Antibacterial Activity of a Cationic Antimicrobial Peptide against Multidrug-Resistant Gram-Negative Clinical Isolates and Their Potential Molecular Targets

Antimicrobial resistance reduces the efficacy of antibiotics. Infections caused by multidrug-resistant (MDR), Gram-negative bacterial strains, such as Klebsiella pneumoniae (MDRKp) and Pseudomonas aeruginosa (MDRPa), are a serious threat to global health. However, cationic antimicrobial peptides (CAMPs) are promising as an alternative therapeutic strategy against MDR strains. In this study, the inhibitory activity of a cationic peptide, derived from cecropin D-like (ΔM2), against MDRKp and MDRPa clinical isolates, and its interaction with membrane models and bacterial genomic DNA were evaluated. In vitro antibacterial activity was determined using the broth microdilution test, whereas interactions with lipids and DNA were studied by differential scanning calorimetry and electronic absorption, respectively. A strong bactericidal effect of ΔM2 against MDR strains, with minimal inhibitory concentration (MIC) and minimal bactericidal concentrations (MBC) between 4 and 16 μg/mL, was observed. The peptide had a pronounced effect on the thermotropic behavior of the 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)/1,2-dimyristoyl-sn-glycero-3-phosphorylglycerol (DMPG) membrane models that mimic bacterial membranes. Finally, the interaction between the peptide and genomic DNA (gDNA) showed a hyperchromic effect, which indicates that ΔM2 can denature bacterial DNA strands via the grooves.


Clinical Isolates: Antibiotic Susceptibility and Resistance Genes
The antibiotic susceptibility and resistance profiles of K. pneumoniae and P. aeruginosa clinical isolates are summarized in Tables 1 and 2, respectively. From 30 K. pneumoniae clinical isolates, 15 strains were WTKp and the remaining were classified as MDRKp because they showed resistance to at least three from the six antimicrobial categories used to construct the susceptibility and resistance profiles ( Table 1). The MDRKp strains showed resistance to penicillin + β-lactamase inhibitors, cephamycins, extended spectrum cephalosporins, aminoglycosides, fluoroquinolones, and carbapenems (Table 1). In particular, the confirmation of the molecular classes of carbapenemases and results of the Carba NP test showed that two MDRKp strains carried only bla KPC for class A serine carbapenemase; six MDRKp clinical isolates were metallo-beta-lactamases-producing (MBL) K. pneumoniae strain with bla NDM , bla VIM and/or bla IMP ; and seven MDR K. pneumoniae strains showed co-production of carbapenemase because they carried a combination of bla KPC , bla NDM , and bla VIM and/or bla IMP (Table 1). Moreover, from 30 P. aeruginosa clinical isolates, 15 were classified as WTPa and 15 as MDRPa, which was resistant to antipseudomonal cephalosporins, fluoroquinolones, penicillin + β-lactamase inhibitors and carbapenems, aminoglycosides, and polymyxins ( Table 2). According to molecular characterization, 13 MDRPa carried bla NDM , and bla VIM and/or bla IMP , which encode MBL enzymes; whereas the two remaining MDRPa strains contained a combination of bla KPC , bla NDM , and bla VIM and/or bla IMP , which indicates the co-production of class A serine carbapenemase and MBL enzymes (Table 2). Thus, the MDRKp and MDRPa strains can be considered "high-risk clones," similar to those that are endemic and circulate in Colombia, and represent a serious threat for public health [14,17]. The presence of MDRKp and MDRPa isolates is consistent with the bacterial resistance situation in Colombia, where a high frequency of MDR K. pneumoniae and P. aeruginosa strains carrying carbapenemase genes has been reported [14,17].

Antibacterial Activity of ∆M2 against Clinical Isolates
Several studies have evaluated the antibacterial activity of CAMPs due to their high potency and rapidity in killing bacterial cells [24,26,31,42]. Here, we investigated the in vitro antibacterial activity of ∆M2 against wild-type and MDR clinical isolates of K. pneumoniae and P. aeruginosa, which is summarized in Table 3. First, the antibacterial activity of ∆M2 against ATTC strains was confirmed and the peptide was evaluated against clinical isolates (Table 3). ∆M2 had antibacterial activity against ATCC strains of Escherichia coli and P. aeruginosa, with minimal inhibitory concentrations (MIC) values between 4 and 8 µg/mL. (Table 3). The effectiveness of this peptide against ATCC Gram-negative strains was consistent with previous reports in which natural cecropins and/or synthetic cecropin-analogs were effective against laboratory strains of E. coli and P. aeruginosa [31,[43][44][45][46][47][48]. Similar to ATCC strains, the clinical isolates of K. pneumoniae and P. aeruginosa showed susceptibility to ∆M2, with MIC values from 4 to 16 µg/mL. (Table 3). However, the MIC values found here for clinical isolates were considerably lower in comparison with those reported previously for other cationic peptides, but were similar to those reported for the cecropin A-melittin hybrid peptide [31,42,49]. Following the comparison of tested isolates, no significant differences were found between wild-type strains of both species, but statistical differences were found between the MDRKp and MDRPa strains (Table 3). Despite these significant differences, ∆M2 exhibited strong activity against all MDR strains, with MIC values between 8 and 16 µg/mL (Table 3 and Figure 1). This efficacy was comparable to that reported previously for other AMPs that have bactericidal activity against MDR strains of Gram-negative bacteria with several resistance mechanisms, including KPC-producing K. pneumoniae and colistin-resistant P. aeruginosa [49][50][51][52][53]. The statistical difference of the effectiveness of ∆M2 between MDR isolates of K. pneumoniae and P. aeruginosa can be explained by the presence of virulence factors and resistance mechanisms in strains. For example, KPC-producing K. pneumoniae strains have a capsule that contributes to their pathogenicity due to the capsular polysaccharides that protect them from the antibacterial activity of ∆M2 [12,41,54,55]. Moreover, the mechanisms and systems associated with colistin resistance in some MDRPa strains can modulate the efficiency of ∆M2 [49]. We found statistical differences in MICs when colistin-susceptible MDR P. aeruginosa and colistin-resistant MDRPa clinical isolates were compared (p = 0.0011) ( Table 2), which suggests that colistin cross-resistance can modulate ∆M2 effectiveness. However, further investigations are necessary in the future. no differences were found between the MICs and MBCs of ΔM2 against all clinical isolates tested in this study, this peptide may be considered a bactericidal agent (Table 3). These findings showed that ΔM2 is a bactericidal peptide active against the clinical isolates of K. pneumoniae and P. aeruginosa, including wild-type and MDR strains. The bactericidal effect of several AMPs, including cecropinderived peptides, cathelicidins, magainins, and nisins, against MDR strains has also been reported [30,31,33,42,49,56]. Although ΔM2 showed slight hemolytic activity in a previous study [44], its therapeutic index (TI) in Gram-negative bacteria was above 128. Taking this and the results obtained into account, ΔM2 may be safe as a topical formulation [57] to treat infectious diseases caused by MDR bacteria.

Interaction of ΔM2 with Model Bacterial Membrane
To identify potential targets of ΔM2 in Gram-negative bacteria, studies on the interaction with model membranes mimicking bacterial membranes were performed. Two endothermic peaks ( Figure  2), one for pre-transition at 13.55 and other for main transition at 22.98 °C, were obtained in the thermograms from membrane models consisting of pure 1,2-dimyristoyl-sn-glycero-3phosphocholine (DMPC)/1,2-dimyristoyl-sn-glycero-3-phosphorylglycerol(DMPG) phospholipids ( Table 4). The transition temperature from gel to crystalline liquid is concordant with that reported by Aragón-Muriel et al. [58]. In contrast, the full width at half maximum of the peak (FWHM) was 0.55 °C (Table 4).  Intraspecific comparisons showed no significant differences between WTKp and MDRKp strains; however, the wild-type clinical isolates showed higher susceptibility with the lowest MIC value of 4 µg/mL ( Figure 1a). Interestingly, these findings suggest that the activity of ∆M2 against K. pneumoniae is independent of its antibiotic resistance pattern (Figure 1a). In contrast, statistical differences were found when susceptible and MDR strains of P. aeruginosa were compared, which can be explained by the several resistance mechanisms in some MDRPa strains that modulate the effectiveness of ∆M2 (Figure 1b). In fact, several MDR P. aeruginosa clinical isolates showed slightly lower susceptibility to ∆M2, with MICs of 16 µg/mL, compared to WTPa strains (Figure 1b). Because no differences were found between the MICs and MBCs of ∆M2 against all clinical isolates tested in this study, this peptide may be considered a bactericidal agent (Table 3). These findings showed that ∆M2 is a bactericidal peptide active against the clinical isolates of K. pneumoniae and P. aeruginosa, including wild-type and MDR strains. The bactericidal effect of several AMPs, including cecropin-derived peptides, cathelicidins, magainins, and nisins, against MDR strains has also been reported [30,31,33,42,49,56].
Although ∆M2 showed slight hemolytic activity in a previous study [44], its therapeutic index (TI) in Gram-negative bacteria was above 128. Taking this and the results obtained into account, ∆M2 may be safe as a topical formulation [57] to treat infectious diseases caused by MDR bacteria.

Interaction of ∆M2 with Model Bacterial Membrane
To identify potential targets of ∆M2 in Gram-negative bacteria, studies on the interaction with model membranes mimicking bacterial membranes were performed. Two endothermic peaks (Figure 2), one for pre-transition at 13.55 and other for main transition at 22.98 • C, were obtained in the thermograms from membrane models consisting of pure 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)/1,2-dimyristoyl-sn-glycero-3-phosphorylglycerol(DMPG) phospholipids ( Table 4). The transition temperature from gel to crystalline liquid is concordant with that reported by Aragón-Muriel et al. [58]. In contrast, the full width at half maximum of the peak (FWHM) was 0.55 • C (Table 4).
Molecules 2020, 25, x FOR PEER REVIEW 10 of 20 After adding the peptide at a 1:10 peptide-lipid molar ratio, a pronounced effect was observed on the thermotropic behavior of the DMPC/DMPG membranes ( Figure 2). The pre-transition peak, which is sensitive to foreign molecules [59], was completely abolished, and the main transition peak visually disappeared ( Figure 2). However, for scaling the Y axis, the formation of two phases was observed (Figure 2), one more fluid than the other, in accordance with a previous report conducted by Oñate-Garzón in 2017, where a similar behavior was observed at 1:25 peptide-lipid ratio [44]. The cationic charge of the peptide could be involved in membrane disturbance, due to the presence of the anionic DMPG in the phospholipids mixture, as the cationic arginine side chain can approach 1,2-dimyristoyl-sn-glycero-3-phosphorylglycerol (DMPG) head groups to 5 Å [60]. Table 4 shows that the transition enthalpy was reduced from 0.32 to 0.03 cal/g after the peptide was added, suggesting that less heat is needed to achieve phase transition because the peptide disrupts interactions between lipid acyl chains, as a result of the interruption of the intra and intermolecular interactions of van der Waals and trans-gauche isomerization [44]. Additionally, FWHM increased considerably from 0.5 to 3.69 °C as a consequence of the interaction between the peptide with the membranes, suggesting a decrease in cooperativity.

Interaction of ΔM2 with Bacterial DNA
AMPs use different modes of action for killing bacteria, including membrane disruption, interaction with intracellular molecules, such as DNA, RNA, or proteins, and influencing biochemical processes that are vital for bacterial survival [7,[26][27][28][29][30]32]. To explore the genomic DNA as a potential target for ΔM2 in Gram-negative bacteria, interaction assays between peptide and bacterial DNA were performed. These analyses could give insights into the mechanisms of action of ΔM2.  After adding the peptide at a 1:10 peptide-lipid molar ratio, a pronounced effect was observed on the thermotropic behavior of the DMPC/DMPG membranes ( Figure 2). The pre-transition peak, which is sensitive to foreign molecules [59], was completely abolished, and the main transition peak visually disappeared ( Figure 2). However, for scaling the Y axis, the formation of two phases was observed (Figure 2), one more fluid than the other, in accordance with a previous report conducted by Oñate-Garzón in 2017, where a similar behavior was observed at 1:25 peptide-lipid ratio [44].

Electronic Absorption Spectra
The cationic charge of the peptide could be involved in membrane disturbance, due to the presence of the anionic DMPG in the phospholipids mixture, as the cationic arginine side chain can approach 1,2-dimyristoyl-sn-glycero-3-phosphorylglycerol (DMPG) head groups to 5 Å [60]. Table 4 shows that the transition enthalpy was reduced from 0.32 to 0.03 cal/g after the peptide was added, suggesting that less heat is needed to achieve phase transition because the peptide disrupts interactions between lipid acyl chains, as a result of the interruption of the intra and intermolecular interactions of van der Waals and trans-gauche isomerization [44]. Additionally, FWHM increased considerably from 0.5 to 3.69 • C as a consequence of the interaction between the peptide with the membranes, suggesting a decrease in cooperativity.

Interaction of ∆M2 with Bacterial DNA
AMPs use different modes of action for killing bacteria, including membrane disruption, interaction with intracellular molecules, such as DNA, RNA, or proteins, and influencing biochemical processes that are vital for bacterial survival [7,[26][27][28][29][30]32]. To explore the genomic DNA as a potential target for ∆M2 in Gram-negative bacteria, interaction assays between peptide and bacterial DNA were performed. These analyses could give insights into the mechanisms of action of ∆M2.

Electronic Absorption Spectra
DNA interaction studies were carried out by monitoring the changes on the electronic absorption spectra of bacterial genomic DNA and calf thymus DNA (CT-DNA) with peptide titration. The effect on the peptide spectrum after the addition of different amounts of CT-DNA and bacterial genomic DNA from susceptible P. aeruginosa ATCC 27853 (gDNA1) and resistant K. pneumoniae ATCC 2146 (gDNA2) is shown in Figure 3. DNA interaction studies were carried out by monitoring the changes on the electronic absorption spectra of bacterial genomic DNA and calf thymus DNA (CT-DNA) with peptide titration. The effect on the peptide spectrum after the addition of different amounts of CT-DNA and bacterial genomic DNA from susceptible P. aeruginosa ATCC 27853 (gDNA1) and resistant K. pneumoniae ATCC 2146 (gDNA2) is shown in Figure 3. A dual effect on the electronic spectrum of the peptide (Figure 3a) was observed after the addition of increasing amounts of eukaryotic CT-DNA. The isosbestic point was observed around 295 nm, indicating the formation of a third, new chemical entity in the binary system [61]. At wavelengths lower than the isosbestic point, a hypochromic effect was observed, whereas hyperchromic behavior was found at wavelengths longer than the isosbestic point. The peptide may be binding CT-DNA in more than one way, with insertion (intercalative and groove binding mode) being the most probable [62]. For bacterial genomic DNA, smaller peptide-gDNA ratios were used. The hyperchromic effect was observed in both cases (Figure 3b,c), which indicates the capability of denaturing smaller DNA strands via groove interactions [61,63].
Data from photometric titration were collected and fitted to the Wolfe-Shimer Equation (1)   A dual effect on the electronic spectrum of the peptide (Figure 3a) was observed after the addition of increasing amounts of eukaryotic CT-DNA. The isosbestic point was observed around 295 nm, indicating the formation of a third, new chemical entity in the binary system [61]. At wavelengths lower than the isosbestic point, a hypochromic effect was observed, whereas hyperchromic behavior was found at wavelengths longer than the isosbestic point. The peptide may be binding CT-DNA in more than one way, with insertion (intercalative and groove binding mode) being the most probable [62]. For bacterial genomic DNA, smaller peptide-gDNA ratios were used. The hyperchromic effect was observed in both cases (Figure 3b,c), which indicates the capability of denaturing smaller DNA strands via groove interactions [61,63].
Data from photometric titration were collected and fitted to the Wolfe-Shimer Equation (1)  [ where [DNA] refers to DNA concentration and ε a , ε b and ε f represent apparent, fully bonded, and free molar extinction coefficients of the peptide, respectively [64][65][66]. The Wolf-Shimer plots, using absorbance of the peptide at 257 nm vs. CT-DNA and gDNA-1gDNA2 concentration, is shown in Figure 4. Kb values obtained were 2 × 10 3 M −1 , 0.5 × 10 6 M −1 , and 2 × 10 6 M −1 for eukaryotic CT-DNA, prokaryotic gDNA1, and prokaryotic gDNA2, respectively. These results show a major effect in the prokaryotic genome compared to the eukaryotic genome, which suggests a selectivity of ∆M2 for bacterial DNA [40,41]. Values of classic intercalators and DNA markers are in the order of 10 6 -10 8 [66,67], which are larger for the CT-DNA binding constant and comparable for gDNA1 and gDNA2. Major binding constants for prokaryotic genomes is attributed to the size and packing properties, in comparison to that in eukaryotic genomes [40,41]. As bacterial DNA is not as well-packaged as eukaryotic DNA, small molecules are likely to interact with it and affect genomic processes at lower concentrations [40,41]. Moreover, the binding constants for both susceptible and resistant bacterial genomes were in the same order of magnitude (Figures 3 and 4). Thus, the interaction between ∆M2 and DNA from Gram-negative bacteria is independent of their antibiotic resistance patterns (Figures 3 and 4).
Molecules 2020, 25, x FOR PEER REVIEW 12 of 20 where [DNA] refers to DNA concentration and εa, εb and εf represent apparent, fully bonded, and free molar extinction coefficients of the peptide, respectively [64][65][66]. The Wolf-Shimer plots, using absorbance of the peptide at 257 nm vs. CT-DNA and gDNA-1gDNA2 concentration, is shown in Figure 4. Kb values obtained were 2 × 10 3 M −1 , 0.5 × 10 6 M −1 , and 2 × 10 6 M −1 for eukaryotic CT-DNA, prokaryotic gDNA1, and prokaryotic gDNA2, respectively. These results show a major effect in the prokaryotic genome compared to the eukaryotic genome, which suggests a selectivity of ΔM2 for bacterial DNA [40,41]. Values of classic intercalators and DNA markers are in the order of 10 6 -10 8 [66,67], which are larger for the CT-DNA binding constant and comparable for gDNA1 and gDNA2. Major binding constants for prokaryotic genomes is attributed to the size and packing properties, in comparison to that in eukaryotic genomes [40,41]. As bacterial DNA is not as well-packaged as eukaryotic DNA, small molecules are likely to interact with it and affect genomic processes at lower concentrations [40,41]. Moreover, the binding constants for both susceptible and resistant bacterial genomes were in the same order of magnitude (Figures 3 and 4). Thus, the interaction between ΔM2 and DNA from Gram-negative bacteria is independent of their antibiotic resistance patterns ( Figures  3 and 4).

Agarose Gel Electrophoresis
To detect the effect of the peptide on bacterial genomic DNA from susceptible and resistant strains, fixed amounts of gDNA1, gDNA2, and pmCherry (as control) were incubated in the presence of variable concentrations of ΔM2 and the systems were electrophoresed. Electrophoretic patterns

Agarose Gel Electrophoresis
To detect the effect of the peptide on bacterial genomic DNA from susceptible and resistant strains, fixed amounts of gDNA1, gDNA2, and pmCherry (as control) were incubated in the presence of variable concentrations of ∆M2 and the systems were electrophoresed. Electrophoretic patterns are shown in Figure 5. Peptide-DNA ratios were fitted to match ratios of different zones in the curves shown in Figure 4. All DNAs were denatured in the evaluated ratios, showing a major effect on bacterial genomes. gDNAs was over 5000 bp and pmCherry vector approximately 2500 bp in size. The phenomenon of DNA aggregation and sedimentation is seen in Figure 5b, as indicated by bromophenol blue loading buffer. Wells with high [P]/[DNA] ratios showed major DNA sedimentation; hence, null mobility and no fluorescent marking (Figure 5a), this confirms the results obtained by UV-vis assays. Thus, denaturation of bacterial DNA may not be via hydrolysis but by non-specific precipitation of a peptide-DNA complex, which cannot run through the gel [68].
Molecules 2020, 25, x FOR PEER REVIEW 13 of 20 are shown in Figure 5. Peptide-DNA ratios were fitted to match ratios of different zones in the curves shown in Figure 4. All DNAs were denatured in the evaluated ratios, showing a major effect on bacterial genomes. gDNAs was over 5000 bp and pmCherry vector approximately 2500 bp in size. The phenomenon of DNA aggregation and sedimentation is seen in Figure 5b, as indicated by bromophenol blue loading buffer. Wells with high [P]/[DNA] ratios showed major DNA sedimentation; hence, null mobility and no fluorescent marking (Figure 5a), this confirms the results obtained by UV-vis assays. Thus, denaturation of bacterial DNA may not be via hydrolysis but by non-specific precipitation of a peptide-DNA complex, which cannot run through the gel [68]. Thus, the peptide-DNA interaction may occur by groove binding and electrostatic interaction. The peptide possesses a net positive charge, which neutralizes the negative charge of the phosphate backbone, inducing precipitation in an aqueous environment. The interaction is concentration dependent, as shown in Figure 4, and small bacterial genomes are more affected, as seen by the higher binding constants [68]. No differences were found when the interactions of gDNA1 and gDNA2 with ΔM2 were compared, suggesting that bacterial DNA interacts with the peptide through electrostatic interactions or groove binding ( Figure 5). Additionally, these results complemented the electronic absorption analysis and confirmed that the effect of the peptide on bacterial genomic DNA is independent of the strains' resistance profile and mechanism used.

Peptide Design and Synthesis
ΔM2 is a CAMP derived from the neutral peptide, cecropin D-like, from G. mellonella, that has antibacterial activity against laboratory strains [44]. This peptide is composed of 39 residues and has Thus, the peptide-DNA interaction may occur by groove binding and electrostatic interaction. The peptide possesses a net positive charge, which neutralizes the negative charge of the phosphate backbone, inducing precipitation in an aqueous environment. The interaction is concentration dependent, as shown in Figure 4, and small bacterial genomes are more affected, as seen by the higher binding constants [68]. No differences were found when the interactions of gDNA1 and gDNA2 with ∆M2 were compared, suggesting that bacterial DNA interacts with the peptide through electrostatic interactions or groove binding ( Figure 5). Additionally, these results complemented the electronic absorption analysis and confirmed that the effect of the peptide on bacterial genomic DNA is independent of the strains' resistance profile and mechanism used.

Peptide Design and Synthesis
∆M2 is a CAMP derived from the neutral peptide, cecropin D-like, from G. mellonella, that has antibacterial activity against laboratory strains [44]. This peptide is composed of 39 residues and has a net charge of +9 at neutral pH. ∆M2 was provided by GenScript Corporation (Piscataway, NJ, USA) at 98% purity. The lyophilized peptide was dissolved in phosphate buffered saline (pH 7.4; 138 mM NaCl, 3 mM KCl, 1.5 mM NaH 2 PO 4 , and 8.1 mM Na 2 HPO 4 ), at an initial concentration of 5000 µg/mL. Dilutions were prepared on the day of use.

Identification of Clinical Isolates
A total of 60 clinical strains were tested in this study: 30 each for K. pneumoniae and P. aeruginosa. All clinical isolates were isolated from clinical specimens, such as urine, secretions, and blood, recovered from two tertiary care hospitals in Cali, Colombia, between 2017 and 2019. All cultures of clinical isolates were sent to Microbiology Laboratory at Laboratorio de Salud Pública Departamental del Valle del Cauca (LSPD-Valle), where bacterial identity was confirmed, and antibiotic susceptibility characterization was performed. Species identification was performed using the automated VITEK ® 2 system, (bioMerieux, 9.02, Marcy l'Etoile, France) with the VITEK ® 2 Gram-Negative Identification card (VITEK ® 2 GN ID), which is based on established biochemical methods and substrates that evaluate the use of carbon, enzymatic activity, and resistance (Ref. 21341, bioMerieux, Marcy l'Etoile, France). All laboratory strains, including E. coli ATCC ® 25922™, K. pneumoniae ATCC ® 2146™, and P. aeruginosa ATCC ® 27853™, were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA).

Antimicrobial Assay of ∆M2
The MIC values of ∆M2 were determined by the broth microdilution test, according to the protocol of CLSI [69,71]. Briefly, pure clinical isolates were grown in brain heart infusion agar and incubated at 37 • C for 18-24 h. A colony from the pure culture was initially resuspended in sterile water to reach a turbidity value of 0.5 McFarland, and the resulting suspension contained approximately 1-4 × 10 8 colony forming units (CFU)/mL. Using this suspension, a final solution was obtained with a concentration of 2-7 × 10 5 CFU/mL in cation-adjusted Mueller-Hinton broth. The bacterial inoculum was incubated with different concentrations of ∆M2. The highest tested concentration of 128 µg/mL was used for serial 1:2 dilutions. In a final volume of 100 µL, peptide and inoculum were incubated in sterile 96-well polypropylene microplates (Sigma-Aldrich, St. Louis, MO, USA) at 37 • C for 18-20 h. A peptide-free control was used for every isolate evaluated. Additionally, the reference ATCC strains were used in each assay as controls to ensure reproducibility. The MIC of ∆M2 for each strain was defined as the lowest concentration that inhibited the visible growth of bacteria after incubation [69,71]. The minimal bactericidal concentrations (MBCs), defined as the lowest concentration of an antibacterial agent required to kill 99.9% of a particular bacterium, were determined by plating the contents of the first three wells showing no visible bacterial growth onto Muller-Hinton agar plates and incubated at 37 • C for 18-20 h [31,72]. MICs and MBCs were determined in duplicate, and at least three independent assays were performed for each isolate.

Differential Scanning Calorimetry
DSC analysis was made using a DSC Q25 (TA instrument, New Castle, DE, USA). MLVs were prepared using 1 mg of lipids to give two peptide-lipid ratios: 1:50 and 1:10. HEPES buffer was used as a reference solution. Samples were encapsulated in standard aluminum DSC pans, and scanning was carried out over a range of 10 • C-40 • C at a heating rate of 1 • C/min. Thermograms were acquired and analyzed using the Trios software package (TA Instruments) to obtain the phase transition temperature (Tm), the transition enthalpy (DH), and the FWHM.

DNA Interaction Assays
DNA interaction studies with the peptide were carried out by electronic absorption experiments and agarose gel electrophoresis. The interaction of ∆M2 with DNA was performed using bacterial genomic DNA from the susceptible P. aeruginosa ATCC 27853 strain (gDNA1) and the resistant K. pneumoniae ATCC 2146 strain (gDNA2). To evaluate the selectivity of the peptide, a control with highly polymerized, lyophilized calf thymus DNA (CT-DNA) from Sigma-Aldrich was used. Genomic DNA from bacterial stains was extracted using the GenoLyse ® kit, according to the manufacturer's instructions (Hain Lifescience-BRUKER, Nehren, Germany). The pmCherry vector from was extracted from E. coli BL21 (DH5α) using Kit Hi-Speed Mini Plasmid from IBI Scientific [73]. All DNA solutions had an A 260 /A 280 value between 1.8 and 1.9, indicating that the DNA was free of RNA and protein.
The DNA was resuspended 10 mmol/L Tris and 1 mmol/L EDTA in deionized water, with the pH adjusted to approximately 7.5. The DNA solutions were stored at −5 • C.

Electronic Absorption Monitoring Assays (UV-Vis)
The electronic absorption spectra were recorded in an UV-visible Evolution 220 Thermo Fisher Scientific spectrophotometer (Waltham, MD, USA), equipped with Single Cell Peltier System for temperature control. The interaction between ∆M2 was carried out with gDNA1, gDNA2, and CT-DNA titrations at a constant concentration of the peptide (300 µM) at 25 • C. After the addition of each titrant, the electronic absorption spectra of the evaluated peptide were measured between 230 nm and 450 nm. DNA absorbance was eliminated by adding an equal amount of DNA to the sample and the standard solution, as reported in the literature [49,52]. A CT-DNA solution 11,000 µM in nucleotides were prepared using ε 260 = 6600 cm −1 ·mol −1 ·L as molar extinction coefficient and measuring the absorbance at 260 nm. The stock solution was stored at −5 • C. The electronic absorption experiments were determined in duplicate, and at least three independent assays were performed for each DNA.

Agarose Gel Electrophoresis
The interaction between ∆M2 and bacterial genomic DNA was studied by agarose gel electrophoresis. Mixtures of the peptide at different concentrations, with gDNA1 and gDNA2 at constant concentrations, in a final volume of 20 µL were incubated in the Eppendorf ThermoMixer ® C under physiological conditions (37 • C and pH = 7.0) for 90 min. The pmCherry vector extracted from E. coli BL21(DH5α) was used as a control. After the incubation was complete, the samples were loaded in 1% agarose gel with 1× TAE buffer and Bioline HyperLadder™ 1 kb as the molecular weight marker. The Owl™ EasyCast™ B1 Mini Gel electrophoresis system was used for electrophoresis. GelGreen™ (GoldBio, St. Louis, MO, USA) was used to visualize DNA on a MaestroGen UltraBright Led 470 nm transilluminator (Hsinchu, Taiwan).

Statistical Analysis
The results were analyzed using descriptive statistical tools with the median. Statistically significant differences in MICs were analyzed and compared using the non-parametric Kruskal-Wallis test with R-Project software Version 1.1.463. p-values ≤ 0.05 were considered statistically significant.

Conclusions
In this study, we have demonstrated that ∆M2 exhibits strong bactericidal activity against both susceptible and MDR clinical isolates of K. pneumoniae and P. aeruginosa. However, significant differences were found between the MDRKp and MDRPa clinical isolates, possibly due to differences in resistance mechanisms between K. pneumoniae and P. aeruginosa, which can modulate ∆M2 effectiveness. Further investigation is warranted in the future. The ∆M2 peptide interacted markedly with the model membrane that mimics bacterial membranes and bacterial genomic DNA from susceptible and resistant strains, thereby providing insights into the mode of action of ∆M2 in interacting with potential molecular targets in Gram-negative bacteria. Considering these results and the TI in Gram-negative bacteria for ∆M2, this peptide may be safe as a topical formulation to treat infectious diseases caused by MDR bacteria.