The Aib residue is a non-natural amino acid mainly used to increase the stability of α-helix conformation (Bellanda et al., 2001). Due to their strong helicogenicity (De Zotti et al., 2012a), when Aib residues are inserted into the primary structure of peptides, they are expected to increase their helical content and to protect them from proteolytic attack (Rink et al., 2010). The usage of Aib residues is also expected to enlarge the spectrum of activity of Esc(1–21). This is because a stabilized α-helix structure is reported to be correlated to the AMPs' activity against Gram-positive bacteria (Giangaspero et al., 2001) toward which Esc(1–21) is not particularly active. The Aib-analog was synthesized by replacing three amino acids in positions 1, 10, and 18 with Aib residues (Figure 1). This replacement was rationally designed on the basis of the following considerations: (i) an Aib residue in position 1 should prevent enzymatic degradation by aminopeptidases, while the protection of the C-terminus from carboxypeptidases would not be necessary, due to the presence of an amidated glycine in Esc(1–21) (Rink et al., 2010); (ii) Aib residues in positions 10 and 18 should contribute to stabilize the α-helix structure, due to their stronger α-helix-promoting activity when placed internally to the primary structure of a peptide; (iii) according to an ideal α-helix folding of Esc(1–21), it was possible to identify an hydrophobic face and an hydrophilic one (Biondi et al., 2016). It is known that in naturally-occurring Aib-rich peptides, such as peptaibiotics, Aib residues are located either within the hydrophobic face or at its boundary with the hydrophilic one (Toniolo et al., 1994; De Zotti et al., 2012b).

The secondary structure of both Esc(1–21) and its [Aib^1,10,18]-Esc(1–21) was initially investigated by circular dichroism (CD) (Biondi et al., 2016) in water and two different membrane-mimicking environments e.g., sodium dodecyl sulfate (SDS) aqueous solution and trifluoroethanol (TFE). The results confirmed that both peptides adopted an unordered conformation in water and an α-helix structure in both SDS and TFE. However, at increasing concentration of TFE (from 20 to 50%) the helical content in [Aib^1,10,18]-Esc(1–21) sharply increased with respect to the parent peptide. The helical and less flexible structure of [Aib^1,10,18]-Esc(1–21) compared to Esc(1–21) was also confirmed by 2D-NMR analysis in TFE solution (Biondi et al., 2016).

Overall, the greater stability and content of α-helix in the Aib-analog were found to influence the biological properties of the peptide. More precisely, the Aib-analog gained an overall higher activity against Gram-positive bacteria, especially those belonging to Staphylococcus genus [MIC of 2–4 vs. 16–64 μM of Esc(1–21)] without losing its efficacy against Gram-negative bacteria and Candida species (Biondi et al., 2016). It is possible that differences in the composition of the membrane or cell wall among these microorganisms account for the different activity of the two esc-peptides against them.

However, a higher α-helicity in [Aib^1,10,18]-Esc(1–21) resulted in increased cytotoxicity against mammalian cell lines (e.g., alveolar epithelial cells and keratinocytes). The experimental data showed that, at the antimicrobial concentrations, Esc(1–21) and [Aib^1,10,18]-Esc(1–21) were harmless to human cells, while at higher concentrations the cytotoxic effect of the Aib-analog became clearly evident in comparison with the parent Esc(1–21) (Biondi et al., 2016). This is consistent with the notion that both α-helix conformation and its stability are crucial parameters for mammalian membrane perturbation and cell lysis, likely assisting the peptide's penetration into the hydrophobic core of phospholipids bilayers (Shai and Oren, 1996).

With the aim to reduce the cytotoxicity and to protect Esc(1–21) from proteolytic degradation, another analog carrying two d-amino acids was synthesized: Esc(1–21)-1c. It was obtained by replacing two l-amino acids i.e., Leu¹⁴ and Ser¹⁷ with the corresponding d-amino acid enantiomers (Figure 1). This diastereomer was rationally designed on the basis of the following considerations: (i) d-amino acids are known to be “α-helix breakers” (Grieco et al., 2013) and a reduction in the α-helix content of the peptide should reduce its propensity to perturb mammalian membranes leading to cell death (Strahilevitz et al., 1994); (ii) previous studies on the shorter analog Esc(1–18) pointed out that in electrically-neutral lipid vesicles the peptide adopted an α-helix conformation at its C-terminal half. With the purpose to disrupt at least the first turn of the α-helix expected to be present in the C-terminal half of Esc(1–21) in mammalian cell membranes, analogously to what found for Esc(1–18), replacement of two l-amino acids with the corresponding d-enantiomers was carried out at position 14 and 17. Note that it was improbable that the C-terminal tail Gly18-Gly21 of Esc(1–21) folded in a stable helical conformation.

The stability of both isomers was initially examined in the presence of 10 and 30% fresh human serum after 24 h incubation at 37°C. The data revealed that in comparison with Esc(1–21) <50% of the diastereomer was degraded (Di Grazia et al., 2015a). Besides, the presence of these two d-amino acids made the peptide significantly more resistant to the proteolytic cleavage caused by both human and bacterial elastases (Cappiello et al., 2016).

When the structure of the two peptides was analyzed by CD in lysophosphatidylcholine (LPC), which simulates the zwitterionic nature of mammalian cell membranes, a loss of α-helix structure was clearly detected for Esc(1–21)-1c (Di Grazia et al., 2015a). In contrast with data obtained for the Aib-analog, this diastereomer was significantly less toxic than Esc(1–21) against mammalian cells, either circulating cells (e.g., erythrocytes, macrophages) or epithelial cells. More precisely, its LD₅₀ was higher than 256 μM in comparison with a LD₅₀ ranging from 64 to 150 μM for the all-l peptide toward macrophages and epithelial cells, respectively (Di Grazia et al., 2015a; Cappiello et al., 2016). Interestingly, the introduction of these two residues in the d-configuration also conferred the peptide: (i) a higher tendency than the all-l counterpart to kill P. aeruginosa biofilms at concentrations lower than 25 μM (despite the diastereomer had a slightly reduced bactericidal activity against the free-living form of this pathogen); and (ii) a higher “wound” healing activity in vitro (Di Grazia et al., 2015a).

Amino acids replacement is not the only strategy to increase the stability of a peptide to proteolytic degradation. Moreover, this approach does not allow a peptide to overcome biological barriers (e.g., mucus, skin layers) before reaching the site of infection at high active concentrations (d'Angelo et al., 2015). A different biochemical approach to also assist drug delivery at effective concentrations is given by its conjugation to nanoparticles (NPs). This would enable not only to protect the drug from the external environment but also to increase its local concentration.

Among the various NPs produced in recent years, AuNPs have attracted most attention due to their small size, high solubility, stability, biocompatibility, and chemical inertness (Connor et al., 2005). They can diffuse through all layers of human skin (Williams et al., 2006) and because of their large surface area, they can be functionalized with a high number of molecules (Yih and Al-Fandi, 2006; Pietro et al., 2016; Soica et al., 2016). Nevertheless, only a limited number of studies has been reported to date on the effects of conjugation of AMPs to AuNPs (Rai et al., 2016). By using Esc(1–21) as a model peptide, it was demonstrated for the first time how a chemical conjugation of an AMP via polyethylene glycol (PEG) linker to AuNPs increases its antimicrobial activity while retaining its mode of action without becoming toxic to human keratinocytes. AuNPs were synthetized by the citrate reduction of gold and stabilized with a bifunctional PEG bearing a thiol and a carboxylic group. The PEG was attached to the AuNPs via a gold-thiol bond (AuNPs@PEG), while the carboxylic group was used for further derivatization with the peptide via carbodiimide-mediated coupling (Casciaro et al., 2017).

Remarkably, the obtained AuNPs@Esc(1–21) resulted to be more active than the free peptide against both planktonic and sessile forms of P. aeruginosa. This was indicated by the corresponding minimal concentrations causing 50% killing of both bacterial phenotypes which were found to be ~15-fold lower than those of the free Esc(1–21). This is presumably due to the higher concentration of peptide molecules at the site of bacterium-NP contact, as visualized by electron microscopy images which evidenced how these AuNPs@Esc(1–21) form clusters at various points on the bacterial surface with disruption of the membrane and leakage of cytosolic material. Otherwise, our unconjugated bare-AuNPs did not show any anti-pseudomonal activity and were not detected around bacterial cells (Casciaro et al., 2017). This is in line with the findings that non-functionalized AuNPs are harmless also to other bacterial pathogens (Williams et al., 2006) and suggests that the cationic AMP represents the driving force allowing AuNPs@Esc(1–21) to reach the target site at high concentration. In addition, AuNPs@Esc(1–21) were resistant to proteolytic degradation preserving their antibacterial activity 2 h after treatment with trypsin (Casciaro et al., 2017). Finally, AuNPs@Esc(1–21) were harmless to keratinocytes and retained the peptide's capability to stimulate migration of keratinocytes in a pseudo-“wound” healing assay. Altogether these findings make AuNPs@Esc(1–21) an attractive nano-formulation for topical treatment of skin infections (Casciaro et al., 2017).