OMN6 a novel bioengineered peptide for the treatment of multidrug resistant Gram negative bacteria

New antimicrobial agents are urgently needed, especially to eliminate multidrug resistant Gram-negative bacteria that stand for most antibiotic-resistant threats. In the following study, we present superior properties of an engineered antimicrobial peptide, OMN6, a 40-amino acid cyclic peptide based on Cecropin A, that presents high efficacy against Gram-negative bacteria with a bactericidal mechanism of action. The target of OMN6 is assumed to be the bacterial membrane in contrast to small molecule-based agents which bind to a specific enzyme or bacterial site. Moreover, OMN6 mechanism of action is effective on Acinetobacter baumannii laboratory strains and clinical isolates, regardless of the bacteria genotype or resistance-phenotype, thus, is by orders-of-magnitude, less likely for mutation-driven development of resistance, recrudescence, or tolerance. OMN6 displays an increase in stability and a significant decrease in proteolytic degradation with full safety margin on erythrocytes and HEK293T cells. Taken together, these results strongly suggest that OMN6 is an efficient, stable, and non-toxic novel antimicrobial agent with the potential to become a therapy for humans.


Methods
Peptide synthesis. OMN6, a C-terminal amidated 40-amino acid long cyclic peptide with a molecular weight of 4339. 36 Da, was synthetized as an acetate salt, using only l-amino acids, by Caslo ApS (Lyngby, Denmark; for stability, E. coli GFPuv, cytotoxicity, and hemolysis experiments) or Ambiopharm Inc. (North Augusta, SC, USA; for antimicrobial activity and time-kill studies). Here below are the details of a representative batch synthesis executed by Caslo ApS. Fmoc solid-phase peptide synthesis strategy on Rink Amide MBHA resin (loading capacity: 0.4 mmol/g) was used. Cyclization was achieved by oxidation in 50% DMSO/H 2 O under peptide concentration of 3 mg/mL. Cyclization success was assessed using ABI Voyager DE-PRO (ThermoFisher Scientific Corp., Waltham, MA, USA) Matrix Assisted Laser Desorption-Ionization Time of Flight Mass Spectrometry (MALDI-TOF-MS) with a matrix of 2,5-dihydroxybenzoic acid, and is presented in Figures S4 and  S5. Peptide samples were purified by preparative reverse-phase high-performance liquid chromatography (RP-HPLC) using Agilent HP 1100 HPLC system (Agilent Technologies, Santa Clara, CA, USA) with an Ultimate C18 column (10 μm, 100 Å, 50 × 250 mm). A solvent system consisting of solvent A (0.1% TFA, 2% CH 3 CN in water) and solvent B (90% CH 3 CN/H 2 O) was used for elution with a gradient of 25-50% B in 60 min, and the absorbance was detected at 220 nm. Solvent was removed by lyophilization to afford a fluffy powder. The salt exchange from trifluoroacetate salt to acetate salt was also achieved by preparative RP-HPLC. The peptide was eluted by 0.02 M NH 4 Ac (approximate pH = 6) and solvent B. The elution solution containing target peptide was collected and then lyophilized. The purity (> 95%) of the purified material was assessed by analytical RP-HPLC with a Kromasil C18 column (4.6 × 250 mm), and the absorbance was detected at 220 nm.
Modeling analysis. Exploratory studies of OMN6 structure were conducted by homology modeling from the helix-hinge-helix structure of papiliocin as was previously described and deposited in the Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB) under ID number 2LA2 25 . For modeling purposes, disulfide bond and structure refinement were estimated using the UCSF Chimera software 26 , as well as the PELE 27 , FG-MD 28 and MRC 29 web servers. Peptide stability. 10   www.nature.com/scientificreports/ independent experiments. The statistical significance of differences between samples was analyzed by unpaired Student's t-tests. Differences with p < 0.05 were considered statistically significant.

Results
Peptide design and stability studies. The generation of new AMPs for therapeutic use was based on the native Cecropin family which serve as potent antimicrobial agents used by insects for millions of years. Biochemical engineering of our lead compound OMN6, a C-terminal amidated 40 amino-acid long cyclic peptide containing only l-amino acids, was based on the sequence of Cecropin A from the Hyalophora cecropia moth. l-Cysteines were incorporated at both the N-and C-terminal ends of the native linear peptide to instigate a disulfide bond formation. By this simple bridge formation, we yielded a cyclic peptide conformation that confers enhanced peptide stability in order to overcome the major hurdle in the development of AMPs as therapeutic agents 14,24 . Also, a l-Methionine amino acid was added at the N-terminal end (Fig. 1A). The goal of the l-Methionine addition was to enable potential peptide production by biomanufacturing using yeast or bacteria, as Methionine is the universal amino acid for initiation of protein synthesis in living organisms 34 . However, OMN6 has been produced so far only by chemical ways using solid-phase peptide synthesis, as the synthetic production process for OMN6 is robust and devoid of any endotoxins from bacterial source that may appear when using biosynthesis 35 . Predictions of OMN6 cyclic conformation were performed by modeling studies (Fig. 1B), according to previous structure studies on the Cecropin family 25 . In Fig. 1C, a helical wheel diagram of OMN6 shows the amphipathicity of the first α-helix (from Lys 3 to Ala 24 ) with hydrophobic residues in the upper region and hydrophilic residues in the lower region. The second α-helix (from Ala 27 to Lys 39 ) is hydrophobic and not amphipathic. As displayed in Fig. 1D, exploratory three-dimension modeling studies suggest that OMN6 has a polar side, in charge of the electrostatic attraction between the negatively charged head groups of membrane lipids and the positively charged amino acids of the peptide. Also, a non-polar side that interact with non-polar tails of membrane lipids are present in the OMN6 peptide.
In order to verify that the cyclization process did improve peptide stability, we compared the native peptide Cecropin A stability to that of OMN6 through Proteinase-K enzymatic degradation-based Western blot. Proteinase-K is a broad-spectrum serine protease that has high proteolytic activity. As shown in Fig. 1E, OMN6 remained intact after exposure to Proteinase-K, while Cecropin A was subjected to major degradation under the same conditions. This test confirms OMN6 increased biostability, when compared to the parent compound Cecropin A.
Selective antimicrobial activity of OMN6. In order to assess OMN6 potential antimicrobial effect and range, we addressed several major global bacterial threats to Minimum Inhibitory Concentration (MIC) in vitro tests according to approved CLSI (Clinical and Laboratory Standards Institute) guidelines 30 . Gram (−) and Gram (+) ATCC strains as well as clinical isolates were selected. Gram (−) Acinetobacter baumannii, Klebsiella pneumoniae, Escherichia coli and Gram (+) Staphylococcus aureus, Bacillus cereus, Enterococcus faecium and Enterococcus faecalis were examined with many of the strains being multidrug resistant ones. As shown in Table 1, OMN6 presents a potent inhibitory effect on Gram-negative bacteria, with MIC of ≤ 32 µg/mL between the different tested bacterial strains, with no regard to their resistance profile. The lowest MIC observed was of the A. baumannii strains, with MICs ranging between 4-8 µg/mL, both in sensitive and multidrug resistant strains. OMN6 presented a MIC of ˃ 256 µg/mL on all tested Gram-positive strains-S. aureus, E. faecium, E. faecalis and B. cereus-showing a less than effective activity margin. All this gives the distinct indication that OMN6 retains an antimicrobial activity after the bioengineering steps required for stabilization have been performed. Thus, OMN6 presents powerful antimicrobial effect against Gram-negative bacteria and is active on A. baumannii regardless of any cross-resistance to other antibiotic drugs.

Mechanism of action.
Membrane disruption is known as one of the characteristics of AMPs mechanism of action (MoA) 36 . This has led us to explore OMN6 induction of membrane disruption via fluorescence assays by utilizing the Gram (−) E. coli GFPuv construct bacteria as a model. As GFP is a large protein of 26.9 kDa, its exit or leakage out of the bacterial cell occurs only upon disruption of the bacterial membrane integrity. In this assay, E. coli GFPuv was incubated with 16 µg/mL OMN6 or double-distilled water (DDW) as the control group for one hour. As shown in Fig. 2A, an example of a bacterium in the control group was unaffected with no evidence of GFP leakage to the surrounding media, with an increase of 8,852 in fluorescence arbitrary units (FU) between T = 0 h to T = 1 h suggesting growth of the bacterial population. Bacteria treated with OMN6 displayed www.nature.com/scientificreports/ significant leakage of GFP out of the bacterial cell into the milieu showing a reduction of 12,000 in FU suggesting disruption of membrane integrity, as was observed with cecropins 37 and other amphipathic peptides 38 . Bacterial death was confirmed at the end of each experiment by seeding all samples on Agar plates. No CFU were present in samples exposed to OMN6, where countless CFU were seen in the control samples (data not shown). Next, we used E. coli GFPuv bacteria treated with DDW or OMN6 in order to visualize the bacteria under UV light. After the end of the treatment, we collected and centrifuged the bacterial samples in order to separate bacteria from media supernatant. In the control tube, fluorescence could be observed only in the pellet suggesting that GFP is restricted to the inside of the bacteria which are still intact. In the OMN6-treated tube, fluorescence was distributed in the supernatant and less bacterial cells were present in the pellet, suggesting GFP leakage has occurred (Fig. 2B). In Fig. 2C, we followed by separation of the supernatants from the pellets. After confirming by absorbance at 600 nm that no significant bacterial presence was present in the supernatants (OD < 0.05, data not shown), the level of fluorescence in the supernatants was quantified as FU. This showed a distinct increase

OMN6 presents a bactericidal effect on Acinetobacter baumannii colistin-sensitive and colistin-resistant strains.
In order to determine the kinetics of bacterial killing induced by OMN6, time-kill assays were performed on two A. baumannii clinical isolates: the colistin-sensitive strain JMI #1088100, and the colistin-resistant strain JMI #1001007. This study was performed according to the CLSI approved guidelines 33 .
Colistin was used as comparator in this study as it kills bacteria by membrane disruption 39 , and as it is used as a last-resort treatment in A. baumannii infections 40 . In order to establish the concentrations to be used in these time-kill assays, MIC of OMN6 and colistin were performed separately to the presented studies. As shown in Table 1, MIC OMN6 value was of 8 µg/mL for the JMI #1088100 colistin-sensitive strain and of 4 µg/mL for the JMI #1001007 colistin-resistant strain. By comparison, MIC colistin value was of 0.5 µg/mL for the JMI #1088100 colistin-sensitive strain and of 8 µg/mL for the JMI #1001007 colistin-resistant strain (data not shown). Figure 3A presents the time course of killing of the colistin-sensitive A. baumannii JMI #1088100 strain by OMN6 and colistin. Both agents showed a bactericidal effect after 2 h of incubation at concentrations superior to 8 µg/mL for OMN6 and 0.25 µg/mL for colistin. No re-growth was observed during the 24 h timeframe. However, on the colistin-resistant strain presented in Fig. 3B, while colistin showed a bactericidal effect only at high concentrations of 16-32 µg/mL, OMN6 retained its bactericidal pattern with similar timeframe and concentrations. To conclude, this experiment shows that OMN6 and colistin act on the same timeframe, but that OMN6 is superior to colistin as it stays active even on colistin-resistant strains. www.nature.com/scientificreports/ Cytotoxicity and hemolytic effect. As OMN6 targets membranes, it is vital to determine, as soon as possible, whether the activity of membrane permeability is selective to bacterial cells and will not affect mammalian cells. To this end, investigative in vitro safety studies were conducted. Firstly, we tested OMN6 for the presence of a cytotoxic effect on HEK293T embryonic cell-line. In this study, HEK293T cells were exposed to increasing concentrations of OMN6 (6.25-200 µM, equivalent to 27-868 µg/mL). Melittin, a 26-amino acid basic peptide that is the major pain producing substance of the honeybee (Apis mellifera) venom 41 , was used as a positive control in this experiment, as it is a non-selective antimicrobial peptide that targets mammalian membranes and bacterial membranes alike 18,42 . As illustrated in Fig. 4A, the survival rate of HEK293T cells was not affected by the presence of the peptide at all concentrations. Moreover, no significant changes in cell death, cell morphology and alteration in survival fractions could be observed in any of the OMN6 groups. On the other hand, more than 90% cell death was observed on cells exposed to 21 μM melittin (equivalent to 60 μg/mL, within the effective range of activity of melittin). Then, the hemolytic effect of OMN6 was assessed by exposure of mouse erythrocytes to similar concentrations of peptide (ranging between 6.25-200 µM, equivalent to 27-868 µg/mL). Tween at 1% caused massive hemolysis and as shown in Fig. 4B, no hemolysis expressed as free Hemoglobin was detected in all OMN6 exposed groups. Together, the observations made here show that the novel AMP OMN6 is a selective new antimicrobial agent presenting high stability and bacteriolytic activity, with no toxicity to mammalian cells even at high doses, thus rendering it acceptable for further development as a therapeutic solution against resistant bacteria.

Discussion
Antimicrobial peptides (AMPs) are a family of potent innate immunity effectors. In insects, Cecropins form a large family of cationic α-helical AMPs with wide antimicrobial and anti-inflammatory effects 10,14,15,17,25,[43][44][45] . Cecropins cause membrane disruption by a unique mode of action that has been described by Gazit et al. 37 , it involves initial adsorption of the peptide onto the membrane surface, driven by the electrostatic attraction www.nature.com/scientificreports/ between the negatively charged head-groups of bacterial membrane lipids and the positively charged amino acids of the peptide. This external interaction is followed by rearrangement of the peptide orientation, so its nonpolar amino acids face the core of the membrane, where they can favorably interact with the nonpolar tails of membrane lipids. When a sufficient number of peptides bind to the membrane, they apply a strain on the membrane that results in its disintegration. This mode of action requires the peptide to be sufficiently amphipathic to be attracted by the bacterial membrane, and hydrophobic enough to partially penetrate the membrane surface and interact with the core.
In the current study, we present OMN6, a new cyclic antimicrobial peptide based on the unique family of Cecropins isolated from the innate immune systems of the Hyalophora cecropia moth. By using initial modeling studies, we observed that the peptide was amphipathic presenting a polar side which includes positively charged amino acids, and a nonpolar hydrophobic side. These preliminary modeling studies results agree with the literature regarding the structure of the parental peptide Cecropin A 43,46 . These results will need to be confirmed by other techniques-such as Nuclear Magnetic Resonance (NMR) spectroscopy or X-ray crystallography-to fully assess the three-dimensional structure of OMN6.
The cyclized form of OMN6 confers high stability to the compound and prevents its proteolytic degradation by commonly present proteases, while upholding its full bioactivity. Our results corroborate with other teams' observations that showed disulfide bridges confer high stability to natural cyclic peptides from the plant defensins or the cyclotide families 47,48 . Likewise, an increased stability was obtained on linear AMPs after their cyclization, while keeping the peptide antimicrobial properties [49][50][51] . HEK293T cells were exposed to increasing concentrations of OMN6, while the control group was treated with double distilled water (DDW). Positive control groups were exposed to 10% DMSO or 21 µM melittin, that caused > 90% cytotoxicity. After 24 h, all experiment groups were subjected to Methylene-Blue assay in order to evaluate cell survival in comparison to the control group at 100%. (NS not significant; ****p < 0.0001). (B) Hemolysis assay via mouse red blood cells. Suspensions of 10% mouse erythrocytes were exposed to increasing concentrations of OMN6. The positive control group was treated with 1% Tween. The negative control group for the peptide treatment and for the Tween treatment were respectively treated with 0.9% saline solution and with PBS × 1. After a one-hour incubation, the relative amount of free hemoglobin (Hgb) was assessed as an indication of erythrocyte hemolysis (NS not significant; ****p < 0.0001). www.nature.com/scientificreports/ Our findings display OMN6 efficient antimicrobial activity against Gram-negative bacteria, including against the concerning ESKAPE pathogens that cause life-threatening nosocomial infections 52 .
At the top of the list of most urgent infectious threat as defined by the CDC 3 and the WHO 53 stands carbapenem-resistant Acinetobacter baumannii. We presented OMN6 ability to eliminate all clinical isolates of A. baumannii whether these strains were sensitive or multidrug resistant to commonly used antimicrobial drugs. In vitro OMN6 superiority on colistin was also assessed by time-kill curve studies in which OMN6 presented a bactericidal effect, with a similar pattern observed on a colistin-sensitive and on a colistin-resistant A. baumannii strain. The promising in vitro profile presented by OMN6 in our studies will need to be confirmed in vivo, as colistin, that is used as a last resort treatment in A. baumannii-based infections 40 , presents high nephrotoxicity and needs to be urgently replaced by safer therapies 54 .
When investigating the in vitro activity of OMN6 against bacterial species, we observed membrane disruption features, with leakage of internal bacterial components to the external media. These results are consistent with the known effects of cecropins and similar membrane-disrupting AMPs, as has been previously described in scientific literature 16,[55][56][57][58] . This distinctive mode of action by membrane disruption bestows key advantages to the AMPs family. Unlike most antibiotics that are dependent on a biochemical specific site of action or interaction, AMPs exert a physical damage on the actual structure of the whole bacterial membrane 14,15 . This mechanism has low propensity to develop antimicrobial resistance both during and after a treatment 59 . Taken together, our results with the previous literature suggest that OMN6 causes membrane disruption and bacterial lysis and death. Though, the precise sequence of events leading to bacterial death after exposition to OMN6 remains to be fully elucidated.
Another important characteristic of the AMPs MoA is their specificity to well organized negatively charged membranes which differentiate bacteria from eukaryotic cells 60 . Our investigations showed that OMN6 presented no toxicity to mammalian cells, even at high concentrations. This therapeutic window, that is larger than tenfold of the MIC value against Gram-negative bacterial species, allows OMN6 to become a therapy for humans.
The selectivity of OMN6 to bacterial membranes is assumed to derive from the peptide net positive charge and from it adopting a biologically active conformation contingent only upon binding to negatively charged bacterial membranes. By contrast, we presume that eukaryote membranes were not targeted by OMN6, since they are nearly electrostatically neutral 61 and present high amounts of cholesterol causing resistance to membrane perturbation 57 . This parameter is not the single one predicting OMN6 effects on bacteria, as Gram-positive bacteria are also negatively charged 62,63 and OMN6 showed no activity against all Gram-positive bacterial strains (Staphylococcus aureus, Bacillus cereus, Enterococcus faecalis and Enterococcus faecium) tested in this study. We assume, as reviewed by Malanovic et al. 64 , that additional membrane features and components may have effects on AMP activity, which could explain the different effects observed with OMN6 on Gram-positive and Gramnegative bacteria. For example, Gram-positive bacteria have a thick peptidoglycan-based cell wall containing lipoteichoic acid that may play as AMP entrapper, resulting in a decrease in local peptide concentration on the cytoplasmic membrane, and finally in protection from membrane disruption caused by AMPs 64,65 . These results would explain why OMN6, like Cecropins A and B 25,44,66,67 , have a better antimicrobial effect on Gram-negative bacteria than on Gram-positive bacteria.

Conclusion
Collectively, our data shows that the cecropin-based novel peptide, OMN6, presents high activity against drug resistant Gram-negative bacteria, employing a bactericidal mechanism of action. OMN6 enhanced resistance to proteolysis and lack of toxicity toward eukaryote mammalian cells, make this novel peptide druggable as well as efficient. All these properties make OMN6 a perfect candidate to be developed as the next generation of anti-infective therapies.