diSC3-5.[21] In brief, EDTA-treated cells were resuspended in HEPES buffer with 0.1 mM EDTA and 4 μM dye, and the cell concentration was adjusted to 5 × 108 CFU/mL (OD650 = 0.05). The cell suspension was incubated at 37°C for 30 min under anaerobic conditions in the dark. During incubation, stable reduction in fluorescence was achieved, implying incorporation of the dye into the bacterial membrane. diSC3-5 release from P. gingivalis cells was monitored for 5 min after the addition of the peptide by measuring the fluorescence intensity at an excitation wavelength of 622 nm and an emission wavelength of 670 nm using a spectrofluorometer (FP-6200; JASCO, Tokyo, Japan), as reported previously.[11] The percentage of diSC3-5 release from P. gingivalis cells was calculated using the following equation:

  • display math

where Fp represents the fluorescence intensity of the cell suspension treated with peptide, Fm is the fluorescent intensity of the sample observed at 5 min after the addition of 20 μM melittin, and F0 is the fluorescence of an untreated sample. The results were expressed as the mean ± SD of three individual experiments.

Preparation of GUVs and Microscopic Observation of Their Morphological Changes

We prepared GUVs composed of DOPE/DOPG in a molar ratio of 7:3 according to the method of Saeki et al.[22] In brief, lipids were dissolved in a mixture of chloroform and methanol with a volume ratio of 2:1 along with 10 mM glucose in methanol. After drying under vacuum for 3 h, the film was rehydrated with 200 μL of 200 mM sucrose solution at 60°C for 2 h. The lipid solution was diluted to a lipid concentration of 200 μM using 200 mM sucrose solution.

The burst of GUVs by CL(14-25) and its alanine-substituted analogs was observed using an inverted phase contrast microscope (IX71, Olympus, Tokyo, Japan), as reported previously.[11] In brief, a GUV suspension (10 μL) was added to 290 μL of 200 mM sucrose solution in a microtube and then used to fill the space created by interposing a U-shaped silicon sheet, with all corners at right angles between a glass slide and cover slip. The peptide solution was then slowly added to the U-shaped space containing the GUV suspension using a microinjector (CellTram® Vario; Eppendorf Japan, Tokyo, Japan) and a handmade glass microcapillary. The peptide–lipid molar ratio was adjusted to 100:1. Phase contrast images of GUVs were recorded using a CCD video camera (Wat-120N+; Watec, Yamagata, Japan). Thirty GUVs were observed for 10 min, and the number of GUVs burst by the addition of CL(14-25) and its alanine-substituted analogs was measured. The percentage of burst GUVs was determined as a function of time.

RESULTS

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

Antimicrobial Activity

In our previous study,[11] we determined the growth–inhibitory activity of CL(14-25) against P. gingivalis JCM 8525. The antimicrobial activity of CL(14-25) increased with increasing peptide concentration, and the IC50 value was calculated to be 0.145 mM for P. gingivalis 8525 as reported previously (Table 1). In addition to CL(14-25) (the parental peptide), Figure 1 also shows the antimicrobial activity of single alanine-substituted CL peptide analogs against P. gingivalis cells. CL(R14A), CL(R15A), and CL(R24A), the analogs obtained by replacing arginine at positions 14, 15, and 24, respectively, and CL(K20A) and CL(K25A), the analogs obtained by replacing lysine at positions 20 and 25, respectively, also exhibited concentration-dependent antimicrobial activity against P. gingivalis cells. The obtained IC50 values are shown in Table 1. With the exception of CL(R15A), their antimicrobial activities were higher than that of CL(14-25). These results showed that each arginine residue at position 14 or 24 (other than 15) and each lysine residue at position 20 or 25 are not necessarily important for exerting antimicrobial activity against P. gingivalis cells. In contrast, the substitution of each cationic amino acid with alanine without charge resulted in the improved antimicrobial activity of the analogs. In particular, the antimicrobial activity of CL(K20A) was 3.8-fold higher than that of CL(14-25).

image

Figure 1. Relative antimicrobial activities of CL(14-25) and its single-alanine substituted CL peptide analogs against P. gingivalis. Symbols: CL(14-25), closed circles; CL(R14A), closed triangles; CL(R15A), open triangles; CL(K20A), open squares; CL(R24A), open diamonds; CL(K25A), closed diamonds.

Download figure to PowerPoint

Figure 2 shows the antimicrobial activity of two alanine-substituted CL peptide analogs against P. gingivalis cells. CL(R14A, R15A), the analog bearing substitution of arginine residues at positions 14 and 15; CL(K20A, K25A), the analog bearing substitution of lysine residues at positions 20 and 25; CL(R14A, K20A) and CL(R14A, K25A), the analogs bearing substitution of each lysine residue at positions 20 and 25, respectively, as well as an arginine residue at position 14; and CL(R24A, K25A), the analog bearing substitution of both an arginine residue at position 24 and a lysine residue at position 25, exhibited concentration-dependent antimicrobial activity against P. gingivalis cells. The IC50 values for CL(R14A, R15A), CL(R14A, K20A), CL(K20A, K25A), and CL(R14A, K25A) are shown in Table 1. Their antimicrobial activities were higher than that of CL(14-25). However, the IC50 value of CL(R24A, K25A) could not be measured. At the concentration of 0.4 mM, the rate of growth inhibition was 36% for CL(R24A, K25A), which was lower than that (78%) for CL(14-25). These results show that the arginine residue at position 24 of the analogs without a C-terminal lysine residue, i.e., CL(K20A, K25A) and CL(R14A, K25A), is essential for exerting the antimicrobial activities against P. gingivalis cells. In fact, of the CL peptide analogs obtained by substituting with two alanine residues, the four analogs with arginine at position 24 possessed higher antimicrobial activity than that of CL(14-25), although the IC50 values varied among them. In particular, the antimicrobial activity of CL(K20A, K25A) was 9.1-fold and 2.4-fold higher than that of CL(14-25) and CL(K20A), respectively.

image

Figure 2. Relative antimicrobial activities of CL(14-25) and its two alanine-substituted CL peptide analogs against P. gingivalis. Symbols: CL(14-25), closed circles; CL(R14A, R15A), closed diamonds; CL(R14A, K20A), closed triangles; CL(K20A, K25A), closed squares; CL(R14A, K25A), open triangles; CL(R24A, K25A), open diamonds.

Download figure to PowerPoint

Hemolytic Activity

The membrane selectivity of AMPs is very important because they will only be useful if they attack invaders without causing toxicity to host cells. The lipid composition of the mammalian cell membrane, which is rich in zwitterionic phosphatidylcholine and sphingomyelin phospholipids, is distinct from that of the bacterial cell membrane.[3, 4, 23, 24] To assess the cytotoxicity of CL(K20A) and CL(K20A, K25A), which displayed higher antimicrobial activity than the other alanine-substituted CL peptide analogs, to the mammalian cell membrane, we measured the hemolysis of sRBCs at various peptide concentrations. CL(14-25) had little or no hemolytic activity toward the mammalian RBCs, as reported previously.[11] Relative rates of hemolysis for CL(K20A) and CL(K20A, K25A), up to a concentration of 1 mM, were less than 1.0% of the rate observed after treating sRBCs with 0.1% Triton-X 100 (data not shown), although they have higher antimicrobial activity than CL(14-25), as described above. These results indicate that the alanine-substituted CL analogs have little or no hemolytic activities toward mammalian RBCs.

Membrane Depolarization of Intact P. gingivalis Cells

To examine whether the antimicrobial activity of alanine-substituted CL peptide analogs depends on their abilities to disrupt bacterial membranes, we measured their abilities to induce the release of a membrane potential-sensitive probe, diSC3-5, from P. gingivalis cells. diSC3-5 is known to distribute between bacterial cells and the surrounding medium depending on the cytoplasmic membrane potential gradient.[25, 26] During a 30-min incubation of a mixture of P. gingivalis cells and diSC3-5, a large portion of diSC3-5 was taken up and concentrated in the cytoplasmic membrane of P. gingivalis cells because of the membrane potential. Thus, diSC3-5 quenched its own fluorescence within 30 min, as reported previously.[11] The addition of each peptide caused membrane lesions, which led to the dissipation of the membrane potential and consequent diSC3-5 release from the membrane, resulting in the rapid increase of fluorescence. We have confirmed that the degree of diSC3-5 release from diSC3-5-containing P. gingivalis cells increased with an increase in CL(14-25) concentration, as reported previously.[11] In this study, the degree of diSC3-5 release by the addition of CL(14-25) at a concentration of 0.435 mM was 55% of that induced by the addition of melittin at a concentration of 20 μM. This result was almost similar to that (58%) obtained in the previous study.[11]

Figure 3 shows the time courses of the increase in diSC3-5 fluorescence intensity induced by the addition of CL(14-25) and its single alanine-substituted analogs to diSC3-5-containing P. gingivalis cells. When CL(K20A) and CL(K25A) were added, the amounts of diSC3-5 released from P. gingivalis cells were 41% and 44% of the amount of diSC3-5 released from the cells by the addition of melittin, respectively. When CL(R14A), CL(R15A), or CL(R24A) was added, the degree of diSC3-5 release from P. gingivalis cells was lower than that of the release induced by the addition of CL(K20A) or CL(K25A). The values are 20–28% of the release induced by the addition of melittin. Among the single alanine-substituted CL peptide analogs tested, CL(R15A) displayed the lowest ability to release diSC3-5 from P. gingivalis cells, a result similar to the ability to inhibit the growth of P. gingivalis, as described above. The results strongly support the assertion that the single alanine-substituted CL peptide analogs have a significant ability to disrupt the cytoplasmic membrane of P. gingivalis cells. However, although the single alanine-substituted CL peptide analogs, except for CL(R15A), exhibited higher antimicrobial activities against P. gingivalis cells than that of CL(14-25), their abilities to release diSC3-5 from P. gingivalis cells were lower than that of CL(14-25).

image

Figure 3. Time courses of diSC3-5 release from P. gingivalis cells induced by CL(14-25) and its single-alanine substituted CL peptide analogs. The concentration of the CL peptides added was 435 μM. The symbols are the same as those shown in Figure 1.

Download figure to PowerPoint

Figure 4 shows the time courses of the increase in diSC3-5 fluorescence intensity induced by the addition of two alanine-substituted CL peptide analogs to diSC3-5-containing P. gingivalis cells. The amounts of diSC3-5 released from P. gingivalis cells by the addition of the CL peptide analogs with a decreased positive charge reduced significantly and the values were in the range of 9.4–22% of the amount of diSC3-5 released from the cells by the addition of melittin. The CL peptide analogs bearing the substitution of arginine and/or lysine with two alanine residues, except for CL(R24A, K25A), exhibited higher antimicrobial activities against P. gingivalis than that of CL(14-25); however, their membrane depolarization activities reduced, probably due to a decrease in the net positive charge. These results on diSC3-5 released from P. gingivalis cells by addition of the CL peptide analogs suggest that the extent of positive charge (cationicity) of CL peptides would play the determining role in mediating its initial interaction with the cytoplasmic membrane of target cells.

image

Figure 4. Time courses of diSC3-5 release from P. gingivalis cells induced by CL(14-25) and its two alanine-substituted CL peptide analogs. The concentration of the CL peptides added was 435 μM. The symbols are the same as those shown in Figure 2.

Download figure to PowerPoint

Relationship between Antimicrobial Activity and Membrane-Disrupting Activity

We investigated the contribution of arginine and lysine residues in CL(14-25) to both the membrane-perturbing ability and antimicrobial activity against P. gingivalis cells. Thus, we evaluated the abilities of the alanine-substituted CL peptide analogs to depolarize the cell membranes of P. gingivalis compared with those of CL(14-25). Figure 5 shows the relationship between IC50 values for the single alanine-substituted analogs (Table 1 and Figure 1) and their membrane depolarization activities (Figure 3). The percentage of diSC3-5 release from P. gingivalis cells induced by the single alanine-substituted analogs gradually increased with a decrease in their IC50 values. Thus, the membrane depolarization activities of the single alanine-substituted analogs were closely related to their antimicrobial activities. Therefore, the antimicrobial activities of the alanine-substituted analogs appear to be mainly based on their membrane-disrupting activity in a manner similar to that of CL(14-25). However, although the antimicrobial activities of the single alanine-substituted analogs, except for CL(R15A), were higher than those of CL(14-25) (Figure 1), their membrane depolarization activities were lower than that of CL(14-25) (Figure 3). The results suggest that diSC3-5 release from P. gingivalis cells is greatly affected by cationicity of the CL peptide analogs added as described above.

image

Figure 5. The relationship between IC50 value and ability to promote diSC3-5 release from P. gingivalis cells for CL(14-25) and its single-alanine substituted CL peptide analogs. The symbols are the same as those shown in Figure 1.

Download figure to PowerPoint

Figure 6 shows the relationship between IC50 values for two alanine-substituted analogs (Table 1 and Figure 2) and their membrane depolarization activities (Figure 4). There was little or no difference in the percentage of diSC3-5 release from P. gingivalis cells among analogs bearing substitution with two alanine residues (CL(R14A, R15A), CL(R14A, K20A), CL(K20A, K25A), and CL(R14A, K25A)), which possessed the antimicrobial activity against P. gingivalis cells (Table 1 and Figure 2). Although the IC50 values were almost identical between CL(R14A, K20A) and CL(K25A) (Table 1), the membrane depolarization activity of CL(R14A, K20A) was one-fourth that of CL(K25A), strongly supporting that the cationicity of CL peptide analogs is highly effective in inducing membrane depolarization. These results suggest that the expression of antimicrobial activity also depends on other factors acting in a concerted manner, including helix formation and hydrophobic partitioning of the peptide into membrane.

image

Figure 6. The relationship between IC50 value and ability to promote diSC3-5 release from P. gingivalis cells for CL(14-25) and its two alanine-substituted CL peptide analogs. The symbols are the same as those shown in Figure 2.

Download figure to PowerPoint

Interaction with Membrane Mimetic GUVs

To obtain a deeper insight into the mechanistic details, we used GUVs composed of DOPE/DOPG in a molar ratio of 7:3, with diameters greater than 10 μm. These GUVs displayed high contrast in a phase contrast microscopy image because of the difference in concentrations between the inside (10 mM glucose) and outside (200 mM sucrose), as reported previously.[11] Figure 7 shows the time courses of the rate of burst GUVs induced by the addition of CL(14-25) and its alanine-substituted analogs including CL(K20A), CL(K20A, K25A), and CL(R24A, K25A). The rate of burst GUVs was 67% at 6 min after the addition of CL(14-25), whereas the rates of burst GUVs induced by the addition of CL(K20A) and CL(K20A, K25A) reached 100% and 84%, respectively, at the same time. However, no ability of CL(R24A, K25A) to disrupt GUVs was observed. The finding corresponds to the lowest antimicrobial activity of CL(R24A, K25A), as described above (Figure 2). The abilities of the alanine-substituted analogs to disrupt GUVs were almost related to their antimicrobial activity. However, the rate of burst GUVs induced by the addition of CL(K20A), which has lower antimicrobial activity than CL(K20A, K25A), remained higher compared with that induced by the addition of CL(K20A, K25A) during the treatment time of 10 min. The results suggest that the abilities of the analogs to disrupt GUVs were dependent on several properties such as cationicity, hydrophobic propensity, and amphipathicity, similar to their antimicrobial activities, as described above.

image

Figure 7. Time courses of percentage of burst GUVs induced by CL(14-25) and its CL peptide analogs obtained by replacing with a single and two alanine residues. Symbols: CL(14-25), closed circles; CL(K20A), open squares; CL(K20A, K25A), closed squares; CL(R24A, K25A), open diamonds.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

CL(14-25), a dodecapeptide, is a novel, short, and cationic α-helical AMP with a net charge of +4. CL(14-25) contains three arginine and two lysine residues (Table 1). Structure−function studies on AMPs indicate that a number of parameters modulate antimicrobial activity, including net positive charge, charge distribution, amphipathicity, and helical propensity.[2, 4, 27-33] Despite the diversity of amino acid sequences in AMPs, most have a high net positive charge resulting from multiple arginine and lysine residues.[2, 3, 7, 23, 34] These cationic properties are very important for binding to negatively charged surfaces of bacterial lipid membranes. On the other hand, the α-helical amphipathic structure is a key factor for facilitating insertion into the cell membrane, ultimately leading to cell death. One way of making AMPs more antimicrobially active is to enhance their positive charge by replacing neutral and acidic amino acids with cationic amino acids, arginine and lysine, thereby resulting in reinforcement of the electrostatic interaction between AMPs and cell membranes. The investigation on magainin 2 amide analogs with cationic charges ranging between +3 and +7 showed that enhancement of the peptide charge up to a threshold value of +5 optimized the antimicrobial activity.[35] The result indicates that the possibility of optimizing the antimicrobial effect via modification of only the peptide charge seems to be limited because modifying the peptide charge results in significant changes in one or more other parameters, including hydrophobicity, amphipathicity, and helical propensity for displaying the antimicrobial activity. A study on the distribution effect of cationic amino acids on the hydrophilic surface of the helical wheel, in which natural AMP-derived α-helical CRAMP18 (GEKLKKIGQKIKNFFQKL) and its analogs with the same net charge of +5 were used, indicated that antimicrobial activity is enhanced by the rearrangement of cationic amino acids to promote dispersed distribution.[36] This result suggests that positive charge distribution in the sequence of AMPs is an important factor for expressing the antimicrobial activity.

In this study, we synthesized 10 types of CL peptide analogs bearing substitution with alanine, an uncharged amino acid, and elucidated the effects of cationicity of CL(14-25) on its antimicrobial activity. The antimicrobial activities of the single alanine-substituted CL peptide analogs against P. gingivalis cells were compared with that of CL(14-25). CL(R15A), the analog bearing replacement of arginine at position 15 with alanine, exhibited concentration-dependent antimicrobial activity against P. gingivalis cells; however, its antimicrobial activity was slightly lower than that of CL(14-25). This decrease shows that the arginine residue at position 15 is important for exerting the antimicrobial activity. This result is identical to that obtained using the truncated CL peptide analogs reported previously.[11] However, the antimicrobial activity of CL(R14A, R15A), the analog bearing replacement of the two N-terminal arginine residues with two alanine residues, was higher than that of CL(14-25), suggesting that the arginine residue at position 15 is not critical for exerting the antimicrobial activity. Furthermore, CL(16–25), the analog missing the two N-terminal arginine residues, rarely exhibited antimicrobial activity as reported previously.[11] These results indicate that the peptide size as well as positive charge is responsible for exerting the antimicrobial activity, although the details remain unclear.

The antimicrobial activity of CL(R24A, K25A), the analog bearing replacement of the C-terminal arginine and lysine residues with two alanine residues, was drastically decreased compared with those of CL(14-25) and its other alanine-substituted analogs, in a manner similar to that of CL(14–23), the analog missing the C-terminal arginine and lysine residues reported previously.[11] Its IC50 value could not be measured. The reduced antimicrobial activity of CL(R24A, K25A) precisely corresponded to its ability to disrupt GUVs (Figure 7). However, CL(R24A), CL(K25A), CL(K20A, K25A), and CL(R14A, K25A), the analogs obtained by replacing either arginine at position 24 or lysine at position 25, exhibited higher antimicrobial activities than that of CL(14-25). In addition, the antimicrobial activity of CL(K25A) was 2.8-fold higher than that of CL(14-25). Therefore, the presence of either arginine at position 24 or lysine at position 25 was particularly important for exerting the antimicrobial activity. These results suggest that the lysine residue at position 25 was not important for exhibiting the antimicrobial activity whenever the peptides contained an arginine residue at position 24. The result on the contribution of the lysine residue at position 25 to the antimicrobial activity is also similar to that obtained using the truncated CL peptide analogs reported previously.[11]

When the CL peptide analogs truncated by removing cationic amino acids (arginine and lysine) at the N- and C-termini of CL(14-25) were used, all their antimicrobial activities were reduced compared with those of CL(14-25). Although one and/or both arginine residues at positions 15 and 24 were found to be important for the antimicrobial activity and the ability to disrupt membranes, it was impossible to obtain insight into enhancement of the antimicrobial activity in the experiments using the truncated CL peptide analogs. Of the single alanine-substituted CL peptide analogs, CL(K20A) possesses the highest growth-inhibitory activity, which was 3.8-fold higher than that of CL(14-25). CL(K20A, K25A) exhibited the highest growth-inhibitory activity among the two alanine-substituted CL peptide analogs. Its antimicrobial activity was 9.1-fold higher than that of CL(14-25). The antimicrobial activity of CL(K20A, K25A) with a net charge of +2 was higher than those of CL(K20A) and CL(14-25) with net charges of +3 and +4, respectively. The hydrophobic moment, that is a quantitative measure of peptide amphipathicity, was used to predict the membrane activities of the alanine-substituted CL peptide analogs. It is defined as the vector sum of the hydrophobicities of the individual amino acids.[37] Mean hydrophobic moment (µH) was calculated by HydroMCalc applet (http://www.bbcm.univ.trieste.it/∼tossi/HydroCalc/HydroMCalc.html). The µH values of CL(14-25), CL(K20A), and CL(K20A, K25A), were −0.18, 0.26, and 0.15, respectively, when they were assumed to have α-helical structures. The µH value of CL(K20A, K25A) with the highest antimicrobial activity was higher than that of CL(14-25), but was lower than that CL(K20A). These results suggest that not only the amphipathic propensity but also one or more other parameters, including cationicity, hydrophobicity, and helical propensity are responsible for antimicrobial activities of the CL peptides. In addition, the helical wheel projections (http://rzlab.ucr.edu/scripts/wheel/wheel.cgi?sequence) of the single alanine-substituted analogs indicate that the hydrophobicity near leucine at position 16 and alanines at positions 18 and/or 21 was enhanced by substituting lysine at positions 20 and 25 with alanine, respectively (Supporting Information Figure S1). The antimicrobial activities of CL(R14A), CL(R15A), and CL(R24A), the single alanine-substituted analogs with replacement of arginine, were lower than those of CL(K20A) and CL(K25A), the single alanine-substituted analogs obtained by replacing lysine, most probably because the primary amine of lysine residue interacts less electrostatically with zwitterionic phospholipids, including phosphatidylethanolamine and phosphatidylcholine, than the guanidinium group of the arginine residue,[38-40] regardless of the position in the sequence of the analogs obtained by replacing with a single alanine residue. In fact, the antimicrobial activities of CL(K20A) and CL(K25A) with three arginine residues were higher than those of CL(R14A), CL(R15A), and CL(R24A) with two arginine residues. The comparison of IC50 values among the two alanine-substituted CL peptide analogs, except for CL(R24A, K25A), also showed that CL(K20A, K25A) containing three arginine residues exhibited the highest antimicrobial activity, while CL(R14A, R15A) containing a single arginine residue exhibited the lowest antimicrobial activity. The contribution of charge distribution in CL peptide analogs with the same net positive charge to their antimicrobial activities was also investigated using the helical wheel projections. However, the antimicrobial activities of CL peptide analogs largely depend on the net positive charge, hydrophobicity, and amphipathicity as described above. The details of the effect of charge distribution on the antimicrobial activities of CL peptide analogs remain unclear.

Although the antimicrobial activities of single alanine-substituted analogs, except for CL(R15A), were higher than that of CL(14-25) (Figure 1), they exhibited lower degrees of diSC3-5 release from P. gingivalis cells than CL(14-25) (Figure 3). Furthermore, the CL peptide analogs bearing substitution of arginine and/or lysine with two alanine residues, except for CL(R24A, K25A), exhibited higher antimicrobial activities against P. gingivalis cells than that of CL(14-25); however, their membrane depolarization activities were significantly reduced. However, among the CL peptide analogs with the same net positive charge, the membrane depolarization activities were related to the antimicrobial activities as described above (Figures 5 and 6). In the previous study,[11] the membrane depolarization activities were also closely related to their antimicrobial activities between the truncated CL peptide analogs with the same net positive charge, in which either a single or two cationic amino acid residues were deleted from the N- and C-termini of CL(14-25). The results suggest that cationicity is mainly a key parameter for releasing diSC3-5 from P. gingivalis cells through membrane depolarization. However, in the assay system used in this study, we cannot rule out the possibility that diSC3-5, a cationic dye, was also released from P. gingivalis cells by the binding of the cationic CL peptides without membrane depolarization, although the details remain unclear. Therefore, other parameters such as hydrophobicity and amphipathicity appear to contribute to the antimicrobial activity of the CL peptides in a concerted manner with cationicity.

The abilities of CL(K20A) and CL(K20A, K25A) to disrupt GUVs were higher than that of CL(14-25). Although the antimicrobial activity of CL(K20A, K25A) was higher than that of CL(K20A) (Table 1), the ability of CL(K20A, K25A) to disrupt GUVs was lower than that of CL(K20A) (Figure 7). In the previous study,[11] fluorescence microscopy image analysis revealed that FITC-conjugated CL(14-25) gradually accumulated on the surface of GUVs, while no fluorescence was detected inside GUVs. Therefore, the decreased ability of CL(K20A, K25A) to disrupt GUVs was not caused by its internalization into GUVs without disruption but by reduced interaction with phospholipids constituting GUVs. Similar to the membrane depolarization activity, the capability to disrupt GUVs seems to depend not only on electrostatic interaction but also on hydrophobicity and amphipathic propensity. The degree of diSC3-5 release from P. gingivalis cells induced by CL(14-25) was higher than that induced by CL(K20A) (Figure 3) or CL(K20A, K25A) (Figure 4). The conflicting findings between the membrane depolarization activity and the ability to disrupt GUVs of the CL peptide analogs appear to be attributable to the difference in phospholipid composition between membranes of P. gingivalis cells and GUVs, although the details of this discrepancy remain unclear. Consequently, the changes in hydrophobicity and amphipathic propensity of CL peptide analogs caused by alanine substitution appear to be mainly responsible for their enhanced antimicrobial activity. To completely comprehend the contribution of the five cationic amino acids in CL(14-25) to the antimicrobial activity, further investigations using analogs bearing the replacement of three and more cationic amino acids with alanine and/or analogs obtained by replacing cationic amino acids with hydrophobic amino acids such as leucine and tryptophan are required in future studies.

CONCLUSIONS

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

In this study, we found that by substituting cationic amino acids in CL(14-25) with alanine, the antimicrobial activity of the obtained CL peptide analogs varied significantly depending on the number and position of alanine and that the alanine-substituted CL peptide analogs, even at a concentration of 1 mM, had no hemolytic activity. Although the net positive charge of CL(K20A) and CL(K20A, K25A) was reduced by alanine substitution compared with that of CL(14-25), their antimicrobial activities against P. gingivalis cells were 3.8-fold and 9.1-fold higher, respectively, compared with that of CL(14-25). The antimicrobial activity of CL(R15A), the analog obtained by replacing arginine at position 15, against P. gingivalis cells, was slightly lower than that of CL(14-25). CL(R14A, K15A), the analog obtained by replacing the two N-terminal arginine residues with two alanine residues had a higher growth-inhibitory activity than that of CL(14-25). These results show that the arginine residue at position 15 is not critical but is important for exerting the antimicrobial activity. In experiments using analogs with the replacement of arginine at position 24 and/or lysine at position 25, we confirmed that the arginine residue at position 24 was crucial for exhibiting the antimicrobial activity whenever the lysine residue at position 25 was substituted with an alanine residue. The helical wheel projections of the single-substituted CL peptide analogs indicate that the hydrophobicity near leucine at position 16 and alanines at positions 18 and/or 21 was enhanced by substituting lysine at positions 20 and 25 with alanine, respectively. The contribution of arginine residues to the antimicrobial activity was more substantial than that of lysine residues, most probably because the primary amine of the lysine residue interacts less electrostatically with zwitterionic phospholipids than the guanidinium group of the arginine residue. The degrees of diSC3-5 release from P. gingivalis cells and disruption of GUVs induced by alanine-substituted CL peptide analogs with different positive charges were not closely related to their antimicrobial activities. In addition, the antimicrobial activities of alanine-substituted CL peptide analogs were not necessarily dependent upon their high cationicity. Therefore, the enhanced antimicrobial activities of the analogs obtained by substitution with alanine appear to be mainly attributable to the changes in properties such as hydrophobicity and amphipathic propensity.

REFERENCES

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
bip22399-sup-0001-suppfig.tif19801KSupplementary Information Figure 1.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.