Most Gram-negative bacteria are susceptible to polymyxin B (PxB), and development of resistance to this cationic lipopeptide is very rare. PxB mechanism of action involves interaction with both the outer membrane (OM) and the inner membrane (IM) of bacteria. For the design of new antibiotics based on the structure of PxB and with improved therapeutic indexes, it is essential to establish the key features of PxB that are important for activity. We have used an approach based on mimicking the outer layers of the OM and the IM of Gram-negative bacteria using monolayers of lipopolysaccharide (LPS) or anionic 1-palmitoyl-2-oleoylglycero-sn-3-phosphoglycerol (POPG), respectively, and using a combination of penetration assay, analysis of pressure/area curves, and Brewster angle microscopy to monitor surface morphology changes. Synthetic analogue sp-B maintains the basic structural characteristics of the natural compound and interacts with the OM and the IM in a similar way. Analogue sp-C, with a mutation of the sequence [d-Phe6-Leu7] into [d-Phe6-Dab7], shows that this hydrophobic domain is involved in LPS binding. The significant role of the positive charges is demonstrated with sp-Dap analogue, where l-alpha,gamma-diaminobutyric acid residues Dab1 and Dab8 are replaced by l-alpha,gamma-diaminopropionic acid (Dap), resulting in lower degrees of insertion in both LPS and PG monolayers. The importance of the N-terminal acyl chain is demonstrated with polymyxin B nonapeptide (PxB-np). PxB-np shows lower affinity for LPS compared to PxB, sp-B, or sp-C, but it does not insert into PG monolayers, although it binds superficially to the anionic film. Since PxB microbial killing appears to be mediated by osmotic instability due to OM-IM phospholipid exchange, the ability of the different peptides to induce membrane-membrane lipid exchange has been studied by use of phospholipid unilamellar vesicles. Results indicate that cationic amphipathicity determines peptide activity.

sP-B is a synthetic analogue of the natural lipopeptide antibiotic polymyxin B (PxB) that maintains the ability of the parent compound to form vesicle-vesicle contacts and induce lipid exchange. Exchange is selective, and only monoanionic phospholipids such as 1-palmitoyl-2-oleoyl-glycero-sn-3-phosphoglycerol (POPG) are transferred, whereas dianionic phospholipids such as 1-palmitoyl-2-oleoyl-glycero-sn-3-phosphate (POPA) are not, as shown by fluorescence experiments based on the excimer/monomer ratio of pyrene-labeled phospholipids. Synthetic fluorescent analogues of sP-B are used to investigate the peptide position and orientation in the intermembrane contacts: sP-Bw, an analogue that contains D-tryptophan (D-Trp) instead of the naturally occurring D-phenylalanine, and sP-Bpy, incorporating a pyrene group at the N-terminus. Tryptophan fluorescence, anisotropy, and quenching measurements performed with sP-Bw indicate that the peptide binds and inserts in anionic vesicles of POPG and POPA. However, significant differences are seen depending on the lipid composition, as also demonstrated by fluorescence resonance energy transfer (FRET) experiments from Trp to 7-nitro-2-1,3-benzoxadiazol (NBD) groups at the interface. Intermolecular FRET using sP-Bw as the donor and sP-Bpy as the acceptor indicates self-association of the peptide, possibly forming dimers, when bound to POPG vesicles at concentrations that induce the vesicle-vesicle contacts.

We have designed synthetic peptides that mimic the primary and secondary structure of the cationic lipopeptide antibiotic polymyxin B (PxB) in order to determine the structural requirements for membrane action and to assess possible therapeutic potential. Two analogues with related sequences to that of PxB, but including synthetic simplifications (disulphide bridge between two cysteines in positions 4 and 10, N-terminal nonanoic acid), have been synthesized. Peptide-lipid interactions have been studied by fluorescence resonance energy transfer between pyrene and 4,4-difluoro-5-methyl-4-bora-3alpha,4alpha-diaza-s-indacene-3-dodecanoyl (BODIPY)probes covalently linked to phospholipids, and the possibility of membrane disruption or permeabilization has been assessed by light scattering and fluorescence quenching assays. The synthetic peptide sP-B, which closely mimics the primary and secondary structures of PxB, binds to vesicles of anionic 1-palmitoyl-2-oleoylglycero-sn-3-phosphoglycerol (POPG) or of lipids extracted from Escherichia coli membranes, and induces apposition of the vesicles and selective lipid exchange without permeabilization of the membrane. We conclude that sP-B forms functional vesicle-vesicle contacts that are selective, as previously described for PxB. The second analogue, sP-C, has a permutation of two amino acids that breaks the hydrophobic patch formed by D-Phe and Leu residues on the cyclic part of the sequence. sP-C lipopeptide is more effective than sP-B in inducing lipid mixing, but shows no selectivity for the lipids that exchange through the vesicle-vesicle contacts, and at high concentrations has a membrane-permeabilizing effect. The deacylated and non-antibiotic derivative PxB-nonapeptide (PxB-NP) does not induce the formation of functional intervesicle contacts in the range of concentrations studied.

Antibiotic resistance is a daunting challenge for public health systems worldwide. A major goal to fight resistant bacteria involves the design, discovery and development of new antibiotics, particularly against multi-drug-resistant strains. Currently, there is renewed interest in polymyxins, an old class of antimicrobial cyclic lipopeptides, highly potent against therapeutically relevant Gramnegative bacteria. Polymyxins are now used as last resort antibiotics in hospitals because of their nephrotoxicity and neurotoxicity that requires careful monitoring of the patient. Our group has embarked on a project to design and develop new polymyxins devoid of toxicity problems using a versatile and chemically accessible scaffold structure [1,2]. Compounds show excellent activity against Gram-negative bacteria. Synergistic and antibiofilm activities have also been recently described in combination with imipenem [3]. Herein, the latest results of our recently designed polymyxin analogs will be presented. Acknowledgments:The research was supported by the University of Barcelona, Fundació Bosch i Gimpera, Xarxa de Referència en Biotecnologia, 2016LLAVO0018 grant (Generalitat de Catalunya) and the European Institute of Innovation and Technology (EIT Health). The authors are members of the ENABLE (European Gramnegative Antibacterial Engine) consortium (IMI-ND4BB, http://www.imi.europa.eu/projects-results/projectfactsheets/enable).