A polyalanine peptide derived from polar fish with anti-infectious activities

Due to the growing concern about antibiotic-resistant microbial infections, increasing support has been given to new drug discovery programs. A promising alternative to counter bacterial infections includes the antimicrobial peptides (AMPs), which have emerged as model molecules for rational design strategies. Here we focused on the study of Pa-MAP 1.9, a rationally designed AMP derived from the polar fish Pleuronectes americanus. Pa-MAP 1.9 was active against Gram-negative planktonic bacteria and biofilms, without being cytotoxic to mammalian cells. By using AFM, leakage assays, CD spectroscopy and in silico tools, we found that Pa-MAP 1.9 may be acting both on intracellular targets and on the bacterial surface, also being more efficient at interacting with anionic LUVs mimicking Gram-negative bacterial surface, where this peptide adopts α-helical conformations, than cholesterol-enriched LUVs mimicking mammalian cells. Thus, as bacteria present varied physiological features that favor antibiotic-resistance, Pa-MAP 1.9 could be a promising candidate in the development of tools against infections caused by pathogenic bacteria.

In recent decades, improvements in the prevention and treatment of infectious diseases caused by pathogenic microorganisms has been of great importance in reducing morbidity and mortality, leading to a better quality of life and longer life expectancy 1 . However, antibiotics have been widely and sometimes indiscriminately used, which has resulted in the emergence of pathogens with multi-drug resistance in a wide range of bacterial species, including Escherichia coli, Staphylococcus aureus, Pseudomonas aeruginosa, Klebsiella pneumoniae and many other species 2 . Conversely, the rate of discovery of new antibiotics has steadily plummeted. Additionally, the occurrence and treatment of biofilm infections has appeared as one of the biggest challenges in the medical field with no available antibiotics that were developed to treat such infections. Biofilms are characterized as being a structured consortium of microorganisms connected by a complex matrix composed of polysaccharide(s), protein and DNA, that grows on biotic or abiotic surfaces via a multistage process 3 . It has been established that pathogenic bacteria are predominantly organized in biofilms, which are the cause of 65% to 80% of all bacterial infections in humans 3,4 . Biofilm growth represents a unique growth state whereby bacteria have major physiological and organizational differences, in particular leading to 10-to 1000-fold increased (adaptive) resistant to conventional antibiotics 3,4 .

Results
Pa-MAP 1.9 synthesis and mass spectrometry analysis. Pa-MAP 1.9 (NH 2 -LAAKLTKAATKLTAALT KLAAALT-COOH) was designed, and synthetized by Fmoc strategy loosely based on the sequence of a previously described multifunctional peptide, Pa-MAP (NH 2 -HTASDAAAAAALTAANAAAAAAASMA-COOH) 12 , in which the net charge and hydrophobic moment were increased, hydrophobic amino acid frequency was decreased (since high hydrophobicity favors toxicity) and alanine residues were distributed along the molecule in order to obtain a linear cationic peptide with predicted helical stretches. MALDI-ToF analysis showed greater than 95% purity and an ion mass of 2668.0 m/z, in agreement with the theoretical calculated molecular mass for this peptide (Fig. S1).
In vitro antimicrobial assays. Pa-MAP 1.9 was evaluated for its ability to inhibit the growth of different bacteria grown planktonically and in biofilms. Pa-MAP 1.9 was able to inhibit the growth of Enterococcus faecalis, E. coli and K. pneumoniae planktonic cells with minimal inhibitory concentrations (MICs) of 1.5, 6-12 and 24-96 μ M, respectively (Table 1). However, no antibacterial activity was observed against P. aeruginosa and S. aureus, even at the maximum concentration used for this assay (115 μ M). Against bacteria in their biofilm growth state, Pa-MAP 1.9 was considerably more potent, with (MBICs) of 3.0 and 1.1 μ M (Table 1) against E. coli and K. pneumoniae biofilms respectively. At these same concentrations, flow cell analysis revealed that pre-formed E. coli and K. pneumoniae biofilms (Fig. 1a,c) were strongly or completely inhibited, with a strong decrease in biofilm volume and height (Fig. 1b,d).
Hemolytic and cytotoxicity assays. Pa-MAP 1.9 did not show neither hemolytic activity against human erythrocytes or cytotoxic effects against RAW 264.7 monocyte cell line up to 115 μ M, the maximum concentration used in the bioassays (Table 1).
Atomic force microscopy analysis. AFM was used to image the possible AMP-induced damage of the Gram-negative bacterial envelope. At 6 μ M (MIC), Pa-MAP 1.9 induced slight morphological changes in the form of increased surface roughness in E. coli (Fig. 2b), relative to the untreated control (Fig. 2a). Damage was more evident at 300 μ M (50-fold the MIC), where substantial bacterial surface disruption was observed (Fig. 2c).
Permeabilization of lipid vesicles. The ability of Pa-MAP 1.9 to disrupt large unilamellar vesicles (LUVs) with compositions roughly mimicking bacterial membranes was also analyzed. Vesicles were prepared with encapsulated carboxyfluorescein (CF), and then incubated with Pa-MAP 1.9 concentrations up to 1.0 μ M. The percentages of leakage were evaluated as previously described 13 , by assessing the differences over time in the emission of CF (which fluoresces more strongly upon leakage from the vesicles), with 100% of leakage corresponding to full disruption of lipid vesicles. The data showed that the peptide induced leakage of vesicles containing anionic lipids such as 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1′ -sn-glycerol) (POPG), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS) or lipopolysaccharide (LPS), requiring only 0.25 μ M peptide for almost complete permeabilization (Fig. 3a). Moreover, at 0.75 μ M of Pa-MAP 1.9, full leakage was reached for all other vesicles containing anionic phospholipids (Fig. 3b). In contrast, zwitterionic vesicles Scientific RepoRts | 6:21385 | DOI: 10.1038/srep21385 of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) or POPC/Cholesterol (Chol) (70:30) clearly demonstrated reduced leakage, even at the maximum concentration used for this experiment (Fig. 3a,b). The results obtained demonstrate the specificity of the AMP for negatively charged membranes, such as those from Gram-negative bacteria and are consistent with the lack of toxicity for mammalian cells, which contain largely neutral or zwitterionic lipids on their external leaflet.
Structural analysis. Circular dichroism spectroscopy studies were performed in water, 2,2,2-trifluoroethanol (TFE) 50% (v:v) and sodium dodecyl sulfate (SDS) 28 mM. When analyzed in water, the peptide showed no stable at pH values ranging from 3 to 10 (Fig. 4a). However, at pH 11 this Pa-MAP 1.9 produced a CD spectrum characteristic of an α -helix conformation, with two negative bands at ~205 and ~222 nm, presenting 14% of ellipticity (Fig. 4a). When analyzed in 50% TFE, Pa-MAP 1.9 adopted a well-defined α -helix structure at all pH values presenting a positive band at 190 nm and two negative bands at ~205 and ~222 nm (Fig. 4b). Similar results were observer for the CD spectrum in anionic lipid-like environment, where the same positive and negative bands could be observed, indicating helical conformation (Fig. 4c).
To obtain insights into the three-dimensional structure of Pa-MAP 1.9, molecular modelling simulations were performed. The lowest free-energy theoretical model for Pa-MAP 1.9 revealed a well-defined α -helical conformation (Fig. 5b), and also revealed an amphipathic character (Fig. 5c) when modelled based on the antifreeze peptide (PDB code: 1wfa) isolated from P. americanus, which presented 58% identity over part of its sequence with the Pa-MAP 1.9 primary sequence (Fig. 5a). When evaluated using PROCHECK 14 , the average score for dihedral angles (φ -ψ , χ and ω ), jointly with the main-chain covalent forces for the best model, was 0.23, which is within the expected range for a reliable structure. Moreover, a Ramachandran plot showed that 100% of the amino acid residues of Pa-MAP 1.9 were located in the most favorable regions. Furthermore, three-dimensional structural superposition (3DSS) analysis revealed that the root mean square deviation (RMSD) between the theoretical and experimental models was equal to 0.567 Å. Based on those data, this model was selected for further in silico studies.
Like the CD analysis, molecular dynamics simulations of the peptide were implemented in three different environments including pure water, water and TFE mixture 50% (v:v) and water where the peptide was in contact with an SDS micelle. The simulations were performed to better understand the behavior of the three-dimensional theoretical structure of the above-described Pa-MAP 1.9 in different environments. After 100 ns of molecular dynamics simulations in water, it was possible to observe high values of RMSD and root mean square fluctuation (RMSF), revealing the instability of Pa-MAP 1.9 in this environment. In addition, it was observed a decrease in the radius of gyration (approximately from 1.3 to 0.6 nm), as well as in the solvent-accessible surface area (SASA) (Fig. 6a). These parameters, allied to the three-dimensional structures observed throughout the simulation, revealed that, consistent with the CD spectra, Pa-MAP 1.9 seems to lose its α -helical structure in water, also tending to hide its hydrophobic amino acid residues by adopting a coil conformation with a short central α -helix since the 20 ns of simulation (Fig. 7a). It can also be associated with a decrease of approximately 0.8 nm 2 in the SASA (Fig. 6a). On the other hand, simulations performed in water and TFE demonstrated that, in this solvent, Pa-MAP 1.9 was able to maintain its initial structure during the 100 ns, presenting only a few structural changes at the N-terminus (Fig. 7b). In comparison with the simulation in water, it was also possible to note decreased RMSD, RMSF, radius of gyration and SASA values, indicating improved structural stability in TFE (Fig. 6b). As expected, simulations in the presence of a micelle containing 128 SDS residues confirmed the preference of Pa-MAP 1.9 for amphipathic anionic environments, preserving its α -helical structure. In this simulation, as for 50% TFE, only a slight RMSD variation of 0.3 nm was observed, stabilizing then between 0.4 and 0.6 nm, indicating few deviations in the input and output structures along the 100 ns (Fig. 6c). Moreover, the radius of gyration was also equivalent to that for TFE simulation, remaining around 1.2 nm (Fig. 6c). In addition, SASA seemed to be conserved in these two simulations, varying from 19 to 23 nm during the 100 ns (Fig. 6c). However, contrary to what was obtained in 50% TFE, a more significant fluctuation could be observed at the C-terminus region of Pa-MAP 1.9 when in contact with SDS micelles, as demonstrated in Figs 6c and7c. By using our best theoretical model for Pa-MAP 1.9, we also predicted the affinity and atomic interactions of this peptide with mimetic membranes containing the same lipid compositions experimentally tested on the leakage studies. Fifty runs of molecular docking were performed, and all peptide/membrane complexes were ranked by their affinity, in kcal.mol −1 . The best affinity values for Pa-MAP 1.9 in POPC/POPS (Fig. 8a) and Pa-MAP 1.9 in POPC/Chol (Fig. 8b) were − 5.4 and − 3.7 kcal.mol −1 , respectively, corroborating our experimental results, wherein Pa-MAP 1.9 seemed to interact better with anionic vesicles and Gram-negative bacteria with anionic lipid compositions. Furthermore, in both complexes, Pa-MAP 1.9 maintained a well-defined α -helical conformation during the simulations (Fig. 8a,b). In the complex of Pa-MAP 1.9 with POPC/POPS it was possible to predict 10 interactions (Fig. 8c) divided into hydrogen bonds (HB), involving nitrogen (N), oxygen (O/OG1) atoms of Leu1, Thr10, Lys11,18 and Ala15,22 from Pa-MAP 1.9, and saline bonds (SB), involving positively charged nitrogen atoms of the side chain (NZ) of Lys4,7,11,18 from Pa-MAP 1.9, ranging from 2.8 to 3.6 Å of distance (Table 2). On the other hand, in the complex of Pa-MAP 1.9 with POPC/Chol, only 6 HBs could be observed (Fig. 8d), involving O/OG1 atoms of Ala2 and Thr6,10,13,17,28 from Pa-MAP 1.9 and oxygen atoms (O3) from cholesterol, with distances between 2.7 and 3.4 (Table 2).

Discussion
Here we describe the functional and structural characterization of Pa-MAP 1.9, a rationally-designed cationic AMP based on a multifunctional peptide analogue from P. americanus, denominated Pa-MAP 12 . In our antibacterial assays, Pa-MAP 1.9 revealed better activities mainly against Gram-negative bacterial strains (E. faecalis, E. coli and K. pneumoniae), with lower MIC against E. coli (6 μ M) when compared with its precursor Pa-MAP (30 μ M) 12 . Similar enhanced activities against Gram-negative strains have also been reported for other AMPs isolated or derived from the winter flounder P. americanus, as it is the case of pleurocidin, a cationic α -helical peptide with high activities against E. coli and Pseudomonas aeruginosa, with MIC values below 1 μ M 16 . Moreover, other fish derived AMPs, such as pardaxin (Purdachirus marmorutus) and piscidins-1, -2, -3 and -4 (Morone chrysops; Morone saxatilis), were also active against the human pathogens E. coli, Acinetobacter calcoaceticus, P. aeruginosa, S. typhimurium and Shigella flexneri 18,20 .
In addition to direct antibacterial assays, we also studied the antimicrobial ability of Pa-MAP 1.9 in preventing biofilm growth, as well as combating pre-formed bacterial biofilms, since different physiological conditions occur in these consortiums when compared to planktonic bacteria, and this leads to increased antibiotic resistance. Previous studies have shown that anti-biofilm activity is independently determined compared to antibacterial activity against planktonic (free-swimming) bacteria. These assays revealed that Pa-MAP 1.9 is a promising anti-biofilm peptide with activity at concentrations 90-fold lower than the MIC against one strain of Klebsiella pneumoniae. Similar to our results, Tao and colleagues 15 described the anti-biofilm potential of another P. americanus derived-peptide (pleurocidin) in combating Streptococcus mutans biofilm, causing a reduction in S. mutans biomass upon treatment with 11.8 μ M of the peptide. In addition, a study performed by Choi and Lee 16 reported that, below 0.7 μ M, pleurocidin could inhibit pre-formed E. coli and P. aeruginosa biofilms, causing biomass reductions of 22.4 and 48.3%, respectively. Interestingly, the concentration of Pa-MAP 1.9 required to combat E. coli and K. pneumoniae biofilms were lower than its MIC values for the two studied strains. Similar findings were previously made for other small synthetic cationic peptides (IDR-1018, and DJK-6), in circumstances where MIC values against K. pneumoniae could not be determined (> 64 μ g.mL −1 ), but relevant anti-biofilm properties were detected at 2-4 μ g.mL −1 (MBIC) 21 . Flow cell experiments showed even better results, with just 2 μ g.mL −1 of IDR-1018 and DJK-6 being necessary to eliminate most pre-existing K. pneumoniae biofilms 21 . These findings were consistent with the concept that anti-biofilm peptides act by several mechanisms that include dispersing bacterial biofilms and triggering death of cells within biofilms, but that these activities involve independent mechanisms from those determining AMP activity.
When evaluated with regard to its potential for causing morphological damage to the surface of E. coli, Pa-MAP 1.9 showed dose-dependent action. Below its MIC, only slight changes were observed, suggesting that Pa-MAP 1.9 might insert into the E. coli cytoplasmic membrane and translocate to act on intracellular targets. However, substantial cell damage was observed upon 50-fold increasing its concentration suggesting lytic activity at concentrations above the MIC 22 . This behavior was also described for Sub3, an AMP optimized based on bovine bactenecin 23 whereby it was found that Sub3 antimicrobial activity is intrinsically related to membrane binding and its cell-penetrating ability, since at 1 to 10 μ M it did not affect the surface of E. coli strains 17 . Nevertheless, increasing the concentration of this peptide to 100 μ M led to evident morphological damage, including bacterial surface disruption. However, as for Pa-MAP 1.9, such a high concentration was not required for the lethal activities of this peptide.
The results obtained in the AFM experiments encouraged us to study the ability of Pa-MAP 1.9 to interact with lipid vesicles mimicking different types of biological membranes. Pa-MAP 1.9 was effective in disrupting   Alignment between the query (Pa-MAP 1.9) and template (PDB: 1wfa) primary sequences, highlighting (yellow) the identical residues (a). Lowest free-energy three-dimensional theoretical model for Pa-MAP 1.9: in white, non-polar residues; in pink, polar residues; in cyan, basic residues (b). Adaptive Poisson-Boltzmann solver (APBS) electrostatic potential of Pa-MAP 1.9; potential ranges from − 10.9 kT/e (red) to + 10.1 kT/e (blue) (c). solely of zwitterionic or neutral phospholipids (e.g., POPC and cholesterol). Furthermore, cholesterol-containing membranes could not be significantly disrupted even at the highest concentrations tested. Apparently, cholesterol which is present in mammalian cells assists in the stability of these membranes 24 . Pa-MAP 1.9 lipid-selective disruption of biomembrane model systems was also consistent with our bioassays, wherein this peptide clearly presented better activities against Gram-negative bacteria, which contains a more anionic surface based on lipid composition. In addition, the observation that Pa-MAP 1.9 revealed no cytotoxicity against different mammalian cell lines could have been related in part to its inability to disrupt neutral (zwitterionic) phospholipids and cholesterol-containing membranes. Lee and colleagues 25 observed similar results in studying detergent-like membrane disruption by four magainin analogues, namely MSI-78, MSI-367, MSI-594 and MSI-843. In this context, by using solid-state nuclear magnetic resonance (NMR), the authors concluded that MSI peptides fragmented LUVs by using a detergent-like process. As described here for Pa-MAP 1.9, MSI peptides were able to disrupt anionic (POPG; POPS) LUVs at much lower concentrations than zwitterionic (POPC) LUVs. However, when cholesterol was present in the lipid vesicle composition, the disruptive potential of these peptides drastically decreased.
Similarly, in molecular docking simulations, the complex of Pa-MAP 1.9 with POPC/POPS revealed improved affinity values and a larger number of atomic interactions of different types (hydrogen and saline bonds), when compared to the Pa-MAP 1.9 complex with POPC/Cholesterol (Table 2). In a recent work, the parent peptide Pa-MAP was predicted to form hydrophobic interactions with a DPPC membrane with a lower affinity value (− 3.1 kcal.mol −1 ), when compared with the results here 22 . Comparing the primary sequences of Pa-MAP and Pa-MAP 1.9, it was possible to observe a decrease in the hydrophobic residue content of Pa-MAP 1.9 (64%) when compared to Pa-MAP (73%), which is due to the substitution of Pa-MAP hydrophobic residues by hydrophilic ones in Pa-MAP 1.9. Similarly this was reflected in the hydrophobic moment of Pa-MAP 1.9, which was higher when compared to Pa-MAP, being 0.26 and 0.10 on the Eisenberg scale, respectively 26 . Furthermore, due to the addition of four lysine residues (Lys4, Lys7, Lys11 and Lys18) 1.9 had a net charge of + 4, which made this peptide much more cationic than its precursor (net charge = − 1), favouring its interaction with bacterial membranes.   In addition, the positioning of the amino acid residues of Pa-MAP 1.9 was designed to allow a more efficient distribution of hydrophobic and positively charged faces, favoring the amphipathicity of the molecule, as well as distributing alanine residues more evenly along the entire peptide, increasing the probability of helical structuring. These characteristics of Pa-MAP 1.9 would provide a major driving force for α -helix insertion into anionic bilayers, causing their perturbation, as well as entropy loss due to peptide loss of water interactions caused by translocation into the membrane and consequent α -helix structuring, as observed previously for other polyalanine peptides 27 .
Structurally, the CD spectra of Pa-MAP 1.9 revealed a well-defined α -helix conformation in the presence of SDS or 50% TFE. Similarly, Perrin and colleagues 28 described the CD spectrum of piscidins 1 and 3 which, in the presence of anionic vesicles (POPC/POPG 3:1 and POPE/POPG 1:1), tend to become arranged into α -helical structures. In water, however, Campagna and co-workers 19 reported that piscidins 1 had an unstructured conformation, also observed by NMR spectroscopy. Consistent with this work, when Pa-MAP 1.9 was solubilized in water, independently of the pH, a random conformation could be observed. These events could also be three-dimensionally visualized in our molecular dynamics simulations, whereby Pa-MAP 1.9 presented a well-defined α -helix conformation when in 50% TFE and in contact with an SDS micelle, but a coil conformation with a small, central α -helix in water. These features are consistent both with other AMPs from polar fishes as with the polyalanine precursor Pa-MAP 12 . Such environment-dependent behavior is important not only for the understanding of membrane-based mechanisms of action, but also are crucial for the elucidation of how AMPs reorient, insert into and translocate across membranes to act on intracellular targets.
In summary, Pa-MAP 1.9 proved to be a useful candidate for the treatment of Gram negative bacteria infections, especially in their biofilm state. It works through a dose-dependent mechanism of action, without being hemolytic and cytotoxic against mammalian cells. We also demonstrated that Pa-MAP 1.9 was able to interact with bacterial mimetic membranes and vesicles, adopting a well-defined α -helical conformation that might favor Pa-MAP 1.9 orientation and insertion into lipid bilayers to perform its activities.

Material and Methods
Pa-MAP 1.9 synthesis and mass spectrometry analysis. Pa-MAP 1.9 was purchased from Peptide 2.0 Incorporated (USA), which synthesized the peptide at 95% purity (after TFA removal) by the stepwise solid-phase method using the N-9-fluorenylmethyloxycarbonyl (Fmoc) chemistry on a Rink amide resin. Pa-MAP 1.   Anti-biofilm assays. Dilutions (1/100) of overnight cultures of E. coli (ATCC O157) and K. pneumoniae (KPC 001825971) grown overnight in LB medium were incubated in basal medium 2 (BM2) [62 mM potassium phosphate buffer (pH 7), 7 mM (NH 4 ) 2 SO 4 , 2 mM MgSO 4 , 10 μ M FeSO 4 , 0.5% glucose], in 96-well cell culture cluster round bottom microplate in the presence of Pa-MAP 1.9 for 24 h at 37 °C. Positive growth control contained only bacteria, and negative control wells contained broth only. Afterwards, planktonic cells were removed and microplate wells were washed twice with deionized water. The attached bacteria remaining were stained with 100 μ L of 0.1% (m/v) crystal violet for 20 min. The microplates was rinsed twice with deionized water, air-dried and resolubilized with 110 μ L of ethanol 70%. The content of the microplate was transferred to a new microplate, and the MBIC of Pa-MAP 1.9 was accessed at 595 nm in a microtiter plate reader (Bio-Tek Instruments Inc., VT).
Flow cell assays. Biofilms were grown for 72 h, at 37 °C in flow chambers with channel dimensions of 1 by 4 by 40 mm, as previously described 29 . For treatment of pre-formed biofilms, bacteria were allowed to develop structured 2-day-old biofilms prior to treatment with Pa-MAP 1.9, for the following 24 h. The medium used to allow biofilm growth was (BM2) containing 0.4% (wt/vol) glucose as a carbon source. Silicone tubing (VWR, 0.062 in ID by 0.125 in OD by 0.032 in wall) was autoclaved, and the system was assembled and sterilized by pumping a 0.5% hypochlorite solution through the system at 6 rpm for 30 min using a Watson Marlow 205S multi-channel peristaltic pump. The system was then rinsed at 6 rpm with sterile water and medium for 30 min each. Flow chambers were inoculated by injecting 400 μ l of an overnight culture diluted at 1:25 (v/v). After inoculation, chambers were left without flow for 2 h, after which medium was pumped through the system at a constant rate of 0.5 rpm for 48 h. After this time, treatments were diluted in BM2 medium at desired concentration and pumped through the system for 24 h. Biofilm cells were stained using the LIVE/DEAD BacLight Bacterial Viability kit (Molecular Probes, Eugene, OR) prior to microscopy experiments. A ratio of SYTO-9 (green fluorescence, live cells) to propidium iodide (PI) (red fluorescence, dead cells) of 1:5 was used. Microscopy was done using a confocal laser-scanning microscope (Olympus, Fluoview FV1000) and three-dimensional reconstructions were generated using the Imaris software package (Bitplane AG). Cytotoxicity assays. For the cytotoxic assays, confluent RAW 264.7 (mouse leukemic monocyte macrophage) cell line was challenged with 0.7 to 115 μ M of Pa-MAP 1.9 and incubated at 37 °C in a 5% CO 2 atmospheres for 48 h, according to previously described methodology 12 . Experiments were performed in triplicate and cells imaged using an inverted optical microscope (Leitz) in order to describe morphological alterations. The neutral red dye-uptake method 30 was used to evaluate cell viability. Cells were incubated in the presence of 0.01% (by weight) neutral red solution for 3 h at 37 °C in a 5% CO 2 atmosphere. After, the medium was removed and the cells were fixed with 4% formalin in phosphate-buffered saline, pH 7.2. The dye, incorporated by the viable cells, was eluted by using a mixture of methanol:acetic acid:water (50:1:49,v:v:v), and the dye uptake was determined by measuring the absorbance at 490 nm in an automatic spectrophotometer (ELx800 TM-Bio-TeK Instruments, Inc.). The cytotoxic assays were performed in triplicate. before measurements. Phosphate buffer saline (20 mM sodium phosphate, 150 mM NaCl) pH 7.4 was used for these measurements. Peptide-induced lipid vesicle leakage was measured as previously described 13 , by fluorescence spectroscopy, on a Varian Cary Eclipse fluorescence spectrophotometer (Mulgrave, AU). In this assay, we monitored the release of 5,(6)-carboxyfluorescein (CF; Sigma, St. Louis, MO), trapped in the LUV. LUVs were prepared as described above, with dried film hydration in phosphate buffered saline (PBS) containing 100 mM CF (pH was adjusted to 7.4 with NaOH). Free CF was removed by passing the suspension through an EconoPac 10 DG column from Bio-Rad (Richmond, CA), where the vesicles are eluted with the void volume 13 . Aliquots of the liposomal stock preparations (diluted to 10 μ M) were incubated with different concentrations of the peptide at 25 °C. Fluorescence was recorded continuously for 30 min, with excitation at 492 nm and emission at 517 nm (5 and 10 nm excitation and emission slits, respectively). Circular dichroism. CD measurements were performed on a JASCO J-815 spectropolarimeter (Easton, MD), equipped with a Peltier-type temperature controller and a thermostable cell holder, also interfaced with a thermostatic bath. Spectra were recorded in 0.1 cm path length quartz cells at a peptide concentration ranging from 0.05 to 0.5 mg.mL −1 in 2 mM Na-acetate buffer at pH 3.0, 2 mM Na-acetate buffer at pH 4.0, 2 mM Na-acetate buffer at pH 5.5, deionized water (Milli-Q), 2 mM Tris-HCl buffer at pH 7.0, 2 mM Tris-HCl buffer at pH 8.5, 2 mM glycine-NaOH buffer at pH 10.0 and 2 mM glycine-NaOH buffer at pH 11.0. Four consecutive scans were accumulated and the average spectra stored. Pa-MAP 1.9 analysis in the presence of SDS and TFE were performed in the same quartz cell with a 0.1 cm path length at 20 °C. The spectra were recorded between 190 and 260 nm at a scan speed of 50 nm.min −1 and four scans were performed per sample. The spectra were recorded in three average environments: distilled water, 28 mM SDS and TFE 50% in water (v:v). The observed ellipticity was converted into the mean residue ellipticity [θ ] based on a mean molecular mass per residue of 115 Da. The data were corrected for the baseline contribution of the buffer and the observed ellipticities at 222 nm were recorded. Molecular Modelling. Initially, BLASTp 35 was performed in order to find the best primary sequence template for molecular modelling. After that, 200 theoretical three-dimensional models were constructed by using Modeller v. 9.12 36 , based on an antifreeze peptide primary sequence (PDB code: 1WFA) isolated from P. americanus. The lowest free-energy theoretical model for Pa-MAP 1.9 was then selected and used for validation procedures. Firstly, PROCHECK 14 evaluated peptide's geometry, stereochemistry and energy distribution, also calculating its average score for dihedral angles, jointly with covalent forces. Moreover, 3DSS 37 was also used for (RMSD) calculation by superimposing Cα -traces and backbones of both theoretical and template tridimensional structures. Molecular Dynamics. The molecular dynamics simulations for the Pa-MAP1.9 were carried out in three different steps. The first one was performed in water; the second one in water and TFE 50%/50% (volume/volume) and, the third one was carried out in water with an anionic environment (SDS micelle). All simulations were performed by using the computational package GROMACS v.4 38 with the GROMOS96 43A1 force field implemented. As initial structure for dynamics simulations the best tridimensional theoretical model of Pa-MAP 1.9 was used, which was immersed in 7,897 water molecules in a cube box with sides measuring 6.22 nm. In order to neutralize the system's charge, chloride ions were also added (four ions in this case). TFE 50% simulation was carried out with 1,892 and 2,023 water molecules, as well as 459 and 354 TFE molecules in a cube box with sides measuring 4.78 nm. The final step (SDS) was carried out in a dodecahedral box with a minimum distance of 0.7 nm between the non-water molecules and the box's frontiers. In this case, Pa-MAP 1.9, a SDS micelle with 128 phospholipid residues, 13,021 water molecules and chlorine ions constituted the system. Geometry of water and water/TFE 50% molecules was constrained using the SETTLE algorithm 39 . Moreover, the LINCS algorithm was used to link all the atom bond length. Particle Mesh Ewald (PME) was also used for electrostatic corrections, with a radius cut-off of 1.4 nm in order to minimize the computational simulation time. The same radius cut off was also used for van der Waals interactions. The list of neighbors of each atom was updated every 10 simulation steps of 2 fs each. A conjugate gradient (2 ns) and the steepest descent algorithms (2 ns) were implemented for energy minimization. After that, the system underwent a normalization of pressure and temperature, using the integrator stochastic dynamics (SD), 2 ns each. The system with minimized energy and balanced temperature and pressure was carried out using a step of position restraint, using the integrator Molecular Dynamics (MD), for 2 ns. The simulations were carried out at 27 °C in silico, aiming to understand the structural conformation of the peptide more nearly to that observed in vitro bioassays. The total time of Pa-MAP 1.9 simulation was 100 ns, in triplicate.

Molecular Docking.
For molecular docking studies, the program AUTODOCK 4.2 40 was used in order to predict the possible modes of interaction between Pa-MAP 1.9 and anionic as well as zwitterionic mimetic membranes, based on what was done in the leakage experiments. For this, the CHARMM-GUI 41 server was used for membrane construction, constituted of POPC/POPS (50:50) (anionic environment) and POPC/cholesterol (70:30) (zwitterionic environment), with 75 × 75 × 45 Å 3 of dimension. Moreover, by using AutoDock Tools, all the hydrogen atoms were added and a grid box of 60 × 60 × 20 points with spacing of 1.0 Å was centered on the membrane surfaces. Furthermore, AutoDock Tools were also used for peptide manipulation, where the maximum freedom to side chains was unlocked. The molecular docking simulations were programed for fifty random runs, with the generated structures being ranked based on their affinity values in kcal.mol −1 . After all