Since the discovery of penicillin by Alexander Fleming in 1929, antibiotics have been regarded as the most important discoveries in modern medicine, and in fact the development of antibiotics has exploded since the Second World War [1-3]. However, antibiotic resistance has been found in the early stages of development, and abuse of antibiotics in humans and livestock has led to the emergence of many multidrug-resistant bacteria or super bacteria [4]. As a result, the development of new antimicrobial agents to replace existing antibiotics that cause resistance problems is urgent, and antimicrobial peptides (AMPs) are emerging as new ones. Most AMPs are classified mainly on the basis of biochemical and structural characteristics, among which cationic AMPs have been found in animals, insects, and plants and have been known to play an important role in their innate host defense [5, 6]. Most cationic AMPs are composed of less than 50 amino acids and are positively charged with multiple lysine and arginine residues and have hydrophobic and amphipathic properties [7, 8]. The first reason for their attention as promising antimicrobial agents is their wide antimicrobial effect, including against gram-positive and gram-negative bacteria, fungi [9], protozoa [10], and enveloped viruses [11] and anticancer activity [12, 13]. The second is that they do not cause bacterial resistance problems because they exhibit antibiotic effects due to nonspecific electrostatic interactions with bacterial membrane lipid components [14, 15]. Although many native AMPs have been discovered and studied so far, they are rarely used as therapeutic agents despite the potential of such peptides. This is because of problems such as low antibiotic activity against fungi and bacteria, in vivo stability issues, or high production costs [16, 17]. Therefore, there is growing interest today in the design and development of new AMPs that have more antibiotic activity and are more secure to host cells, based on existing native AMPs.

We have identified the relationship between the structure and activity of lactophoricin, a natural AMP found in bovine milk [18-20], and successfully designed, purified, and characterized the YK3 peptide, a novel AMP with shorter length and improved activity than lactophoricin [21-23]. In this study, we developed another new design based on the following three principles to develop other novel AMPs with a more effective and less toxic antimicrobial capacity, using the sequence of the YK3 peptide. First, by displacing the uncharged serine residue of the hydrophilic region with a positively charged lysine residue, the net charge of the entire peptide is increased to promote electrostatic binding with the negatively charged bacterial cell wall [24, 25]. Second, the tryptophan residue, which can interact with the membrane through the indole side chain and attach the peptide to the bilayer surface, was positioned at the interface between the hydrophilic and the hydrophobic regions to aid membrane association [26-28]. Finally, the polar residue was replaced with a nonpolar surface to give imperfect amphipathic properties. This method has recently been shown in many studies to maintain antimicrobial activity and reduce hemolytic activity [24, 29, 30]. Based on the methods mentioned briefly above, we designed a total of nine analogs with amino acid substitutions and deletions. Disc agar diffusion and micro broth dilution tests were performed for nine analogs, and three peptides with the highest antimicrobial activity were selected and named as YK5, YK8, and YK11, respectively. Each peptide showed superior antimicrobial activity compared with YK3 and showed better activity, especially against gram-negative bacteria. We also expressed and purified each peptide to obtain a high-purity peptide, and characterization was performed using various spectroscopic analytical techniques.

Preparation of Newly Designed Peptides

The newly designed analog peptides were synthesized using the solid-phase synthesis method and purified by reverse-phase high-performance liquid chromatography (RP-HPLC) by Peptron (Korea). Synthetic peptides with greater than 95% purity were dissolved in distilled deionized water and stored at -20°C until use in antimicrobial activity tests. The antimicrobial activity of a total of nine substitution and deletion analog peptides was evaluated by disc agar diffusion assay. Among them, three analog peptides with remarkable antimicrobial activity were selected, and a recombinant bacterial expression system was used to produce these peptides. These were named YK5, YK8, and YK11, respectively.

Antimicrobial Activity

The antimicrobial activity of each peptide was tested by disc agar diffusion and micro broth dilution tests, with gram-positive (Listeria innocua MC2 KCTC 3658 and Staphylococcus aureus subsp. aureus Rosenbach ATCC 6538) and gram-negative (Pseudomonas aeruginosa (Schroeter) Migula ATCC 27853, Salmonella enterica subsp. enterica (ex Kauffmann and Edwards) Le Minor and Popoff serovar Typhi ATCC 19430, and Escherichia coli KCTC 1682) bacteria. For both types of test, the five strains were cultured overnight in 3.7% BHI (Bacto brain heart infusion; Beckton Dickinson and Company, USA) medium at 37°C and 240 rpm. The culture was transferred to fresh pre-warmed 3.7% BHI medium and incubated for 2 h with shaking to the mid-exponential growth phase and diluted to a bacterial concentration of 1 × 10⁸ CFU (colony formation units)/ml using 0.037% BHI medium.

The primary antimicrobial activities of the analog peptides were detected by disc agar diffusion test using 3.7% BHI agar plates. Prepared bacterial cultures (1 × 10⁸ CFU/ml) were added to a sterile 3.7% BHI agar medium plate and maintained at 37°C for 30 min. Sterile paper discs (6 mm) were prepared and soaked with 20 μl of each analog peptide solution, and placed on the inoculated plates. After incubation in an oven at 37°C overnight, the diameter of the inhibition zone around each disc was measured to confirm the antimicrobial activity.

For the MIC (minimum inhibitory concentration) determination, a micro broth dilution test was performed using sterile 96-well microtiter plates. Briefly, serial 2-fold dilutions (1.8-900 μM) of each analog peptide were added to 50 μl of bacterial suspensions and incubated at 37°C overnight. The lowest peptide concentration (when no bacterial growth was visible) was defined as the MIC and measured at 600 nm with a spectrophotometer (Thermo Multiskan FC; Thermo Fisher Scientific, USA). All antimicrobial activity test methods were performed according to the Clinical & Laboratory Standards Institute (CLSI) guideline, ‘Method for testing antimicrobial susceptibility to aerobic bacteria; Approved Standard–Ninth Edition’.

Expression of Three Selected Analog Peptides

Oligonucleotides encoding the respective peptides were chemically synthesized by Integrated DNA Technologies (USA) and cloned into pET31b (+) expression vector (Novagen, USA). For optimal peptide expression, the constructed plasmids were transformed into various mutant Escherichia coli competent cells (OverExpress C41(DE3) and C43(DE3) Competent Cells; Lucigen, USA), resulting in C43(DE3) being selected for YK5 and YK8, and C41(DE3) being selected for YK11. The ¹⁵N-enriched peptides used in the NMR structure studies were produced from M9 minimal medium using ¹⁵NH₄Cl as the nitrogen source. A 10 ml volume of the culture solution grown in LB medium containing carbenicillin was transferred to 1 L of M9 minimal medium and incubated at 37°C with shaking at 250 rpm. When the cell culture reached 0.5 at 600 nm optical density (OD₆₀₀), recombinant fusion protein was induced by addition of IPTG to 1 mM final concentration. After 16 h of induction, fully grown cells were harvested by centrifugation and stored in a -80°C freezer for more than 3 h for effective cell lysis.

Purification and Isolation of Expressed Analog Peptides

The harvested cells were resuspended in lysis buffer (20 mM Tris, 500 mM NaCl, 15% glycerol) containing lysozyme (Sigma, USA) for purification and lysed by ultrasonication. The insoluble inclusion bodies containing fusion proteins were separated by centrifugation at 4°C and 13,200 rpm for 30 min and dissolved in Ni-NTA binding buffer (20 mM Tris, 500 mM NaCl, 5 mM imidazole, pH 8.0) of 6 M guanidine-HCl. Prior to application to Ni-NTA resin (Qiagen, Germany), the resulting protein solution clarified by centrifugation was purified using nickel-loaded agarose resin and eluted with elution buffer containing 500 mM imidazole. The 6 M guanidine-HCl as denaturants was removed by dialysis and lyophilized to chemically cleave the fusion partner and target protein using 70% formic acid and cyanogen bromide (CNBr; Sigma). Final purification of the analog peptides was performed by preparative RP-HPLC on a Delta Pak C18 column (7.8 × 300 mm; Waters, USA) with a multistep gradient of 5-75%acetonitrile containing 0.1% trifluoroacetic acid (TFA) at a flow rate of 3 ml/min. The absorbance at 220 and 280 nm was monitored.

Structural Studies of Expressed Analog Peptides

We used matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS, 4800plus Analyzer; AB Sciex, USA) for primarily molecular mass characterization of the analog peptides. Each of the purified and lyophilized peptides was dissolved in 30% acetonitrile/0.1% TFA, mixed with α-cyano-4-hydroxycinnamic acid (Sigma) matrix saturated solution and loaded on the stainless steel sample plate and dried. The instrument equipped with a 355 nm Nd:YAG laser was performed with an accelerating voltage of 20 kV and calibrated using AB Sciex’s peptide mass standard calibration kit. The mass spectra were obtained in the positive-ion reflectron mode, and several spotted samples were individually examined for reproducibility.

The secondary structures of the expressed analog peptides were measured by circular dichroism (CD) spectroscopy. For the analysis, 400 μl samples were measured on a Jasco model J-710 spectropolarimeter (JASCO Corporation, Japan) in a quartz cell with a path length of 1 mm. The spectra were averaged over five scans per sample and were measured over a wavelength range of 190-260 nm with 20 mdeg sensitivity. Peptides were measured in a water environment and a dodecylphosphocholine (DPC; Cambridge Isotope Laboratories, USA) micelle environment, respectively, and the DPC micelle concentrations were varied at 20, 40, 60, 80, and 100 mM. The spectra of the solvent alone were measured under the same conditions and removed from the sample spectra, and all measurements were performed at room temperature.

Cytotoxicity Assay

The cytotoxicity test of the expressed and purified peptides was performed by Drug Discovery Platform Technology Group (Korea). Cell viability was determined using the Cyto X cell viability assay kit (DoGenBio, Korea) for five mammalian cell lines. This is done by using water-soluble tetrazolium salt (Wst-8) and measuring the absorbance at 450 nm after a short reaction time of 1-4h. The mammalian cells VERO, HFL-1, L929, NIH 3T3, and CHO-K1 were used.

Statistical Analysis

In this study, all experiments were performed at least three to five times and the data were presented as the mean ± SD.

Design and Preparation of Peptides

On the basis of the YK3 peptide, which exhibits the best antimicrobial effect, we proceeded to design new AMPs with shorter lengths, better activity, and less toxicity. Since we have already confirmed that the N-terminal part of the YK series is important for antimicrobial activity [18, 20], we retained that part and have designed it according to the following principles. It is known that the majority of AMPs have a net charge of at least +2, and this property is very important for antimicrobial activity because AMPs can bind with negatively charged bacterial phospholipid membranes by electrical interaction [31, 32]. We therefore mutated the serine residue in the hydrophilic moiety to lysine to increase the net charge of the YK3 peptide, allowing it to bind more selectively to the bacterial surface. The tryptophan residue, which has quite a large indole side chain, has both hydrophilic and hydrophobic properties and is known as a typical amino acid affecting lipid polymorphism [26, 27]. These properties can interact with the membrane interface to anchor the peptide on the membrane surface and have the advantage of facilitating hydrophobic and electrostatic interactions between the peptide and the membrane when placed in an amphipathic interface [28, 33]. Thus, we confirmed the helical wheel projection of the YK3 peptide and expected to increase the antibiotic activity by placing the tryptophan residue at the hydrophilic and hydrophobic interface. Previously, AMPs have been described as having characterized amphiphilic forms, but it is known that unnecessarily high amphipathic properties result in rather increased toxicity or reduced cell selectivity [15, 24, 29]. For example, in the case of magainin 2 peptide analogs, the bactericidal activity and cytotoxicity were increased when the amphipathicity of the peptide was increased. Similarly, gramicidin S analogs with b-sheet conformations also showed high hemolytic activity and low antimicrobial activity [34, 35]. Therefore, we mutated the leucine residue of the nonpolar face with a polar lysine residue to improve the antimicrobial activity and minimize cytotoxicity by disrupting the α-helical amphipathic structure of the YK3 peptide. Analogs designed by deletion, substitution, and length control of individual amino acids by the above principles are shown in Table 1. A total of nine analog peptides based on the sequence of the YK3 peptide were designed. From gradual adjustment of the length of the peptides, YK4 to YK8 consist of 15 amino acids, YK9 to YK11 consist of 13 amino acids, and YK12 consists of 11 amino acids. Each peptide was synthesized by solid-phase synthesis to confirm its antimicrobial activity, and the disc agar diffusion test was performed to select the final three peptides having remarkable antimicrobial activity.

Table 1 . Sequences and structural parameters of YK3 and its analog peptides..

Peptide	Sequencea	Length	Net charge	Grand average of hydropathicity (GRAVYb)	Molecular Weight (Calcc)
YK3	NKVKE WIKYL KSLFS	15	+3	-0.487	1883.26
YK4	NKVKE WWKWL KSLFS	15	+3	-0.820	1979.35
YK5	NKVKE WIKYL KSLFK	15	+4	-0.693	1924.36
YK6	NKVKE WIKYL KSKFS	15	+4	-1.000	1898.28
YK7	NKVKE WWKWL KSLFK	15	+4	-1.027	2020.45
YK8	NKVKE WIKYL KSKFK	15	+5	-1.207	1939.37
YK9	NKVKE WWKWL KSL	13	+3	-1.100	1745.10
YK10	NKVKE WIKYL KKL	13	+4	-0.954	1690.11
YK11	NKVKE WWKWL KKL	13	+4	-1.338	1786.20
YK12	NKVKE WWKWL K	11	+3	-1.573	1544.86

^aMutated amino acids are bold and underlined..

^bThe GRAVY value is calculated as the sum of the hydropathy values of all amino acids in the sequence divided by the number of sequences, resulting in an average hydropathy value of the peptide..

^cCalculated by using the ExPASy ProtParam tool..

Antimicrobial Activity

The antimicrobial activity of the designed analog peptides was confirmed using a standardized disc agar diffusion test. This method is very simple, reproducible, and economical, and is therefore suitable for determining the antimicrobial susceptibility of many candidate analogs. As shown in Table 2, the inhibition zone diameters of the analog peptides were measured for gram-positive and gram-negative bacteria. All of the designed analog peptides have similar inhibitory domains to both gram-positive and gram-negative bacteria, particularly those that form an effective inhibitor zone against gram-negative bacteria. Of these, YK4, YK5, YK8, and YK11 formed inhibition zones similar to YK3, and YK10 especially was more effective than YK3 against Salmonella typhimurium and Escherichia coli among gram-negative bacteria. Staphylococcus aureus, Salmonella typhimurium, and Escherichia coli were similarly affected even when the peptide sequence was changed. However, for Listeria innocua and Pseudomonas aeruginosa, the antimicrobial activity and inactivity depended on the length of the peptide sequence. We selected YK5, YK8, and YK11 in the order of antimicrobial effect and proceeded to the next experiment.

Table 2 . Inhibition zone diameters of YK3 and its analog peptides..

Peptide	Zone of inhibition (mm)
	Gram-positive		Gram-negative
	Listeria innocua	Staphylococcus aureus	Pseudomonas aeruginosa	Salmonella typhimurium	Escherichia coli
YK3	13	14	11	14	12
YK4	9	13	7	10	8
YK5	10	12	10	12	11
YK6	-a	10	-	13	11
YK7	-	12	11	10	11
YK8	9	9	9	14	11
YK9	-	11	9	11	10
YK10	-	9	-	14	13
YK11	9	12	10	12	11
YK12	-	12	9	10	9

^aDashes indicate no detectable inhibitor zone..

Expression and Purification of Analog Peptides

The vector used contains the ketosteroid isomerase (KSI) fusion protein and the hexa-histidine tag to facilitate purification, and the methionine residue is located downstream of the KSI sequence and upstream of the hexa-histidine tag sequence, enabling the separation of the target peptide through CNBr cleavage. The resulting vectors containing the selected YK5, YK8, and YK11 DNA sequences were successfully transformed into expression host Escherichia coli competent cells. Then, 12% Tris-Tricine polyacrylamide gel electrophoresis (PAGE) was used to monitor the band thickness of the target fusion protein induced in each of the expression host cell lines (Figs. 1A-1C). In Figs. 1A-1C, lanes 2, 5, and 8 show the expression before induction, lanes 3, 6, and 9 show that 3 h after induction, and lanes 4, 7, and 10 show that after 16 h induction. As a result, YK5 and YK8 selected the C43 (DE3) cell and YK11 selected the C41 (DE3) cell and proceeded to high-level expression. We performed high-level expression of the target peptides using LB medium and M9 minimal medium as previously described [18-23]. Because the KSI fusion protein was completely insoluble in E. coli, fully grown cells were harvested by centrifugation and the inclusion bodies were isolated by cell disruption using lysozyme and ultrasonication. The collected fusion proteins were dissolved using a denaturant and then separated by Ni-NTA affinity chromatography using a hexa-histidine tag. The fusion proteins, from which the denaturant had been removed by dialysis, was subjected to CNBr cleavage in the dark room for 5 h to separate the KSI fusion partner, target peptide, and hexa-histidine tag. These expression and purification steps were confirmed by 12% Tris-Tricine PAGE analysis, and D-F in Fig. 1 shows YK5, YK8, and YK11, respectively. After CNBr cleavage in lane 7 of Figs. 1D-1F, the isolated peptides can be identified. Final purification was performed by RP-HPLC using a C18 reverse-phase column with acetonitrile/water gradient (Fig. 2). All peaks in the chromatogram were identified using 12% Tris-Tricine PAGE and the broad peak at 60-70 min was the KSI protein. The elution times of each target peptide were as follows: YK5 at 47-49 min, YK8 at 39-42 min, and YK11 at 42-45 min. The final yield was 6-8 mg per liter, and the main peptides were collected and lyophilized, followed by experiments for identification and characterization.

Figure 1. Tris-Tricine PAGE analysis of recombinant protein expression and production. (A-C) 12% Tris-Tricine PAGE analysis that confirmed the expression levels of expression host cells containing a single insert of YK5, YK8, and YK11, respectively. Lane 1: molecular weight marker; Lanes 2-4: BL21(DE3)pLysS; Lanes 5-7: C41(DE3); Lanes 8-10: C43(DE3). (D-F) 12% Tris- Tricine PAGE showing the expression and purification steps of analog peptides. Lane 1: molecular weight marker; Lane 2: before induction, Lane 3: after induction; Lane 4: supernatant of cell lysis; Lane 5: pellet of cell lysis; Lane 6: after Ni-NTA affinity chromatography; Lane 7: after CNBr cleavage.

Figure 2. Reverse-phase HPLC purification of analog peptides after chemical cleavage. (A, B, and C) HPLC chromatograms of YK5, YK9, and YK11, respectively. Each numbered peak can be identified by 12% Tris-Tricine PAGE. The high absorbance line was observed at 220 nm and the low absorbance line was observed at 280 nm. In each chromatogram, YK5 was the second peak, YK8 was the first peak, and YK11 was the second peak. These peaks were collected and used for various spectroscopic analytical experiments.

Identification and Structural Characterization of Peptides Purification and identification of purified peptide expression were performed using MALDI-TOF mass spectroscopy. After CNBr cleavage, the exact mass of homoserine lactone forms of each peptide was 2,007.41 Da for YK5, 2,022.43 Da for YK8, and 1,869.25 Da for YK11. The exact mass of homoserine free acid form was 2,025.43 Da for YK5, 2,040.44 Da for YK8, and 1,887.27 Da for YK11. As can be seen in Fig. 3, the observed values of the mass spectra obtained using the final purified peptides showed agreement with the theoretical molecular weight and high purity. We also used CD spectroscopy to determine the secondary structure of analog peptides in aqueous solutions with or without DPC micelles and to investigate possible conformations of the AMPs in membrane mimetic environments. The DPC micelles were chosen because they are suitable membrane environments for subsequent 3D structure study using NMR experiments (Fig. 4). As can be seen from each CD spectrum, peptides in aqueous solution without a membrane environment exhibit a strong negative band at 200 nm, indicating a random coil form with unordered structure. YK11 (Fig. 4C) showed an additional negative band around 220-230 nm, which is the exciton coupling band due to the Trp-Trp interaction seen in the tryptophan-rich peptide, indicating the presence of turn conformation [36, 37]. Since YK11 has a total of three tryptophan residues in the sequence, the seventh isoleucine and the ninth tyrosine are mutated into tryptophan, which may affect the secondary CD spectral characteristics of the peptide. The DPC concentration was measured at 20, 40, 60, 80, and 100 mM, and no significant difference was observed (data not shown). The spectra in Fig. 4 are all in the presence of 100 mM DPC.

Figure 3. MALDI-TOF mass spectra of analog peptides. The spectra were measured in positive ion reflector mode and were measured several times with different spots in the same sample. (A, B, and C) Mass spectra of YK5, YK9, and YK11, respectively. The spectra show that each peptide has high purity and no degradation of the peptide. The observed molecular mass for YK5 is 2,007.17 Da, YK8 is 2,022.37 Da, and YK11 is 1,869.17 Da, and these values agree with theoretical values.

Figure 4. CD spectra of analog peptides. (A, B, and C) CD spectra of YK5, YK9, and YK11, respectively. The peptides were measured in an aqueous environment with or without dodecylphosphocholine (DPC) micelles and showed significant structural changes. It has been confirmed that it has an alpha-helical structure in the membrane environment as in the case of the previous YK series of antimicrobial peptides.

Antimicrobial Activity and Cytotoxicity of Expressed and Purified Analog Peptides

The analog peptides expressed in E. coli and purified were used to confirm antimicrobial activity and cytotoxicity. In addition to disc agar diffusion testing previously performed to screen for peptides, a micro broth dilution test was performed to determine the MIC₅₀ concentration for bacterial cells (Table 3). Additionally, cytotoxicity studies on mammalian cell lines were performed to confirm the safety of the peptides (Table 4). Disc agar diffusion test results using the expressed and purified peptides showed similar results to those performed to screen for analog peptides, which means that the efficiency of antimicrobial activity of the synthesized peptide and the expressed peptide was the same. This method is easy and convenient, and it is inexpensive and results can be seen at a glance, but it has a disadvantage that it cannot measure values like MIC or the minimum bactericidal concentration because it is not quantitative. Therefore, we performed a micro broth dilution test to determine the quantitative antimicrobial activity of each of the analog peptides, and the MIC₅₀ values required for 50% growth inhibition of the bacteria are shown in Table 3. The values in the table indicate that YK3 is less effective against Listeria innocua, Salmonella typhimurium, and Escherichia coli than against Staphylococcus aureus and Pseudomonas aeruginosa, whereas YK5, YK8, and YK11 can inhibit the growth of all five bacteria in very small amounts. We next examined the cytotoxicity of the peptides to mammalian cells. As shown in Table 4, with YK5 and YK8, 50% of the cells were alive except for CHO-K, even if the peptide concentration was 100 μM or more and that result implies the analog peptides have low cytotoxicity to mammalian cells.

Table 3 . Inhibition zone diameters and MIC₅₀ of expressed and purified YK3 and its analog peptides..

Peptide			YK3	YK5	YK8	YK11
Zone of Inhibition (mm)	Gram-positive	Listeria innocua	12	12	12	12
		Staphylococcus aureus	13	12	9	13
	Gram-negative	Pseudomonas aeruginosa	11	11	9	10
		Salmonella typhimurium	12	12	15	14
		Escherichia coli	11	13	13	12
MIC50 (μM)	Gram-positive	Listeria innocua	20.0	5.0	3.0	6.5
		Staphylococcus aureus	2.5	3.2	2.0	3.0
	Gram-negative	Pseudomonas aeruginosa	2.5	6.2	5.0	2.5
		Salmonella typhimurium	17.0	5.5	2.5	4.7
		Escherichia coli	35.0	12.0	4.5	4.5

Table 4 . IC₅₀ values in various mammalian cell lines..

Cell lines a	Peptide	YK3	YK5	YK8	YK11
IC50 (μM)b	VERO	99.1	>100	>100	83.7
	HFL-1	94.0	>100	>100	>100
	L929	>100	>100	>100	>100
	NIH 3T3	76.0	>100	>100	96.6
	CHO-K	56.2	57.0	71.9	86.8

^aVERO: African green monkey kidney cell line; HFL-1: Human embryonic lung cell line; L929: NCTC clone 929, mouse fibroblast cell line; NIH 3T3: Mouse embryonic fibroblast cell line; CHO-K: Chinese hamster ovary cell line..

^bThe half maximal inhibitory concentration..

AMPs vary in their structure, sequence, and origin, but they generally have a positive charge and are well divided into hydrophilic/hydrophobic regions and folded into amphipathic conformations to interact with the bacterial membrane. These features allow AMPs to bind through the electrostatic interaction with negatively charged bacterial membranes and to induce antimicrobial activity by allowing them to be inserted into the membrane through hydrophobic interaction. In previous studies, YK1, 2, and 3 peptides based on lactophoricin were newly designed, and their antimicrobial activity and structure studies were carried out. These peptides had further enhanced amphiphilicity by substituting amino acids for better interaction with bacterial membranes via helical wheel projection [21-23].

We have attempted to design other AMPs based on YK3, which has antimicrobial activity but some cytotoxicity. In order to achieve better antimicrobial activity and safety, analog peptides were designed based on the importance of individual amino acids. A total of nine analog peptides from YK4 to YK12 were designed and disc agar diffusion test was used to select the three most potent; YK5, YK8, and YK11. Each of the peptides was successfully transformed into E. coli competent cell for high-level expression, and after many optimization steps, the final high-purity peptide was obtained. Expressed and purified peptides were clearly identified by PAGE and mass spectroscopy. In particular, it was confirmed that after CNBr cleavage, the homoserine lactone formed peptide c-terminal was dominant through molecular weight check. As a result of confirming the secondary structure of the peptide by CD spectroscopy, it was shown that the analog peptides have an alpha-helical structure under membrane conditions, and in the absence of a membrane environment, the random coil form is locally increased in turn conformation as the tryptophan residue increases. We also evaluated the antimicrobial activity and cytotoxicity of the expressed analog peptides and examined the possibility of AMPs as a substitute for existing antibiotics. Although the degree of activity varies depending on the method of antimicrobial activity test, the analog peptides have antimicrobial activity against both gram-spositive and gram-negative bacteria, and are particularly effective against gram-negative bacteria. Since YK3 has been verified for activity against fungi such as Candida albicans through the antifungal susceptibility test, the analog peptides will also be tested against various microorganisms.

However, despite these AMPs having broad antimicrobial activity and cell stability, they are unstable owing to the in vivo proteases, and thus there are many reports that they are difficult to apply to therapeutic applications. In other words, it is easily decomposed by endogenous proteases from human skin and proteases from invading microbes and does not show sufficient activity. Fortunately, however, a variety of methods have been reported to improve the protease resistance by modifying sequences of AMPs, such as incorporation of non-natural amino acids, tryptophan substitutions, N-/C-terminal protection, and cyclization [38, 39]. In fact, although recent studies have shown that YK3 is degraded by proteases such as trypsin and chymotrypsin, based on these references it is expected that YK11, in which the amino acids in the sequence are replaced by tryptophan, will be resistant to degradation by these proteases. In conclusion, we have developed new AMPs that have better antimicrobial activity, low cytotoxicity, and short lengths, and are likely to be used as new antimicrobial agents. In our study, we have confirmed that single amino acid substitutions can affect not only the structure of the peptide but also its activity. In the future, we will study clinical applicability through sequence optimization, along with evaluation of activity against various microorganisms.

This work was supported by the HUFS research fund of 2017 and Basic Science Research Program through the National Research Foundation of Korea(NRF) funded by the Ministry of Education (2017012599).

The authors have no financial conflicts of interest to declare.