|Home | About | Journals | Submit | Contact Us | Français|
Cationic antimicrobial peptides (AMPs) have attracted a great deal of interest as a promising candidate for a novel class of antibiotics that might effectively treat recalcitrant infections caused by a variety of microbes that are resistant to currently available drugs. However, the AMPs are inherently limited in that they are inevitably susceptible to attacks by proteases generated by human and pathogenic microbes; this vulnerability severely hinders their pharmaceutical use in human therapeutic protocols. In this study, we report that a halocidin-derived AMP, designated HG1, was found to be resistant to proteolytic degradation. As a result of its unique structural features, HG1 proved capable of preserving its antimicrobial activity after incubation with trypsin, chymotrypsin, and human matrix metalloprotease 7 (MMP-7). Additionally, HG1 was observed to exhibit profound antimicrobial activity in the presence of fluid from human skin wounds or proteins extracted from the culture supernatants of Staphylococcus aureus and Pseudomonas aeruginosa. Greater understanding of the structural motifs of HG1 required for its protease resistance might provide feasible ways to solve the problems intrinsic to the development of an AMP-based antibiotic.
Over the last 2 decades, cationic antimicrobial peptides (AMPs) have been considered as promising candidates for a novel type of antibiotic, as many AMPs have been shown to have broad-spectrum antimicrobial activity and have also been demonstrated to kill microbes via a mode of action that differs distinctly from those of conventional drugs (11, 34, 39). Among the hundreds of natural AMPs thus far discovered, several peptides (23, 30), including magainin (14) and cathelicidins (26, 29), have been employed as templates for the development of new antibiotics. The majority of these developments have been targeted at the topical treatment of recalcitrant infection, as AMPs have shown some evidence of potential toxicities under systemic usage conditions but no topical toxicity (28). However, several problems continue to hinder the exploitation of the potential of this class of molecules in clinical settings. Many of the AMPs that demonstrate significant in vitro antimicrobial activity tend to lose that activity under physiological conditions, such as elevated salt concentration and/or in the presence of zwitterions (19, 49, 50). Additionally, it has been demonstrated that AMPs are susceptible to damage by proteases from human skin, as well as from invading microbes (24, 47), which are likely to be the most obvious causes of poor or incomplete in vivo activity. In particular, it has been well demonstrated that AMPs such as LL-37 were destroyed by proteases secreted from Staphylococcus aureus V8 (40) and Pseudomonas aeruginosa (37). Therefore, AMP-based antibiotics should be considered to be susceptible to inactivation via proteolytic degradation upon administration to skin infection sites, wherein large quantities of microbial and skin proteases tend to exist.
Among the hundreds of cationic AMPs found in nature, very little information has been discovered that might be relevant to surmounting the problems of protease liability. It has recently been proposed that AMP, which exhibits a dimeric structure stabilized by an intermolecular disulfide bond, will prove resistant to proteases (3, 5). The cryptdin-related sequence (CRS) peptides have been identified as members of a family of AMPs occurring in the small intestines of mice (32), and some CRS peptides have been isolated and determined to be covalent dimers whose components each contain 38 amino acids (13). These CRS peptides were considered to play an important role in the killing of a number of microbes invading the small intestines of mice. Hence, it was suggested that they were capable of exerting their antimicrobial activity in environments containing an abundance of digestive proteases (22), although their protease resistance has not yet been clearly demonstrated. Distinctin was another dimeric AMP, which was obtained from the skin glands of a tree frog (1). It was reported that this peptide showed evidence of resistance against proteolytic degradation by trypsin, and its resistance was attributed to an interdisulfide bond linking two monomers (5). As a natural AMP detected in the tunicate Halocynthia aurantium, halocidin has also been identified as a dimeric peptide consisting of two monomers of 18 and 15 amino acids (15). Previously, several synthetic halocidin analogs were designed via the addition and/or substitution of amino acid residues and assessed with regard to their antimicrobial activity under a variety of conditions. As a result, di-K19Hc, referred to herein as HG1, was selected as the most appropriate candidate (17). HG1 exhibited a significant degree of antimicrobial activity against a wide range of pathogenic microbes under elevated salt concentrations or in the presence of Ca2+ or Mg2+ (16, 17). However, it has yet to be determined whether or not HG1 could maintain its metabolic stability against protease-mediated proteolytic degradation. In this study, we assessed the resistance of HG1 to trypsin, chymotrypsin, and matrix metalloprotease 7 (MMP-7), which are known to be abundant in human skin. We have endeavored to demonstrate its protease resistance abilities by elucidating in detail the structural changes in HG1 resulting from treatment with three enzymes. Additionally, we have attempted to determine whether or not HG1 could preserve its antimicrobial activity in the presence of human wound fluid (HWF) and proteins extracted from culture supernatants of S. aureus and P. aeruginosa, both of which are known to be major causal agents of human external infections. We hoped that this study might provide critical clues by which the therapeutic potential of HG1 for the treatment of superficial infections caused by rapidly emerging antibiotic-resistant microbes might be judged.
Table Table11 shows the amino acid sequences of the AMPs utilized in this study. Each of the peptides was synthesized using an automated solid-phase peptide synthesizer (Pioneer Applied Biosystems, Foster City, CA) at Peptron, Inc. (Daejon, Korea), and then purified to over 95% with a C18 reversed-phase high-performance liquid chromatography (RP-HPLC) column (Vydac 218TP54; The Separation Group, Hesperia, CA). In an effort to prepare HG1, HG34, and Khal, each of the monomers was exposed to 0.1 M ammonium bicarbonate (1 mg/ml, pH 8.7) for 3 days at room temperature and then subjected to RP-HPLC to purify the resultant dimer peptide (17). Three other control AMPs (MSI-78, LL-37, and indolicidin) were also synthesized in accordance with their amino acid sequences. All synthetic peptides were verified by measuring their molecular masses via matrix-assisted laser desorption ionization (MALDI) mass analysis (4700-MALDI tandem time of flight [TOF/TOF] mass spectrometry; Pioneer Applied Biosystems, Foster City, CA) at the Korea Basic Science Institute (Daejon, Korea). The following four commercial antibiotics were utilized for antimicrobial tests performed for comparison with the antimicrobial activities of AMPs: ceftazidime (C3809; Sigma), teicoplanin (50-980-654; Fisher Scientific), aztreonam (A6848; Sigma), and ceftriaxone (C5793; Sigma).
In an effort to obtain bacterial culture supernatants (CS), we utilized S. aureus V8 (ATCC 49775) and multidrug-resistant P. aeruginosa (MDRPA) strain CCARM 2161, provided by the Culture Collection of Antibiotic Resistant Microbes (CCARM) at Seoul Women's University in Korea. A culture of S. aureus V8 or MDRPA was incubated for 20 h in tryptic soy broth (TSB; Difco) and centrifuged for 30 min at 10,000 × g. The CS of each strain was concentrated via ultrafiltration using a 10-kDa-cutoff membrane (Vivaflow 200, no. 05VF20045; Viva Science, Ltd., United Kingdom). Proteins in the concentrated CS were precipitated via 60% saturation with ammonium sulfate. After 30 min of centrifugation at 10,000 × g, the precipitate was resuspended in a small volume of 10 mM sodium phosphate buffer (NaP) at pH 7.4 and transferred into dialysis tubing. The sample was then dialyzed against NaP for 18 h at 4°C. The protein concentration was measured via bicinchoninic acid assay (Pierce, Rockford, IL), using bovine serum albumin as a standard protein. The human wound fluid (HWF) was obtained from wound beds at the end of resuturing for dehiscent wounds, which were kept open for several days due to wound seroma. The HWF collected from four patients was pooled, centrifuged, aliquoted, and stored at −80°C. The use of HWF was consented to by the patients.
Three types of antimicrobial assays were utilized to assess the anti-methicillin-resistant S. aureus (MRSA) activity of samples. The MICs of six AMPs and four conventional antibiotics were determined via a broth dilution assay that was conducted in accordance with recommendations of the Clinical and Laboratory Standards Institute (4). In brief, bacteria were cultured overnight to stationary phase in Mueller-Hinton broth (MHB). The cultures were diluted in fresh MHB to a final concentration of 2 × 105 CFU/ml. Stock solutions of each peptide and antibiotic were prepared in acidified water (0.01% acetic acid) at 640 μg/ml in polypropylene microtubes, and the samples were 2-fold diluted serially in acidified water to 10 μg/ml. One-hundred-microliter aliquots of the bacterial suspension were dispensed into each well of a 96-well microtiter plate (Costar 3790; Corning), and then 11 μl of sample solution was added, thereby adjusting to total volume of 111 μl. The antibacterial activity of the samples was evaluated by visible turbidity in each well after 18 h of incubation at 37°C. The MICs were expressed as the minimum concentration of each sample required for a visible inhibition of growth. The experiment was conducted in triplicate. For the colony count assay, 100 μl of a test sample was mixed with a 100-μl aliquot of methicillin-resistant Staphylococcus aureus (MRSA) strain CCARM 3696 in NaP. The mixture was subsequently incubated for 5 min at 37°C, and then 50-μl aliquots were plated, either directly or after dilution, on tryptic soy agar (TSA) consisting of 1.5% agarose and 3% TSB powder in NaP, at neutral pH. After overnight incubation at 37°C, the resultant viable colonies were counted. The data were expressed as the number of recovered CFU per ml of sample or as the survival ratio to CFU of the control sample that was not treated with AMPs. Radial diffusion assays were conducted as previously described (41). In brief, the underlay gel consisted of 1% agarose and 0.03% TSB in a buffer (pH 7.4) containing 9 mM NaP and 1 mM sodium citrate, and the overlay gel consisted of 1% agarose and 6% TSB in distilled water. The underlay solution (10 ml) was mixed with washed, mid-logarithmic-phase MRSA (4 × 106 CFU) and poured into 100-cm2 square petri dishes. Five-microliter aliquots of peptide samples adjusted to concentration of 200 μg/ml in 0.01% acetic acid were subsequently loaded into a series of 3-mm-diameter wells made in the underlay gel. After 3 h of incubation at 37°C, 8 to 10 ml overlay gel was poured on the underlay gel. After the plate was incubated overnight, the clear zone diameters were measured to the nearest 0.1 mm and were expressed in units, where 0.1 mm = 1 unit.
Twenty micrograms of each AMP was mixed with 100 μl of 10 mM NaP buffer containing 50 nM trypsin (Sigma catalog no. T0303) or chymotrypsin (Sigma catalog no. C7762). In the case of MMP-7 (catalog no. 444270; CalBiochem, La Jolla, CA), the protease was dissolved in 10 mM HEPES (pH 7.4) containing 150 mM NaCl and 5 mM CaCl2 and adjusted to 50 nM. After 5, 10, 20, or 30 min of incubation and an additional 1 h at 37°C, the sample was heat treated for 10 min at 80°C to halt the enzyme reaction. After the samples were cooled to room temperature, 100 μl of MRSA solution (2 × 107 CFU/ml) was added to each sample. The mixture was then incubated for 5 min, and 50 μl of the sample was plated on TSA for the colony count assay. The experiments were repeated three times on different days. In an effort to further evaluate the resistance of HG1 and HG34 to trypsin, chymotrypsin, and MMP-7, 100 μg of HG1 or HG34 was treated with each of enzymes at 25 nM in 1 ml of NaP or HEPES buffer for 10 or 30 min or for 1 h at 37°C. After incubation for the given time, 1 ml of solution A (0.1% trifluoroacetic acid [TFA]) in RP-HPLC was added to the sample. The mixture was then subjected to a C18 RP-HPLC column, and the eluted peaks were analyzed via MALDI mass spectrometry at the Korea Basic Science Institute. The primary sequences of the digestion fragments were predicted via comparison of the measured values with the theoretical average masses, and each of the fragments was assessed for its antimicrobial activity against MRSA via radial diffusion assay.
Protease activity of the bacterial CS proteins was monitored by using an EnzCheck protease assay kit (catalog no. E6638; Invitrogen, United Kingdom) in accordance with the manufacturer's instructions. In brief, 500 μl of NaP buffer containing the concentrated CS proteins of each microbe (200 μg for S. aureus V8 or 50 μg for MDRPA) was mixed with 500 μl of BODIPY-FL-casein substrate in NaP buffer and incubated at 37°C for 24 h. Alternatively, prior to incubation with BODIPY-FL-casein substrate, each CS-containing NaP buffer was incubated for 15 min at 37°C with the following protease inhibitors: 100 μl/ml protease inhibitor cocktail (Sigma catalog no. P2714), 200 μM AEBSF [4-(2-aminoethyl) benzenesulfonyl fluoride] (Sigma catalog no. A8456), 40 μg/ml of aprotinin (Sigma catalog no. A1153), 40 μg/ml of E-64 (Sigma catalog no. E3132), 40 μg/ml of pepstatin (Sigma catalog no. P9375), 400 μg/ml of bestatin (Sigma catalog no. B8385), and 2 mM EDTA (Sigma catalog no. ED2SS). We monitored protease activity as an increase in fluorescence with spectrophotometer (Cary Eclipse; Varian). In order to detect protease activity of HWF and to determine the nature of the proteolytic activities produced by HWF, casein zymography was performed. The HWF sample (2 μl) was mixed with sample buffer for SDS-PAGE and loaded onto 12% SDS-PAGE gel containing 0.1% (wt/vol) casein. After electrophoresis under cooling conditions, the gel was washed twice in the equilibration buffer (0.05 M Tris-HCl, 2.5% Triton X-100, pH 7.6) for 2 h with gentle shaking. The gel lanes were then dissected, and each gel slice was incubated at 37°C for 18 h in enzyme reaction buffers (0.05 M Tris-HCl, 0.15 M NaCl, pH 7.6), each containing a protease inhibitor (100 μl/ml inhibitor cocktail, 200 μM AEBSF, 10 μM aprotinin, 40 μg/ml E-64, 40 μg/ml pepstatin, 100 μg/ml bestatin, or 20 mM EDTA). Zones of proteolysis were visualized after Coomassie blue staining and destaining.
In an effort to determine the effect of proteases in two bacterial CS on the anti-MRSA activity of the peptides, 20 μl of each peptide solution (1 mg/ml) in NaP buffer was mixed with 80 μl of NaP buffer containing the concentrated CS proteins of each microbe (200 μg for S. aureus V8 or 50 μg for MDRPA). After 1 h of incubation at 37°C, 100 μl of MRSA suspension (2 × 107 CFU/ml) was added to the mixture for the colony count assay. Alternatively, prior to incubation with peptides, each CS-containing NaP buffer (72 μl) was incubated for 10 min at 37°C with 8 μl of inhibitor cocktail solution, which was prepared in accordance with the supplier's recommendations. Additionally, the HG1 or HG34 samples incubated with the proteins of each bacterial CS were heat treated for 10 min at 80°C in order to stop the protease reaction and then subjected to MALDI mass analysis. The primary sequences of the digestion products were identified via comparison of the measured value with the theoretical average mass. In an effort to assess the effects of HWF on each peptide, 20 μl of peptide solution (1 mg/ml) was mixed with 80 μl of Hank's balanced salt solution (HBSS) buffer containing 20 μl of HWF. After 1 h of incubation at 37°C, the mixture was subjected to a colony count assay, as described above. Alternatively, the HWF-containing HBSS buffer was also pretreated with inhibitor cocktail via the same procedure as that utilized for bacterial CS. Additionally, the HG1 and HG34 samples treated with HWF for different times (1 min, 30 min, 1 h, 2 h, and 4 h) were subjected to Tricine-SDS-PAGE analysis under nonreducing conditions. In the case of HG1 samples, the HG1 band was consistently detected on PAGE gel after treatment with HWF (see Fig. Fig.6C).6C). Accordingly, the longest-postincubation sample (4 h) was subjected to RP-HPLC to determine whether the peak corresponding to HG1 could be detected. For HPLC analysis, the HG1 sample incubated with HWF was mixed with an identical volume of 20% acetic acid and incubated overnight with gentle stirring. The sample was centrifuged for 20 min at 10,000 × g, and the supernatant was injected into a C18 RP-HPLC column. For controls, 100 μl of HBSS containing 20 μl of HWF was subjected to the same procedure.
In order to obtain each of the three dimeric peptides included in peak a in Fig. Fig.11 B, we synthesized two types of monomers (a 14-mer and a 15-mer). Two homodimers were generated via the same procedure as was utilized in the case of HG1. In the case of heterodimeric peptides consisting of the 14-mer and 15-mer, two different monomers were mixed in equal amounts in 0.1 M ammonium bicarbonate (pH 8.7). After 3 days of incubation at room temperature, the mixture was added to the same volume of 20% acetic acid and then subjected to preparative acid urea (AU)-PAGE (12). The AU-PAGE eluates were collected in 1-ml fractions, every second of which was subjected to analytical AU-PAGE to identify the heterodimer-containing fractions. The heterodimers were purified further via C18 RP-HPLC. Three purified dimeric peptides (14-14-mer, 14-15-mer, and 15-15-mer) were finally confirmed by measuring their molecular masses via MALDI mass analysis. A model Mini-Protean 3 cell (Bio-Rad) was used to perform analytical AU-PAGE and SDS-PAGE. A Mini Prep cell (catalog no. 170-2908; Bio-Rad) was utilized for preparative AU-PAGE.
Table Table22 shows MIC values of three halocidin-derived peptides (HG1, HG34, and Khal), three control AMPs (MSI-78, LL-37, and indolicidin), and four currently available antibiotics. Whereas all AMPs, with the exception of LL-37, have demonstrated potent antibacterial activity in a selected range of concentrations (1 to 64 μg/ml) against all tested bacterial strains, each of the commercial antibiotics showed evidence of a biased activity against certain bacteria. Among six AMPs, HG34 consistently exhibited the most profound antimicrobial activity against all bacterial strains. It was worth noting that the MICs of three halocidin-derived peptides were equivalent to or lower than those of MSI-78, which had entered phase III clinical trials for development as a topical antibiotic. Overall, three peptides (HG1, HG34, and Khal) demonstrated more potent antimicrobial activity and broader antimicrobial spectra than the three control AMPs, in addition to the four commercial antibiotics.
Colony count analysis showed that the anti-MRSA activity of four AMPs (HG34, MSI-78, LL-37, and indolicidin) was impaired within 10 min and was abolished completely after 20 min of incubation with trypsin (Fig. (Fig.1A).1A). In contrast, HG1 was determined to be capable of exerting its anti-MRSA activity for up to 1 h of incubation. This result led us to postulate that either the entire structure of HG1 may be resistant to trypsin digestion or some of the fragments generated by trypsin digestion may retain their anti-MRSA activity. In an effort to clarify this, HG1 was treated with trypsin for 10 min, 30 min, and 1 h, and the mixture was then subjected to RP-HPLC, after which the emerging peaks were separately isolated. In the case of the 10-min-incubation sample, two HPLC peaks (a and b) were assessed (the HG1 peak, *, was not) (Fig. (Fig.1B).1B). Peak b was found to harbor a heterodimeric peptide (15-19-mer) in which the C-terminal 4 amino acid residues (GVLA) were removed from one of the two chains. Additionally, peak a was shown to harbor digestion fragments mixed with three types of peptides, which were identified as two homodimers (14-14-mer and 15-15-mer) and a heterodimer (14-15-mer) (Table (Table3).3). However, we proved unable to separately isolate three peptides in peak a with our HPLC system. Thus, we prepared each of the three dimers using two types of synthetic monomers (a 14-mer and 15-mer) and subsequently purified each of them as described in Materials and Methods. Each of the purified dimers was monitored via AU-PAGE analysis (inset in the bottom profile of Fig. Fig.1B).1B). In the 30-min and 1-h incubation samples, the results of HPLC analyses showed that only peak a was eluted as a major peak. Even after overnight incubation with trypsin, peak a was found to harbor all three peptide types (data not shown), thereby indicating that trypsin worked incompletely as an exopeptidase. Radial diffusion assays showed that the heterodimer (15-19-mer) of peak b and the homodimer (15-15-mer) and heterodimer (14-15-mer) of peak a retained their anti-MRSA activity, although it was somewhat reduced compared to that of HG1 (Table (Table33 and see Fig. Fig.4).4). In contrast, the smaller homodimer consisting of two 14-mers exhibited no activity (Table (Table33 and Fig. Fig.44).
For the control, after the trypsin digestion of HG34, the resultant products were also evaluated via the same procedure as in the case of HG1 (Fig. (Fig.1C).1C). Peak q was determined to harbor a heterodimer consisting of an intact 19-mer and a 15-mer that was missing the N-terminal 4-amino-acid residue (KWKK). This peptide fragment was confirmed to retain anti-MRSA activity comparable to that of HG34 (Table (Table3).3). However, the digestion fragments in another five peaks (peaks l, m, n, o, and p) were shown to exhibit no anti-MRSA activity. From this result, it was determined that the anti-MRSA activity detected in the HG34 sample after 10 min of incubation (Fig. (Fig.1A)1A) was attributed to the peak q peptide and the remaining HG34 (peak **). However, peak q and HG34 were not detected after 30 min of incubation, which explained why the HG34 sample did not demonstrate its anti-MRSA activity after 30 min of trypsin treatment (Fig. (Fig.1A).1A). The trypsin scissile points of HG1 and HG34 were indicated as arrows on their amino acid sequences (Fig. (Fig.1D).1D). Our results demonstrated that HG1 could maintain its anti-MRSA activity if the N-terminal K residue was kept in either of the two chains even after the C-terminal four amino acid residues (GVLA) were excised from both chains. In contrast, HG34 easily abrogated its antimicrobial activity upon trypsin digestion, since it harbors two amino acid residues (K3 and K4) in each chain that are vulnerable to trypsin-mediated endopeptidic degradation.
The resistance of AMPs to chymotrypsin was also evaluated via the same procedure as in the experiment performed with trypsin. As shown in Fig. Fig.22 A, it was determined that the HG1 and HG34 samples maintained their profound anti-MRSA activity during 1 h of chymotrypsin incubation. Generally, it has been demonstrated that the preference of chymotrypsin for the cleavage of the peptide bonds was just past the amino acid residues with large and hydrophobic side chains, such as Trp (W), Leu (L), Tyr (Y), Phe (F), and Met (M). In our experiment, we selected an α-chymotrypsin, which has previously been shown to cleave peptide bonds selectively on the C-terminal side of W and L (http://merops.sanger.ac.uk), as HG1 harbors one W residue and five L residues on its amino acid sequence, but no Y, F, or M residues. Therefore, at the inception of this experiment, it was postulated that each chain of HG1 may be cleaved at six positions by α-chymotrypsin, which probably generates a host of complicated and small fragments. However, during our entire incubation time, only peaks c, d, and e together with an intact HG1 peak were detected as discrete peaks (Fig. (Fig.2B).2B). As shown in Table Table3,3, peak e was determined to harbor a heterodimer consisting of an intact chain (19-mer) and a 12-mer that was missing seven amino acid residues from the N terminus. This 7-mer peptide was detected as peak d. Peak c, which was confirmed to harbor a homodimer consisting of two 12-mer chains, could be readily detected after 30 min of incubation. Additionally, our antimicrobial assay showed that whereas peptides in peaks c and d demonstrated no anti-MRSA activity, the peptides in peak e retained an anti-MRSA activity comparable to that of HG1 (Table (Table33 and Fig. Fig.44).
For the controls, we also conducted the same experiment with HG34, as in the case of HG1. As a consequence, it was determined that HG34 was cleaved at three scissile bonds (W2-K3, L6-L7, and L7-H8) on one of two chains after 10 min of α-chymotrypsin treatment. Peak s, the tallest of the three peaks, was generated via a cleavage at L7-H8 on one of two chains, as was the case with peak e of the HG1 sample. However, unlike peak e of the HG1 sample, peak s was diminished significantly at 1 h postincubation. According to these results, we concluded that the chymotrypsin digestion at L7-H8 proceeded much faster for HG34 than for HG1. Additionally, it was determined that an intact chain of the peak s peptide could also be cleaved at W2-K3, apart from L7-H8, after 30 min and 1 h of incubation. As a consequence, another dimeric peptide was generated and detected as peak r (Fig. (Fig.2C2C and Table Table3),3), the counterpart of which was not detected in the HPLC profile generated with the HG1 sample. The anti-MRSA assay showed that three types of heterodimeric peptides (peak s, t, and u peptides) with an intact chain had anti-MRSA activity, but another heterodimer (peak r) missing two amino acids from the N terminus of an intact chain of the peptide in peak s showed no evidence of anti-MRSA activity (Table (Table33 and Fig. Fig.44).
Human MMP-7 is known to be constitutively produced by all skin and is upregulated upon bacterial exposure (25, 36). It was surmised that HG1 might be consistently affected by MMP-7 upon administration onto skin infection sites, and its therapeutic potency might be abolished. Therefore, we evaluated the resistance of HG1 and HG34 to MMP-7 digestion according to the same procedures used in the experiments performed with trypsin and chymotrypsin. As a result, both peptides were confirmed to preserve their profound antimicrobial activity during 1 h of incubation (Fig. (Fig.33 A). In contrast, the antimicrobial activities of three other control AMPs were severely impaired within 5 min after incubation with MMP-7. In consecutive experiments, HG1 was treated with MMP-7 for 10 min, 30 min, and 1 h, and the mixture was then subjected to RP-HPLC analyses. As shown in Fig. Fig.3B,3B, the 1-h-incubation sample generated six peaks, apart from that of intact HG1. Mass analyses revealed that peaks j and k harbored heterodimeric peptides consisting of an intact chain (19-mer) and 13-mer or 14-mer segments that were missing the N-terminal 6- or 5-amino acid residue, respectively (Table (Table3).3). Both heterodimeric peptides were confirmed to retain their anti-MRSA activity (Table (Table33 and Fig. Fig.4).4). The intact chain of these two heterodimers was then cleaved at the same sites as in the first cleavages (L6-L7 and A5-L6). As a result, another three types of digestion fragment were generated and detected in peaks g, h, and i (Fig. (Fig.3B3B and Table Table3).3). These truncated peptides demonstrated no anti-MRSA activity (Fig. (Fig.4).4). Therefore, it was concluded that the anti-MRSA activity of the HG1 sample detected after incubation with MMP-7 was attributable to the peptides corresponding to peaks j and k, as well as intact HG1. The MMP-7 scissile points of HG1 were indicated as arrows on its amino acid sequence (Fig. (Fig.3D).3D). Unlike HG1, HG34 was not affected by MMP-7 digestion during 1 h of incubation.
To identify protease activity in the CS of S. aureus and P. aeruginosa, we conducted a fluorescence-conjugated casein-based assay. Protease activity was detected as an increase in fluorescence. To determine the nature of these proteases, a panel of inhibitors that distinguish between the various major protease families was included. For S. aureus CS, its proteolytic activity was reduced by AEBSF (serine protease inhibitor), aprotinin (serine protease inhibitor), and E-64 (cysteine protease inhibitor), but was only minimally affected by pepstatin (acidic protease), bestatin (amino peptidase inhibitor), and EDTA (metalloprotease inhibitor) (Fig. (Fig.55 A). In the case of P. aeruginosa CS, its protease activity was inhibited by AEBSF, aprotinin, and EDTA (Fig. (Fig.5B).5B). Therefore, it was suggested that serine proteases such as trypsin and chymotrypsin were included in S. aureus and P. aeruginosa CS, and cysteine protease and metalloprotease were included in S. aureus or P. aeruginosa CS, respectively. Figure 5C and D show the anti-MRSA activities of five AMPs detected at 1 h after incubation with proteins extracted from the CS of S. aureus and P. aeruginosa, respectively. In this test, it was noted that less than 0.1% of MRSA was recovered from the HG1-containing samples, thereby indicating that the anti-MRSA activity of HG1 was only slightly affected by the extracellular proteins secreted from two types of microbes. In contrast, the anti-MRSA activities of the other four AMPs were severely impaired after 1 h of incubation with proteins from the CS of two bacteria. Additionally, upon the incubation of AMPs with CS proteins treated with a protease inhibitor cocktail, their antimicrobial activities were observed to improve compared to those of control samples. From these results, it has become apparent that the antimicrobial activities of four AMPs (all except HG1) were damaged by proteases in the microbial CS, although there is some probability that another unknown factor in bacterial CS may interfere with the antimicrobial activity of AMPs. MALDI mass analysis conducted with HG1 samples incubated with CS proteins revealed that a peptide fragment with a molecular mass of 3,571 Da was detected in the HG1 sample after 1 h of incubation with the CS proteins of both bacteria (Fig. 5E and F), and the intact form of HG1 (4,113 Da) was also detected in the HG1 sample incubated with the CS proteins of P. aeruginosa (Fig. (Fig.5F).5F). Predicting via comparison of the detected mass to the calculated one, somewhat surprisingly, we noticed that the digestion product was equivalent to the Khal peptide, which was studied previously in our laboratory (Tables (Tables11 and and4)4) (18). This unanticipated result led us to conduct the same experiment using Khal, in an effort to verify its resistance against proteins from two bacterial CS. As a result, it was confirmed that Khal maintained its intact form after 1 h of incubation with two bacterial CS proteins (Fig. (Fig.5G).5G). On the other hand, the other two peptides predicted from their detected masses (3,143.70 and 3,030.71 Da) (Table (Table4)4) were observed to exhibit no anti-MRSA activity (data not shown). Consequently, it was concluded that the anti-MRSA activity of the HG1 sample detected after 1 h of incubation with the proteins of S. aureus or P. aeruginosa CS was attributable either to Khal or to both Khal and the remaining HG1, respectively. In contrast, unlike HG1 and Khal, no mass peak was detected in the range of 3.0 to 4.5 kDa after the incubation of HG34 with two bacterial CS proteins (lower profiles in Fig. 5E and F). According to these results, we concluded that the N-terminal portion (K3-K4) of HG34, unlike HG1, made HG34 vulnerable to proteolytic degradation, thereby rendering its anti-MRSA activity abolished upon incubation with bacterial CS proteins.
We also evaluated the resistance of AMPs against the proteases detected in wound fluids obtained from human skin. We first attempted to identify the protease activity in HWF via casein zymography. A variety of inhibitors were used to determine the nature of the proteolytic activities in HWF. As shown in Fig. Fig.66 A, at least two types of proteases were detected in the HWF samples. The upper and lower ones were shown to be inhibited by aprotinin and EDTA, respectively. Therefore, it was concluded that HWF contained trypsin- and chymotrypsin-like proteases and metalloprotease. As in the experiments conducted with the bacterial CS proteins, the anti-MRSA activity of each peptide was assessed after 1 h of incubation with HWF. As shown in Fig. Fig.6B,6B, HG1 was demonstrated to maintain its anti-MRSA activity in the presence of HWF. In contrast, the other four AMPs were shown to abrogate their anti-MRSA activity following incubation with HWF. Additionally, when they were incubated with HWF that had been pretreated with a protease inhibitor cocktail, their anti-MRSA activity was shown to be improved. Therefore, we concluded that the anti-MRSA activities of four AMPs were affected by proteases occurring in HWF. Also, as in the cases of bacterial CS, it was surmised that their antimicrobial activity might be impaired by other unknown factors in HWF. Furthermore, HG1 and HG34 samples at different times after incubation with HWF were subjected to Tricine-SDS-PAGE analysis. As shown in Fig. Fig.6C,6C, HG1 was detected consistently during the entire incubation time. In contrast, HG34 was shown to be damaged within 30 min after incubation and disappeared from the gel thereafter. In the following experiment, a sample of HG1 incubated for 4 h with HWF was injected into a C18 HPLC column to determine definitively whether a discrete peak emerged at the same retention time as the original HG1. As can be seen from the HPLC profiles, the HWF sample incubated with HG1 generated a distinct peak corresponding to HG1 (lower profile in Fig. Fig.6D),6D), which was not detected in the control HWF sample, which contained no HG1 (upper profile in Fig. Fig.6D).6D). Collectively, our results demonstrated that HG1 could preserve its intact form over at least 4 h of incubation with HWF and was thereby capable of exerting its anti-MRSA activity in the presence of HWF.
A number of studies have been conducted in an attempt to develop a new AMP-based antibiotic drug with pharmaceutical potential. Although AMPs have several advantages over conventional antibiotics, in terms of both their mode of microbicidal action as well as their broad antimicrobial spectra, there remain sizeable obstacles such as protease instability, thereby preventing their therapeutic use (27, 46). Trypsin and chymotrypsin are notable endopeptidases, which demonstrate specificity for the cleavage of the C-terminal side of amino acids with cationic residues and hydrophobic residues, respectively. Therefore, cationic AMPs are undoubtedly good substrates for these enzymes, as the positively charged and hydrophobic amino acid residues are crucial to their antimicrobial activity. Furthermore, it has been reported that trypsin- and chymotrypsin-like proteases abound in the human skin. In addition to these enzymes, matrix metalloproteases (MMPs) are known to be constitutively expressed in human skin and upregulated upon wound infection (36, 44), which was also confirmed in this study (Fig. (Fig.6A).6A). There is a strong possibility that AMPs administered onto skin infection sites will lose their therapeutic activity as the result of MMP degradation. Accordingly, it has been suggested that the susceptibility of AMP to proteolytic degradations by these three enzymes (trypsin, chymotrypsin, and MMP) may pose a major limitation to its use as a human therapeutic modality for the treatment of skin infections (10, 20, 42), although the peptide showed evidence of excellent in vitro antimicrobial property, to a degree commensurate with that of an effective antibiotic drug. In an effort to resolve this problem, a variety of strategies have been proposed (28), including the replacement of l-amino acids with d-amino acids (7, 9), the use of peptidomimetics (8), and selective fluoridation of peptides (31). However, much remains to be elucidated with regard to the pharmacokinetics, antimicrobial activity, and toxicity of AMP derivatives designed via the use of unusual amino acids. Therefore, controversy continues to rage over whether these modifications could improve the drug potential of certain AMPs. The cyclization of linear AMP has also been attempted in order to augment its protease resistance. This modification was achieved via the formation of an intramolecular disulfide bond between two Cys residues inserted or substituted into the (near) N and C terminals of AMPs (2, 35, 43). The effects of cyclization on protease resistance and antimicrobial activity have been extensively examined for an indolicidin analog (CP-11) (35). As compared to the linear peptide, the cyclic derivative of CP-11 was confirmed to be resistant to trypsin digestion, although it showed a slight reduction in antimicrobial activity. It was suggested that the cyclization of AMPs might prove an appropriate strategy to ensure higher protease stability and to improve the potential utility of AMPs. However, no subsequent studies have, until now, been reported on drug development using cyclic AMPs.
Rather than designing candidate molecules via somewhat troublesome and as-yet-uncertain modifications, the finding of natural AMP that is inherently resistant to proteases may provide a feasible prototype for the development of an effective therapeutic antibiotic. In this context, dimeric AMPs consisting of two peptide chains connected by an interdisulfide bond have recently emerged as convincing templates with higher protease resistance. Furthermore, it has been reported that dimerized AMP analogs show evidence of more favorable antimicrobial activity than do monomers. They also demonstrated enhanced binding to microbial membranes and greater pore formation capability (6, 45). Consequently, dimeric AMPs including synthetic derivatives were confirmed to have lower MIC values and to kill microbes much more quickly than their monomeric versions (5, 6, 16, 17). Dimeric AMPs have, thus far, rarely been found in nature. They include distinctin (1), CRS peptides (13), cathelicidin CAP11 (48), PMAP36 (38), β-defensin-related peptide (Defr1) (5), dicynthaurin (21), and halocidin (15). The protease resistance of dimeric AMP was first described for distinctin, a heterodimeric peptide consisting of two different chains of 22 and 25 amino acid residues (1, 33). Distinctin, but not each monomeric analog nor the two homodimeric analogs, was demonstrated to be resistant to trypsin digestion (5). Distinctin has been shown to be able to aggregate in a stable oligomeric form via intermolecular noncovalent interactions, and this was not observed with the monomeric or homodimeric peptides. Accordingly, it was proposed that this compact quaternary structure of distinctin is responsible for trypsin resistance.
Among the 7 above-mentioned natural dimeric AMPs, halocidin is the smallest peptide and HG1 is a homodimeric version of a 19-mer peptide designed via the addition of a lysine (K) residue to the N terminus of the longer subunit of halocidin. Previously, HG1 was shown to kill microbes within 30 s (17) and to have a broad antimicrobial spectrum (16, 17), as has also been reported in studies conducted with other cationic AMPs (11, 19). In this study, it was determined that HG1 could retain its active form even after 1 h of trypsin, α-chymotrypsin, or MMP-7 digestion. Furthermore, we demonstrated that the antimicrobial activity of HG1 was resistant to proteases occurring in the fluid from human skin wounds, as well as in the CS of two bacterial strains. We first reasoned that the resistance of HG1 to proteolytic degradation may be attributed to a compact oligomeric structure formed by noncovalent interactions between peptide chains in aqueous solution, as in the case of distinctin. However, via the size exclusion chromatography analysis that was conducted in order to confirm our assumptions, we noted that HG1 and HG34 were not eluted at retention times corresponding to those of oligomeric peptides (data not shown). Therefore, it was concluded that the possible oligomerization of dimeric AMP did not occur with HG1 and also that it was not responsible for the protease resistance of HG1. As noted from the amino acid sequence of HG1, each chain of HG1 harbors only one cleavage site (K15-G16) for endopeptidic digestion by trypsin. Furthermore, this scissile bond appears to be protected to some degree against trypsin attack, as it is located near an intermolecular disulfide bond which could lead to more compact packing of nearby side chains. Indeed, in the case of HG34 with two additional tryptic scissile bonds near the N terminus, it was confirmed that trypsin attack was initiated at sites remote from the disulfide bond (Fig. (Fig.1C1C and Table Table3).3). Likewise, it was demonstrated that the chymotryptic scissile site (L11-N12) located near the disulfide bond was not affected upon the treatment of HG1 and HG34 with α-chymotrypsin. Additionally, the fact that intra- or interdisulfide bonds are associated with the protease stability of the peptide was previously described in a study of cyclic derivatives of an indolicidin analog (35) and Defr1 (5). The N-terminal part of HG1 or HG34, which is comprised of four residues, was confirmed to be a requisite for their antimicrobial activity. This was corroborated by the following two results: First, anti-MRSA activity was not detected in peak a3 and r peptides (Table (Table3),3), as well as in two peptides with masses of 3,143 and 3,030 Da (Table (Table4),4), the N-terminal parts of which were impaired and/or removed from both chains. Second, in the case of fragments of HG1 (peak a2, e, j, and k peptides) and HG34 (peak q, s, t, and u peptides) that conserved their intact N-terminal parts in one of two chains, they all demonstrated anti-MRSA activity (Table (Table3).3). Therefore, it was considered that the fortitude of those parts against proteolytic degradation would become a favorable feature, which would enable these peptides to maintain their antimicrobial activity upon protease attack. In the case of HG1, it was observed that at least one of its chains could preserve its N-terminal part for up to 1 h of incubation with trypsin, α-chymotrypsin, and MMP-7, despite the fact that it harbors two cleavage sites (W2-L3 and L3-N4) for α-chymotrypsin. In contrast, the N-terminal part of HG34 was readily affected upon treatment with trypsin; thus, its antimicrobial activity was damaged quite rapidly compared to that of HG1. Accordingly, it was proposed that this vulnerable part of HG34 might account for the susceptibility of its antimicrobial activity to trypsin-like protease found in bacterial CS and HWF. Collectively, it was determined that the protease resistance of HG1 is attributable to three structural features: (i) the dimeric structure owing to an interdisulfide bond, (ii) the position of a basic residue (K) on its amino acid sequence, and (iii) the relative rigidity of the N-terminal part to protease attack.
In conclusion, as a cationic AMP which harbors neither unusual amino acids nor modified molecules, HG1 demonstrated more profound protease resistance in the presence of HWF and bacterial CS proteins, thereby suggesting that it might be used to overcome a serious problem that currently prevents the clinical usage of AMP. Taken together with previous studies that have demonstrated its profound antimicrobial activity and broad spectrum (16, 17), the present study strongly bolstered our expectations that HG1 might become a novel therapeutic agent for the treatment of superficial infections caused by antibiotic-resistant microbes.
This work was supported by a grant from the World-Class 2030 Project of Hoseo University.
Published ahead of print on 12 April 2010.