N-Terminally Modified Linear and Branched Spermine Backbone Dipeptidomimetics against Planktonic and Sessile Methicillin-Resistant Staphylococcus aureus

Abstract

Toward the discovery of useful therapeutic molecules, we report the design and synthesis of a focused library of new ultrashort N-terminally modified dipeptidomimetics, with or without modifications in the spermine backbone leading to linear (series 1) or branched (series 2) tryptophans, as antimicrobial agents. Eight peptidomimetics in the library showed good antibacterial activity (MICs of 1.77 to 14.2 μg/ml) against methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-resistant Staphylococcus epidermidis bacterial strains. Tryptophan fluorescence measurements on artificial bacterial or mammalian mimic membranes and assessment of the MRSA potential depolarization ability of the designed compounds revealed membrane interactions dependent on tryptophan positioning and N-terminal tagging. Among active peptidomimetics, compounds 1c and 1d were found to be nonhemolytic, displaying rapid bactericidal activity (at 4× MIC) against exponentially growing MRSA. Further, scanning electron microscopy of peptidomimetic 1c- and 1d-treated MRSA showed morphological changes with damage to cell walls, defining a membrane-active mode of action. Moreover, peptidomimetics 1c and 1d did not induce significant drug resistance in MRSA even after 17 passages. We also investigated the activity of these molecules against MRSA biofilms. At sub-MIC levels (∼2 to 4 μg/ml), both peptidomimetics inhibited biofilm formation. At concentrations higher than the MIC (35 to 140 μg/ml), peptidomimetics 1c and 1d significantly reduced the metabolic activity and biomass of mature (24-h) MRSA biofilms. These results were corroborated by confocal laser scanning microscopy (live/dead assay). The in vitro protease stability and lower cytotoxicity of peptidomimetics against peripheral blood mononuclear cells (PBMCs) support them being novel staphylocidal peptidomimetics. In conclusion, this study provides two peptidomimetics as potential leads for treatment of staphylococcal infections under planktonic and sessile conditions.

INTRODUCTION

Infectious diseases represent a major global health care concern due to escalating multidrug resistance (MDR) against currently available antibiotics (1). Multidrug-resistant strains such as methicillin-resistant Staphylococcus aureus (MRSA), vancomycin (VAN)-resistant enterococci (VRE), and carbapenem-resistant Enterobacteriaceae (CRE) in communities and nosocomial environments are rendering antibiotic therapy more difficult and costly at an unprecedented rate (2, 3). The development of resistance is aggravated by the irrational use of antibiotics in livestock and medical practices, which has armed microbes with a multitude of novel drug resistance mechanisms. In the present scenario, no class of antibiotics with a fixed metabolic target in microbes is free from the resistance development problem, as microbes are able to reinvent themselves, acquiring gene-encoded or plasmid-mediated drug resistance leading to better survival chances. A passive known contributory lifestyle approach toward resistance development involves slow growth and a heterogeneous microbial population, phenotypically as well as genetically, in the form of biofilms (4, 5).

Biofilms are microbial communities adhering to surfaces or floating at air-water interphases in which the microbes are embedded in a self-produced exopolymeric substance (EPS), which is composed largely of proteins, DNA, and different extracellular polysaccharides (6). Novel agents and strategies are needed to eradicate biofilms, as they play a major role in almost 80% of infections, including cystic fibrosis, dental plaques, chronic wounds, and infections involving implanted medical devices (7). Most antibiotics are active against log-phase bacterial cells, as they target metabolic processes in bacteria to inhibit growth. However, biofilms act as recalcitrant infection reservoirs and contribute to virulence, since the exopolymeric matrix and retarded metabolic activity inside biofilm communities lead to increased persistence of biofilms (8). Additionally, it is known that bacteria in biofilms generally tolerate antibiotic treatment, and antibiotics can even produce a trigger for biofilm formation (9).

As an answer to MDR microbes, host defense cationic peptides (HDCPs) (12- to 60-mers) and their mimics, with a multitude of novel mechanisms, are commercial candidates that hold potential to circumvent drug resistance (10, 11). HDCPs are produced by almost all living organisms as a first line of defense against invading microbes. Owing to global amphiphilicity, i.e., the balance between positive charge at physiological pH and hydrophobicity, HDCPs predominantly exhibit membrane-disruptive modes of action, although they have also been reported to be metabolic inhibitors in microbes (11). The positive charge on HDCPs helps them to become attracted to negatively charged surfaces of bacterial cells, facilitating primary interactions. After initial attachment, by virtue of their amphipathic nature, HDCPs are able to cause lipid clustering and segregation of domains, leading to bacterial cell death (10). It is difficult for bacteria to develop resistance to HDCPs because most HDCPs kill bacterial cells quickly through their actions on the entire cytoplasm, acting as pore formers (11). HDCPs have been reported to efficiently eradicate slow-growing cells from planktonic and biofilm cultures and thus have been proposed as promising alternative agents for the cure of biofilm-associated multidrug-resistant infections (12). However, the challenges in the application of HDCPs have been their high cost, protease instability, reduced activity in the presence of salts, and poor bioavailability (13). Over the past decades, attempts have been made to mimic the structures and functions of HDCPs, leading to the design of potent synthetic mimics such as oligoacyllysines (OAKs), cationic steroid antibiotics (CSAs), and cyclic cationic peptides, some of which are presently undergoing clinical trials as antibacterial agents (14, 15).

The aim of the present study was to optimize our previously designed N-terminally tagged dipeptide spermidine template (16). Toward this goal, the roles of hydrophobicity and charge distribution in activity have been assessed with different positioning of tryptophan residues on the spermine backbone. Furthermore, the mode of action and efficacy of the lead molecule to eradicate clinically relevant MRSA biofilms have been determined.

MATERIALS AND METHODS

Chemicals.

9-Fluorenylmethoxy carbonyl (Fmoc)-protected amino acids and resins were purchased from Novabiochem (Darmstadt, Germany), and N,N-diisopropylcarbodiimide (DIPCDI) (catalog no. D12,540-7), 1-hydroxybenzotrizole (HOBt) (catalog no. 54804), diisopropylethylamine (DIPEA) (catalog no. D-3887), N-methylpyrrolidinone (NMP) (catalog no. 494496), piperidine (catalog no. 411027), spermine (catalog no. S3256), triisopropylsilane (catalog no. 23378-1), crystal violet (CV) (catalog no. C3886), glucose (catalog no. G7528), hydrazine (catalog no. 225819), 3,3′-dipropylthiadicarbocyanine iodide (DiSC₃5) (catalog no. 43608), and the Tox7 kit (lactate dehydrogenase [LDH] release assay kit) were obtained from Sigma-Aldrich. Trifluoroacetic acid (TFA) (catalog no. 80826005001730) and 2-acetyldimedone (Dde-OH) (catalog no. 8.51015.0005) were purchased from Merck. All of the moieties used as N-terminal tags were purchased from Sigma-Aldrich. Tryptone soy broth (TSB) (catalog no. M011-500G) was purchased from HiMedia (India), and Mueller-Hinton broth (MHB) and Mueller-Hinton agar were purchased from Difco (Franklin Lakes, NJ). The alamarBlue reagent (catalog no. DAL 1025) and the Molecular Probes Live/Dead BacLight assay kit (L7012) were procured from Invitrogen (Eugene, OR). High-performance liquid chromatography (HPLC)-grade solvents were obtained from Merck (Germany). Dimethylformamide (DMF) and dichloromethane (DCM) were obtained from Merck (Mumbai, India). DMF was double distilled prior to its use.

Synthesis and purification of peptidomimetics.

All peptidomimetics were synthesized on 2-chlorotrityl chloride resin as a solid support, as described previously, with slight modifications (17). Briefly, on preswelled resin, 5 eq of spermine dissolved in dichloromethane was added under an inert atmosphere for 4 h. Completion of the reaction was monitored by positive Kaiser test results (18). After coupling, capping of unreacted resin with methanol was performed for 45 min. The primary amino group of spermine was protected through overnight reaction with 2 eq of Dde-OH in DMF. After protection of the primary amino group, the secondary amino groups were protected through reaction for 4 h with 6 eq of t-butoxycarbonyl (Boc)-anhydride in the presence of DIPEA. Then Dde-OH protection of primary amines was removed using 2% (wt/vol) hydrazine in DMF. Two additional couplings were performed with Fmoc-Trp(Boc)-OH in the presence of HOBt and DIPCDI in DCM-DMF (1:1). The N-terminal tagging was performed with 4 eq of unnatural tag, HOBt, and DIPCDI in DCM-DMF (1:1), leading to peptidomimetics 1a to 1f (Fig. 1). For synthesis of peptidomimetics 2a to 2f, Dde-OH-protected resin was coupled with 4 eq of Boc-Trp(Boc)-OH, HOBt, and DIPCDI. Then, deprotection of the primary amino group was performed with 2% (wt/vol) hydrazine in DMF. The N-terminal tagging was performed by a procedure similar to that described above. Final deprotection of peptidomimetics from the resin in both series was performed using a cleavage cocktail (DCM, TFA, ethanedithiol, triisopropylsilane, phenol, and water in a ratio of 65:30:2:1:1:1). The solution was filtered, and cold ether was added to the filtrate to precipitate the crude product, which was filtered and washed with cold ether (2 × 25 ml). The solid was dissolved in methanol and desalted using an LH-20 Sephadex column (Sigma). The peptidomimetics were further purified by reverse-phase (RP)-HPLC, using a semipreparative column (7.8 by 300 mm, 125-Å pore size, 10-μm particle size) with a gradient of 10 to 90% buffer 2 over 45 min; buffer 1 was water with 0.1% TFA and buffer 2 was acetonitrile with 0.1% TFA. After purification, the peptidomimetics were confirmed by either liquid chromatography-tandem mass spectrometry (LC-MS/MS) (Quattro Micro API; Waters) or ultra-high-performance liquid chromatography (UHPLC) (Dionex, Germany) with LTQ Orbitrap XL (Thermo Fisher Scientific) mass determination. Analytical HPLC traces and mass spectra of representative peptidomimetics are provided in Fig. S1 and S2 in the supplemental material.

FIG 1.

Reagents and conditions. Reaction 1, 5 eq spermine in DCM, 3 h; reaction 2, methanol, 30 min; reaction 3, 2 eq Dde-OH in DMF, overnight; reaction 4, 6 eq (Boc)₂O in DCM-DMF (1:1), 3 h; reaction 5, Boc-Trp(Boc)-OH, HOBt, and DIPCDI in DCM-DMF (1:1), overnight; reaction 6, 2% hydrazine in DMF; reaction 7, Fmoc-Trp(Boc)-COOH, HOBt, and DIPCDI in DCM-DMF (1:1), 1.5 h; reaction 8, 20% piperidine in DMF; reaction 9, 3 eq R-COOH, HOBt, and DIPCDI in DCM-DMF (1:1), overnight; reaction 10, 30% TFA in DCM.

Antibacterial activity under planktonic conditions.

The antibacterial activities of the designed peptidomimetics were evaluated by using a modified serial broth dilution method, as reported previously (19, 20). The following bacterial strains were used in this study: S. aureus (ATCC 29213), methicillin-resistant S. aureus (ATCC 33591), Staphylococcus epidermidis (ATCC 12228), methicillin-resistant S. epidermidis (ATCC 51625), Enterococcus faecalis (ATCC 7080), Escherichia coli (ATCC 11775), and Acinetobacter baumannii (ATCC 19606). The inocula were prepared from mid-log-phase bacterial cultures. Each well of the first 11 columns of 96-well polypropylene microtiter plates was inoculated with 100 μl of approximately 10⁵ CFU/ml of bacterial suspension in Mueller-Hinton broth (MHB) (Difco). Then 10 μl of serially diluted peptidomimetic in 0.01% (vol/vol) acetic acid and 0.2% bovine serum albumin (Sigma), over the desired concentration range, was added to the wells of the microtiter plates. The microtiter plates were incubated overnight at 37°C, with agitation (200 rpm). After 18 h, absorbance was measured at 630 nm. Cultures without test peptidomimetics were used as positive controls. Uninoculated MHB was used as a negative control. Tests were carried out in duplicate on three different days. MIC was defined as the lowest concentration of peptidomimetic that completely inhibited growth. For comparison purposes, the standard peptide antibiotics VAN and polymyxin B (PMB) were assayed under identical conditions. The antibacterial activities of peptidomimetics and the standard antibiotic VAN were evaluated against MRSA strain 33591 in the presence of 25% (vol/vol) human serum and fetal bovine serum (FBS) in biofilm growth medium (tryptone soy broth [TSB] supplemented with 0.5% NaCl and 0.25% glucose). A protocol similar to that described previously was used (21). Briefly, MHB was adjusted to 25% (vol/vol) of a heat-inactivated human serum pool obtained from two healthy volunteers. Growth control experiments were conducted using MHB with and without 25% serum. MICs were determined as described above, according to CLSI standard methods.

Hemolytic activity.

The hemolytic activities of the peptidomimetics were evaluated with human red blood cells (hRBCs). Briefly, 100 μl of a fresh 4% (vol/vol) suspension of hRBCs in NaCl-P_i (35 mM phosphate buffer [35 mM Na₂HPO₄ and 35 mM NaH₂PO₄·2H₂O], 150 mM NaCl [pH 7.2]) was placed in a 96-well plate. After incubation of the peptidomimetics (100 μl) in the hRBC suspension for 1 h at 37°C, the plates were centrifuged, and the supernatant (100 μl) was transferred to a fresh 96-well plate. Absorbance was read at 540 nm using an enzyme-linked immunosorbent assay (ELISA) plate reader (Molecular Devices). Percent hemolysis was calculated using the following formula: % hemolysis = 100[(A − A₀)/(A_t − A₀)], where A represents the absorbance of sample wells at 540 nm and A₀ and A_t represent 0% and 100% hemolysis, respectively, determined in NaCl-P_i with 1% Triton X-100.

Cytotoxicity assay in peripheral blood mononuclear cells.

Blood from healthy human donors was collected in tubes containing the anticoagulant sodium heparin, in accordance with institutional guidelines. The blood was diluted 1:1 with NaCl-P_i (35 mM phosphate buffer, 150 mM NaCl [pH 7.2]). Blood cells were separated over Histopaque separation medium (Sigma-Aldrich) by centrifugation at 1,200 rpm for 30 min. The peripheral blood mononuclear cells (PBMCs) were collected and washed twice with NaCl-P_i (35 mM phosphate buffer, 150 mM NaCl [pH 7.2]). The cells were then resuspended in complete RPMI 1640 medium (HiMedia) supplemented with 10% FBS (Sigma) and were quantified by trypan blue exclusion, with microscopic assessment. PBMCs (1 × 10⁶ cells/ml) in complete medium were seeded in a 24-well plate and left in the incubator in 5% CO₂ for 2 h at 37°C. The cells were then treated with peptidomimetic 1c, peptidomimetic 1d, or VAN at the desired concentrations (20 μg/ml and 50 μg/ml). Triton X-100 (2%) was used as a negative control. After 24 h of incubation, the contents of each well were transferred to sterile 1.5-ml Eppendorf tubes, and cells were pelleted at 2,000 rpm for 10 min. The supernatants were assessed for the release of LDH by using the Tox7 kit (Sigma), as described previously (22, 23). The experiments were carried out in duplicate on three different days, and data are presented as mean ± standard deviation (SD).

Tryptophan fluorescence.

Small unilamellar vesicles (SUVs), which were prepared following the standard method (20), were used for the experiment. Briefly, dry lipids 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and 1,2-dipalmitoyl-sn-glycero-3-phospho(1′-rac-glycerol) (sodium salt) (DPPG) (7:3 [wt/wt]) to mimic bacterial membranes or DPPC to mimic mammalian membranes were dissolved in a chloroform-methanol mixture in a 150-ml round-bottom flask. The solvent was removed with a stream of nitrogen gas, to allow formation of a thin lipid film on the walls of the glass vessel. The lipid film thus obtained was lyophilized for 6 h to remove traces of solvent. Dried thin films were resuspended in 10 mM Tris buffer (0.1 mM EDTA, 150 mM NaCl [pH 7.4]) preheated at 60°C, with vortex mixing. The lipid dispersions were then sonicated on ice for 15 to 20 min using a titanium-tip ultrasonicator, with burst and rest times of 30 s and 10 s, respectively, until the solutions became opalescent. Titanium debris was removed by centrifugation. Each peptidomimetic (final concentration, 5 μg/ml) was added to 500 μl of 10 mM Tris buffer (0.1 mM EDTA, 150 mM NaCl [pH 7.4]) or 0.5 μg/ml bacterial or mammalian mimic SUVs, and the peptidomimetic-lipid mixture was allowed to interact at 25°C for 2 min in a cuvette. The fluorescence measurements were performed with a Fluorolog spectrofluorometer (Jobin Yuvon, Horiba, Japan). Samples were excited at 280 nm, and the emission was scanned from 300 to 400 nm, with a 5-nm slit width for both excitation and emission. The experiment was repeated twice on the same day, and representative data are presented here.

Membrane depolarization.

The evaluation of membrane depolarization of MRSA was performed as described previously (16). Briefly, MRSA that had been grown overnight was subcultured in MHB for 2 to 3 h at 37°C to obtain mid-log-phase cultures. The cells were centrifuged at 4,000 rpm for 10 min at 25°C, washed, and resuspended in respiration buffer (5 mM HEPES, 20 mM glucose [pH 7.4]) to obtain a diluted suspension of optical density at 600 nm (OD₆₀₀) of ∼0.05. The membrane potential-sensitive dye 3,3′-dipropylthiadicarbocyanine iodide (DiSC₃5) (0.18 μM in dimethyl sulfoxide [DMSO]) was added to 500-μl aliquots of resuspended cells, and the mixtures were allowed to equilibrate for 1 h. Baseline fluorescence was assessed using an Edinburg F900 spectrofluorometer, with excitation at 622 nm and emission at 670 nm, in a cuvette with a 1-cm path length. A bandwidth of 5 nm was employed for excitation and emission. Subsequently, increasing concentrations of test peptidomimetics were added to the equilibrated cells, and the increase in fluorescence resulting from dequenching of the DiSC₃5 dye was measured every 2 min, to obtain the maximal depolarization. Increases in relative fluorescence units (RFU) were plotted against increasing concentrations of different peptidomimetics or PMB.

Bactericidal kinetics.

The kinetics of bacterial killing of a MRSA strain (ATCC 33591) by peptidomimetics at 2× MIC and 4× MIC were determined and compared with those of VAN as described previously (24). Log-phase bacteria (1.2 × 10⁷ to 3.0 × 10⁷ CFU/ml) were incubated with peptidomimetic 1c, peptidomimetic 1d, or VAN at 2× MIC or 4× MIC in MHB. Aliquots were removed after 0.5, 1, 2, 3, and 6 h and diluted in sterile normal saline solution before plating on Mueller-Hinton II agar; CFU were counted after 24 h of incubation at 37°C. To decrease the limit of detection, larger aliquots were removed and centrifuged to remove antibacterial agent carryover. The experiment was repeated on three different days, and curves were plotted for log₁₀ CFU/ml versus time.

Scanning electron microscopy.

For electron microscopy, samples were prepared by following previously reported protocols (16, 25). Briefly, freshly inoculated MRSA (ATCC 33591) was grown on MHB up to an OD₆₀₀ of ∼0.5 (corresponding to 10⁸ CFU/ml). Bacterial cells were then centrifuged at 4,000 rpm for 15 min, washed three times with NaCl-P_i (10 mM phosphate buffer, 150 mM NaCl [pH 7.4]), and resuspended in an equal volume of NaCl-P_i. For scanning electron microscopy (SEM) experiments, larger bacterial inocula (10⁸ CFU/ml) were used; therefore, the cells were incubated with test peptidomimetic 1c, peptidomimetic 1d, or VAN at 10× MIC for 30 min. Controls were run in the absence of antibacterial agents. After 30 min, the cells were centrifuged and washed three times with NaCl-P_i. For cell fixation, the washed bacterial pallet was resuspended in 0.5 ml of 2.5% paraformaldehyde in NaCl-P_i and incubated overnight at 4°C. After fixation, cells were centrifuged, washed twice with 0.1 M sodium cacodylate buffer, and fixed with 1% osmium tetraoxide in 0.1 M sodium cacodylate buffer for 40 min at room temperature (RT) in the dark. The samples were then dehydrated in a series of graded ethanol solutions (30% to 100%) and finally dried in desiccators under reduced pressure. Upon dehydration, the cells were air dried for 15 min at RT in the dark after immersion in hexamethyldisilazane. An automatic sputter coater (Quorum SC7640) was used to coat the specimens with gold particles at a thickness of 30 Å. Then samples were imaged via scanning electron microscopy (Zeiss EVO LS15).

Drug resistance study.

The initial MICs against MRSA of peptidomimetics and the control antibiotics VAN and ciprofloxacin (CIP) were determined as described above. Bacterial suspensions (100 μl) from duplicate wells at sub-MIC concentrations were then used to inoculate fresh cultures. The cultures was grown to yield approximately 10⁵ CFU/ml for the next experiment. These bacterial suspensions were then incubated with the desired concentrations of antibacterial agents for 18 h to determine new MICs. The same subculturing protocol was used for the next 16 passages, and MICs were determined using OD₆₃₀ values as described previously (23).

Biofilm susceptibility assay.

For the biofilm inhibition assay, the standard protocol was used as reported previously (26). Briefly, freshly inoculated MRSA (ATCC 33591) was grown overnight on biofilm growth medium (TSB supplemented with 0.5% [wt/vol] NaCl and 0.25% [wt/vol] glucose). The next day, the cultures were diluted to 10⁵ CFU/ml in fresh biofilm growth medium. Two hundred microliters of diluted culture was dispensed into wells of a 96-well polystyrene plate for biofilm formation. To evaluate the inhibition of biofilm formation, antibacterial agents at the planktonic MIC in biofilm medium (MIC_b) and sub-MIC_b concentrations were added initially to diluted cultures following incubation at 37°C without shaking. Another set of experiments was performed with the addition of fresh medium containing antibacterial agents at 10× MIC_b and 20× MIC_b to 24-h-preformed biofilm, after gentle washing with sterile NaCl-P_i (35 mM phosphate buffer, 150 mM NaCl [pH 7.4]). Biofilm cultures were reincubated at 37°C for 24 h. After removal of the medium, the biofilms were washed twice with sterile NaCl-P_i and assessed for metabolic activity (alamarBlue assay) and biomass quantities (crystal violet assay), as follows.

For determination of metabolic activity, the plates were sonicated in a ultrasonic bath (Elmasonic, Germany) for 5 min at 37°C, with sonication at 30 kHz, to ensure detachment of bacteria from the biofilms before the addition of 10% (vol/vol) alamarBlue reagent (according to the manufacturer's instructions). The plates were further incubated at 37°C for 2 h. After 2 h, absorbance was measured at 570 nm and 600 nm, and the percent reduction of alamarBlue (cell viability) was calculated by using a formula provided in the manufacturer's protocol. The experiment was repeated three times on three different days, and results are given as mean ± SD.

For biomass quantification, the crystal violet (CV) staining protocol was used as reported previously (27). Slime and adherent cells were fixed for 20 min with 1 ml of 99% methanol and then stained for 20 min with 200 μl of 0.1% crystal violet. Excess stain was removed by washing the coverslips with NaCl-P_i, and then the coverslips were air dried. The stained dye was redissolved with the addition of 33% acetic acid and incubation for 1 h at room temperature without shaking. The optical density at 570 nm (OD₅₇₀) was measured spectrophotometrically, and data are presented as percent biomass in comparison with the positive control.

Confocal laser scanning microscopy of biofilms.

For confocal microscopy, biofilm formation was induced on glass coverslips in a 6-well plate, following a reported procedure (27). Briefly, overnight cultures of MRSA were diluted to 10⁵ CFU/ml, and 3-ml volumes of this suspension were used to grow biofilms on glass coverslips in the wells of a 6-well plate at 37°C. Biofilm growth conditions and treatment of biofilms with antibacterial agents were as described above for the alamarBlue and crystal violet assays. Then the coverslips were washed twice with sterile NaCl-P_i and stained with reagent from the Molecular Probes Live/Dead kit (Invitrogen, Eugene, OR), following the manufacturer's instructions. This stain contains the DNA-binding dyes SYTO 9 (green fluorescence) and propidium iodide (PI) (red fluorescence). When used alone, SYTO 9 stains all bacteria in a population, i.e., those with intact or damaged membranes. In contrast, PI penetrates only bacteria with damaged membranes, causing a reduction in the SYTO 9 staining (green fluorescence). The biofilms were examined with an Olympus FluoView FV1000 confocal laser scanning microscope. For detection of SYTO 9 (green channel) and PI (red channel), 488-nm and 561-nm lasers, respectively, were used. For measurement of biofilm depths, z-stack images were acquired at approximately 0.4-μm intervals, using a 100× HCX PL APO oil immersion lens (numerical aperture, 1.2); image analyses and export were performed with FV10-ASW-1.7 software. For each sample, at least five different regions on a single coverslip were scanned. The experiment was repeated three times on three different days, and representative data are presented here.

RESULTS

Rational design and synthesis of peptidomimetics.

Antimicrobial peptidomimetics based on the defined pharmacophore with at least +2 charges at physiological pH and hydrophobicity have been designed by various research groups (28, 29). Recently, we reported a small series of potent peptidomimetics with broad-spectrum antibacterial activity based on a template containing N-terminally tagged dipeptidomimetics conjugated with spermidine (16). In the present work, new N-terminal tags and cationic spermine at the C terminus were conjugated to the same template to expand and optimize the library. In the template peptidomimetic 1a, two Trp residues were attached to the spermine moiety. The hydrophobic bulk, aromatic π electron cloud, and lipid membrane anchorage ability of Trp residues have made Trp a suitable residue for incorporation in novel antibacterial peptidomimetics (28, 30). In peptidomimetics 1b to 1f (series 1), different N-terminal tags, i.e., caffeic acid, 4-(trifluoromethyl)phenylacetic acid, decanoic acid, lauric acid, and linoleic acid, were used to vary the relative hydrophobicity (Fig. 1). The peptidomimetics in series 2 (peptidomimetics 2a to 2f) were synthesized to investigate the effects of Trp positioning on the spermine backbone on activity and therapeutic index values. In peptidomimetics 2a to 2f, the secondary N atoms of spermine were coupled with the carboxylic acid end of Trp residues, leaving the alpha-amino group of Trp residues ionizable at physiological pH (Fig. 1). All of the designed peptidomimetics were >80% pure, and their masses were in the range of 575 to 850 Da (Table 1).

TABLE 1.

Purity, proportion of acetonitrile for RP-HPLC elution, and molecular masses of designed peptidomimetics

Peptidomimetic	Purity (%)	Acetonitrile (%)a	Mass ([M+H]+) (Da)
			Calculated	Observed
1a	95	17.41	575.3816	575.3808
1b	99	46.42	737.4133	737.4139
1c	95	54.72	761.4109	761.4110
1d	95	61.57	729.5174	729.5178
1e	95	65.21	757.5487	757.5489
1f	98	70.36	837.6113	837.6097
2a	80	12.30	575.3816	575.3815
2b	80	44.34	737.4133	737.4140
2c	83	49.85	761.4109	761.4118
2d	99	57.92	729.5174	729.5181
2e	99	62.63	757.5487	757.5495
2f	99	69.78	837.6113	837.6113

^aPercentage of acetonitrile for RP-HPLC elution.

Biological activities of designed peptidomimetics.

The antibacterial activities of the designed peptidomimetics against five Gram-positive bacterial strains and two Gram-negative bacterial strains were evaluated using the serial broth dilution method (Table 2). The template peptidomimetic 1a showed moderate activity against Gram-positive bacterial strains, while peptidomimetics 1b to 1f displayed good activity against Gram-positive bacterial strains, with MICs of <10 μg/ml against all tested strains except E. faecalis. Peptidomimetics in series 1 also showed activity against E. coli, with MICs in the range of 14.2 to 56.8 μg/ml. Similarly, in series 2, peptidomimetics 2a and 2b showed negligible growth inhibition against all tested bacterial strains up to 454.4 μg/ml, while peptidomimetic 2c showed moderate activity and peptidomimetics 2d to 2f exhibited good growth inhibition (MICs of 0.8 to 28.4 μg/ml) of all of the bacterial stains except A. baumannii. PMB showed relatively poor activity against Staphylococcus species, although it showed excellent growth inhibition of Gram-negative bacterial strains. VAN showed potent growth inhibition of Staphylococcus species but was ineffective against Gram-negative strains under the experimental conditions.

TABLE 2.

Antibacterial activities of peptidomimetics against Gram-positive and Gram-negative bacterial strains and cytotoxicity in blood cells

Peptidomimetic	MIC (μg/ml) of:							Hemolysis (%)b	LDH release (%)c
	S. aureus (ATCC 29213)	MRSA (ATCC 33591)	S. epidermidis (ATCC 12228)	MRSEa (ATCC 51625)	E. faecalis (ATCC 7080)	E. coli (ATCC 11775)	A. baumannii (ATCC 19606)
1a	113.6	227.2	113.6	NDd	454.5	ND	ND	4	ND
1b	3.5	7.1	3.5	7.1	113.6	14.2	ND	16	ND
1c	1.7	3.5	1.7	3.5	28.4	56.8	28.4	2	5.78
1d	1.7	1.7	1.7	1.7	3.5	14.2	113.6	9	17.5
1e	1.7	3.5	1.7	1.7	7.1	14.2	56.8	31	ND
1f	7.1	3.5	1.7	7.1	28.4	28.4	ND	30	ND
2a	>454.4	>227.2	>454.4	227.2	ND	>454.4	ND	0	ND
2b	>454.4	454.4	ND	ND	ND	>454.4	ND	5	ND
2c	14.2	28.4	7.1	14.2	ND	113.6	113.6	1	ND
2d	0.8	1.7	0.8	1.7	28.4	28.4	113.6	83	ND
2e	0.8	1.7	0.8	1.7	7.1	28.4	113.6	96	ND
2f	0.8	3.5	0.8	1.7	14.2	28.4	56.8	88	ND
PMB	14.2	28.4	7.1	28.4	113.6	0.4	ND	ND	ND
VAN	0.4	0.8	0.4	0.8	ND	113.6	56.8	ND	ND

^aMRSE, methicillin-resistant Staphylococcus epidermidis.

^bHemolysis at 250 μg/ml.

^cLDH release at 20 μg/ml.

^dND, not determined.

The cell selectivity of the designed peptidomimetics on enucleated hRBCs was evaluated (Table 2). Most of the peptidomimetics, including peptidomimetics 1a to 1d and 2a to 2c, were found to cause minimal hemolysis up to the maximal concentration tested of 250 μg/ml. Peptidomimetics 1e and 1f caused 31% and 30% hemolysis, respectively, at 250 μg/ml. Peptidomimetics 2d, 2e, and 2f caused significant hemolysis, leading to 83%, 96%, and 88% damage to hRBCs, respectively, at 250 μg/ml.

The antibacterial activities of nonhemolytic peptidomimetics 1c and 1d were also evaluated against MRSA in the presence of 25% (vol/vol) human serum or bovine serum. Fourfold and 8-fold increases in MICs were observed for peptidomimetics 1c and 1d, respectively, with human serum (Table 3).

TABLE 3.

Effects of salt concentrations and serum on antibacterial activities of compounds

Compound	MIC (μg/ml) against MRSA 33591 in:
	Biofilm medium (low salt)a	TSB with high saltb	MHB with human serumc	MHB with FBSd
Peptidomimetic 1c	7.1	28.4	14.1	7.1
Peptidomimetic 1d	3.5	7.1	14.1	7.1
VAN	0.8	1.7	0.8	0.8

^aTSB supplemented with 0.5% (wt/vol) NaCl and 0.25% (wt/vol) glucose (i.e., MIC_b).

^bTSB supplemented with 3% (wt/vol) NaCl and 0.5% (wt/vol) glucose.

^cMHB with 25% human serum added.

^dMHB with 25% FBS added.

The LDH release assay with PBMCs demonstrated 5.78% ± 6.58% and 17.56% ± 10.15% LDH release caused by peptidomimetics 1c and 1d, respectively, at 20 μg/ml. At 50 μg/ml, the release was 20.81% ± 5.4% and 21.62% ± 5.04% with peptidomimetics 1c and 1d, respectively.

Membrane insertion and depolarization potential of designed peptidomimetics.

Trp fluorescence was used as a probe to evaluate the effects of Trp positioning on the insertion depth of designed peptidomimetics in bacterial and mammalian mimic membranes. In buffer, all peptidomimetics showed fluorescence emission maxima in the range of 356 to 362 nm (Table 4). In bacterial mimic SUVs (DPPC-DPPG, 7:3 [wt/wt]), blue shifts in emission maxima in the range of 5 to 12 nm, concomitant with increases in fluorescence intensity, in comparison with buffer, were observed for all of the peptidomimetics in series 1. In mammalian mimic DPPC SUVs, blue shifts in the range of 1 to 7 nm were observed for peptidomimetics 1a to 1f. The emission maxima for series 2 peptidomimetics (peptidomimetics 2a to 2f) shifted more toward blue wavelengths than did peptidomimetics 1a to 1f in both bacterial mimic and mammalian mimic membranes. Noticeably, for peptidomimetics 2a to 2f in bacterial mimic membranes, significant blue shifts (2 to 16 nm) were observed subsequent to partitioning, and the concomitant increase in emission intensity was not observed for peptidomimetics 2a, 2c, and 2d, in comparison with buffer (see Fig. S3 in the supplemental material).

TABLE 4.

Tryptophan fluorescence emission maxima of designed peptidomimetics in buffer, DPPC SUVs, or DPPC-DPPG SUVs

Peptidomimetic	Emission maximum (nm)a
	Bufferb	DPPC	DPPC-DPPG (7:3 [wt/wt])
1a	361	358 (3)	351 (10)
1c	356	355 (1)	351 (5)
1d	362	355 (7)	350 (12)
1e	357	350 (7)	348 (9)
1f	356	352 (4)	351 (5)
2a	359	358 (1)	357 (2)
2c	360	354 (6)	353 (7)
2d	358	350 (8)	343 (15)
2e	357	347 (10)	341 (16)
2f	354	350 (4)	348 (6)

^aBlue shifts are indicated in parentheses.

^bThe buffer contains 0.1 mM EDTA and 150 mM NaCl (pH 7.4).

Next, the ability of the designed peptidomimetics to compromise the membrane potential in MRSA was evaluated by using the membrane potential-sensitive dye DiSC₃5. Upon partitioning in the membranes of live cells at sufficiently high concentrations, DiSC₃5 self-quenches its fluorescence. Under the influence of a membrane-depolarizing agent, there is dye release with a significant increase in dye fluorescence, which is measured fluorometrically. For peptidomimetics 1a and 2a, no significant increases in relative fluorescence units (RFU) were observed up to the maximum concentrations tested, suggesting an inability of these peptidomimetics to alter membrane potential at concentrations below the MIC (data not shown). For peptidomimetics 1c and 2c, with aromatic N-terminal tags, only marginal changes in RFU were observed up to the highest concentrations tested (Fig. 2). Intermediate changes in fluorescence intensity were observed for peptidomimetics 1d and 2d, whereas significant changes in RFU were observed for peptidomimetics 1e, 1f, 2e, and 2f. The increases in fluorescence with lipid-tagged peptidomimetics were concentration dependent up to 9.9 μg/ml and then were saturated, resulting in plateau-like dose-response curves. The experiment was repeated twice on two consecutive days, with similar results. Representative results from one assay are presented here. Furthermore, interaction studies were performed with peptidomimetics 1c and 1d, which are active and cell-selective peptidomimetics from series 1.

FIG 2.

Concentration-dependent cell membrane depolarization assessed with the potential-sensitive dye DiSC₃5.

Bactericidal kinetics and membrane-disruptive mode of action.

Bactericidal kinetic experiments with peptidomimetic 1c, peptidomimetic 1d, and VAN at 2 times and 4 times their respective planktonic MICs were performed with exponentially growing S. aureus ATCC 33591 (Fig. 3). At 2× MIC, both peptidomimetics produced ≥3-log₁₀ CFU/ml reductions within 3 h of incubation; at 4× MIC, bactericidal effects with >4-log₁₀ CFU/ml reductions within 30 min of incubation were observed. VAN did not show any bactericidal activity even upon incubation at 4× MIC under the same conditions and produced only 2-log₁₀ CFU/ml reductions over a period of 6 h. The lower limit of detection was determined to be 100 CFU/ml, and bactericidal activity was defined as a 3-log₁₀ CFU/ml decrease, in comparison with the time zero value.

FIG 3.

Time-kill curves for S. aureus strain ATCC 33591 incubated with 2× MIC (A) or 4× MIC (B) levels of peptidomimetic 1c, peptidomimetic 1d, or VAN and sampled at the indicated time points. The curves were plotted for log₁₀ CFU/ml versus time as described in Materials and Methods. The data shown are from one of three independent experiments with similar results.

Further, we visualized the effects on MRSA of 30-min incubations with peptidomimetic 1c, peptidomimetic 1d, and VAN, at 10× MIC, using SEM. Control MRSA cells exhibited a bright, smooth appearance, with intact cell membranes (Fig. 4A). Peptidomimetic 1c treatment caused rough damaged surfaces, cell bursting, leakage, and string-like substances, which are considered to be cellular debris arising from cell lysis (Fig. 4B). Peptidomimetic 1d-treated cells appeared distorted, with depressions and hole formation (Fig. 4C), indicating the membrane-active mode of action for the designed peptidomimetics. Surprisingly, VAN-treated cells mostly retained their smooth appearance, albeit with slight deformations in shape, compared with control cells (Fig. 4D).

FIG 4.

Scanning electron microscopic images of MRSA. (A) Untreated bacterial cells. (B) Cells treated with peptidomimetic 1c. (C) Cells treated with peptidomimetic 1d. (D) Cells treated with VAN. The cells were exposed to various agents for 30 min at 10 times their respective planktonic MICs. Arrows, morphological alterations produced. Insets, higher-magnification images (magnification, ×150,000).

Resistance development against peptidomimetics in MRSA.

The ability of active peptidomimetics 1c and 1d to induce resistance development in MRSA strain ATCC 33591 in 17 sub-MIC serial passages was evaluated (Fig. 5). Fourfold and 2-fold increases in MIC values were observed for peptidomimetics 1c and 1d, respectively. The MIC of the standard antibiotic VAN was increased 4-fold after 17 passages, whereas a radical 256-fold change in the MIC was observed for ciprofloxacin (CIP).

FIG 5.

Resistance development induced by antibacterial agents in S. aureus (ATCC 33591) after 17 serial passages with sub-MIC levels of peptidomimetic 1c, peptidomimetic 1d, or antibiotic. The fold change in MIC is the ratio of the MIC after 17 passages to the MIC before the first passage.

Activity against MRSA biofilms. (i) Quantification of viability and reduction in biomass.

After establishing their antibacterial activity and mode of action on planktonic cells, we further evaluated the efficacy of peptidomimetics 1c and 1d to prevent the formation of biofilms and to eradicate preformed MRSA biofilms (24 h) by using alamarBlue as a redox indicator for assessment of metabolic activity and crystal violet for biomass quantification. A well-characterized biofilm-producing reference strain of MRSA (ATCC 33591) was used for the experiments. Prior to this experiment, the MICs of peptidomimetic 1c, peptidomimetic 1d, and VAN in biofilm growth medium (TSB with 0.5% NaCl and 0.25% glucose) were evaluated. The results showed pronounced effects of a high salt concentration (supplemented with 3% NaCl) on MICs, whereas 2-fold increases in the MICs of peptidomimetics 1c and 1d in low-salt medium (supplemented with 0.5% NaCl) were observed. The MIC for VAN was increased only 2-fold even in the high salt concentration. All biofilm-related experiments were performed with MIC_b measurements; MIC_b values were the planktonic MICs of peptidomimetics and VAN in biofilm growth medium (Table 3). For the biofilm formation inhibition assay, initial inocula were added with sub-MIC_b and MIC_b concentrations of the tested agents (Fig. 6A and B). Peptidomimetics 1c and 1d were able to inhibit biofilm formation at sub-MIC_b concentrations (∼4 μg/ml and ∼2 μg/ml, respectively), causing reductions in metabolic activity of up to 33.1% ± 5.7% and 26.4% ± 3.3%, respectively. Under identical treatment conditions, biomass reductions were found to be 19.8% ± 5.6% and 28.2% ± 11.1% for peptidomimetics 1c and 1d, respectively. At MIC_b, both peptidomimetics were able to inhibit the adhesion of biofilm, causing >90% reductions in measured viability and biomass quantity. Metabolic activity and biomass quantity were not reduced significantly with VAN at sub-MIC_b concentrations (∼0.5 μg/ml), compared with control values, whereas VAN at MIC_b concentrations (∼1 μg/ml) inhibited biomass quantity to 27.4% ± 1.3%.

FIG 6.

(A and B) Inhibition of MRSA biofilm formation by different agents using the alamarBlue assay (A) and biomass quantification using the crystal violet staining assay (B). (C and D) Metabolic activity of 24-h mature biofilm-embedded MRSA using the alamarBlue assay (C) and biomass quantification using the crystal violet assay (D). The MIC_b values for peptidomimetic 1c, peptidomimetic 1d, and VAN were 7.1 μg/ml, 3.5 μg/ml, and 0.8 μg/ml, respectively. For all experiments, data are expressed as mean ± SD. Statistical differences from control values were determined by one-way analysis of variance (ANOVA) with Tukey's multiple-comparison post hoc tests. All differences between the control and treated biofilms were considered statistically significant (P < 0.001).

The effects of peptidomimetics on the viability of 24-h-preformed mature biofilms were also evaluated at concentrations higher than MIC_b. At 20× MIC_b, the designed peptidomimetics 1c (140 μg/ml) and 1d (70 μg/ml) showed better killing profiles than did VAN (20 μg/ml), showing 6.4% ± 0.2% and 10.1% ± 7.8% viable cells, respectively, versus 77.7% ± 7.0% viable cells for VAN at the indicated concentration (Fig. 6C).

In parallel with viability results, assessments of reductions in the biomass quantities of 24-h mature MRSA biofilms showed a reduction in biomass to 24.0% ± 13.4% with peptidomimetic 1c at 140 μg/ml (Fig. 6D). For peptidomimetic 1d, significant differences in biomass quantities, in comparison with control values, were observed at both tested concentrations (35 μg/ml and 70 μg/ml, corresponding to 10× MIC_b and 20× MIC_b), i.e., 66.7% ± 8.2% and 21.4% ± 9.2%, respectively. For VAN-treated biofilms, the observed biomass quantities were 119.3% ± 17.5% at 10× MIC_b (10 μg/ml) and 83.7% ± 24.1% at 20× MIC_b (20 μg/ml).

(ii) Visualization of effects of designed peptidomimetics on biofilms using confocal laser scanning microscopy.

We next visualized the effects of the designed peptidomimetics and VAN on biofilm-embedded MRSA, making use of the membrane permeability-sensitive, DNA-binding dyes SYTO 9 and propidium iodide as markers. As a measure of biofilm formation/growth inhibition, the thickness of biofilm was measured using confocal microscopy. The control biofilm (24 h) showed a lawn of viable (green) cells, with an average thickness of 14.3 ± 1.4 μm (Fig. 7A; also see Table S1 in the supplemental material). At MIC_b, peptidomimetics 1c and 1d prevented the formation of biofilm; very few cells adhered to the substratum, with observed average thicknesses of 3.9 ± 1.1 μm and 3.5 ± 0.6 μm, respectively. Furthermore, at sub-MIC_b concentrations, the observed thicknesses were 5.2 ± 0.3 μm and 5.8 ± 0.4 μm, respectively (Fig. 7Ab and Ad). The measured biofilm thickness was 11.4 ± 2.9 μm with VAN at MIC_b (Fig. 7Ag), whereas VAN was unable to reduce the biofilm thickness at sub-MIC_b levels.

FIG 7.

Three-dimensional images of MRSA biofilms. (A) Effects of antibacterial agents on biofilm formation of MRSA, assessed using confocal laser scanning microscopy. (a) Control; (b) peptidomimetic 1c at sub-MIC_b level; (c) peptidomimetic 1c at MIC_b; (d) peptidomimetic 1d at sub-MIC_b level; (e) peptidomimetic 1d at MIC_b; (f) VAN at sub-MIC_b level; (g) VAN at MIC_b. (B) Effects of antibacterial agents against 24-h-preformed mature MRSA biofilms, assessed using confocal laser scanning microscopy. (a) Control; (b) peptidomimetic 1c at 10× MIC_b; (c) peptidomimetic 1c at 20× MIC_b; (d) peptidomimetic 1d at 10× MIC_b; (e) peptidomimetic 1d at 20× MIC_b; (f) VAN at 10× MIC_b; (g) VAN at 20× MIC_b. After treatment at different concentrations, the biofilms were stained with SYTO 9 (green; viable cells) and propidium iodide (red; dead cells), as described in the manufacturers' protocol.

Untreated 48-h mature biofilm (24 h plus 24 h) showed a lawn of viable (green) cells with an average thickness of 23.6 ± 2.5 μm (Fig. 7B; also see Table S2 in the supplemental material). Subsequent to treatment with peptidomimetics 1c and 1d at 10× MIC_b, there were visible decreases in the numbers of live cells and thickness was reduced to 7.1 ± 1.5 and 7.0 ± 1.0 μm, respectively (Fig. 7Bb and Bd). With peptidomimetics 1c and 1d, most of the cells lost their integrity at 20× MIC_b, appearing red (Fig. 7Bc and Be), and a smear of permeabilized cells was observed. Upon VAN treatment, no significant differences in the numbers of live cells were observed, inasmuch as mixed bacterial populations stained green were visible at both tested concentrations. VAN had little effect on 24-h biofilm at 10× MIC_b, with no distinction between control biofilm and VAN-treated biofilms being visible. Only with VAN at 20× MIC_b was a slight decrease in the height of mature biofilm observed (Fig. 7Bf and Bg). The confocal imaging experiments were repeated three times on three different days, and similar results were obtained (representative data from one set is shown here).

DISCUSSION

Polyamines (putrescine, spermidine, and spermine) modulate various processes in cells, including nucleic acid packaging, DNA replication, transcription, and translation, and thus are required for optimal growth in prokaryotes and eukaryotes (31). Polyamines and their analogues exhibit versatile biological activities, including anticancer, antiparasitic, antiendotoxin, and antibacterial activities (32–34). Squalamine (from dogfish shark) and cinodine (from Nocardia spp.) are natural antibiotics with a polyamine backbone in their structures (35, 36). The role of polyamine conjugation in improving activity for a number of synthetic antibacterial agents, such as ceragenins, acylpolyamines, and caffeoyl polyamines, has been reported (15, 33, 37). Synergistic effects of exogenous polyamines (added to growth medium) and various antibiotics have also been investigated, and it was shown that 1 mM spermine caused up to 200-fold reductions in the MIC of oxacillin against MRSA strain Mu50 under test conditions (38). Interestingly, it was recently reported that S. aureus lacks identifiable genes for polyamine biosynthesis and consequently produces no spermine or spermidine or their precursors; therefore, polyamines and their conjugates act as toxins to S. aureus (39). Supporting this, in a recent report, the exceptional virulence of MRSA strain USA300 was ascribed to the arginine catabolic mobile element (ACME), which harbors the spermidine acetyltransferase gene (speG), imparting resistance to spermidine and other polyamines (40). Therefore, for polyamine-sensitive MRSA, conjugation of spermine might be a robust strategy to overcome this deadly strain.

Various valuable structure-activity relationships for antibacterial peptidomimetics have been reported, and modifications in charge distribution or hydrophobicity have led to optimization of molecules for therapeutic applications (41–44). Extending our previous findings with the N-terminally tagged dipeptide spermidine template, in the present work we designed two series of peptidomimetics (series 1 and 2) with linear or branched arrangements of Trp residues on the spermine backbone to explore the effects on antibacterial activity and selectivity and the mode of action. In our previous work, we established 50 to 70% hydrophobicity (based on reverse-phase [RP]-HPLC retention times) and at least +2 charges to be crucial for antibacterial activity and cell selectivity (16). The comparative MIC data for series 1 and 2 showed that peptidomimetics 1b and 1c with linear arrangements of tryptophan were more active than the corresponding peptidomimetics with branched arrangements of tryptophan (i.e., peptidomimetics 2b and 2c) against all of the Gram-positive bacterial strains tested. For lipidated N-tagged peptidomimetics, series 1 and 2 showed comparable inhibitory effects on all Gram-positive bacterial strains; however, series 2 was more hemolytic than series 1, although the hydrophobicity ranges for the two series were the same. As reported in the literature and observed in our previous study, hydrophobicity above a threshold range plays a crucial role in increased hemolytic activity (16). This finding also holds true in the present study, since overall, the charges were the same and differences in hydrophobicity among corresponding pairs in series 1 and 2 were not very significant (<0.6 to 6%, as measured by RP-HPLC). Therefore, the key determinant for activity and selectivity in the present study, besides hydrophobicity, was the placement of Trp residues at different positions in the template, which indicates the role of the amino groups of spermine in activity.

To further elucidate the role of Trp moieties in the results described above, we performed interaction studies and evaluated the mode of action of these peptidomimetics with artificial membranes and intact bacterial cells. Trp fluorescence measurements have been used as a sensitive tool to probe the interactions of peptides with artificial bacterial or mammalian mimic membranes. Partitioning of Trp residues into the hydrophobic membrane environment has been reported to result in blue shifts accompanied by increases in emission intensity (45), as was observed for all peptidomimetics in series 1. In series 2, however, the emission intensity in mammalian mimic SUVs revealed more-pronounced increases for all of the peptidomimetics. Interestingly, for peptidomimetics 2a, 2c, and 2d in bacterial mimic membranes, blue shifts were observed without concomitant increases in fluorescence intensity, compared with buffer (see Fig. S3 in the supplemental material). Similar observations of blue shifts subsequent to peptide-lipid interactions without increases in emission intensity were reported previously for the antimicrobial peptides temporin L and nisin; the authors attributed the decreases in Trp fluorescence intensity to quenching due to aggregation of peptide in the vicinity of membranes or due to the quenching properties of the negatively charged lipid head groups, which can interact directly with π orbitals of the indole ring in the Trp residue (46, 47). In the present study also, the positive charge distribution in these peptidomimetics led to better electrostatic interactions with the negatively charged bacterial mimic membranes and vicinity-induced aggregation, causing decreases in fluorescence intensity in bacterial mimic membranes.

Further, the dependence of membrane depolarization ability on Trp branching and N-terminal tagging was evident from the results (Fig. 2). The untagged template peptidomimetics (+4 charges and <20% hydrophobicity) were unable to alter membrane potential up to the highest concentrations tested. Despite good activity against MRSA, aromatic tagging in peptidomimetics 1c and 2c did not allow significant changes in membrane potential, which might have resulted from poor insertion of these peptidomimetics into hydrophobic membrane interiors. The effect of hydrophobicity on membrane insertion was evident as more-hydrophobic peptidomimetics (peptidomimetics 1e, 1f, 2e, and 2f) in both series 1 and series 2 were better able to alter membrane potential than were less-hydrophobic peptidomimetics (peptidomimetics 1d and 2d). It was evident from Trp fluorescence and membrane depolarization experiments that series 2 peptidomimetics with branched Trp residues on the spermine backbone although more potent, were more prone to cause nondifferential interactions, whereas series 1 peptidomimetics were potent, cell-selective, membrane-depolarizing agents. In series 1, peptidomimetics 1c and 1d were found to be more active and cell selective; therefore, further studies were carried out with these two molecules for optimization.

To establish whether bactericidal ability is inherent in the present designed peptidomimetics, the time course of bacterial killing was studied by exposing MRSA to 2× MIC and 4× MIC levels of peptidomimetics 1c and 1d (Fig. 3). At 4× MIC, which is a therapeutically relevant concentration, most of the bacteria were killed within 30 min, as extremely rapid bactericidal effects are often seen for antimicrobial peptides (20). The fast bacterial killing suggests that, at these concentrations, the antibacterial effects are mediated through significant permeabilization or lysis of bacterial membranes, which was corroborated by scanning electron microscopic images of MRSA showing distinct membrane damage at 10× MIC with greater numbers of bacteria (10⁸ CFU/ml) for both peptidomimetics (Fig. 4).

Several reports suggested that antimicrobial peptides and their analogue peptidomimetics have novel membrane-active modes of action with multiple nonspecific targets, resulting in the development of resistance to bacteria (23, 48). The results for peptidomimetics 1c and 1d after 17 serial passages at sub-MIC doses could not demonstrate resistance for MRSA. Poor serum protease stability limits most of the developed antimicrobial peptides to topical application. Peptidomimetics 1c and 1d were found to kill MRSA in the presence of human serum (25% [vol/vol]), with 4- and 8-fold increases in their MIC values, respectively (Table 3). The increases in MICs observed in the present study corroborated the previous report of short cationic antimicrobial peptides binding with serum protein albumin (49). Further, to explain the increases in the MIC values of peptidomimetics, the stability of peptidomimetics 1c and 1d in human serum was evaluated with RP-HPLC, and the data demonstrated that no degradation was found for the peptidomimetics even with 72 h of incubation (see Fig. S4 in the supplemental material). Peptidomimetics 1c and 1d were further assessed for cytotoxicity against primary PBMCs and demonstrated mostly favorable nontoxic profiles at concentrations (20 μg/ml) higher than the MICs (Table 2).

MRSA is an extraordinary etiological agent due to its virulence, multidrug-resistant profile, and increasing prevalence in community and health care settings. Biofilm formation is a particularly virulent mechanism for Staphylococcus species that renders treatment and cure difficult with invasion, with associated mortality rates in severe cases of MRSA infections being about 20% (50). Various strategies have been proposed to either kill microbes or drive them out of biofilms. Among these strategies, targeting quorum sensing and designing antiadhesion agents and antimicrobial peptides are a few effective means that are currently being explored (51, 52). Intrigued by the success of lipopeptide daptomycin, oritavancin, and other membrane-active peptidomimetics with membrane depolarization and disruption abilities against biofilm-embedded MRSA (26, 53), we extended our studies against MRSA biofilms and compared the activities of peptidomimetics 1c and 1d with that of the standard drug VAN.

As a standard protocol for determination of biofilm formation/killing abilities, we used a combination of the alamarBlue assay (for measurement of viability) and crystal violet assay (for quantification of biomass) (27). In the biofilm assay, no perfect correlation between cell viability and biomass quantity was observed, although similar patterns were seen in both experiments. At sub-MIC_b levels, the peptidomimetics 1c and 1d decreased biofilm formation, indicating the potential of these molecules to prevent MRSA adhesion to surfaces. For 24-h mature biofilm, peptidomimetics 1c and 1d were more effective in reducing viability and biomass than was VAN at 20× MIC_b (Fig. 6C and D). Although VAN showed better ability to inhibit growth in planktonic cultures of MRSA under sessile conditions with 24-h mature biofilms, the designed peptidomimetics 1c and 1d proved to be more efficacious at the tested concentrations, exhibiting significant decreases in viability versus the positive control (P < 0.001).

Further, the effects of peptidomimetic treatment on biofilm formation and killing were visualized using confocal microscopy, which is a well-known method to assess such effects (26, 27). The results revealed a marked difference in the viability of 24-h mature biofilms with peptidomimetics 1c and 1d versus VAN (Fig. 7). For VAN, a subpopulation of viable, predominantly green cells was observed. VAN has been reported to be less membrane depolarizing and less effective in reducing the viability of biofilm-embedded S. aureus, due to the slow growth of bacterial cells under biofilm conditions (26). Making use of live/dead cell staining, it was shown that VAN, even at a high concentration of 500 μg/ml, was unable to cause growth depletion of Staphylococcus haemolyticus biofilms (27).

In summary, we designed new ultrashort N-terminally modified dipeptidomimetics with or without modifications on the spermine backbone leading to linear or branched tryptophans, which could effectively inhibit the growth of Gram-positive and Gram-negative bacterial strains under planktonic conditions. Direct effects of Trp positioning on the depth of insertion in artificial membranes were observed. Furthermore, disruption of membrane potential in intact MRSA pointed to different charge-hydrophobicity interactions leading to a lack of cell selectivity for series 2 peptidomimetics. We found the linear arrangement of Trp residues without backbone spermine modification to be better for therapeutically viable antibacterial peptidomimetics. Interestingly, under identical experimental conditions, with the dual modes of action of membrane depolarization and disruption, peptidomimetics 1c and 1d showed better efficacy than the conventional antibiotic VAN against biofilm formation and eradication of 24-h mature MRSA biofilms. These findings highlight the potential of membrane-active antibacterial peptidomimetics as useful tools to eradicate clinically relevant biofilms. Overall, our present work provides an impetus for the design of better membrane-active, spermine-based, antibacterial peptidomimetics to treat recalcitrant biofilm communities of MRSA. At present, we are exploring the ability of the most active peptidomimetics to hamper biofilm formation on solid supports, which would broaden the therapeutic applications of these peptidomimetics in clinical settings.