The Oligo-Acyl Lysyl Antimicrobial Peptide C₁₂K-2β₁₂ Exhibits a Dual Mechanism of Action and Demonstrates Strong In Vivo Efficacy against Helicobacter pylori

Abstract

Helicobacter pylori has developed antimicrobial resistance to virtually all current antibiotics. Thus, there is a pressing need to develop new anti-H. pylori therapies. We recently described a novel oligo-acyl-lysyl (OAK) antimicrobial peptidomimetic, C₁₂K-2β₁₂, that shows potent in vitro bactericidal activity against H. pylori. Herein, we define the mechanism of action and evaluate the in vivo efficacy of C₁₂K-2β₁₂ against H. pylori after experimental infection of Mongolian gerbils. We demonstrate using a 1-N-phenylnaphthylamine (fluorescent probe) uptake assay and electron microscopy that C₁₂K-2β₁₂ rapidly permeabilizes the bacterial membrane and creates pores that cause bacterial cell lysis. Furthermore, using nucleic acid binding assays, Western blots, and confocal microscopy, we show that C₁₂K-2β₁₂ can cross the bacterial membranes into the cytoplasm and tightly bind to bacterial DNA, RNA, and proteins, a property that may result in inhibition of enzymatic activities and macromolecule synthesis. To define the in vivo efficacy of C₁₂K-2β₁₂, H. pylori-infected gerbils were orogastrically treated with increasing doses and concentrations of C₁₂K-2β₁₂ 1 day or 1 week postinfection. The efficacy of C₁₂K-2β₁₂ was strongest in animals that received the largest number of doses at the highest concentration, indicating dose-dependent activity of the peptide (P < 0.001 by analysis of variance [ANOVA]) regardless of the timing of the treatment with C₁₂K-2β₁₂. Overall, our results demonstrate a dual mode of action of C₁₂K-2β₁₂ against the H. pylori membrane and cytoplasmic components. Moreover, and consistent with the previously reported in vitro efficacy, C₁₂K-2β₁₂ shows significant in vivo efficacy against H. pylori when used as monotherapy. Therefore, OAK peptides may be a valuable resource for therapeutic treatment of H. pylori infection.

INTRODUCTION

Helicobacter pylori is a spiral-shaped, flagellated, Gram-negative bacterium that has adapted to infect the human gastric mucosa. Persistent infection leads to the development of gastroduodenal diseases that include chronic gastritis, peptic ulcer disease, and gastric cancer (31). H. pylori is carried by half of the world's population, and its prevalence is estimated to exceed 90% in some developing countries. In the United States, it is estimated that 30 to 40% of the population is infected with H. pylori (8). In addition, individuals infected with H. pylori are at a greater risk of developing gastric cancer (33). This fact led the World Health Organization (WHO) International Agency for Research on Cancer to classify H. pylori as a group I carcinogen (9). Given the association with gastric disease, together with complicated diagnosis procedures and aggregate health care expenditures, the annual health care costs attributed to H. pylori infection are estimated at about $10 billion in the United States alone (8, 66). Thus, H. pylori infection presents a major global public health burden, and eradication of the bacterium from infected symptomatic individuals is still the preferred choice of treatment and management.

In the 3 decades since the discovery of the bacterium, treatment strategies for H. pylori infection have undergone tremendous evolutionary changes that span from monotherapy to multidrug therapy for extended periods of treatment (15). Though initially favored, monotherapy and dual therapy for 4 to 7 days were the first treatments to be shown to result in eradication failure (21). As a result, treatment strategies currently involve combination therapies administered for 10 to 14 days and include triple, quadruple, and sequential therapies.

Despite complex treatment strategies, during the last 4 to 6 years, the efficacy of triple and quadruple therapies has been reduced dramatically, and eradication rates of below 50 to 75% have been reported in some areas (21, 25). Eradication failure is mainly attributed to the development of antimicrobial resistance and to noncompliance by patients who do not adhere to the complicated and sometimes side effect-inducing multidrug therapy. Given the immense challenge in rising antimicrobial resistance (36), there is a need for new antibiotics for the treatment of H. pylori infection. Moreover, given the complicated nature of the H. pylori treatment regimens, it is envisioned that drugs that could be used as monotherapy or in combination with new classes of antibiotics would be the most helpful.

Recently, antimicrobial peptides (AMPs) have been scrutinized as potential novel antimicrobial agents against human pathogens (63, 68). AMPs are usually small molecules comprised of 12 to 60 amino acids with molecular masses of 10 kDa or less. However, much bigger polypeptides have recently been discovered; examples are lactoffericin (12), kinocidins (microbicidal chemokines) (55), and complement-derived peptides (82). AMPs display immense diversity in sequence, secondary structural motifs, charge, and/or the abundance of certain specific amino acids (72). However, the formation of amphipathic structures and the presence of multiple basic amino acid residues are highly conserved (14, 74) and are the characteristic features that endow cationicity and enable the AMPs to interact selectively with anionic bacterial membranes.

The use of AMPs as potential therapeutics has been particularly attractive because of their ability to be promptly synthesized by the host and their capacity to lyse the cell membranes of pathogens via direct interaction. However, several studies have demonstrated that pathogens are able to elude the effects of natural AMPs by production of proteolytic peptidases (24, 75), by recognition and extracellular capture of the AMPs, by active extrusion of AMPs from the bacterial cell wall (35), or by reduction of the net anionic charge of the bacterial cell envelope (1, 59, 61). Thus, research has turned to the production of synthetic AMPs in order to circumvent the possibility of microbial resistance. Oligomers of acylated lysines (oligo-acyl-lysyls [OAKs]) are novel synthetic peptidomimetics that consist of alternating amino acyl chains and cationic amino acids that are arranged to create an optimal molecular charge and hydrophobicity for enhanced potency (65, 71). Like most natural AMPs, the OAKs are cationic and form amphipathic structures that associate with one another to protect the hydrophobic side and expose the cations to interact with the bacterial membrane (87). However, in contrast to natural AMPs, the OAK design allows for the formation of a simple structure, which allows for the establishment of structure-activity relationships and prevents the formation of secondary structures (65). These properties are in turn associated with the lack of development of antimicrobial resistance (37, 65). In addition, the OAK peptides are resistant to bacterial enzymatic degradation (13, 67) and exhibit no or minimal cytotoxicity to mammalian cells (64, 67). Furthermore, their structures can be fine tuned to optimize antibacterial activity (67), and they have been shown to possess significant in vivo efficacy against Escherichia coli (65, 67) and Staphylococcus aureus (43, 70) in mouse models of infection. Thus, OAKs display characteristic features that are attractive for the development of a potent new class of therapeutic drugs.

We recently demonstrated the strong in vitro efficacy of the OAK C₁₂K-2β₁₂ against H. pylori (47). C₁₂K-2β₁₂ showed significant postantibiotic effects (PAE) and significant thermal and pH stability, suggesting that it might be a potential new therapeutic agent. Given the strong in vitro efficacy we previously reported, here we addressed two major questions. (i) How does C₁₂K-2β₁₂ exert its effect on the H. pylori cell? (ii) Could monotreatment with C₁₂K-2β₁₂ be used to decrease the H. pylori colonization load in an animal model of infection? Collectively, our data demonstrate that C₁₂K-2β₁₂ exerts its bactericidal effects through a dual mode of action; at low concentrations, C₁₂K-2β₁₂ binds nucleic acids and proteins, while at higher concentrations, the OAK extensively disrupts the bacterial membrane leading to cell lysis by pore formation. Furthermore, we found that monotherapeutic dosing with C₁₂K-2β₁₂ resulted in significant clearance of H. pylori in the gerbil model of infection. These results suggest that C₁₂K-2β₁₂ and other OAK peptides may represent a potential new class of antibiotics for treatment of H. pylori infection.

MATERIALS AND METHODS

Peptide synthesis, reagents, and antibiotics.

The C₁₂K-2β₁₂ and K-5α₁₂ oligo-acyl-lysyls (OAKs) used in this study were synthesized as previously described (47, 63–65). K-5α₁₂ is a control OAK peptide that has no antibacterial activity (data not shown). The composition of purified peptides was confirmed using amino acid analysis and electrospray mass spectrometry. Peptides were stored as lyophilized powder at −20°C. Before each experiment, a fresh solution containing the appropriate concentration of peptide was reconstituted from a stock solution (1 mg/ml) that was prepared by two cycles, with one cycle consisting of 2 min of sonication in a water bath followed by brief vortexing. The fresh peptide solution was centrifuged and then used in subsequent efficacy studies and mechanism of action experiments. Amoxicillin (Sigma, St. Louis, MO) and vancomycin (USB Corporation, Cleveland, OH) were reconstituted according to the manufacturer's instructions and used at the indicated concentrations.

Bacterial strains and growth conditions.

H. pylori strains G27 (17) and 7.13 (18) were used to determine the C₁₂K-2β₁₂ mechanism of action and to test for in vivo efficacy, respectively. Both strains were cryopreserved at −80°C in brain heart infusion medium supplemented with 20% glycerol and 10% fetal bovine serum (FBS). Bacteria were cultured under microaerobic and medium conditions as previously described (47), and the cultures were maintained with shaking in brucella broth (Neogen-Acumedia) containing 10 μg/ml vancomycin (Sigma) and 10% fetal bovine serum (Invitrogen Gibco-BRL, Carlsbad, CA) in a microaerobic environment. An Anoxomat (Spiral Biotech, Norwood, MA) was used to achieve the microaerobic atmospheric condition of 5% O₂, 10% CO₂, and 85% N₂ gas mixture.

Determination of membrane perturbation and translocation into bacterial cytoplasm by C₁₂K-2β₁₂. (i) NPN membrane permeation assay.

The ability of C₁₂K-2β₁₂ to permeabilize the bacterial membrane was determined using intact H. pylori cells and 1-N-phenylnaphthylamine (NPN). NPN (Sigma) is a membrane potential-sensitive fluorescent probe that shows increased fluorescence upon exposure to a hydrophobic environment (44). Briefly, a starter culture of H. pylori G27 was grown for 16 to 18 h (optical density at 600 nm [OD₆₀₀] of 0.4 to 0.5) as described above. At this point, the starter culture was used to inoculate independent test cultures to an OD₆₀₀ of 0.05. These cultures were then incubated for 6 to 8 h in the presence or absence of C₁₂K-2β₁₂ (5 μM, 10 μM, 20 μM, 40 μM, and 80 μM) or amoxicillin (100 μM). Approximately 5 × 10⁶ bacterial cells were then sampled from each culture to determine NPN uptake at 15 min, 30 min, 1 h, 4 h, and 6 h. The cells were washed and resuspended in PBS (pH 7.2) to achieve a final volume of 250 μl in the presence of 22 μg/ml of NPN. Cells grown in the absence of drugs and 80 μM C₁₂K-2β₁₂ resuspended in phosphate-buffered saline (PBS) alone in the absence of H. pylori cells were used as controls. The NPN fluorescence intensity emitted (excitation λ [λ_ex] of 350 nm, emission λ [λ_em] of 420 nm, and a slit width of 5 nm) was recorded after 3 min as arbitrary units (AU) on a Spectramax M2 spectrofluorimeter (Molecular Devices, Sunnyvale, CA) and was considered a correlate of membrane potential and permeation. The NPN assays were performed at room temperature and were repeated three times. Additionally, to determine the kinetics of H. pylori death in relation to membrane permeation, samples exposed to 20 μM C₁₂K-2β₁₂ were plated to determine the viable CFU in relation to NPN fluorescence.

(ii) Electron microscopy studies.

To investigate whether C₁₂K-2β₁₂ affected the overall cell morphology and to visualize the bacterial membrane, we studied bacterial ultrastructure using transmission electron microscopy (TEM). Briefly, a starter culture of H. pylori G27 was grown for 16 to 18 h (OD₆₀₀ of 0.4 to 0.5) in liquid medium as described above. At this point, the starter culture was used to inoculate independent test cultures at an OD₆₀₀ of 0.05. These cultures were then incubated with 20 μM K-5α₁₂ (control peptide that has no antibacterial activity; data not shown), 20 μM C₁₂K-2β₁₂, 100 μM amoxicillin, or PBS as an untreated control. At 2 h and 16 h of culture, approximately 1.5 × 10⁷ bacterial cells were collected and subsequently washed twice in PBS by centrifugation and resuspension. The cells were then fixed and processed for TEM using previously described methods (50, 85) with minor modifications. Briefly, the cells were exposed to dehydration in a graduated series of ethanol followed by resin polymerization at 65°C for 3 days. A Leica EM UC6 ultramicrotome was used to slice thin sections of 70- to 80-nm thickness from the polymerized blocks, which were then loaded on 400-mesh copper grids. The sections were poststained with 2% uranyl acetate for 15 min and Reynold's lead citrate for 5 min before being mounted for observation on a Philips CM100 transmission electron microscope (FEI Company, Hillsboro, OR) operating at 80 keV. Images were recorded with a 4MP SPOT Insight charge-coupled-device (CCD) camera (Diagnostic Instruments Inc., Sterling Heights, MI) using SPOT Insight software (Diagnostic Instruments). The morphology of the cells was observed at low-power magnification (×13,500), and the cell wall ultrastructure was observed at high-power magnification (×64,000 or ×130,000). Using the digital camera, images were acquired in the thin sections randomly across multiple grids. Each grid was examined in a similar systematic manner.

(iii) Confocal microscopy studies.

In order to determine whether C₁₂K-2β₁₂ had the ability to traverse the bacterial membrane and enter the cytoplasm, DyLight 488 fluorescent dye (Thermo Fisher Scientific-Pierce Biotechnology, Rockford, IL) was conjugated to C₁₂K-2β₁₂. Labeled peptide was purified to remove unreacted dye according to the manufacturer's protocol in the DyLight 488 antibody labeling kit. It is noteworthy that fluorescent labeling did not affect the bactericidal activity of the peptide. Bacterial samples were inoculated into 2-ml liquid cultures at an OD₆₀₀ of 0.05 and then were incubated with 20 μM, 40 μM, or 80 μM DyLight 488-conjugated C₁₂K-2β₁₂ for 30 min. Bacterial cell samples incubated with unconjugated DyLight 488 or PBS were used as controls. The bacterial cells were washed twice in PBS (pH 6.5) and resuspended in 2% paraformaldehyde–PBS fixative followed by incubation in the dark for 30 min on ice. The bacterial cells were then washed twice to remove traces of fixative, and the DNA was stained with 4′,6′-diamidino-2-phenylindole (DAPI) for 20 min to mark the bacterial cytoplasm. The cells were then resuspended in PBS and spotted onto gelatin-coated slides treated with vector-shield mounting reagent. Fluorescent images of bacterial cells were visualized at a magnification of ×100, and confocal images were taken with a Zeiss LSM 710 confocal microscope (Carl Zeiss MicroImaging, LLC, Thornwood, NY) using two lasers with peak wavelengths allowing excitation of green and blue fluorochromes. The images were processed and analyzed with Zeiss LSM Image 5 Browser software.

Determination of C₁₂K-2β₁₂ interaction with bacterial nucleic acids and proteins.

All experiments for peptide-nucleic acid interactions were performed with peptides that were preheated to 95°C.

(i) Binding to DNA.

The ability of C₁₂K-2β₁₂ to bind DNA was investigated by incubation of the peptide with ³²P-labeled (end-labeled) DNA probes, followed by analysis of subsequent DNA mobility using electrophoretic mobility shift assay (EMSA) gels as previously described (5, 20). Briefly, regions of the oorD and rpoB promoters were PCR amplified from the H. pylori G27 strain, and the PCR products were purified by gel extraction. Additionally, the E. coli chloramphenicol acetyltransferase gene and human antibody light-chain DNA fragment were included to test the specificity of the binding.

For the EMSA, 1 ng of an end-labeled DNA fragment was mixed with C₁₂K-2β₁₂ in a final volume of 20 μl binding buffer (24% glycerol, 40 mM Tris [pH 8.0], 150 mM KCl, 2 mM dithiothreitol [DTT], 600 μg/ml bovine serum albumin, 200 μM EDTA, and 0.1 mg/ml sheared salmon sperm DNA). Final OAK peptide concentrations of 1.25 μM, 2.5 μM, 5 μM, 10 μM, and 20 μM C₁₂K-2β₁₂ were used. A tube containing no peptide and a tube containing 20 μM K-5α₁₂ control peptide were included as control reactions. All reaction mixtures were incubated at 37°C for 1 h. Following incubation, the reaction products were separated on a 5% polyacrylamide gel (5% 19:1 acrylamide, Tris-glycine buffer, 2.5% glycerol) at 100 V in Tris-glycine buffer. The gels were then exposed to a phosphor screen and scanned on a FLA-5100 multifunctional phosphorimager (Fuji Medical Systems, Hanover Park, IL). Analysis was performed using Multi-Gauge version 3.0 software (Fujifilm Medical Systems). The reported results are representative of five independent experiments.

(ii) Binding to RNA.

Similar to binding to DNA, the ability of C₁₂K-2β₁₂ to bind to RNA was determined by examining changes in the electrophoretic mobility of ureA or oorD RNA probes that were preincubated with C₁₂K-2β₁₂. RNA riboprobes labeled with [α-³²P]UTP (Perkin Elmer) were synthesized by in vitro transcription as previously described (6, 78). Briefly, 1 μg of ureA or oorD RNA riboprobes were mixed with C₁₂K-2β₁₂ (0 μg, 1 μg, or 5 μg) or with 5 μg of K-5α₁₂ in a final volume of 50 μl. The reaction mixtures were incubated at 37°C for 1 h. The mixtures were then resolved on 5% acrylamide–Tris-borate-EDTA–8 M urea denaturing gels to determine the mobility of the RNA riboprobes. The gels were exposed to phosphor screens, which were subsequently scanned using a FLA-5100 multifunctional scanner (Fujifilm Medical Systems). The images were analyzed using Multi-Gauge software version 3.0 (Fujifilm Medical Systems). The reported results are representative of three independent experiments.

(iii) Binding to H. pylori proteins by SDS-PAGE and Western blot analysis.

The ability of the C₁₂K-2β₁₂ peptide to bind to H. pylori proteins was investigated using a modified Western blot protocol where blots were probed with DyLight 488-conjugated C₁₂K-2β₁₂. Briefly, bacterial cell lysates were prepared from liquid cultured bacterial pellets using lysis buffer (150 mM NaCl, 50 mM Tris-HCl [pH 8.0], 5 mM EDTA, 1% sodium dodecyl sulfate [SDS], 10% glycerol containing protease inhibitor cocktail [Roche]). Proteins from these lysates, as well as purified recombinant H. pylori Fur protein, were separated via SDS-polyacrylamide gel electrophoresis (PAGE) using a 4% stacking gel and an 18% separating gel and were then electrotransferred onto a nitrocellulose membrane by semidry blotting (Owl separation system; ThermoScientific). After transfer, residual proteins in the gel were detected with Coomassie brilliant blue R250. The membranes were washed in PBS, blocked for 1 h in PBS containing 3% bovine serum albumin, and then washed again 3 times for 5 min each time in PBS containing 0.05% Tween 20 (PBS-T). The membranes were then probed with 20 μM DyLight 488-conjugated C₁₂K-2β₁₂ (ThermoScientific/Pierce), and direct protein–C₁₂K-2β₁₂ binding was revealed by fluorescence analysis using a LAS-3000 instrument with LAS-3000 Lite capture software (FujiFilm).

Functional inhibition of restriction digestion by C₁₂K-2β₁₂.

The ability of C₁₂K-2β₁₂ to block restriction enzyme activity via DNA binding was examined using the 10,031-bp endogenous plasmid from H. pylori G27 and five restriction enzymes predicted to cut multiple times within the plasmid sequence. Briefly, the H. pylori G27 endogenous plasmid was isolated using the QIAprep spin miniprep kit (Qiagen). For each restriction enzyme examined (MfeI, HindIII, EcoRI, SpeI, and XmnI [New England BioLabs]), two reaction mixtures containing 500 ng plasmid were preincubated in the presence or absence of 40 μM C₁₂K-2β₁₂ at 37°C for 1 h in a total volume of 20 μl. Plasmids preincubated with 40 μM or 80 μM K-5α₁₂ were also included as controls. Following incubation, the appropriate restriction enzyme buffer was added, and the reaction mixtures were split such that one half of each tube was exposed to 5,000 U of MfeI, HindIII, EcoRI, SpeRI, or XmnI, and the second half of each tube was treated as an undigested native control. The digestion reaction mixtures were incubated at 37°C for 1 h, and the plasmid DNA was analyzed on a 1% agarose gel stained with 5 μg/ml ethidium bromide (Sigma). The experiments were repeated three times.

Determination of the in vivo efficacy of C₁₂K-2β₁₂ in H. pylori-infected gerbils.

A total of 90 animals that had been allowed water but no food for 12 h were inoculated via intragastric gavage with approximately 8 × 10⁸ CFU of H. pylori. Five animals were then assigned to a total of 18 possible groups. Briefly, animals were grouped into two major categories: (i) animals treated with C₁₂K-2β₁₂ 1 day after H. pylori infection (n = 45) and (ii) animals treated with C₁₂K-2β₁₂ 7 days postinfection (n = 45). Each of these two categories was further subdivided into three dose treatment groups based on the following C₁₂K-2β₁₂ dose schedule: (i) those given a single dose (n = 15), (ii) those given three sequential once daily doses (n = 15), and (iii) those given five sequential once daily doses of the peptide (n = 15). Each dose treatment group was additionally divided into three regimen treatment groups: (i) those treated with 1 mg C₁₂K-2β₁₂/kg body weight (n = 5), (ii) those treated with 4 mg C₁₂K-2β₁₂/kg body weight (n = 5), and (iii) those treated with PBS rather than the peptide. The animals were weighed immediately before dosing, and all treatments were orogastrically administered following a 12-hour fasting period during which animals were allowed water. One week after the last treatment, the animals were sacrificed, and their stomachs were harvested, weighed, and homogenized. Stomach homogenates were then plated to determine H. pylori CFU per gram of stomach tissue. Animal experiments were carried out according to strict guidelines and recommendations in the Guide for the Care and Use of Laboratory Animals (54a), and the study protocols were approved by the Uniformed Services University of Health Sciences Institutional Animal Care and Use Committee (approved protocol numbers MIC-10-774 and MIC-07-503).

Statistical analysis.

To determine whether a significant difference in the number of colonizing bacteria existed after C₁₂K-2β₁₂ treatment of H. pylori-challenged gerbils, colonization levels across various peptide treatment groups were analyzed using one-way analysis of variance (ANOVA) analysis. The ANOVA analyses were completed using SPSS 16.0 statistical software (SPSS Inc., IBM Corporation, Somers, NY). Where significant treatment effects were observed (P < 0.05), the least significant difference (LSD) analyses were computed to permit separation and comparison of the means of the C₁₂K-2β₁₂-treated groups and the corresponding PBS control. A P value of <0.05 was considered a significant difference.

RESULTS

C₁₂K-2β₁₂ rapidly permeabilizes the H. pylori membrane and triggers bactericidal effects.

The mechanisms by which antimicrobial agents act are varied and may involve membrane disruption or functional inhibition of key pathways that are essential for microbial growth and/or survival. Cationic antimicrobial peptides have been shown to target and disrupt bacterial membranes and/or interact with nucleic acids in a manner that could potentially block macromolecule synthesis (19, 57, 69). Thus, to determine whether the effect of the C₁₂K-2β₁₂ OAK peptide on H. pylori involved disruption of the cell membrane, we assessed membrane integrity using 1-N-phenylnaphthylamine (NPN). NPN is a neutral hydrophobic fluorescent probe whose quantum yield increases in hydrophobic environments compared to aqueous environments (44). Normally, NPN is excluded by the intact bacterial cell membrane. However, when the outer membrane (OM) structure is damaged, NPN can partition into the hydrophobic interior of the OM or plasma membrane, leading to a dramatic increase in fluorescence. Therefore, we used changes in NPN fluorescence intensity as a direct indicator of cell membrane permeability (26).

Incubation of H. pylori with 5 μM C₁₂K-2β₁₂, which is below the previously demonstrated MIC of 6.5 to 7.0 μM for the G27 strain, resulted in a minor change in NPN fluorescence that was similar to the control H. pylori exposed to PBS (Fig. 1A). However, utilization of concentrations of C₁₂K-2β₁₂ above the MIC resulted in significant changes in fluorescence as early as 15 min after exposure to the peptide. This response was dose dependent, since increasing concentrations of C₁₂K-2β₁₂ (5 μM, 10 μM, 20 μM, 40 μM, and 80 μM) resulted in a more rapid increase in NPN fluorescence, and it was specific to C₁₂K-2β₁₂ interaction with H. pylori, since utilization of related OAK peptides that do not affect H. pylori growth or survival, 2β₁₂ (47) and K-5α₁₂ (data not shown), did not affect NPN fluorescence. Incubation of C₁₂K-2β₁₂ alone without H. pylori cells also did not affect NPN fluorescence, indicating no or minimal C₁₂K-2β₁₂ autofluorescence. However, the response could be mimicked through the addition of amoxicillin (AMX), which acts by inhibiting cell wall mucopeptide synthesis and is known to disrupt bacterial membrane integrity through inhibition of peptidoglycan polymerization (2). It is noteworthy that 10 μM C₁₂K-2β₁₂ showed greater and more rapid membrane permeabilization effects than a 10-fold higher concentration of AMX (100 μM), consistent with the previously reported superior effects of C₁₂K-2β₁₂ over AMX (47). To determine whether membrane permeation was coincident with or a consequence of bacterial death, the NPN assay was repeated, and the bacteria were monitored by plating. As shown in Fig. 1B, treatment of H. pylori with 20 μM C₁₂K-2β₁₂ resulted in dramatic increases in NPN fluorescence prior to significant killing; therefore, membrane permeation precedes cell death. Overall, these data suggest that C₁₂K-2β₁₂ has the ability to rapidly disrupt the bacterial outer membrane and indicate that our previously demonstrated in vitro killing of H. pylori is likely due to induction of bacterial membrane damage.

Fig 1.

C₁₂K-2β₁₂ rapidly permeabilizes the H. pylori membrane in a dose-dependent manner that correlates with bactericidal activity. (A) Membrane permeabilization was measured by increased fluorescence due to uptake of 1-N-phenylnaphthylamine (NPN), and (B) a parallel bactericidal assay was performed with 20 μM C₁₂K-2β₁₂. H. pylori (Hp) G27 was cultured for 24 h in the presence of C₁₂K-2β₁₂ (5 μM, 10 μM, 20 μM, 40 μM, and 80 μM) or amoxicillin (100 μM). Cells grown in the absence of drugs but exposed to PBS or 80 μM C₁₂K-2β₁₂ alone (without cells) were used to determine background fluorescence. Approximately 5 × 10⁶ bacterial cells were sampled at 15 min, 30 min, 1 h, 4 h, and 6 h to determine NPN uptake and bactericidal activity. NPN fluorescence intensity emitted (excitation λ of 350 nm and emission λ of 420 nm) was recorded in arbitrary units (AU) as a correlate of membrane permeation. Bactericidal activity was expressed as percent survival of starting CFU. The data shown are representative of five independent experiments. In panel B, NPN fluorescence is shown on the left y axis and depicted by lines, while percent survival is shown on the right y axis and indicated by the bars.

C₁₂K-2β₁₂ dramatically disrupts H. pylori membranes, causes lysis through pore formation, and traverses into the cytoplasm.

The dose-dependent increase in NPN fluorescence (Fig. 1) suggested that C₁₂K-2β₁₂ was inducing membrane damage. To further confirm and more thoroughly investigate this possibility, we next examined the ultrastructural morphology of the bacterial cell after C₁₂K-2β₁₂ exposure using transmission electron microscopy (TEM). PBS-treated cells, cells treated with the OAK peptide K-5α₁₂ (47), and cells treated with amoxicillin were included as controls. H. pylori cells were exposed to the indicated treatments for 2 or 16 h. The cells were then harvested, fixed, stained with uranyl acetate, and observed as thin sections (70 to 80 nm). As shown in electron micrographs taken at low magnification (×13,500) 16 h posttreatment (Fig. 2), no dramatic differences in cell morphology were observed for PBS-treated bacterial cells (Fig. 2A) or those exposed to the K-5α₁₂ control peptide (Fig. 2C). However, amoxicillin-treated H. pylori cells developed an enlarged coccoid morphology (Fig. 2B). These enlarged cells were typified by clear separation of the plasma membrane from the outer membrane. Additionally, 5 to 10% of the cells appeared as ghost cells, which showed a complete loss of cytoplasmic contents and appeared as empty cell membrane outlines similar to previously described H. pylori ghost cells (56). Save for the appearance of the ghost cells, these data are consistent with the previously reported effects of amoxicillin on H. pylori (2); the minor difference is likely due to amoxicillin concentration differences.

Fig 2.

Thin-section transmission electron micrograph of H. pylori G27 reveals dramatic changes in morphology following 16-h treatment with C₁₂K-2β₁₂. Approximately 1.5 × 10⁷ H. pylori cells were sampled following 16 h of exposure to PBS (A), 100 μM amoxicillin (B), 20 μM K-5α₁₂ (C), or 20 μM C₁₂K-2β₁₂ (D). The cells were examined at a low magnification (×13,500) on a transmission electron microscope (TEM) equipped with a 4MP SPOT Insight CCD camera. (A and C) Normal spiral- or comma-shaped rod morphology of H. pylori cells and intact cell membranes following treatment with PBS and the K-5α₁₂ control peptide (negative controls), respectively. (B) Amoxicillin-treated cells show characteristically swollen cells (white arrows) and detachment of the inner membrane from the outer membrane (white arrowhead). (D) C₁₂K-2β₁₂-treated cells show the appearance of numerous ghost cells (white arrows) and the formation of electron-dense structures (white arrowhead) within the extracellular medium and inside the cells. The data are representative images from two independent experiments.

H. pylori samples that were exposed to the C₁₂K-2β₁₂ OAK peptide (Fig. 2D) showed a significantly altered morphological phenotype compared to control cells. Approximately 20 to 35% of the C₁₂K-2β₁₂-treated cells appeared as ghost cells, and the media and intact cells became filled with extracellular electron-dense structures. The appearance of the electron-dense structures, which were absent in PBS- or K-5α₁₂-treated samples, is consistent with extravasation of the bacterial cytoplasmic contents due to cell lysis and/or the presence of membrane pores. Furthermore, the appearance of the electron-dense material in nonghost intact H. pylori cells suggested translocation of C₁₂K-2β₁₂ into the bacterial cytoplasm and the subsequent formation of aggregate molecules with the cytoplasmic contents. Additionally, treatment of cells with C₁₂K-2β₁₂ resulted in changes in overall bacterial shape; the typical spiral and comma rod-shaped bacterial cells seen in the controls (Fig. 2A and C) were not observed in C₁₂K-2β₁₂-treated cells (Fig. 2D). Instead, consistent with modification of the cell wall and/or cytoskeleton structure (2), these cells became coccoid. The notable lack of intracellular and extracellular electron-dense structures in amoxicillin-treated control H. pylori cells and the presence of significantly higher numbers of ghost cells in the C₁₂K-2β₁₂-treated samples suggest differences in the modes of action of the two drugs. In order to better visualize and analyze the ultrastructural morphology of the cell wall and cytoplasmic changes observed at lower magnification (×13,500; Fig. 2), we additionally examined each sample at higher magnification following treatment of bacteria with the peptide for 16 h. We observed extensive damage of the cell wall, and these results are presented as Fig. S1 in the supplemental material.

The formation of ghost cells by the C₁₂K-2β₁₂-treated H. pylori samples suggested that the OAK peptide was inducing the formation of pores in the bacterial membrane that allowed cellular contents to leak out. However, no visible pores in the cell walls were observed at the 16-hour time point (see Fig. S1 in the supplemental material). Given the fast killing rate of C₁₂K-2β₁₂, the observation that no viable bacteria were evident after treatment with C₁₂K-2β₁₂ for 6 to 8 h (47), and the rapid permeation of the membrane demonstrated by the NPN assays (Fig. 1), we hypothesized that the extent of cell damage was too great at the late time point to see changes in the membrane. Thus, in order to investigate the early events related to treatment of H. pylori with C₁₂K-2β₁₂, we next used TEM to examine the ultrastructure of the cells harvested 2 h posttreatment with the OAK peptide. Analysis of low-magnification (×13,500) images revealed that 4 to 8% of the H. pylori samples appeared as ghost cells at this early time point (Fig. 3A). Additionally, we visualized the presence of intracellular and extracellular electron-dense structures, as well as less dramatic changes in cell morphology. The C₁₂K-2β₁₂ effects at this early time point are consistent with the observed NPN membrane permeabilization within 15 to 30 min (Fig. 1) and our previously reported in vitro killing kinetics of 90% of the bacterial population within 2 to 3 h of OAK treatment (47). At higher magnifications (×64,000 and ×130,000), we visualized extensive membrane perturbation of H. pylori following two hours of exposure to C₁₂K-2β₁₂ (Fig. 3B, C, and D). Consistent with the appearance of ghost cells, cell lysis, and leakage of cellular contents observed at later time points (Fig. 2D), membrane damage could be clearly visualized by 2 h after exposure to the OAK peptide. Outer membrane damage was evidenced as the formation of pore-like structures with a width measured between 50 nm (Fig. 3B, black arrows) and over 400 nm (Fig. 3C, black arrowheads). In addition, we observed large areas of membrane sloughing (S), which led to complete separation and peeling away of the OM from the plasma membrane (Fig. 3D). Taken en masse, these data indicate that the C₁₂K-2β₁₂ OAK peptide induces H. pylori cell death through dramatic disruption of the bacterial cell membrane, which in turn leads to leakage of the cytoplasmic contents and bacterial cell lysis.

Fig 3.

Distinct classical pores and sloughing of the outer membrane of C₁₂K-2β₁₂-treated H. pylori cells as visualized with a TEM. H. pylori G27 was incubated with 20 μM C₁₂K-2β₁₂ for 2 h. (A) TEM analysis at a low magnification (×13,500) shows morphological changes at this early time point. (B to D) Representative images showing H. pylori cells observed at a higher magnification (×130,000). The black arrows in panel B indicate the loss of a small section of outer membrane of the bacterium, while the black arrowheads in panel C indicate the borders of a large 400-nm section of missing membrane. In panel D, sloughing (S) of a large section of the outer membrane and electron-dense structures (E) inside the cell cytoplasm are indicated. The data are representative results from two independent experiments.

There is evidence that several cationic antimicrobial peptides exert their effects through interaction with intracellular nucleic acids (4, 32). Moreover, it has been reported that the hexamer OAK peptide C₁₂K-5α₈ is able to traverse the bacterial membrane and to bind to bacterial DNA without overt disruption of the cell wall (67). Given that we saw virtually no change in NPN fluorescence at low concentrations of C₁₂K-2β₁₂ (Fig. 1A), we next considered the possibility that at lower concentrations, C₁₂K-2β₁₂ might have the ability to permeate the bacterial cell membrane into the cytoplasm. Since we observed little immediate killing of samples treated with 20 μM C₁₂K-2β₁₂ (Fig. 1B), we utilized confocal microscopy and this concentration of fluorescently labeled C₁₂K-2β₁₂ to study localization of the peptide 30 min after exposure. Analysis of single-plane sections through the center of the bacteria (Fig. 4A) and series of z-stack images taken from the surface of the bacterial cell moving through the interior of the cell (Fig. 4B) revealed that little C₁₂K-2β₁₂ was visible near the cell surface but was instead colocalized with bacterial cytoplasmic DNA prior to cell death. These data indicate that the peptide has the ability to cross the bacterial membrane and enter the cytosol.

Fig 4.

DyLight 488-conjugated C₁₂K-2β₁₂ rapidly colocalizes with DNA in the bacterial cytoplasm upon treatment of H. pylori with the peptide. Bacterial cells were treated with DyLight 488-conjugated C₁₂K-2β₁₂ (green) for 30 min. The cells were washed and fixed, and cytoplasmic DNA was stained with DAPI (blue). The cells were then visualized by confocal microscopy. Images represent a single plane taken through the center of the bacterium (A) or six sequential z-stack slices starting near the surface of bacteria and moving toward the middle (B). Both verify the presence of C₁₂K-2β₁₂ in the bacterial cytoplasm, as significant green fluorescence is seen only within the cell center. Differential interference contrast (DIC) and overlay images are included for clarity. All images were taken with a magnification of ×400. The images are representative of three experiments.

C₁₂K-2β₁₂ binds to bacterial nucleic acids (DNA and RNA) and proteins.

Given that C₁₂K-2β₁₂ had the ability to traverse into the cytoplasm and appeared to colocalize with DNA, we hypothesized that C₁₂K-2β₁₂ may directly interact with cytoplasmic nucleic acids and proteins. To determine whether C₁₂K-2β₁₂ could interact with DNA, we first investigated whether preincubation of DNA with the peptide could prevent restriction enzyme cleavage of the H. pylori G27 endogenous plasmid. Undigested plasmid and plasmid preincubated with K-5α₁₂ were included as negative controls. As shown in Fig. 5, undigested native plasmid ran as two bands (nicked and supercoiled species) (lane 2), and digestion with the MfeI restriction enzyme (lane 4) resulted in the appearance of linearized bands. The overall pattern of plasmid migration was slightly affected by preincubation with C₁₂K-2β₁₂ (Fig. 5, lane 3); some DNA failed to migrate into the gel, and bands appeared less distinct, suggesting the formation of DNA aggregates. Moreover, MfeI digestion was completely blocked upon preincubation with C₁₂K-2β₁₂ (lane 5). The blockage of enzymatic activity was specific to C₁₂K-2β₁₂, since preincubation with the control peptide K-5α₁₂ (lanes 6 and 7) resulted in a digestion pattern similar to that of nontreated DNA (lane 4). Similar results were also obtained using HindIII, EcoRI, SpeI, and XmnI restriction enzymes (data not shown). Taken together, these results suggest that C₁₂K-2β₁₂ is able to tightly bind to bacterial nucleic acid and that this binding inhibits restriction enzyme digestion.

Fig 5.

Functional inhibition of restriction enzyme activity by C₁₂K-2β₁₂. The H. pylori G27 strain endogenous plasmid, which contains MfeI restriction enzyme-cut sites, was incubated for 1 h at 37°C in the presence (+) or absence (−) of 40 or 80 μM K-5α₁₂ (control peptide) or 40 μM C₁₂K-2β₁₂ prior to digestion with MfeI (+). Agarose gel electrophoresis (1%) was then used to analyze the digestion. Lane 1, molecular weight markers; lane 2, native undigested plasmid; lane 3, C₁₂K-2β₁₂-treated native plasmid; lane 4, MfeI-digested plasmid; lane 5, C₁₂K-2β₁₂-treated MfeI-digested plasmid; lane 6, 80 μM K-5α₁₂-treated MfeI-digested plasmid; lane 7, 40 μM K-5α₁₂-treated MfeI-digested plasmid. Since the banding patterns for native plasmid (lane 2), native plasmid incubated with C₁₂K-2β₁₂ (lane 3), and C₁₂K-2β₁₂-treated MfeI-digested plasmid (lane 5) are similar, C₁₂K-2β₁₂ incubation protected the plasmid DNA from restriction enzyme digestion. This gel is a representative of the gels in three independent experiments.

To conclusively prove interaction of C₁₂K-2β₁₂ with DNA and to rule out the possibility that enzymatic activity was inhibited via direct interaction of C₁₂K-2β₁₂ with the restriction enzymes, we also investigated the ability of the peptide to directly bind to DNA. To this end, the ability of the peptide to change the electrophoretic mobility of DNA was examined by electrophoretic mobility shift assay (EMSA). As shown in Fig. 6A, EMSA analysis of a labeled PCR product representing a portion of the H. pylori oorD gene (377 bp) demonstrated that in the absence of C₁₂K-2β₁₂ (lane 2), oorD migrated as a single 377-bp band. However, upon the addition of C₁₂K-2β₁₂ (lane 3 to 7), this pattern was altered such that the oorD fragment was no longer able to migrate into the gel. Similar results were obtained when a labeled fragment of the H. pylori rpoB gene (142 bp) (Fig. 6A), the E. coli chloramphenicol acetyltransferase gene (data not shown), or human antibody light-chain DNA fragment (data not shown) was used as the DNA target. This suggests that C₁₂K-2β₁₂ is able to efficiently bind to DNA in a sequence-independent manner. DNA binding was specific to C₁₂K-2β₁₂, since incubation of all of the fragments with the K-5α₁₂ control peptide (Fig. 6A, lanes 1 for both DNA fragments and data not shown) resulted in no change in DNA migration. Thus, the ability of C₁₂K-2β₁₂ to bind DNA may be due to specific structure-function relationships to the target and not merely electrostatic attraction to anionic DNA.

Fig 6.

Gel retardation and modified Western blot assays demonstrate the ability of C₁₂K-2β₁₂ to bind H. pylori nucleic acids and proteins. (A) oorD or rpoB DNA was end labeled with [γ-³²P]ATP according to standard procedures (see Materials and Methods) and was incubated for 1 h at 37°C in the presence of different concentrations (0 μM [−], 1.3 μM, 2.5 μM, 5 μM, 10 μM, and 20 μM) of C₁₂K-2β₁₂. Twenty micromolar K-5α₁₂, an OAK peptide that has no activity against H. pylori, was included as a control (lane 1) for each DNA sample. (B) ureA or oorD mRNA was end labeled with [α-³²P]UTP, according to standard procedures (see Materials and Methods), and was incubated for 1 h at 37°C in the presence of increasing amounts of C₁₂K-2β₁₂ (0 μg, 1.0 μg, or 5 μg). Five micrograms of K-5α₁₂ was included as a control. Increased retention of DNA and RNA in the wells corresponding to increasing concentrations of C₁₂K-2β₁₂ demonstrate binding of the peptide to the nucleic acids. (C) Analysis of H. pylori proteins in Coomassie blue-stained SDS-polyacrylamide gel (left panel) and modified Western blot (right panel). Following separation and transfer of proteins, the blot was probed with fluorescently labeled C₁₂K-2β₁₂ (DyLight 488) to identify binding to purified H. pylori Fur and lysate protein dilutions of 1:2, 1:4, and 1:8. PBS containing no protein was used as a negative control. These are representative gels and blots from three independent experiments.

Because C₁₂K-2β₁₂ strongly interacted with DNA, we next sought to determine whether C₁₂K-2β₁₂ could also bind to RNA. Thus, the above-described assays were repeated using in vitro-transcribed ureA and oorD riboprobes instead of PCR products. As shown in Fig. 6B, approximately 60% of the ureA RNA probe was efficiently shifted by the addition of C₁₂K-2β₁₂ at a peptide/RNA ratio (wt/wt) of 1:1 (lane 3). Nearly 90% of the RNA was bound when the ratio (wt/wt) was increased to 5:1 (lane 4), thus indicating concentration-dependent binding of C₁₂K-2β₁₂ to the RNA target. Similar patterns were observed when the oorD RNA probe was used (Fig. 6B). Importantly, the addition of the K-5α₁₂ control peptide resulted in no change in migration of either riboprobes (lanes 2 of both RNA fragments in Fig. 6B). Taken en masse, these results suggest that in addition to bacterial lysis due to membrane pore formation, at low concentrations, C₁₂K-2β₁₂ may also exert an antimicrobial effect by traversing the membrane and binding to both DNA and RNA.

Finally, we monitored the ability of C₁₂K-2β₁₂ to bind H. pylori proteins by using a modified Western blot method and fluorescently labeled peptide. The separation of purified H. pylori Fur protein or total cell lysates on an SDS-polyacrylamide gel followed by probing with fluorescent C₁₂K-2β₁₂ revealed that the peptide binds numerous bacterial proteins (Fig. 6C). Given that proteins and nucleic acids are key molecules involved in replication and synthesis of micro- and macromolecules, we speculate that the binding of C₁₂K-2β₁₂ to nucleic acid and proteins may also have significant implications for survival of H. pylori.

C₁₂K-2β₁₂ shows strong in vivo efficacy against H. pylori in the gerbil model of infection.

Having previously demonstrated strong in vitro efficacy of C₁₂K-2β₁₂ against H. pylori (47) and having shown two possible mechanisms of action of the peptide (described above), we next sought to determine the in vivo efficacy of C₁₂K-2β₁₂ as a therapeutic agent against H. pylori. To this end, we utilized the Mongolian gerbil model of infection, which has become a popular animal model to study H. pylori colonization and disease due to the similarity of clinical pathophysiology found in this model compared to infected humans (18). Figure 7 presents a schematic of the challenge and treatment schedule followed during the study and shows the sampling times of the animal stomach tissue for determination of viable CFU. Animals infected with H. pylori strain 7.13 (18) were treated with one, three, or five doses of C₁₂K-2β₁₂ orogastrically either 1 day (Fig. 8A) or 1 week (Fig. 8B) postinfection. Two concentrations of drug were administered by applying the following dosing regimen: 1 mg C₁₂K-2β₁₂/kg body weight or 4 mg C₁₂K-2β₁₂/kg body weight. The animals were sacrificed 1 week after the end of treatment, and the density of H. pylori colonization was determined by plating and enumeration of viable CFU. The CFU per gram of stomach in the C₁₂K-2β₁₂-treated gerbils was then compared to counts determined from the corresponding control animals that were administered PBS instead of the C₁₂K-2β₁₂ peptide.

Fig 7.

Gerbil infection and treatment schedule utilized to investigate in vivo efficacy of C₁₂K-2β₁₂ against H. pylori. All gerbils were infected with approximately 8 × 10⁸ CFU of H. pylori strain 7.13 on day 0 (black arrowhead) and divided into two categories: (i) animals that received treatment 1 day postinfection (schedule shown above the black line in the middle of the figure) and (ii) animals that received treatment 1 week postinfection (schedule shown below the black line in the middle of the figure). Each category consisted of three subgroups of dose treatments: (i) single dose, (ii) three doses, or (iii) five doses. In each dose treatment subgroup, there were three groups of animals based on treatment regimen; animals were orogastrically administered 4 mg C₁₂K-2β₁₂/kg body weight or 1 mg C₁₂K-2β₁₂/kg body weight or PBS as a placebo control. All treatment doses were administered sequentially once daily by oral gavage. The initiation and completion of treatment in all groups for each dose treatment are shown (long thin black arrows). Shorter thicker black arrows represent 1 week after completion of treatment and represent the day gerbils were sacrificed for stomach tissue sampling for surviving CFU.

Fig 8.

Therapeutic efficacy of C₁₂K-2β₁₂ against H. pylori in the gerbil model of colonization. Ninety gerbils divided into 18 groups of five gerbils per group were orogastrically infected with approximately 8 × 10⁸ CFU of H. pylori strain 7.13. The gerbils were then randomly assigned into two equal categories based on whether treatment was initiated 1 day postinfection (A) or 1 week postinfection (B) as described in the legend to Fig. 7. Animals were administered a single dose, 3 doses, or 5 doses of 1 mg C₁₂K-2β₁₂/kg body weight or 4 mg C₁₂K-2β₁₂/kg body weight or PBS as a placebo control. One week posttreatment, the gerbils were sacrificed. The stomachs of the gerbils were harvested, weighed, and homogenized, and the homogenates were plated to determine the CFU of H. pylori colonization. Data are expressed as CFU/gram of stomach, and each symbol represents the value for a single animal. The geometric mean for each group is plotted for comparison (short black bar). The P values are from one-way ANOVA analyses of H. pylori colonization levels across each dose treatment group. P values greater than 0.05 were considered nonsignificant difference (NS) in the means of bacterial colonization levels as determined by comparison of C₁₂K-2β₁₂-treated group and the corresponding PBS-treated control animals.

Figure 8A presents colonization levels for animals whose treatment was initiated 1 day postinfection. Animals that received a single dose of either 1 mg/kg body weight or 4 mg/kg body weight C₁₂K-2β₁₂ showed significantly lower levels of H. pylori colonization compared to PBS-treated control animals (P = 0.034 and P = 0.023 by one-way ANOVA, respectively). Similarly, we observed significantly lower levels of colonization in the stomachs of gerbils treated with 5 doses of 4 mg/kg body weight and 1 mg/kg body weight C₁₂K-2β₁₂ compared to PBS-treated control gerbils (P = 0.0001 and P = 0.01 by one-factor ANOVA, respectively) (Fig. 8A). Although the differences in the bacterial colonization density of gerbils that received three doses of either regimens of C₁₂K-2β₁₂ (1 mg/kg or 4 mg/kg) were not statistically significant, the trends were the same as the other dosing regimens; there was a reduction in the bacterial load in the C₁₂K-2β₁₂-treated animals compared to the PBS-treated animal control.

In order to more closely imitate a clinical situation in which H. pylori infection is established before commencement of treatment, we next challenged animals 1 week postinfection with H. pylori; at this point, the bacterial infection is established in the stomach, and H. pylori has typically achieved a near-maximal bacterial load within the gerbil gastric mucosa (51). In contrast to the treatment 1 day postinfection, C₁₂K-2β₁₂ showed no effect on H. pylori colonization density in animals that received a single dose of 1 mg/kg or 4 mg/kg of the peptide (Fig. 8B). Similarly, treatment of gerbils with three doses of 1 mg of C₁₂K-2β₁₂/kg of body weight showed no significant effect on colonization levels compared to the PBS-treated control gerbils (Fig. 8B). However, comparison of bacterial colonization levels in animals that received three doses of 4 mg/kg of C₁₂K-2β₁₂ with the corresponding PBS-treated animal controls revealed a significant reduction in the H. pylori colonization density (P < 0.043 by one-factor ANOVA) (Fig. 8B). Notably, animals that received 5 doses of either 1 mg/kg or 4 mg/kg of C₁₂K-2β₁₂/kg showed a significant reduction in H. pylori colonization compared to PBS-treated control animals (P = 0.021 and P = 0.001 by one-way ANOVA, respectively) (Fig. 8B). Taken together, these data indicate that therapeutic treatment of gerbils with five doses of C₁₂K-2β₁₂ showed the greatest effect in reduction of H. pylori colonization whether the treatment was initiated 1 day (Fig. 8A) or 1 week (Fig. 8B) postinfection. Similarly, administration of more concentrated doses (4 mg/kg) of C₁₂K-2β₁₂ resulted in lower H. pylori colonization than lower doses, regardless of the number of treatments. Thus, the results are consistent with dose- and concentration-dependent effects and demonstrate strong in vivo efficacy of C₁₂K-2β₁₂ against H. pylori in the gerbil model of infection.

DISCUSSION

In the past half century, only three new classes of antibiotics have entered the clinics, lipopeptides, oxazolidinones, and streptogramins (11, 48, 52), and each of these drugs targets Gram-positive bacterial infections. This trend signifies the slow pace of development of new drugs and highlights the dire lack of novel drugs against Gram-negative bacteria such as H. pylori. Indeed, H. pylori has developed resistance to nearly all conventional antibiotics currently used in the clinic for treatment of the infection. Furthermore, the requirement for the use of multiple drugs for extended periods of time represents an increased cost to the patient and often results in negative antibiotic-associated side effects; collectively, these factors often lead to noncompliance by patients. Therefore, H. pylori clearly represents a bacterium that demands that new antimicrobial strategies be explored for the effective future treatment of the infection. To this end, we recently screened a panel of peptidomimetic antimicrobial peptides and found that the OAK peptide C₁₂K-2β₁₂ demonstrated strong in vitro efficacy against numerous strains of H. pylori (47). In vitro, C₁₂K-2β₁₂ was effective at low pH and against nonreplicating or slow-growing bacteria. Furthermore, the OAK peptide demonstrated strong irreversible PAE and bactericidal effects against H. pylori (47). As an extension of that previous work, we sought to characterize the mechanism of action of C₁₂K-2β₁₂ and to determine the in vivo efficacy of the OAK peptide. Herein, we show that the bactericidal effect of the drug is exerted through the induction of distinct classical pores in the H. pylori cell membrane. Furthermore, at lower concentrations, the peptide is able to cross bacterial membranes, enter the cytoplasm, and tightly bind to bacterial nucleic acids and proteins, which we speculate may block bacte-rial replication, transcription, and translation. To our knowledge, this study is the first to demonstrate action of an OAK peptide against numerous bacterial targets.

Transmission electron microscopy revealed the presence of pores, ghost cells, membrane blebs, and areas of sloughing of the outer membranes of H. pylori cells treated with C₁₂K-2β₁₂. Interestingly, these results are similar to the reported effects of α_s2 casein-derived peptide on E. coli (45) and suggest that pore formation and cell lysis may be a general mechanism of action of antimicrobial peptides. Furthermore, these observations largely explain the previously described irreversible PAE and rapid in vitro bactericidal effects of C₁₂K-2β₁₂ (47).

Broadly speaking, one of the most attractive advantages of antimicrobial peptides is their ability to exert antibacterial activity extremely rapidly, which is not the case for many of the currently used antibiotics. For instance, the beta-lactam antibiotic amoxicillin targets the bacterial cell wall by inhibiting synthesis of peptidoglycan. However, the drug requires active replication to be the most effective. While amoxicillin is effective against replicating H. pylori (73), it lacks significant PAE against the bacterium (27, 47). Moreover, it exhibits relatively slow antibacterial activity against H. pylori (22, 29, 47), which is likely due to the fact that the in vitro division time of H. pylori is long, approximately 3 to 7 h (23, 62). In vivo, not only is the doubling rate likely longer, but amoxicillin efficacy is further compromised by the gastric acidic pH (7, 16). Thus, the long division time combined with the requirement for active replication likely helps explain the long lag phase in activity. Similarly, during stationary phase, relatively slower activity is also evidenced for antibiotics such as penicillins and vancomycin, which target the bacterial cell wall (53) via mechanisms involving peptidases and peptidoglycans, respectively, that must be synthesized before they can be properly targeted (77, 83).

As a potential class of new drugs, antimicrobial peptides do not suffer from some of these limitations due to the fact that they typically have a broader more nonspecific target within the bacterial cell. Indeed, C₁₂K-2β₁₂ permeation of the bacterial membrane was very rapid; significant dose-dependent membrane disruption was observed within 15 min and reached a maximum within 30 min at the higher concentrations (Fig. 1). These permeation kinetics correlate with our previously reported in vitro killing kinetics of C₁₂K-2β₁₂ (47). Moreover, compared to amoxicillin, C₁₂K-2β₁₂ showed superior ability to permeate the bacterial membrane; similar membrane permeation levels by amoxicillin required 10-fold the concentration of C₁₂K-2β₁₂, supporting differential modes of action.

In many cases, the activity of antimicrobial peptides is dependent on charge, hydrophobicity, and amphipathic structure (4). Indeed, cationic peptides such as cecropin A and magainin 2 have been shown to utilize electrostatic attractions between the peptide and the negatively charged molecules on the surface of bacterial membrane as a means of facilitating activity (76, 88). Given the cationic nature of the C₁₂K-2β₁₂ OAK peptide, a similar mechanism is likely used in interaction with the bacterial membrane; phospholipids and acidic lipopolysaccharide (LPS) molecules are likely targeted (46). However, the fact that we observed a dose-dependent response in the permeation of the bacterial cell wall by C₁₂K-2β₁₂ (Fig. 1) suggests that a critical ratio between peptide molecules and target cell membrane molecules exists. C₁₂K-2β₁₂, like other cationic peptides, is amphipathic, a property that likely allows the peptide molecules to orient and insert themselves into the bacterial membrane (60, 86). Typically, this insertion is thought to be responsible for the formation of transmembrane pores via the interaction of hydrophobic regions of the peptide with lipid components of the membrane and association of hydrophilic regions of the peptide with phospholipid head groups within the membrane (4). Thus, perhaps a particular concentration of C₁₂K-2β₁₂ is required to effectively facilitate these types of interactions and to induce pore formation within the H. pylori membrane. Though the exact mechanism by which antimicrobial peptides interact with target membranes is not clearly understood, electrostatic interactions between the cationic peptide and anionic bacterial cell membrane are no doubt crucial. Furthermore, the exact amino acid composition, amphipathicity, size, and cationic charge of the peptide no doubt affect how an individual peptide would attach to and insert into a membrane bilayer to form transmembrane pores (38).

Permeation of the bacterial cell membranes was previously believed to be the sole mechanism of action of antibacterial peptides. However, recent studies have demonstrated that some antimicrobial peptides, including certain OAK peptides, achieve their effects through alternative modes of action (34, 58, 67). Some peptides have the ability to traverse to the bacterial cytoplasm and to bind to intracellular targets such as nucleic acids (DNA and RNA) (67) and proteins (58); these interactions may result in bacterial death (3, 4, 57). Indeed, human defensin HNP-1 (human neutrophil peptide 1), cathelicidin, magainin 2, and buforin II, a 21-amino-acid peptide derived from a natural peptide isolated from the Asian toad, have the ability to permeate membranes without significant disruption and traverse into the cytoplasm to bind bacterial nucleic acids (49, 57). We found that C₁₂K-2β₁₂ tightly bound to DNA and RNA (Fig. 6), whereas similar OAK peptides that were not active against H. pylori, K-5α₁₂ (Fig. 5 and 6) and 2β₁₂ (data not shown), did not bind either nucleic acid. This dramatic difference between OAKs perhaps suggests that protein and DNA/RNA binding is not a mere consequence of electrostatic interactions between the positively charged peptide and the polyanionic DNA or RNA. Conversely, binding may be attributed to a specific structure-activity relationship of each of the OAKs and is consistent with the structures of C₁₂K-2β₁₂, K-5α₁₂, and 2β₁₂ (47, 64). Equally consistent with this idea is lactoferricin B, a 25-amino-acid-residue lactoferrin-derived peptide with a net charge of +8 that does not bind to DNA (80, 81), whereas liver-expressed antimicrobial peptide 2 (LEAP-2), an antibacterial peptide with a net charge of +9 binds efficiently to DNA (30). Thus, a positive charge is not sufficient, and perhaps distinct differences in overall structural attributes are required to promote nucleic acid binding. That overall peptide structure might be more important than net charge and perhaps hydrophobicity for a functional mode of action of the OAK peptides is also evident when one contrasts the biophysical and chemical properties of C₁₂K-2β₁₂ and 2β₁₂. Both are linear OAK peptides that display similar net charges and hydrophobicities (+5 and 51 for C₁₂K-2β₁₂ and +5 and 38.1 for 2β₁₂, respectively). However, despite this similarity, 2β₁₂ lacks antimicrobial activity against H. pylori (47). Regardless of the mechanism of interaction, the ability of C₁₂K-2β₁₂ to bind H. pylori nucleic acids and proteins at low concentrations suggests that this OAK peptide may have the ability to inhibit DNA, RNA, and protein synthesis, which would in turn lead to inhibition of bacterial growth and compromise bacterial survival similar to what has been seen with indolicidin-treated S. aureus (54). Indeed, the cytoplasmic electron-dense structures we observed (Fig. 2 and 3) may be the result of C₁₂K-2β₁₂-induced aggregation of nucleic acids and protein. The fact that C₁₂K-2β₁₂ binding to DNA showed no sequence specificity (Fig. 6) suggests that the peptide may have broad-spectrum activity and will likely affect the normal gut flora. Conversely, binding of the peptide to human DNA should not be a concern due to the selective nature of amphipathic peptides against anionic bacterial membranes; they exhibit no or low affinity for the zwitterionic membranes of mammalian cells (74).

In vivo testing of the efficacy of the C₁₂K-2β₁₂ OAK peptide against H. pylori in the gerbil model of infection demonstrated that therapeutic monotreatment of animals with C₁₂K-2β₁₂ significantly reduced H. pylori colonization (Fig. 8). This suggests that, despite the presence of a physical mucous barrier and the harsh acidic and proteolytic environment of the stomach, which are commonly associated with reduced bioavailability of some antibiotics, C₁₂K-2β₁₂ is stable and active in this environment. Furthermore, C₁₂K-2β₁₂ must be able to penetrate the mucous layer to reach the mucosal surface where the bacterium resides; however, the mechanism by which this is accomplished is not currently clear. Indeed, orogastrically administered C₁₂K-2β₁₂ could reach and eventually interact with the bacteria in vivo following absorption via two possible ways: (i) directly within the stomach where the peptide simply diffuses across the mucous layer to reach the gastric epithelium and/or (ii) indirectly via absorption in the intestine and permeation through the vasculature back to the gastric epithelium. Regardless of the mechanism by which it reaches the bacteria, C₁₂K-2β₁₂ is clearly effective in vivo.

Interestingly, therapeutic treatment of animals 1 week postinfection with 3 doses of 4 mg C₁₂K-2β₁₂/kg body weight resulted in significant reduction in stomach colonization (P = 0.043 by one-factor ANOVA) (Fig. 8B). However, the same dose and concentration administered 1 day postinfection lacked this magnitude of therapeutic effect; this may suggest interplay of C₁₂K-2β₁₂ with the immune response similar to what has been reported with other AMPs (10, 40, 41, 79). Indeed, antimicrobial peptides play a multifactorial role in the body, including modulation of the immune response. Cathelicidins and β-defensins have been shown to have the ability to recruit leukocytes directly or to stimulate the production of cytokines such as interleukin 8 (IL-8), monocyte chemotactic peptide 1 (MCP-1), and gamma interferon (IFN-γ), thereby indirectly recruiting other effector cells (28, 41, 42). We postulate that allowing infection to establish for 1 week before the initiation of C₁₂K-2β₁₂ treatment may have resulted in increased bacterial numbers that resulted in the induction of an immune response that acted synergistically with the effects of C₁₂K-2β₁₂ to further reduce colonization. Indeed, our explanation is corroborated by previous studies which demonstrated the appearance of serum immunoglobulin G immune response to H. pylori antigens at 1 week postinfection of Mongolian gerbils (39, 84). In humans, H. pylori typically colonizes for extended periods prior to clinical symptoms. Thus, it would be of interest to determine whether chronic colonization in the face of inflammation results in adaptation of the bacteria to become more resistant to conventional antibiotics and C₁₂K-2β₁₂.

One of the significant limitations in the development of AMPs for use in the clinic is cytotoxicity to host cells. We observed no mortality or obvious observable side effects due to administration of C₁₂K-2β₁₂, suggesting no or minimal physiological toxicity. Furthermore, these data agree with our previously reported in vitro results in which C₁₂K-2β₁₂ showed nonhemolytic properties (47). However, clearly detailed toxicology, pharmacokinetic, and pharmacodynamic studies are required to determine systemic or gastrointestinal tract bioavailability and potential safety parameters of the peptide.

In conclusion, our data demonstrate a dual mechanism of action of C₁₂K-2β₁₂ against H. pylori and suggest that the previously observed in vitro bactericidal effect of the peptide against H. pylori (47) is largely due to bacterial lysis as a result of pore formation within the bacterial membrane as well as binding of the peptide to bacterial proteins and nucleic acids, which in turn may inhibit critical cellular processes that may be important for bacterial survival. The data further demonstrate strong in vivo efficacy of C₁₂K-2β₁₂ when used as monotherapy against H. pylori in the gerbil model of infection. En masse, these results support the potential development of C₁₂K-2β₁₂ as a novel class of anti-H. pylori therapeutics.