ll-37-derived cyclic antimicrobial drug...

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 281

LL-37-derived cyclicantimicrobial drug leads

Design, synthesis, activity and different ways ofcreating them

TAJ MUHAMMAD

ISSN 1651-6192ISBN 978-91-513-0813-5urn:nbn:se:uu:diva-397191

Dissertation presented at Uppsala University to be publicly examined in A1:107a, Biomedical Centrum (BMC), Uppsala, Thursday, 19 December 2019 at 09:15 for the degree of Doctor of Philosophy (Faculty of Pharmacy). The examination will be conducted in English. Faculty examiner: Associate Professor Wang Guangshun (University of Nebraska Medical Center).

AbstractMuhammad, T. 2019. LL-37-derived cyclic antimicrobial drug leads. Design,synthesis, activity and different ways of creating them . Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy 281. 65 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0813-5.

In an era where last-line antibiotics are failing, one of the powerful approaches to develop novel therapeutic agents is to turn back to nature in order to identify possible drug candidates. Among the potential candidates, antimicrobial peptides (AMPs) have garnered much attention as an antimicrobial. These are broad spectrum host defense molecules produced by all living organisms. LL-37 is such a multitask human defense peptide that mediates various host immune responses and also exerts antimicrobial activity. However, the direct use of this 37-amino acid long α-helical peptide is hampered by protease susceptibility, in particular for antimicrobial applications. A small 12-residues peptide, referred as KR-12, derived from LL-37, has been reported to have selective toxic effect on bacteria.

Analogues of KR-12 were generated in the form of Alanine and Lysine scans to find out the positions important for improved activity and selectivity. Backbone-cyclised dimers based on KR-12 and KR-12 analogues, tethered by linkers of two to four amino acid residues, were synthesised to explore the concept of cyclisation, dimerisation and cross-linking as means to enhance peptide stability and activity. Antimicrobial activities of the linear peptides and cyclic dimers were assayed against human pathogens, in buffer and/or physiological conditions. Proteolytic stability, permeabilisation efficacy on microbial membranes and, their structures were also characterised.

From Ala and Lys scans, it was possible to identify two key positions for the enhanced broad-spectrum antibacterial activity: replacement of Gln5 with Lys, and Asp9 with either Ala or Lys. In serum stability assay, KR-12 and analogues were found to be unstable. The backbone-cyclised KR-12 dimers showed improved antimicrobial activity and increased stability compared to monomeric KR-12. KR-12 monomers adopt a well-defined α-helical structure in membrane-mimicking environment, while cyclised dimers were unstructured in solution judged by NMR. The KR-12 (Q5K, D9A) cyclised dimers retained antimicrobial activity in physiological conditions. Circular dichroism showed that the cyclic dimer, cd4-PP, had 77% helical content when bound to lyso-phosphatidylglycerol micelles.

Moreover, the limits of cyanobactin-macrocyclase PatGmac were explored to cyclise peptides larger than their natural substrates, namely the PawS derived peptide Sunflower Trypsin Inhibitor-1 (SFTI-1) and the cyclotide kalata B1. PatGmac was used very efficiently to cyclise SFTI-1. In addition, semi-pure butelase 1, isolated from Clitoria ternatea seeds, was immobilised on NHS column. The immobilised column was then used to produce substrates ranging from 16 to 34 varying length.

Keywords: antimicrobial peptide, host defense, antimicrobial, LL-37, KR-12, peptide cyclisation, peptide dimerisation, croos-linking, cyanobactin-macrocyclase, PatGmac, butelase 1, enzymatic cyclisation, immobilisation

Taj Muhammad, Department of Medicinal Chemistry, Farmakognosi, Box 574, Uppsala University, SE-751 23 Uppsala, Sweden.

© Taj Muhammad 2019

ISSN 1651-6192ISBN 978-91-513-0813-5urn:nbn:se:uu:diva-397191 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-397191)

To my family past, present & future

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Gunasekera, S*., Muhammad, T*., Strömstedt, A. A*., Rosen-gren, K. J., Göransson, U. (2018) Alanine and lysine scans of the LL-37 derived peptide fragment KR-12 reveal key residues forantimicrobial activity. ChemBioChem, 19(9):931-939.

II Gunasekera, S*., Muhammad, T*., Strömstedt, A. A*., Rosen-gren, K. J., Göransson, U., Backbone cyclisation and dimerisa-tion of LL-37-derived peptides enhance antimicrobial activityand proteolytic stability. (Submitted)

III Muhammad, T., Strömstedt, A. A., Skogman, M., Fallarero, A., Gunasekera, S., Göransson, U. Transformation of LL-37 based cross-linked cyclic dimers into stable and potent antimicrobial drug leads. (Manuscript)

IV Muhammad, T., Houssen, W. E., Thomas, L., Alexandru-Crivac, C. N., Gunasekera, S., Jaspars, M., Göransson, U. Ex-ploring the limits of cyanobactin macrocyclase PatGmac: Cy-clisation of PawS derived peptide sunflower trypsin inhibitor-1 and cyclotide kalata B1. (Submitted)

V Muhammad, T., Shafiullah, Md., Gunasekera, S., Göransson, U. Immobilisation of semi-pure butelase 1 and its applications toproduce head-to-tail cyclic peptides. (Manuscript)

Reprints were made with permission from the respective publishers. *These authors contributed equally.

Additional Publications

• Shaikh, J. U., Muhammad, T., Shafiullah, Md., Slazak, B., Raouf, R., Göransson U. (2017) Single-step purification of cyclotides using affinity chromatography. Biopolymers, 108, e23010.

• Hellinger, R., Thell1, K., Vasileva1, M., Muhammad, T., Gunasekera, S., Daniel Kümmel, D., Göransson, U., Becker, C. W., Gruber, C. W. (2017) Chemical Proteomics for target discovery of head-to-tail cyclized mini-proteins. Frontiers in Chemistry, 5: 73.

• Mohotti, S., Rajendran, S., Muhammad, T., Strömstedt, A. A. Adhi-kari, A., Burman, R., de Silva E. D., Göransson, U. Hettiarachchi, C. M., Gunasekera, S. (2019) Screening for bioactive secondary metab-olites in Sri Lankan medicinal plants by microfractionation and tar-geted isolation of antimicrobial flavonoids from Derris scandens. Journal of Ethnopharmacology, 246:112158.

Contents

Abbreviations ............................................................................................ ix

Introduction ...............................................................................................11 The discovery of antimicrobial peptides ................................................11 Antimicrobial Peptides ..........................................................................11 Mechanism of action of AMPs ..............................................................12 AMPs as therapeutic agents ...................................................................13 Obstacles to develop AMPs for clinical applications ..............................14 Chemical strategies to improve AMPs as drug leads ..............................15

Aim ...........................................................................................................16

LL-37-based drug design (I-III) .................................................................17 Background...........................................................................................17 Paper I ..................................................................................................19

Design Strategy ................................................................................19 Results and Discussion .....................................................................19

Paper II .................................................................................................24 Design Strategy ................................................................................24 Results and Discussion .....................................................................24

Paper III ................................................................................................29 Design Strategy ................................................................................29 Results and Discussion .....................................................................29

Enzymatic cyclisation (IV-V) ....................................................................35 Background...........................................................................................35 Paper IV................................................................................................36

Cyanobactin macrocyclase PatGmac .................................................36 Cyclisation Strategy ..........................................................................36 Results and Discussion .....................................................................37

Paper V .................................................................................................40 Butelase 1 .........................................................................................40 Cyclisation Strategy ..........................................................................40 Results and Discussion .....................................................................41

Concluding Remarks .................................................................................45

Perspectives and Proposals ........................................................................47

Popular Science Summary .........................................................................51

Acknowledgements ...................................................................................54

References .................................................................................................57

Abbreviations

AEPs Asparaginyl endopeptidases AMPs Antimicrobial peptides Ala Alanine ATCC American type cell culture Boc t-butyloxycarbonyl CD Circular dichroism Cys Cysteine cyO19 Cycloviolacin O19 C. albicans Candida albicans C. ternatea Clitoria ternatea Dbz Di-Fmoc-3,4-diaminobenzoic acid DMF Dimethylformamide E. coli Escherichia coli FMCA Fluorometric microculture cytotoxicity assay Fmoc Fluorenylmethoxycarbonyl hLF Human lactoferricin kB1 Kalata B1 kB7 Kalata B7 LC-MS Liquid chromatography mass spectrometry LTA Lipoteichoic acid LPS Lipopolysaccharides MIC Minimum inhibitory concentration MHB Müller-Hinton Broth MS Mass spectrometry NMR Nuclear magnetic resonance Nbz N-benzymidazolinone NCL Native chemical ligation NK cells Natural killer cells PatGmac Patellamide G macrocyclase PBS Phosphate buffer saline Phe Phenyalanine PS Phosphatidylserine PG Phosphatidylglycerol PI Phosphatidylinositol P. aeruginosa Pseudomonas aeruginosa POPB Prolyl oligopeptidase

PCY1 Peptide cyclase 1 S. aureus Staphylococcus aureus SFTI-1 Sunflower trypsin inhibitor-1 SDS Sodium dodecyl sulphate SrtA Sortase A SPPS Solid phase peptide synthesis TOCSY Total correlation spectroscopy TSB Trypticase Soy Broth

11

Introduction

The discovery of antimicrobial peptides The discovery of antimicrobial peptides and proteins (AMPs) goes back to the end of 1920’s when Alexander Fleming reported bactericidal and bacterio-static activities from nasal secretions of a patient suffering from nasal coryza and named the activity, lysozyme, the first known human antimicrobial pro-tein (Fleming 1922). Meanwhile, nisin, the first gene-encoded post-transla-tionally modified AMP, was reported in 1928 in fermented milk cultures and later on, in 1969, nisin was approved as a food additive (Rogers and Whittier 1928, Shin et al. 2016). Then, with the advent of penicillin and streptomycin, the Golden Age of antibiotics began, which led to a rapid decline in the dis-covery and loss of interest in the therapeutic potential of antimicrobial pep-tides (Zaffiri, Gardner and Toledo-Pereyra 2012). The emergence of multi-drug-resistant pathogens in the early 1960s again awakened interest in the AMPs as a potential antibiotic and a new era of research started with the dis-covery of cationic proteins and peptides from guinea-pig polymorphonuclear leukocytes with antimicrobial properties (Zeya and Spitznagel 1963). Later on, the discoveries of cecropins from moth (Hultmark et al. 1982), magainins from amphibians (Zasloff 1987) and defensins from rabbit (Selsted et al. 1985) in the 1980s, and cathelicidins (Turner et al. 1998, Bals et al. 1998b, Agerberth et al. 1995, Agerberth et al. 2000) and defensins (Bals et al. 1998a) from hu-mans in 1990s laid the foundation of antimicrobial field.

Antimicrobial Peptides Antimicrobial peptides are endogenous polypeptides, and represent an ancient form of defense strategy against infections of most organisms, and form a cru-cial component of their innate immune system (Zasloff 2002). As of Novem-ber 2019, 3136 AMPs have been reported from all life kingdoms, including archaea, bacteria, fungi, plants and animals, and some synthetic peptides, ac-cording to the online antimicrobial peptide database (http://aps.unmc.edu/AP/) (Wang, Li and Wang 2016, Wang, Li and Wang 2009). In humans, 134 AMPs have been identified in various tissues, includ-ing lysozyme, defensins, histatins, cathelicidins, lactoferricin, kinocidins, and dermcidin. Majority of AMPs are gene-encoded polypeptides but some

12

bacteria and fungi utilize non-ribosomal pathways, by using assembly lines of multiple enzyme, to produce peptide antibiotics with chemical modifications (Hancock and Chapple 1999). The vast majority of these peptides are cationic, having a net positive charge conferred by lysine and arginine residues, and only few are anionic. They have typically distinct amphiphilic properties, which are fundamental for their membranolytic and cell-penetrating proper-ties (Hancock and Sahl 2006, Yeung, Gellatly and Hancock 2011).

The structure of AMPs displays a large variety in the primary sequence but the major factors, which influence the antimicrobial activity, are net charge, hydrophobicity, position of cationic residues, conformation and amphipathi-city. AMPs can be divided into different structural groups based on the con-formation adopted by a peptide on interaction with bacterial membrane or in solution: • The alpha family is composed of AMPs with a-helical structures, such as

magainin and human cathelicidin LL-37 (Zasloff 1987, Agerberth et al. 2000).

• The beta family is composed of AMPs with b-sheets structures and ma-jority of them are cyclic, with one or more disulfide bonds. The best stud-ied peptides in this group are a-defensins (Agerberth et al. 2000), proteg-rins (Zhao, Liu and Lehrer 1994), and polyphemusin (Powers, Rozek and Hancock 2004).

• The non-alpha beta family consists of AMPs that are enriched in certain amino acids such as histidine, glycine, tryptophan, arginine or proline such as indolicidin (Sitaram, Subbalakshmi and Nagaraj 2003).

• The alpha-beta family is composed of AMPs with both a and b-sheets structures such as b-defensins (Hancock and Lehrer 1998), and cyclotides (Pränting et al. 2010, Malik et al. 2017).

Mechanism of action of AMPs The classical mode of action of cationic AMPs is direct and rapid by destroy-ing physical integrity of the microbial membrane. The crucial factor for initial interaction is the electrostatic forces between positively charged AMP and negatively charged bacterial membranes, whereas in anionic AMPs, hydro-phobicity governs the initial interaction. After membrane binding and reach-ing critical concentration, the antimicrobial peptides contribute to possible al-terations of the membrane structure such as thinning, pore formation, altered curvatures and more. This results in loss of integrity of the membrane which eventually leads to cell death. Various models have been proposed for pore formation and membrane permeabilisation of AMPs, i.e. the barrel-stave, two-state, toroidal pore, interfacial activity, and sinking raft models (Huang 2000, Pokorny, Birkbeck and Almeida 2002, Wimley 2010).

13

The preferential interaction with bacterial membranes and killing of bacte-ria by membrane disruption is due to the differences in the lipid composition of the bacterial and mammalian cell membranes (Sohlenkamp and Geiger 2016). The cytoplasmic membranes of bacteria are rich in negatively charged phospholipids such as phosphatidylglycerol (PG), phosphatidylinositol (PI), phosphatidylserine (PS), cardiolipin (CL) and zwitterionic phosphatidyleth-anolamine (PE). In addition, the outer membrane of Gram-negative bacteria contains negatively charged lipopolysaccharides (LPS) that provides addi-tional negative charge to bacterial surface and also acts as a steric barrier by reducing the accessibility to the cytoplasmic membrane. In Gram-positive bacteria, the presence of lipoteichoic acid (LTA) in the cell wall provides ad-ditional similar functionality (Epand and Epand 2009).

In contrast, mammalian cytoplasmic membranes are predominant in zwit-terionic phospholipids with net neutral charge, including, phosphatidylcholine (PC), PE, and sphingomyelin. The PI lipids and most of the phospholipids with anionic head groups are on the inner leaflets facing the cytoplasm in mammalian membranes. Moreover, cholesterol is also present in significant amounts (up to 25%) in mammalian membranes that affect the fluidity of the phospholipids in the membrane and increases the cohesiveness and thickness of the lipid bilayer (Yeaman and Yount 2003).

AMPs as therapeutic agents From early 1960s, AMPs have been in limelight to be developed into novel antibiotics, especially after the discovery of cationic bactericidal peptides in polymorphonuclear leukocytes (Zeya and Spitznagel 1963). During the 1990s after the discovery of new classes of AMPs in humans, including the b-defen-sins (Bensch et al. 1995) and the cathelicidin (Agerberth et al. 1995, Gudmundsson et al. 1996), the AMPs research mushroomed, and preclinical studies were conducted on potential candidates, leading to clinical studies and applications were filed for approval with FDA.

There was a major setback for the AMPs drug development field in 1999 when FDA approval was denied for the broad-spectrum antimicrobial peptide, pexiganan (Lamb 1998) for the topical treatment of the diabetic foot ulcers. Pexiganan is a 22-amino acid analogue of magainin 2 (Zasloff 1987) (a cati-onic peptide isolated in the skin of the African clawed frog Xenopus laevis).

Currently, there are numerous AMPs in clinical development to treat infec-tions associated with pathogenic microbes and most of these are intended for topical applications (Koo and Seo 2019). Majority of the AMPs, considered for systemic applications, failed at preclinical trials or discovery phase due to proteolytic degradation or toxicity to the host cells, and also the systemic use of the two peptide antibiotics (polymyxin B and colistin) is restricted due to in vivo toxicity (Ryan et al. 1969, Falagas and Kasiakou 2006, Falagas,

14

Kasiakou and Saravolatz 2005). The peptide, hLF1-11, a cationic fragment comprising N-terminal amino acids 1-11 of human lactoferricin, had been in clinical trial for intravenous use, to treat bacterial or fungal infections associ-ated with hematopoietic stem cell transplantation (Velden et al. 2009).

The human cathelicidin LL-37 is in current development phase, and its safety and tolerability has been studied in a phase IIa clinical trials in patients with venous leg ulcer (Grönberg et al. 2014) (https://www.promorepharma.com). OP-145, previously termed as P60.4Ac, a synthetic antimicrobial peptide derived from LL-37 binds to LPS and LTA was formulated in eardrops to treat chronic bacterial middle ear infections was in phase II clinical trials (Peek et al. 2009, Malanovic et al. 2015).

Other AMPs of animal origin were under clinical development for the treat-ment against various bacterial pathogens include pexiganan, iseganan and omiganan (Lamb 1998, Jeffcoate et al. 2008, Fritsche et al. 2008, Rubinchik et al. 2009). Lytix Biopharma was developing LTX-109, a synthetic AMP to prevent nasal carriage of methicillin-resistant and -sensitive Staphylococcus aureus (Nilsson et al. 2015).

Despite no success stories, AMPs have a number of potential advantages as future therapeutics. In addition to their broad spectrum of activity and rapid killing of microbes, they frequently neutralise endotoxins released by the path-ogens and have the ability to modulate the host immune response to infection (Yeaman and Yount 2003, Hancock and Sahl 2006, Yeung et al. 2011). Sig-nificantly, their initial typical target is the microbial membrane. Thus, a prom-ising potential application of AMPs is to facilitate the entry of conventional antibiotic into the microbial cell, resulting in synergetic therapeutic effects.

Obstacles to develop AMPs for clinical applications As mentioned earlier, a high number of AMPs have been discovered with promising and potent activities but their translation into the clinic has been difficult. During the design and optimization stage, antimicrobial activities are usually assessed without mimicking/considering in vivo environment using standard assays, mostly in buffer conditions. The antimicrobial activity of the AMPs are sensitive to the environmental conditions, that means that physio-logical levels of salt, presence of proteases or serum protein, or pH can influ-ence the activity, which results in discrepancies between in vitro vs in vivo efficacy (Mahlapuu et al. 2016).

Adding to the above, one of the greatest challenges restricting development of AMPs into therapeutics is their low metabolic stability. The oral admin-istration of peptide drugs, in general, is hampered due to low oral bioavaila-bility, because of degradations by proteolytic enzymes of the digestive system as well as poor penetration across the intestinal mucosa (Vlieghe et al. 2010). Furthermore, intravenous administration is also limited by a short half-life due

15

to rapid cleavage by proteolytic enzymes in the blood plasma, poor distribu-tion to the surrounding tissues and rapid removal from the circulation by ex-cretory organs (liver and kidneys) (Vlieghe et al. 2010).

Another limiting factor to develop AMPs into therapeutic is the cost of pro-duction. Currently, AMPs can be produced by three ways, i.e., isolation from the natural source, chemical synthesis and expression using biological sys-tems. Isolation from natural sources is beneficial for peptide discovery but is not viable to produce in bulk quantities for clinical trials and use. An alterna-tive approach is chemical synthesis using solid phase peptide synthesis (SPPS) based on Boc or Fmoc-chemistry, which is generally considered as the most mature technology available to produce linear peptides up to 50 residues with sufficient yields (Münzker, Oddo and Hansen 2017). The expression sys-tem can also be used to produce AMPs in bacteria, yeast, insect, plants and mammalian cells but with lower yields and with limitations to introduce chem-ical modifications into the peptide sequence (Parachin et al. 2012).

Chemical strategies to improve AMPs as drug leads To overcome some of the hurdles like high biodegradability and rapid removal from circulation, new strategies are needed to facilitate peptide antibiotic de-velopment. One powerful approach is to borrow ‘‘nature’s wisdom’’ to design AMP-based antibiotics with stability, selectivity, and conformational rigidity. Nature is still our main source of novel chemistries. Of these, ribosomally synthesised and post-translationally modified peptides including AMPs, the so-called RiPPs, are a large group of natural products with a high degree of structural diversity (Wang 2012, Arnison et al. 2013).

Inspired by nature, several of the post-translational modification strategies have been used in peptide engineering. Head-to-tail cyclisation and incorpo-ration of D-amino acids are strategies to improve stability, antimicrobial ac-tivity and reduce salts sensitivity of various AMPs (Dathe et al. 2004, Yu, Lehrer and Tam 2000, Fernandez-Lopez et al. 2001, Wessolowski, Bienert and Dathe 2004, Rosengren et al. 2004). Introduction of disulfide-bonds is another approach for the structural stability and biological activity of bioactive peptides (Tam, Lu and Yang 2002). Substitution of disulfide bonds with other intermolecular bridges has also been used to further stabilise structures under reducing conditions. An example of this is cystathionine and lanthionine bridges that have the ability to improve stability while imposing minimal structural alteration in the peptide (Knerr et al. 2011) . Peptidomimetics and peptide stapling have been explored as potential strategies to improve some of the inherent problems associated with AMPs (Borowski et al. 2018) (Migoń, Neubauer and Kamysz 2018, Mourtada et al. 2019).

16

Aim

The misuse of antibiotics since the discovery in the 1920s, has led to the emer-gence of multi-drug resistant (MDR) pathogens, a serious global health prob-lem. There is a pressing need for anti-infective agents with novel chemistries and mechanisms of action, and naturally occurring AMPs and their derivatives are promising alternatives.

The overall aim of my thesis has been to design stable and potent backbone cyclic dimers based on minimalised antimicrobial region of LL-37, named KR-12, and explore nature based enzymatic cyclisation strategies to create them. In particular, the projects were based on the concepts of peptide stabili-sation by backbone cyclisation and cross-linking via a disulfide bond, an idea originating from nature, using chemical synthesis and/or cyclisation enzymes.

The specific aims were to: • analyse the relationship between the primary structure and bioactivity of

KR-12, and to identify modifications that would improve activity and se-lectivity paper I.

• evaluate the impact of dimerisation, backbone cyclisation and cross-link-ing on secondary structure, proteolytic stability and bioactivity in biolog-ical relevant conditions, on KR-12 and KR-12(Q5K, D9A)-derived di-mers paper II-III.

• understand the limitations of PatGmac in cyclising longer sequences con-taining disulfide bonds, paper IV.

• expand the scope of immobilised semi-pure butelase 1 approach to cyclise substrates of diverse nature, paper V.

17

LL-37-based drug design (I-III)

Background In humans, the cathelicidin gene encodes an inactive precursor, human cati-onic antimicrobial protein (hCAP-18) that releases the active 37-residue pep-tide upon processing by proteinase 3, a serine proteinase, from the C-terminal portion of the hCAP-18 (Zanetti, Gennaro and Romeo 1995, Agerberth et al. 1995). The LL-37 production is triggered by the presence of pathogenic bac-teria in the body or by immune stimuli such as vitamin D or sunlight (Gombart 2009).

LL-37 is the only human cathelicidin identified, and can be found in a va-riety of cells such as B cells, NK cells, T cells, monocytes, mast cells and immature neutrophils (Sørensen et al. 1997, Bals et al. 1998b, Agerberth et al. 2000). Human LL-37 plays important roles in angiogenesis, wound healing, chemotaxis, signal transduction and immune modulation (Scott et al. 2002, Tjabringa et al. 2003, Zanetti 2004). As a consequence, there is an interest in developing novel peptides therapeutics based on LL-37.

The peptide itself is a 37-residue linear, cationic (net charge of +6 at phys-iological pH), a-helical peptide, which is amphiphilic in nature. LL-37 adopts a-helical structure upon interaction with the bacterial cell membranes (Wang 2008). It has broad-spectrum activity against both Gram-positive and Gram-negative bacteria, including multi-resistant strains as well as pathogenic fungi and viruses (Bals et al. 1998b, Durr, Sudheendra and Ramamoorthy 2006, Vandamme et al. 2012). Besides its direct antimicrobial activity, LL-37 has been reported to bind and neutralise LPS and LTA, and to protect the host from endotoxin shock (Larrick et al. 1995, Nell et al. 2005).

Despite the interesting bioactivities of LL-37, its direct utility as a drug lead has been hindered because of its size (and thus cost of production), non-spe-cific cytotoxicity to normal human cells and instability in human serum as it harbors several proteolytic cleavage sites (Sieprawska-Lupa et al. 2004, Rapala-Kozik et al. 2015). For those reasons, LL-37 has been truncated into smaller fragments to identify the bioactive domain/s using variety of strate-gies. Sieprawska-Lupa et al. hydrolysed the Glu16-Phe17 peptide bond using V8 proteinase, to give LL-37 (17-37) fragment, which is as active as the parent peptide itself against Staphylococcus aureus (Sieprawska-Lupa et al. 2004). Li and colleague identified two smaller fragments, LL-37 (17-32) and LL-37 (17-29) possessing antibacterial and anticancer activities using a structure-

18

based approach by trimming non-essential regions guided by total correlated spectroscopy (a “TOCSY-trim”), (Li et al. 2006). Then, the smallest bioactive domain of LL-37 has been identified as LL-37 (18-29), a peptide referred as KR-12 (Wang 2008). Interestingly, KR-12 has a minimum inhibitory concen-tration (MIC) almost as low as parent LL-37, but is devoid of the cytotoxicity shown by LL-37 (Wang 2008). This makes KR-12 an attractive target for fine-tuning its antibacterial activity and stability.

Figure 1. The design strategy of paper I, II and III. The cyclic analogues were syn-thesised from a truncated region of LL-37, named KR-12. A) The three-dimensional structure of LL-37 (PDB code 2K60). B) The minimal antibacterial region KR-12. C) Ala/Lys scans and double/triple mutants, paper I. D) A schematic representation of the backbone cyclised KR-12 dimer, paper II. E) Optimised peptide KR-12 (Q5K, D9A) from paper I. F) A schematic representation of the backbone cyclised KR-12 (Q5K, D9A) dimer, paper III.

19

Paper I Design Strategy In paper I, two series of 12 and 10 mutants of KR-12 were synthesised, by systematically replacing Ala or Lys to identify the residues important bioac-tivity (Figure 1). In addition, double and triple mutants were also synthesised to identify additive or synergistic effects between key positions. All peptides were assembled using fluorenylmethyloxycarbonyl (Fmoc)-based solid-phase peptide synthesis with piperidine (20% v/v in DMF) as the Fmoc deprotecting agent, on a CEM Liberty 1 automated microwave-assisted peptide synthesiser.

Results and Discussion

Ala and Lys scans identified key positions In paper I, KR-12 analogues were evaluated for antimicrobial activity using the two-step microdilution assay optimised for AMP-activity testing (Strömstedt et al. 2017) against human pathogens.

From Ala/Lys scans, residues Asp9 and Gln5 in KR-12 stand out as key positions suitable to modulate the activity (Figure 2). Asp9 was the only resi-due for which overall activity increased was seen in Ala scanning, against all organisms. Upon replacement with Lys, the resulting peptide D9K displayed four-to eightfold improvement on all organisms. Q5K also showed increased antibacterial activity against all organisms. Substitution of Gln with Ala, had a marginal effect on activity. KR-12 adopts a well-defined three-turn am-phiphilic a-helical structure in membrane mimicking environment, displaying distinct hydrophilic and -phobic surfaces (Wang 2008). Asp9 and Gln5 are located on the polar side of KR-12, replacement of these with Ala or Lys had a major influence on cationic movement.

It is well known that an overall positive charge is generally important for the bioactivity of antimicrobial peptides, thus it was perhaps not surprising to see that substitution of any of the Lys or Arg residue with Ala resulting in reduced activity.

Ala substitutions of large hydrophobic residues in the hydrophobic face also reduced activity: notably at Ile7, Phe10 and Leu11, indicating the im-portance of hydrophobicity for KR-12.

The importance of the position of substitution was best demonstrated by the I3K and I7K variants, which have the same overall charge and hydropathy index (+5, -1.408), but differ substantially in antibacterial activity. Introduc-tion of a positive charge in the middle of the hydrophobic surface, at position 7, was detrimental for activity.

Relative to the activity of single mutants, marginal improvements in po-tency was observed in double/triple mutants. The complementary effects from

20

both mutations was evident in Q5K, D9A (Ala and Lys mutant combined), which displayed eight-fold improvement in activity compared to KR-12 for most of the organisms. The Q5K, D9K mutant, which confers an even higher increase in dipole moment, was slightly less active against S. aureus and C. albicans as compared to Q5K, D9A. This may suggest that too high density of the repulsive positive charges, which is detrimental to helix formation (Murai and Sugai 1974), also has a negative influence on bioactivity.

Figure 2. Relative MICs for LL-37 and KR-12 analogues. The relative MIC was de-rived by dividing the median MIC value for each peptide variant against a specific organism by the median MIC of native KR-12 against that organism. Each organism is represented by a colour: Blue - Escherichia coli (EC), yellow – Pseudomonas ae-ruginosa (PA), black - Staphylococcus aureus (SA), grey - Candida albicans (CA). A) Ala substitutions, B) Lys substitutions, C) double and triple Ala/Lys substitutions.MIC values of KR-12 are E. coli: 2.5 µM; P. aeruginosa: 10; S. aureus: 10; C. albi-cans 10. For Q9A, 1.25, 1.25, 1.25 ,5 µM. D9K, values are 0.625, 1.25, 1.25, and 2.5µM, respectively.

Activity on liposomes suggests variations in modes of action A liposome leakage assay was used to assess the effects on membrane disrup-tion by the amino acid substitutions (Figure 3). The activity of KR-12 was weaker on the E. coli liposomes than that on the corresponding bacteria. KR-12 was also less potent than LL-37 on this phospholipid composition. the Sim-ilar to the results from bacterial assay, substitutions of Gln5 to Lys and Asp9

21

to either Ala or Lys increased potency most. When hydrophobic residues Ile7, Leu11 and Phe10, were substituted with Ala, a substantial decrease in leakage was observed. Substitutions of cationic residues by Ala did not decrease ac-tivity to the same extent.

The Lys scan of KR-12 presented similar effects as the Ala scan when hy-drophobic residues like Ile7 and Leu11 were substituted showing them to be the most important for membrane activity. Arg to Lys substitutions showed no marked difference in leakage.

In double/triple substitutions, no synergetic or additive effect on activity was observed in terms of membrane permeabilisation.

Figure 3. E. coli liposome leakage activity. Liposomes made from E. coli polar lipid extract were monitored for fluorescent marker leakage. The data represent the leakage levels reached after 45 min of peptide incubation. A) Ala substitutions B) Lys substi-tutions C) double/triple substitutions. Substitutions of Gln5 to Lys and Asp9 to either Ala or Lys increased potency most.

KR-12 analogues retained low cytotoxic activity KR-12 and analogues showed low toxicity against the human lymphoma cell line used but the parent peptide LL-37 was highly toxic. Similarly, KR-12 and analogues were non-hemolytic on human erythrocytes. This agrees with pre-vious studies that have attributed low levels of hemolytic activity (10%

22

hemolysis at 100 µM) and toxicity against mammalian cells (TC50 of >63 µM) to KR-12 (Wang, Watson and Buckheit 2008, Mishra et al. 2013).

Cancer cells are often more sensitive to AMPs than healthy cells, due to abundance of anionic phosphatidylserine in the outer leaflet, (Utsugi et al. 1991, Riedl, Zweytick and Lohner 2011) making them easier to target by cat-ionic peptides. Importantly, for lymphoma cell line, the low toxicity of KR-12 was retained also for the Lys and Ala double mutants, highlighting that the KR-12 template is amenable to modifications that improve activity without compromising target selectivity (Figure 4).

Figure 4. Cytotoxicity data for LL-37 and KR-12 analogues. The cell toxicity data is generated from hydrolysis of the probe fluorescein diacetate by esterases in cells with intact plasma membranes using the human lymphoma cell line U-937 GTB.

Instability of KR-12 analogues in human serum Proteolytic stability of KR-12 and mutants were evaluated in diluted human serum (1:4 v/v serum: PBS) at 37 °C followed by LC-MS analysis. KR-12 itself was found to have poor stability in human serum, but Ala substitutions at the susceptible cleavage site, R12A, displayed enhanced biological stabil-ity. These results indicate the need and potential to improve stability of KR-12 (Figure 5).

23

Figure 5. Serum stability profile of selected KR-12 analogues with corresponding sites of proteolytic digestion. A) Stability profile. B) Fragments identified by LC-MS analysis revealed that the amide bond immediately before or after an Arg is the most susceptible position of degradation (e.g. R2, R6 and R12 in KR-12). R12A was rela-tively stable.

From paper I it is concluded that the amphiphilic nature of KR-12 is very important for activity. Gln5 and Asp9 were identified as key positions to mod-ulate antimicrobial activity and positions, Ile3 and Val4, that can be explored to control selectivity towards different microbial strains. KR-12 and analogues were unstable in proteolytic environment.

Can the activity and stability of KR-12 be improved further? In paper II mac-rocyclisation and dimerisation strategies were explored to test this.

24

Paper II Design Strategy The concept of making α-helical peptide KR-12 into a dimer was explored in paper II, which in turn was made into a macrocycle to address the inherent challenge of stability. Peptides were designed to determine the importance of linker length (two, three or four residues) joining the monomers and how it affects the efficiency of peptide cyclisation and activity (Figure 6).

Figure 6. Synthesis of cyclic peptides. All the peptides intended for cyclisation were assembled as linear precursors using Fmoc-chemistry, and cyclised using native chemical ligation (NCL). A) Di-Fmoc-3,4-diaminobenzoic acid (Dbz) was coupled to Rink-amide-resin manually. B) Coupling of first residue was performed manually fol-lowed by peptide chain elongation, the last amino acid used was a Boc-protected cys-teine. C) Acylation and activation of the resin bound Dbz-precursor to yield the N-acyl urea peptide (Nbz-peptide). D) Nbz-peptide is then fully deprotected and cleaved from resin by strong acid cleavage. E) Peptide cyclisation (via thioesterification, S, N-intramolecular acyl shift and NCL leads to amide formation to yield cyclic prod-uct), the Cys residue needed for cyclisation reaction was subsequently converted toAlanine (Wan and Danishefsky 2007) to prevent intermolecular dimerisation.


Dimerisation of KR-12 increases antimicrobial activity The KR-12 dimers showed a substantial increase in antimicrobial activity compared to monomer KR-12, irrespective of whether they were head-to-tail

25

cyclic or linear. Most dimers reached an activity level equivalent to, or sur-passing, LL-37. Out of all the dimers synthesised, cd4(Q5K, D9K) stood out as the most broad-spectrum antimicrobial lead, with 16-fold lower MIC against both P. aeruginosa and S. aureus and 8-fold lower MIC against C. albicans (Figure 7).

Figure 7. In the naming of the dimers, either the prefix ‘c’ or ‘l’ were used to indicate that the dimers were cyclic or linear, respectively. Peptides cd3 and cd4 were similar to cd2 except that the size of the linker is three and four residues respectively. The term ‘retro’ was included in the name to highlight those analogues with two mono-meric units adjoined in retro direction to each other. MIC values of LL-37 are E. coli: 0.312 µM; P. aeruginosa: 0.625; S. aureus: 1.25; C. albicans 1.25. For cd4(Q5K, D9K), values are 0.625, 0.625, 0.625, and 1.25 µM, respectively.

Dimers disrupt lipid membranes with enhanced efficacy Escherichia coli and Saccharomyces cerevisiae membrane lipid extracts were used to study membrane disruption of KR-12 dimers. The membrane lipids were reconstituted into large unilamellar vesicles, trapping a self-quenching fluorescent marker (Figure 8). The dimers were about 50-fold more potent at bacterial membrane permeabilisation than the monomeric peptides. The cd4dimers and LL-37 exhibited almost identical activity on liposomes of fungal and bacterial lipid extracts. Peptide cd4(Q5K, D9K) stood out as the most active one from membrane studies with an EC50 of 0.3 µM.

26

Figure 8. Peptide induced liposome leakage. Bacterial liposomes based on E. coli polar lipid extract (A and B), and fungal liposomes of S. cerevisiae polar lipid extract and ergosterol (7:3) C), were monitored for fluorescent marker leakage over 45 minutes of incubation with peptides. cd4(Q5K, D9K) stood out most active peptide on membranes.

Dimerisation leads to cancer cell toxicity In paper I monomer KR-12 lacked cytotoxicity against the human lymphoma cell-line with only 13% loss of cell viability at highest tested concentration of 80 µM. In comparison, LL-37 was cytotoxic to lymphoma cells with an IC50 of 10 µM (Figure 9). Dimerisation resulted in cancer cell toxicity (1-5 µM) irrespective of whether the dimers were linear or cyclic. Notably, cd4(Q5K, D9K) showed the highest cytotoxicity (1 µM of IC50) of all the analogues tested in fluorometric cytotoxicity assay (FMCA).

In the assay for hemolysis, the dimers were 7-10-fold less toxic to healthy erythrocytes compared to lymphoma cells. Both KR-12 and LL-37 showed no hemolysis at the highest concentrations tested (>80 µM).

27

Figure 9. Cytotoxicity on a human lymphoma cell line (U-937 GTB), the cell viabil-ity derives from the hydrolysis of the probe fluorescein diacetate by esterases in cells with intact plasma membranes. cd4(Q5K, D9K) showed the highest cytotoxicity (1 µM of IC50).

Backbone cyclisation improves proteolytic stability The proteolytic stability of the KR-12 dimers were evaluated in human serum followed by LCMS analysis (Figure 10). The KR-12 cyclic dimers were stable even after six hours of incubating in serum and around 50% of the cyclic peptides, cd4 and cd4(Q5K, D9K) were remaining. LL-37 and ld4 peptides gave weak MS signals at time 0, indicating that there was little free peptide in the sample either due to the peptides being rapidly degraded or binding to serum compo-nents (Wang et al. 1998), e.g. albumin or other proteins.

Figure 10. Serum stability profile of selected KR-12 dimers. KR-12 and retro-KR-12 breakdown in 10 minutes and cyclic dimers were stable even after 6 hrs.

28

KR-12 dimers show high a-helical content NMR spectroscopy was used to determine structures of KR-12 monomers and cyclic peptides. In membrane simulated environment KR-12 and retro-KR-12 adopted well-defined α-helical conformations. The data quality for cyclic di-mers cd4, cd4(Q5K, D9K), and two retro-cd4, was poor with limited disper-sion and broad lines, indicative of disordered conformations. Addition of SDS and lyso-phosphatidylglycerol micelles did not improve the quality of the NMR spectra either.

The circular dichroism (CD) spectroscopy measurements for the peptides were investigated in 10mM Tris-buffer, pH 7.4 or with lyso-phosphatidylglyc-erol micelles (representing the negatively charged microbial membrane envi-ronment) (Figure 11). All the peptide adopted a clear random coil confirma-tion in Tris-buffer. In contrast, the CD spectra of all the dimers show clear a-helical structure and 49-65% α-helical content (Figure 11).

Figure 11.Peptide general structure from circular dichroism. A) peptides in Tris buffer only and B) peptides in 16:0 lyso-phosphatidylglycerol (lyso-PG) micelles at a 1:1 peptide-to-micelle ratio (eq. to a 1:200 peptide-to-phospholipid ratio). The amount of α-helical content. C) is determined from the signal intensity at 225 nm using a poly Lys reference. Lys modifications on cd4(Q5K, D9K) have negative effect on helicity. cd4 peptide stood out as the most helical dimers (65% a-helical content).

From paper II, it is concluded that backbone cyclisation of KR-12 dimers substantially improved biological stability of monomer KR-12, and dimerisa-tion facilitated a higher local concentration of the peptide to mediate enhanced antimicrobial activity. Cyclic dimers adopt alpha helical structures in mem-brane environment and increased cationicity was detrimental to hemolysis.

tajmu645

Cross-Out

29

Paper III Design Strategy The cyclic dimers designed in paper II were unstructured in solution as judged by NMR, and the most active cyclic dimers were hemolytic. In paper III, the backbone cyclic analogues were designed by connecting two KR-12(Q5K, D9A) monomer via four residue long linkers (Figure 1). To stabilise the peptide chains, the two KR-12(Q5K, D9A) subunits were cross-linked via a disulfide bond. The cysteine residues were introduced by replacing Q5 from unit 1 and D9 from unit 2 with cysteines. A proline residue was incorporated in the 4-residues linker region to induce helix-turn-helix motif with KR-12(Q5K, D9A) subunits in uninterrupted helical confirmation (cd4-PP and cd4-CCPP). All the peptides intended for cyclisation were synthesised using the protocol as mentioned in paper II, and the peptides containing two cyste-ines were oxidised in 0.1 M ammonium bicarbonate, pH 8.2 buffer.


Cyclic dimers were more potent as compared to the LL-37 The cyclic dimers were particularly active against S. aureus, and C. albicans, surpassing, LL-37 (Figure 12).

Figure 12. Relative MICs of the KR-12 dimers. MIC was derived from three inde-pendent triplicate experiments as compared to KR-12. Each organism is shown with different color pattern. MIC values of LL-37 are E. coli: 0.312 µM; P. aeruginosa: 0.625; S. aureus: 1.25; C. albicans 1.25. For cd4-PP, values are 0.312, 0.625, 0.312, and 0.312 µM, respectively.

The disulfide bond introduced had no negative impact on activity in cross-linked di-mers, i.e. cd4-CC and cd4-CCPP. The cyclic dimers from the current study were as potent as cd4(Q5K, D9K), the most potent analogue from previous study, carrying a charge of +14, paper II.

30

Influence of physiological salt concentrations on antimicrobial activity The antimicrobial activity of all the peptides were evaluated in the presence of monovalent, divalent and trivalent salt (NaCl 150mM, NH4Cl 6µM, KCl 4.5µM, CaCl2 2.5mM, MgCl2 1mM, FeCl3 4µM i.e. physiological ionic strength). The presence of salts did not affect the antimicrobial activities of the cyclic dimers as well as LL-37 peptide against Gram-negative bacteria, E. coli and P. aeruginosa to any major extent.

Most pronounced loss in activity, 2-to-16-fold, for the cyclic peptides were observed against, Gram-positive bacteria S. aureus and the fungal strain, C. albicans, as compared to the MIC from two-step microdilution assay.

Cyclic analogues retained antimicrobial activity in the presence of human serum One of the limitations to the use of AMPs, including LL-37, for local and in-vivo applications are due to their inactivation by proteases (human or bacterial origin) or binding with serum components (Wang et al. 1998). The antimicro-bial activities of all the peptides were evaluated in the presence of diluted se-rum. The linear peptides including LL-37 completely became inactive at the highest tested concentration (MIC >40 µM) against all bacterial strains. The cyclic peptides were still active with a slight drop in activity 8-to-16-fold MIC against Gram-negative bacteria and 32-to-64-fold MIC against S. aureus as compared to MIC from two-step microdilution assay (Table 1).

Table 1. Minimum Inhibitory Concentrations (MIC) of cyclic dimers in the presence of diluted serum (25%) and Tris-buffer (two-step microdilution assay)

MIC (μM) Peptides 25% serum Tris-buffer (10mM)

E C P A S A CA E C P A S A C A

LL-37 >40 >40 >40 NP 0.312 0.625 1.25 2.5 KR-12 >40 >40 >40 NP 2.5 10 10 10 Q5K, D9A >40 >40 >40 NP 1.25 1.25 1.25 1.25 cd4-PP 5 2.5 20 NP 0.312 0.625 0.312 0.312 cd4-CC 5 2.5 20 NP 0.312 0.625 0.625 0.625 cd4-CCPP 2.5 1.25 10 NP 0.625 0.625 0.312 0.625 MIC were determined as the lowest concentration of the peptide inhibited microbial growth. E. coli (EC), P. aeruginosa (PA), S. aureus (SA) C. albicans (CA), Not per-formed (NP).

The loss in activity for the linear peptides (KR-12 and Q5K, D9A) could be due to instability, as we had shown in paper I that KR-12 and its mutant breakdown within minutes in serum. Previously, it has been reported that LL-37 loss antimicrobial activity when tested in the presence of human serum by binding to serum components, i.e., albumin or apolipoprotein A-1 (Wang et

31

al. 1998). Likewise, it is concluded that the drop-in activity of LL-37 and cy-clic dimers was due to binding to serum components as no cleavage products were identified during stability assay.

Proteolytic stability in human serum and S. aureus protease The proteolytic stability of the cyclic dimers was also evaluated in diluted human serum and in bacterial protease, aureolysin (Staphylococcus au-reus metalloproteinase). The recovery rate of LL-37 and cyclic dimers were low in serum as compared to the peptides in PBS. This phenomenon was also observed for linear dimers and LL-37 in paper II. In serum assay during the precipitation step, using trifluoracetic acid, the cyclic dimers get precipitated with the serum proteins (personal observation). with the serum proteins (personal observation).

The LC-MS analysis of LL-37 incubation mixture with aureolysin indi-cated that peptide was efficiently cleaved by the enzyme with in first few minutes (Figure 13). LL-37 peptide degradation apparently occurred by sim-ultaneous hydrolysis of the Leu31-Val32 peptide bond just after incubation.The cyclic dimer, cd4-CCPP, was intact up to 8 hrs.

Figure 13. Proteolytic stability of cyclic peptides in relevant biological conditions. Cleavage sites of linear peptides (LL-37, KR-12, Q5K, D9A) by aureolysin. A), Solid arrows indicate cleavage site identified in this study and dotted arrow represent cleav-age sites reported earlier by aureolysin. B) Stability of peptides in presence of aure-olysin. Kalata B7, cyclotide was stable up to 24hrs.

Antimicrobial activity against resistant strains As mentioned in the introduction of this thesis, severe infections caused by multidrug-resistant bacteria have become a major challenge for conventional antibiotic treatments. Hence, to probe the efficacy of cyclic dimers against resistant bacteria, their antimicrobial activity against Salmonella enterica serovar Typhimurium (wild type DA6192) and LL-37-resistant mutants (DN-22427-waaY and DN-23307-phoP) and E. coli MG1655 (wild type) and cy-clotides-resistant mutants (DA54114 and DA57105) were tested.

32

All cyclic peptides exhibited higher (4-to-8-fold) antimicrobial activity against all the tested S. typhimurium strains than LL-37 (Table 2), while 4-to-16 fold more active as compared to LL-37 against cyclotides resistant E. coli mutants.

Table 2. Minimum Inhibitory Concentrations (MICs) of cyclic dimers against resistant strains Peptides MIC (μM)

S. typhimurium E. coliWT waaY phoP WT DA54114 DA57107

LL-37 2.5 5 10 1 1 2.25 KR-12 20 20 20 1 1 1 Q5K, D9A 10 10 10 1 1 1 cd4-PP 1.25 1.25 1.25 0.37 0.25 0.25 cd4-CCPP 2.5 1.25 1.25 0.37 0.25 0.25 cd4-CC 1.25 2.5 1.25 NP NP NP

The S. typhimurium strains used in this study were LL-37-resistant and had mutations in the genes waaY and phoP, which are connected with LPS modi-fications (Lofton et al. 2013b). In my hands, the MICs obtained for LL-37 peptide against S. typhimurium wild type and resistant mutants were almost similar with a marginal difference 2-4-fold concentration. This low difference in MIC could be due to the assay method used, as microdilution assay (MIC) is not sensitive enough to detect such a small mutation (a time kill assay is more sensitive enough detect the difference).

The E. coli MG1655 strains used in this study were resistant to cyclotides (cycloviolacin O19 and kalata B7). The E. coli DA54114 was CyO19 resistant and had a mutation in gene prc, which has a role in osmotic stress. The kB7-DA57107 resistant mutant used in this study was un-characterised, during whole genome sequencing and it was not possible to detect mutation (Malik 2017).

Quantifying the membrane disrupting mechanism Again, the liposome assay was used to determine the membrane permeabili-sation. Here, E. coli polar lipid extract and a POPC/cholesterol system was used in the form of large unilamellar liposomes to mimic bacterial or human cell membranes, respectively.

33

Figure 14. Liposome leakage levels as a function of peptide concentration. The lipo-some systems used were: A) E. coli polar lipid extract as a generic bacterial composi-tion and B) POPC:cholesterol (3:2 molar ratio) as a simplified human model. Com-pared to LL-37, cyclic dimers showed higher selectivity for the E. coli membrane.

As demonstrated in (Figure 14), the relative level of membrane leakage within each individual lipid system mirror the order exhibited in the two-step micro-dilution assay (i.e. the AMPs adapted assay method, Figure 12) in regards to the antimicrobial activities for these peptides. Even though cd4-PP stood outto be the most potent dimer on the E. coli liposomes whereas the two cross-linked counterparts have the higher target selectivity. In comparison, the first-generation cyclic dimers were less membrane active on E. coli (cd4).

Cross-linked cyclic dimers possess potent anti-biofilm activity The ability of the peptides to inhibit the formation of bacterial biofilm was also investigated against S. aureus 25923 biofilms. In the pre-exposure para-digm, cd4-PP at 100 μM showed significant effect by completely inhibiting viability of biofilms and also significantly reduced biomass production. On the other hand, LL-37 displayed biofilm-inhibiting effects over the activity threshold (50% inhibition) but such effect was not coupled to a significant reduction on the biofilm biomass (Skogman, Vuorela and Fallarero 2012).

Structural characterisation NMR spectroscopy was also recorded for the cyclic dimers in membrane mim-icking environment (Lyso-PG), however the spectra showed limited disper-sion and broad lines for all peptides. Introduction of a disulfide bond and pro-line residue at linker region did not stabilise KR-12 units.

CD analysis showed that cyclic dimer, cd4-PP, has the highest helicity as compared to other peptides, surpassing LL-37. In cd4-CC and cd4-CCPP cy-clic dimers, containing cysteine residues had a detrimental effect on the helic-ity (Figure 15).

34

Figure 15. The CD spectra of parent LL-37 and KR12(Q5K, D9A) cyclic dimers. cd4-PP showed more a-helical content (78%) in lyso-PG as compared to LL-37 (72%). A) peptides in Tris buffer only and B) peptides in lyso-phosphatidylglycerol. C) theamount of α-helical content.

It is concluded from paper III that multiple parameters need to be assessed in the design and development of AMPs-based antimicrobials for local as well as systemic applications. The cyclic dimers showed potent antimicrobial ac-tivities in biological relevant conditions proven to be challenging to linear counter parts and were less-hemolytic compared to the dimers from first gen-erations cyclic dimer in paper II.

35

Enzymatic cyclisation (IV-V)

Background

The macrocycles designed in paper II-III showed great potential as therapeu-tics with increase proteolytic stability and potent activity. Likewise, backbone cyclic disulfide-rich macrocycles such as cyclotides, and the PawS derived peptides sunflower trypsin inhibitor (SFTI-1), are particularly promising clas-ses of compounds for grafting bioactive epitopes onto their scaffolds (Chan et al. 2011, Burman et al. 2014, Franke, Mylne and Rosengren 2018, Craik and Du 2017).

Despite their pharmaceutical potential, large-scale synthesis with better yields is a significant challenge. I used NCL to produce cyclic dimers, but requirement of N-terminal Cys, unstable thioester intermediate during cyclisa-tion, and the lack of scalability, often associated with low yields, limit chem-ical-based cyclisation on larger scale (Gunasekera et al. 2013). Thus, more efficient, safe, environment friendly and broad-spectrum methods are highly desired which can be scaled up to industrial settings.

One approach is to turn to nature to look which type of chemistry nature is using to produce head-to-tail cyclic peptides. How can we utilise these en-zymes to produce back-bone cyclic peptides in good yields in a laboratory or industrial setting? Is there a way to avoid repeated extraction of enzymes and reutilise the enzyme in an efficient way?

In recent years, a handful of naturally occurring enzymes have been re-ported for the production of macrocycles. Of these, one of the best character-ised enzymes is the serine protease PatG, which cyclises the bacterial cyano-bactins (Koehnke et al. 2014). In plants, another serine protease, peptide cyclase 1 PCY, cyclises the segetalins; cyclic peptides from Caryllophylaceae (Barber et al. 2013). Prolyl oligopeptidase (POPB), a cyclase from the familyof serine proteases, isolated from posionous mashrooms Amanita phalloides, catalyses the cyclisation of amatoxins (Luo et al. 2014). More recently, aspar-aginyl endopeptidases (AEPs) such as butelase-1 (Nguyen et al. 2014a) and OaAEP1b (Harris et al. 2015) were reported. Engineered subtilisin such aspeptiligase variants have been shown to produce cyclic peptides (Schmidt et al. 2017). The serine-protease trypsin were reported to cyclise SFTI-1 (Marx et al. 2003) and MCoTI I/II, the trypsin-inhibitor cyclotides (Thongyoo, Tate and Leatherbarrow 2006). Sortase A (SrtA), a transpeptidase enzyme from S.

36

aureus has been widely used for cyclic peptide synthesis but leaves a footprint in the form of a five-residue sequence LPTXT at the ligation site (Wu, Guo and Guo 2011, Jia et al. 2014, Stanger et al. 2014). In addition, various recom-binant expression systems have also been reported for cyclic peptide produc-tion (Aboye and Camarero 2012) (Yilong Li, Tao Bi and Camarero 2015).

In paper IV-V, I addressed some of above questions to develop enzymatic macrocyclisation as a way to generate cyclic peptides.

Paper IV Cyanobactin macrocyclase PatGmac PatG is a subtilisin-like macrocyclase, identiffied from patellamide pathway. Patellamide are small cyclic peptides, produced by the cyanobacteria Prochloron didemni, a symbiont of the sea squirt Lissoclinum patella (Koehnke et al. 2012). PatGmac requires a four residues (-AYDG) motif at C-terminal. PatGmac natural substrates are 8-11 residues in length.

Cyclisation Strategy Can PatGmac also cyclise the larger, plant derived, cyclic peptides in their native form, such as the 14 residues long SFTI-1, or the 29 residues long kalata B1 (kB1)? To address the above hypothesis, SFTI and kB1 precusors were synthesised contaning -AYDG tag and tested PatG macrocyclisation (Figure 16).

Figure 16. Macrocyclisation by PatGmac. A) PatGmac requries a proline,proceeded by a recongnitation tag AYDG at the C-terminal for macrocyclisation, B) synthesised precursor peptides with ligation points.

37


PatGmac cyclisation of 14 residues long SFTI-peptides In SFTI-peptides, two ligation points, one between Pro13-Asp14 in the cy-clisation loop and the other between Pro8-Pro9 in the trypsin binding loop were tested for PatG macrocyclisation.

Using PatGmac native SFTIP13-D14 and a variant, SFTIP13-D14G, were cyclised efficiently. LC-MS analysis confirmed to contain both cyclic and hydrolysed products. The co-elution pattern of the enzyme cyclised SFTIP13-D14 was similar to the native peptide isolated from sunflower seeds (Figure 17).

Figure 17. LC-MS characterisation of PatGmac cyclised SFTI-1 peptides. A) Enzyme cyclised SFTIP13-D14 B) Native SFTI-1 isolated from sunflower seeds. C) Co-injection of enzyme cyclised SFTIP13-D14 and native SFTI-1. D) Cyclised SFTIP13-D14G (letter ‘C’ indicates cyclic peak and ‘L’ linear).

PatGmac cyclised SFTI-peptides were reduced, alkylated and cleaved with trypsin and cyclic backbone was unequivocally confirmed by MS/MS frag-mentation (Figure 18). PatGmac macrocyclisation appeared to be better for ligation point in the cyclisation loop and the cyclic SFTIP13-D14 was obtained in a reasonable yield of 42%. When a non-bulky glycine residue was incor-porated at the N-terminal in SFTIP13-D14G, a slight increase in yield (46%) was achieved relative to native sequence which has Asp, a bulky residue.

38

Figure 18. MS/MS spectra for backbone cyclised SFTI-peptides. A) SFTIP13-G14 frag-ment (SIPPICFPDGR) with an observed monoisotopic mass of 1326.66 Da spans the amide bond at the ligation site between P13 and D14. B) SFTIP13-D14G fragment (SIP-PICFPGGR) with an observed monoisotopic mass of 1268.62 Da spans the amide bond at the ligation site between P13 and D14G.

PatGmac cyclisation of 29 residues long kalata B1(kB1)-peptides In kB1-peptides, two ligation points, one between Pro28 and Val29 in loop 6 and the other between Pro20 and Val21 in loop 5 were chosen to cyclise with PatGmac.

LC-MS analysis of PatGmac cyclised kB-peptides confirmed the removal of recognition tag (-AYDG) and a loss of water molecule, which was an indi-cation for head-to-tail cyclisation. However, neither of the PatGmac cyclised kB1-peptides did co-elute with native kB1 peptide isolated from Oldenlandia affinis plant, indicating that the disulfide connectivity of these peptides was non-native despite the formation of cyclic backbone (Figure 19). When kB1-peptides were alkylated, the peptides co-eluted with alkylated native kB1, con-firming similar structure.

Figure 19. PatGmac cyclised kalata B1-peptides. A) Enzyme cyclised kB1P3-V4 B) Co-injection C) extracted kB1P3-V4 and native kB1 co-injection. D) Native kB1 peptide isolated from O. affinis. (letter ‘C´’ indicates PatG cyclised kB1 ‘C’ native kB1, elute at different time)

39

PatGmac cyclisation reactions were carried out in reaction buffer containing 500 mM NaCl, 10 mM bicine pH 7.5 and 5% DMSO. Kalata B1 contains six cystines, which form three disulfide bonds with a connectivity pattern of I-IV, II-V, III-VI. Previously, Aboye et al. reported the use of DMSO containingbuffer for the oxidative folding of kB1 and that it was not possible to achievenative disulfide fold (Aboye et al. 2011). Thus, it appears that the buffer con-ditions used to cyclise kB1 using PatGmac were non-optimal to achieve nativedisulfide connectivity in kB1-peptides.

In contrast to SFTI-peptides, the PatGmac macrocyclisation was not effi-cient for kB1-peptides presumably due to its large size and multiple disulfide bonds. When Jia et al. used SrtA to cyclise reduced kB1, it was not possible to achieve correctly folded kB1 and cyclisation yield was also low (Jia et al. 2014). However, when oxidised peptide was used for cyclisation correctly folded peptide with higher yields was produced (Jia et al. 2014).

From paper IV, it is concluded that macrocyclisation of disulfide rich pep-tides by PatGmac depends on the peptide length, nature of the residue at the ligation site, number of the disulfide bonds and the position of the ligation site in the sequence.

40

Paper V

Butelase 1

Butelase 1 is a naturally occurring asparagine/aspartate (Asx)-specific ligase, isolated from cyclotides producing tropical plant Clitoria ternatea of the Fa-baceae family (Cao et al. 2015, Nguyen et al. 2016). Butelase 1 recognizes a specific tripeptide recognition motif, N-HV or D-HV at the C-terminus and accepts any amino acid (except P, D and E) followed by a hydrophobic residue at the N-terminus (I, L, V, and C) (Nguyen et al. 2015). During biocatalysis, the recognition motif, HV dipeptide at the P1´ and P2´ positions is cleaved, forming an acyl-enzyme intermediate, then followed by ligation of Asx through the nucleophilic attack of the N-terminal amino acid to form cyclic peptide.

Use of butelase as a cyclisation tool on large scale is still restricted because of low availability. The extraction protocol reported by Tam et al. to isolate pure butelase from butterfly pea is not straight forward due to multiple steps and its utility in laboratory or industrial settings is limited due to low yields and high extraction cost. An alternative approach is to express butelase in bi-ological system to lower the production cost and for continuous supply of it. Attempts have been made to express butelase in E. coli system (James et al. 2019), but its availability is still restricted.

Cyclisation Strategy

In paper V, different substrates were synthesised based on SFTID14N-G1 andKR-12 sequence (Figure 20A) and cyclisation efficiency was monitored using semi-pure butelase 1. To reduce the enzyme isolation cost, semi-pure butelase 1 was immobilised on NHS column and cyclisation efficiency was monitored (Figure 20B and C).

41

Figure 20. Schematic illustration of butelase 1-mediated cyclisation. A) synthesisedprecursor peptides with ligation sites B) the typical butelase-mediated cyclisation of a peptide carrying a C-terminal Asn-His-Val (NHV) recognition motif C) butelase 1 immobilised column and interaction with the substrate.


Isolation of semi-pure butelase 1 Dry seeds of C. ternatea were used to isolate butelase 1. A series of extraction and precipitation procedures were employed, followed by the separation of crude extract through anion exchange chromatography which led to the col-lection of semi-pure enzyme fractions. The enzyme containing fractions from ion exchange column were identified by overnight incubation with SFTID14-G1 substrate and monitoring the formation of cyclic product on LC-MS.

Substrate to enzyme ratio in solution Two sets of reactions were conducted at different enzyme volumes involving two different concentration, 1mg/ml and 0.1mg/ml. A fixed volume (10 µL) and concentration (1mg/ml) of SFTI-NHV were maintained. 10 µL of unpuri-fied Butelase 1 at 0.1mg/ml appears to be the minimum volume and concen-tration required to produce cyclic product of SFTI-NHV.

Cyclisation of SFTID14-G1 in solution SFTID14-G1 peptide was incubated in semi-pure butelase fraction and reac-tion mixture was analysed on LC-MS (Figure 21). The LC-MS chromatogram shows cyclic, linear as well as precursor peptide masses, an indication for in-complete reaction.

42

Figure 21. Butelase-mediated cyclisation of SFTID14-G1. Enzyme processed SFTID14-G1 (mass spectra of the cyclic peak (marked by*).

The LC-MS data indicates that enzyme had processed the KR-12monomer-1 to a hydrolysed product with a loss of -HV as well as to a cyclic product (very small peak). The butelase mediated cyclisation appeared better for KR-12monomer-2 as a distinct peak was observed for cyclic peptide (Figure 22).

Figure 22. Butelase-mediated cyclisation of KR-12 peptides (A) KR-12monomer-1 precursor (B) KR-12monomer-1 cyclised (C) KR-12monomer-2 precursor (D) KR-12monomer-2 cyclised, mass spectra of the cyclic peak (marked by*)

In dimer peptides butelase1 cyclisation appeared to be better for KR-12 dimer-2 with a Gly and Ile in the N-terminal and a very small cyclic peak was ob-served. However, the major part of the mixture contained hydrolysed mass. Interestingly, neither cyclic nor hydrolysed mass for KR-12 dimer-1 could be identified.

Production of cyclic peptides using butelase-immobilised column Semi-pure butelase 1 enzyme (1-6mg) was immobilised on HiTrap NHS-ac-tivated column with the loading capacity of 5ml. The enzyme-coupled column was washed with different buffers.

After immobilisation of the enzyme, cyclisation reactions were conducted using different substrates. First, cyclisation activity was monitored without

43

circulation of substrate. Secondly, the substrate was recirculated through the column over a 24 hrs period using a medium pressure pump.

SFTID14-G1 and KR-12monomer-2 (approx. 0.25 mg each) were loaded on the column and cyclisation was performed sequentially. Samples were col-lected at different times points and analysed on LC-UV-MS to confirm the cyclisation efficiency over time.

In SFTID14-G1 peptide, the cyclic mass appeared from 6 hrs and over time the intensity of the cyclic peak increased (Figure 23). The 24hrs recirculated sample showed that the immobilised enzyme had processed SFTID14-G1 pre-cursor completely to the cyclic product. It was not possible to integrate UV data to confirm the percentage yield, as the concentration of the sample was not enough to get a good signal and in-over time samples both the cyclic as well as acyclic products co-eluted. When SFTID14-G1 sample at stationary position was analysed after 24 hrs, the cyclic as well as acyclic masses were observed.

Figure 23. LC-UV-MS analysis of butelase immobilised SFTID14-G1. A) recircu-lated SFTID14-G1 precursor peptide at 0hr B) reaction mixture 6 hrs, C) cyclic SFTI, 24hrs, the cyclisation efficiency increased overtime, and 24hrs sample revealed cyclic mass, but it was not possible to quantify as intensity was weak, mass spectra of the cyclic peaks (marked by*).

The analysis of KR-12monomer-2 showed that the cyclic product appeared from 8hrs and over time the intensity increased. The LCMS analysis of the 24hrs recirculated sample showed both cyclic and acyclic masses (Figure 24).

44

Figure 24. LC-UV-MS analysis of butelase immobilised-cyclised KR12monomer-2. (A) recirculated KR12monomer-2 at 0hr (B) reaction mixture 6 hrs (C) reaction mix-ture 24hrs, the cyclisation efficiency increased overtime, and 24hrs sample revealed cyclic and acyclic masses, but was not possible to separate and quantify cyclic fromacyclic mass and mass spectra of the cyclic peaks (marked by*).

From paper V, it is concluded that immobilised semi-pure butelase 1 was ac-tive for 2 weeks and macrocyclisation was achieved for two substrates of var-ying length and nature.

In the long run, the ultimate aim is to produce cyclic peptides using mac-rocyclisation enzymes, immobilised on a solid support, as a way to reduce peptide production cost.

The covalent immobilisation of butelase 1 on NHS column is a promising approach to produce head-to-tail cyclic as well as to ligate peptides and open up new perspectives for further applications.

45

Concluding Remarks

This thesis describes attempts to develop stable cyclic dimers based on the minimalised antimicrobial fragment of LL-37, named KR-12, and also ex-plores possibilities of producing these cyclic dimers using naturally occurring enzymes. In particular, my focus has been on the exploration of the potential to combine dimerisation, head-to-tail cyclisation and cross-linking, as an ef-fective strategy to impart desired properties to the protease-labile KR-12. Fur-thermore, the limits of PatG macrocyclase to cyclise disulfide rich peptides of varying length were demonstrated; and finally, semi pure butelase 1 was im-mobilized on column and shown to produce head-to-tail cyclic peptides over a period of two weeks.

In paper I, key factors to modulate the antimicrobial activity of KR-12 were explored. 1) Structure-activity relationship studies were performed on KR-12 to figure out the balance of hydrophobic to cationic residues. 2) Loca-tion of the specific amino acids in the three-dimensional structure of the KR-12 peptide and their contribution to the peptide helicity were explored. The amphiphilic nature of KR-12 and amino acid residues Gln5 and Asp9 were found to be significant for the antimicrobial activity. Besides that, KR-12 and analogues were found to be unstable in human serum in this study.

Therefore, in paper II the concept of backbone cyclisation and dimerisa-tion were investigated, as means to improve peptide stability and activity, in the design of a potent and stable antimicrobial peptide based on KR-12. These attempts resulted in substantially improved antimicrobial activity and biolog-ical stability in human serum as compared to linear counterparts. Some of the cyclic dimers also showed potent activity against lymphoma cell line. In ad-dition, the cyclic dimers were unstructured in solution judged by NMR. However, in CD studies the helical propensity of KR-12 was found to be maintained in the cyclic dimers in micelles.

In paper-III, a second generation of cyclic dimers were synthesised based on the positions identified in paper I i.e. KR-12 (Q5K, D9A), and in attempt to further stabilize dimer structures. For this, a proline residue was introduced in the linker region to induce “helix-turn-helix”-motif and helices were cross linked with a disulfide bond. Thereafter, the activity of the cross-linked cyclic analogues was compared with the parent antimicrobial peptide LL-37 under physiologically relevant conditions that are proven to be challenging for most antimicrobial peptides. The cyclic dimers maintained antibacterial activity under these conditions and showed

46

selectivity against Gram-negative bacteria in rich growth media. The cross-linked cyclic analogues were found to be non-hemolytic in this study.

Though the cyclic dimers designed in Paper II-III proved to be potent and stable enough to be a drug lead e.g. 32-fold more potency than monomer, ex-ceptional stability in serum and retaining of antimicrobial activity in biologi-cal conditions yet their large-scale synthesis remains quite challenging due to lack of scalability and low yield. So, alternative methods are required which can be scalable, efficient and safe. In paper IV and V enzymatic methods for cyclisation were explored. First, (cyanobactin macrocyclase enzyme PatGmac was employed to macrocyclise head-to-tail disulfide rich peptides of varying length in Paper IV. It was shown to cyclized 14 amino acid residue long SFTI-peptides were with 42-46% yield, but cyclisation was less effective for the longer peptides tried.

Similarly, in paper V semi-pure butelase 1 enzyme was used to produce cyclic peptides as an approach towards green chemistry and lower peptide production cost. Semi pure butelase was immobilised on NHS column and a wide range of substrates of varying length were recirculated. The immobilised butelase retained enzymatic activity over a period of 2-weeks and was used to successively cyclise different substrates. KR-12monomer-2 was cyclised us-ing butelase in solution form as well as immobilised column. However, bute-lase-mediated cyclisation was non-optimal for the dimer peptides. A small cy-clic peak for KR-12dimer-2 was observed in one trial, however, this needs to be further confirmed in the future studies. The concentration of precursor pep-tides re-circulated on butelase-immobilised column was too low to get a good signal on LC-UV-MS. If a concentrated enzyme is immobilised in NHS col-umn in the future, the cyclisation efficiently would be improved and mean time large quantity of substrate can be re-circulated.

The dimer peptides used for butelase-mediated cyclisation had proline and glycine residues in linker region. In the future, it would be worth to try bute-lase mediated cyclisation on cross-linked dimers.

47

Perspectives and Proposals

Naturally derived medicinal agents have been used as remedies for human diseases from centuries. The discovery of penicillin was the greatest invention in modern medicine history to treat infections. The major classes of the anti-biotic were discovered during the golden era between 1940-1960. Most of currently available antibiotics are derived from microorganisms. The genome mining strategy was also extensively used to discover new antibiotics. In the last four decades antibiotic research has mainly been focused on tweaking of traditional antibiotics discovered during golden era. The extensive misuse of antibiotics for human medicines and animal farming industry lead to antimi-crobial resistance. Bacteria inactive conventional antibiotics by different ways either by genetic mutations, enzymatic degradation, restrict antibiotic entry or pump out from the cell (Zaman et al. 2017). Ideally to develop a new antibi-otic, the chemistry needs to be novel with a new target. With an alarming sit-uation of antibiotic resistance, where last resort antibiotics are failing, it’s time now to turn to nature for the novel compounds in order to save human race.

Pharmacognosy is an inter-disciplinary science that explores bioactive sub-stances of natural origins. One of the focus points of research in the Pharma-cognosy group of Uppsala University is to identify organisms of interest, iso-late the active constituents and study the pharmacological properties. The re-search work in this thesis mainly covers all these areas of Pharmacognosy.

Antimicrobial peptides are nature’s antibiotics produced by all living or-ganisms with promising therapeutic potential targeting microbial membranes (Zasloff 2002). In this thesis, KR-12 was chosen as a template for peptide design because of its size, a-structure and its selective activity towards bacte-ria and its low hemolytic activity (HC50>80 µM) (Wang 2008, Mishra et al. 2013, Mishra and Wang 2017). The parent peptide of KR-12, i.e. LL-37, is known to be hemolytic (Ciornei et al. 2005). However, both LL-37 and its minimalised version KR-12 harbor a number of cleavage sites and easily de-grade in human serum and bacterial proteases (Strömstedt et al. 2009, Mishra et al. 2016). There are various strategies available to impart stability to a linear peptide and this issue was addressed by using backbone cyclisation as an ap-proach to prevent KR-12 degradation.

The cyclic dimers developed in our studies displayed extraordinary stabil-ity in human serum and bacterial proteases. In comparison to KR-12, which degrades within 10 minutes, the cyclic peptides were stable for up to 6 hours.

48

During serum stability assays in current settings, it was not possible to de-tect LL-37 or linear dimer ld1 in serum even at time 0, demonstrating their rapid degradation or binding to serum protein (Pasupuleti et al. 2009).

It is known that AMPs form oligomers in aqueous solution before binding to the bacterial membrane (Wimley 2010). The preformed dimers facilitated a higher local concentration of the peptide to mediate enhanced antimicrobial activity. The increased activity may be explained by the fact that dimerisation is a source of not only a two-fold increase of the number of active peptides in the system, but also to artificially increase the local peptide concentration on the membrane. The bactericidal and fungicidal properties of KR-12 and cyclic dimers were also confirmed and the killing effects were very rapid in buffer conditions while in serum were delayed.

The delicate balance of cationic charge vs hydrophobicity was evident in this thesis for cyclic dimers. The true picture could be viewed in physiological conditions. The most active dimer cd4 (Q5K, D9K) from paper II in the an-timicrobial, liposome leakage and cytotoxicity assays indicate the importance of cationic charge interaction with the negatively charged membranes. All di-mers displayed higher antimicrobial activity than KR-12 and they are also more potent in the liposome assay. This indicates that membrane permeabili-sation is the main mechanism responsible for bactericidal and fungicidal ac-tivity by the dimers though the internal targets cannot be ruled out.

From drug development point of view, AMPs have to withstand physiolog-ical conditions i.e. salts, pH and serum binding to exert its activity against microbes. It is known that human beta-defensin (hBD-1) failed to kill the P. aeruginosa at high NaCl concentration (Goldman et al. 1997). The parent pep-tide LL-37 has also been reported to loss activity at pH 6.8 or after binding to human serum component (Wang et al. 1998) (Abou Alaiwa et al. 2014). But the cyclic peptide remained active in physiological conditions tested. The par-ent peptide LL-37 including KR-12 lost antibacterial activity. The loss in ac-tivity in the presence of serum for linear peptides is associated with peptide stability and binding to serum proteins. Likewise, we speculate that the loss in activity of cyclic dimers were due to binding to serum components as we were not able to identify any cleavage product on LCMS analysis. The loss in ac-tivity against bacterial strain in the presence of divalent cation (Mg2+ and Ca2+) could be explained by binding of divalent cations with the bacterial mem-branes. As reported previously, divalent cations bridge lipopolysaccharides on the bacterial surfaces, and are very important for membrane integrity and serve as competitive inhibitors to cationic AMPs (Xie and Yang 2016).

The strategies used in this thesis i.e., back-bone cyclisation, dimerisation and cross-linking to design stable and potent KR-12-based cyclic dimers can be used for other peptides as well. In therapeutic context, cross-linking via a disulfide bond is prone to reducing agents found in cellular environment and could be improved in future by utilising posttranslational modifications like thioether bridges (Moll et al. 2009, Wang 2012). Though the cyclic dimers

49

showed high alpha helical content in micelles and their a room to be improved in future studies by incorporating helix-inducing or helix-stabilisation modi-fications (Moiola, Memeo and Quadrelli 2019).

The enzymatic cyclisation strategies used in this thesis may offer alterna-tive methods for head-to-tail cyclisation of biologically active peptides to in-crease proteolytic stability and conformational rigidity. The catalytic effi-ciency of PatGmac is slow in vitro and is not viable to produce cyclic peptides on large scale (Koehnke et al. 2012). Mutagenesis of PatGmac, might enhance its ability to macrocyclise longer sequences, a new variant of PatGmac has been reported to catalyse a diverse set of substrates (Oueis et al. 2017). Most likely, there is a room to address the macrocyclisation efficiency of PatGmac to get better yield for kB1 like peptides with multiple disulfide bonds by mod-ifying the reaction conditions or employing oxidation prior to cyclisation ap-proach. In the long run covalent immobilisation of butelase like enzyme is a promising approach to produce macrocycles in industrial settings to reduce production cost, a step towards green chemistry.

Antimicrobial peptides present significant potential as antimicrobial agents for addressing the growing burden of drug resistant infections. As the designed cyclic dimers showed potent activity against LL-37 resistant S. typhimurium strains (Lofton et al. 2013a) and Cyclotide resistant E. coli MG1655 strains (Malik 2017). It is possible to envision potential applications of the potent and stable dimers to treat infections associated with resistant strains alone or in combination with conventional antibiotics. LL-37 is currently in phase IIa clinical trial in patients with chronic venous leg ulcers (Gronberg et al. 2014). Likewise, cyclic dimers can also be employed for topical treatment of chronic wounds and also for coating the dressings of wounds. Furthermore, the high prevalence of biofilm infections associated with surgical devices (catheters) and implants have created a demand for stable antimicrobials indwelling de-vice coatings. Thus, the cyclic dimers developed in this thesis may offer po-tential applications to coat surgical devices and implants to prevent biofilm formation.

However, these antimicrobial peptides are part of our innate immunity. If pathogens develop resistance against these peptides, it can cause lethal conse-quences (Andersson, Hughes and Kubicek-Sutherland 2016). Therefore, we need to be very careful while managing drugs based on these natural peptides unlike the case of conventional antibiotics. At peak of the golden era of cur-rently available antibiotics, an American epidemiologist William Stewart overlooking the future antibiotic resistance said in 1967 that “The time has come to close the book on infectious diseases”. Time has proved him wrong as pathogens developed resistance against the conventional drugs. So, we have to be cautious before repeating the same mistake.

At the start of this thesis, the aims were to design stable and potent back-bone cyclic dimers based on truncated antimicrobial region of LL-37 i.e. KR-12 and explore the posttranslational machinery to create them. During this

50

thesis work, key positions for increased antimicrobial activity on KR-12 were identified, the cyclic dimers were synthesised to increase stability and were also evaluated for efficacy and enzymatic cyclisations were explored to syn-thesize these cyclic dimers.

51

Popular Science Summary

The misuse of conventional antibiotics, since their discovery in the 1920s, has led to the emergence of multi-drug resistant pathogens as a serious global health problem. If no action is taken against this, it can cause 10 million deaths each year by 2050. So, there is an urgent need for new anti-infective agents with novel chemistries. One way to discover such new molecules is to turn to nature. Nature has been a rich source for drugs with unique properties and majority of the antibiotics available in the market nowadays are derived from nature.

One of nature’s untapped sources of anti-infective agents are Antimicrobial Peptides (AMPs). All living organisms, from unicellular to multicellular, pro-duce these molecules for defense purposes. We, as humans, also produce such molecules to fight pathogens. Majority of these AMPs kill bacteria by binding to their membranes though a few have internal targets. These peptides are ac-tually made up of amino acids as building blocks. These amino acids carry positive, negative or neutral charges which in turn render an overall charge to these AMPs. And, most of the AMPs have a net positive charge while a few have negative charge. The mode of action of these positively charged AMPs is that they bind to the bacterial membrane due to the opposite negative charge on the membrane surface. The bacteria then die due to lysis of the membranes. This specific interaction, with the bacterial membrane and subsequent death of bacteria by membrane disruption, is due to the differences in the lipid com-position of the prokaryotic bacterial. Eukaryotic mammalian cell membranes. Eukaryotic mammalian membranes are rich in zwitterionic phospholipids with net neutral charge unlike the negatively charged lipids in prokaryotes.

The AMPs have a number of potential advantages as future therapeutics. In addition to their broad spectrum of activity and rapid killing of microbes, they frequently neutralise endotoxins released by pathogens and also have the ability to modulate the host immune response to infection. A high number of promising AMPs have been discovered from nature with potent activities but translation of these molecules into clinics has been hampered due to their pro-teolytic stability, high cost of production and loss of activity in vivo.

A number of human based AMPs have been characterised and among these, the only peptide from cathelicidin polypeptide family, LL-37, is a potential molecule to develop into drug leads. The LL-37 peptide and its truncated ver-sion are currently in clinical trials for the treatment of middle ear infections

52

and other chronic infections related to burn wounds. Besides direct antimicro-bial activity, LL-37 plays a major role in regenerating blood vessels, wound healing, chemotaxis, signal transduction and immune modulation. Hence, there is great interest in developing novel peptides therapeutic based on LL-37.

In this thesis, I designed cyclic dimers based on minimalised antimicrobial region of LL-37, which is called KR-12. This thesis is mainly based on the concepts of peptide stabilisation by backbone cyclisation and cross-linking via a disulfide bond, an idea originating from nature, using chemical synthesis and/or cyclisation enzymes.

First, Alanine/Lysine scans on KR-12 peptide were conducted to find out the key positions important for the antimicrobial activity. Based on those scans, double/ triple analogues were designed to see the additive or synergistic effect on human pathogens. The identified key positions to increase antimi-crobial activity of KR-12 were Q5 and D9. The overall charge of the peptide sequence seems to be very important for activity as when Arginine or Lysine amino acid was replaced by Alanine, a reduction in the activity was noticed. Alanine substitutions of large hydrophobic residues in the hydrophobic face also reduced activity: notably at Ile7, Phe10 and Leu11. KR-12 and analogues were not cytotoxic but they breakdown within 10 minutes in human serum. It is thus concluded from these results that multiple parameters needs to be as-sessed and optimized to order to design AMPs based drug lead.

In order to improve antimicrobial potency and biological stability of KR-12, a series of back-bone cyclic KR-12 dimers were synthesised by joining two KR-12 units. The KR-12 dimers irrespective of whether they were head-to-tail cyclised or linear showed a substantial increase in antimicrobial activity compared to KR-12 against tested bacterial strains. The dimers were about 50-fold more potent at bacterial membrane permeabilisation than the monomeric peptides. The KR-12 cyclic dimers were stable even after 6 hours of incubat-ing in serum and around 50% of the cyclic peptides, cd4 and cd4 (Q5K, D9K) were remaining.

The problems identified with the KR-12 dimers were that they were un-structured and haemolytic. To overcome this issue, new dimers were designed by connecting two Q5K, D9A units via a four-residue linker peptide. A proline residue was introduced in the 4-residue linker region to induce turn-helix-turn. Two analogues (cd4-CC and cd4-CCPP) were cross-linked via a disulfide bond by replacing Q5 from unit 1 and D9 from unit 2. The cross-linked cyclic analogues showed potent activity in two-step microdilution assay. Cyclic di-mers were active in physiological salt concentrations and human serum as compared to KR-12 and Q5K, D9A. The cross-linked cyclic dimers were sta-ble for up to 8hrs in Staphylococcus aureus protease while LL-37 breaks down within minutes.

The cyclic dimers synthesised earlier were based on chemical methods like Native Chemical Ligation. These methods require lots of chemicals and the

53

yields are usually low. To upscale the cyclisation step in industrial set up, other cheaper and safer methods are required. An alternative nature-based ap-proach can be enzymatic cyclization. Thus, naturally derived enzymes i.e PatGmac and Butelase 1 were used to cyclise different substrates like SFTI-1, kalata B1 and KR-12-based precursors. Using PatGmac, SFTI peptides were cyclised efficiently with sufficient yields but for kB1 PatGmac cyclisation seems non-optimal due to the size and number of disulfide bonds. Using semi pure butelase, cyclic SFTI-NHV and KR-12 (monomer as well as dimers) were produced in solution. Butelase 1 was immobilized on NHS column and the column was then used to cyclise different substrates for two weeks. Using immobilized column, both SFTI and KR-12 peptides were successfully syn-thesised. My thesis thus contributes, generally, to the natural products-based antibiotics discovery process and specifically to the development of stable cyclic antimi-crobial peptides using different strategies.

54

Acknowledgements

I would like to convey my sincere gratitude to everyone who has generously contributed to make this thesis possible:

First and foremost, my main supervisor Ulf Göransson, I thank you whole-heartedly for your continuous guidance and tremendous academic support during this whole period of my PhD studies. I really appreciate your trust in me especially when I wanted to try any of my crude ideas during my project work. It has been a wonderful experience and I learned a lot from you. I am also grateful for making me part of a number of international collaborations, helping me while collaborating with my home country (Pakistan) and also hosting Jawad and Neelam.

Similar profound gratitude goes to Sunithi Gunasekera and Adam A. Strömstedt, my co-supervisors, who have been truly dedicated mentors. The Sri Lankan experience with you guys is unforgettable and I would rate it as one of the best trips of my life. Thank you Sunithi for taking care of me like a family member in Sri Lanka and also Amma and Thaththi for their love, deli-cious homemade food and quality time during my visit! Very special thanks to Adam for being a nice officemate, for the chit-chats and also for the fruitful project-related discussions!

My huge thanks to the whole Sri Lankan team: Machans (Supun Mohotti and Sanjevaan Rajendran), for familiarizing me with the Sri Lankan culture, train rides in Sri Lanka and amazing time also here in Sweden. I am really grateful to your families for preparing authentic Sri Lankan food for us. Cha-mari Hettiarachchi, for hosting me at the University of Colombo, facilitating my lab work there and also for the Mac-rice. Professor Dilip de Silva, for the valuable discussion about natural products!

Thanks to all my co-authors and collaborators, Johan Rosengren (Australia), Wael Houssen and Marcel Jaspars (UK), Adeyare Fallarero and Malena Skogman (Finland), Zafar Iqbal (Pakistan), Shaikh Jamal U din (Bangla-desh), Roland Hellinger and Christian Gruber (Austria), Björn Hellman (Uppsala)!

55

Thank you, my fellow PhD students, Erik for helping me with the lab instru-ment Qtof and also nice time since UGSPR time, Camilla for being a good friend and also maintaining the lab, Momo for nice memories in Finland, Astrid for your concern in the last days of my PhD, Karin for being a good safety officer, Juma for our discussion about African politics with Mahmoud and Quentin Good luck with the synthesizer! (I can feel your painJ)

I would also like to thank the following former members of pharmacognosy group: Robert for helping me in the cytotoxicity lab, Sungkyu for being the mastermind in mafia game, Stefan for the nice time during course lab, Delgerbet for being a good friend and introducing me to the Mongolian cul-ture, Amit for the lecture about the specialties of Bhutan during yearly Christ-mas dinners (I hope that you understand what I am taking about J), Anna Koptina for being a family friend and all the nice memories we had during our cruise trip to Finland, culture evening in Stockholm, dinner at your place and many others, Sohaib for your expert opinion on Microbiology and helping me out with the assays, Elisabet for all the fikas, Mahmoud I miss our dis-cussion during the lunch time bro!

To all pharmacognocists: Lars and Anders for being so nice and kind, Chris-tina for the truffle lunch; Paco for your enthusiasm and passion about sponges, Hesham for your interest in natural products and also late office hours! I would also like to thank the current members of pharmacognosy group for your presence and making work environment so pleasant: Chi, Håkan, Blazej, Nils-Otto, Luke, Scott, Mingshu, Lakmini, Omar, Eslam, Doaa, Martin, Vasiliki.

To all the students who I supervised over the years: Alan, Mia, Jean Baptiste, Liselott, Martin, Jia, Anna, Sacha, Gabriella, Oliver, Aila, Noura, Lau, Francesca, Pouya, Johan, Kathrina, Hassan Thank you guys, you have been amazing students and also your contributions in my project work! To all the administration staff for taking care of all the practical matters, Tusen Tack! J

To all the travel grants, awards and scholarships specifically from “Apotekarsocieteten” for making my participation in different valuable con-ferences possible!

To all the nice people I met in Uppsala: Ram bhai, Altai, Jhanghir, Ahmed, Aslam bhai, Wali and Ding. Thanks for all the fun time!

To my family: my grandmother (Late) and grandfather for all the sacrifices you made to provide good education to your children and grandchildren, my

56

father K. M. Khan for being my mentor and my ideal! my mother and siblings for their love and support, Special thanks to my uncle M. W. Khan for your support when I had first decided to travel to Sweden for higher studies!

To my in-laws especially my father-in-law, F. A. Khan for your encourage-ment, endless support and fatherly love all this time!

To my two twin daughters Ella and Mila, for coming in my life and making my life more cheerful. You are the best things that have ever been mine-your daddy loves you! J

Last but not least my wife Faheema, for taking care of our twins in the last days of my PhD. For your love, care and support!

57

References

Abou Alaiwa, M. H., L. R. Reznikov, N. D. Gansemer, K. A. Sheets, A. R. Horswill, D. A. Stoltz, J. Zabner & M. J. Welsh (2014) pH modulates the activity and synergism of the airway surface liquid antimicrobials beta-defensin-3 and LL-37. Proc Natl Acad Sci U S A, 111, 18703-8.

Aboye, L. T. & J. A. Camarero (2012) Biological synthesis of Circular Polypeptides. J Biol Chem, 287, 27026-27032.

Aboye, T. L., R. J. Clark, R. Burman, M. B. Roig, D. J. Craik & U. Göransson (2011) Interlocking Disulfides in Circular Proteins: Toward Efficient Oxidative Folding of Cyclotides. Antioxidants & Redox Signaling, 14, 77-86.

Agerberth, B., J. Charo, J. Werr, B. Olsson, F. Idali, L. Lindbom, R. Kiessling, H. Jörnvall & H. Wigzell (2000) The human antimicrobial and chemotactic peptides LL-37 and a-defensins HNP 1-3 are expressed by specific lymphocyte and monocyte populations. Blood, 96, 3086-3093.

Agerberth, B., H. Gunne, J. Odeberg, P. Kogner, H. G. Boman & G. H. Gudmundsson (1995) FALL-39, a putative human peptide antibiotic, is cysteine-free and expressed in bone marrow and testis. PNAS, 92, 195-199.

Andersson, D. I., D. Hughes & J. Z. Kubicek-Sutherland (2016) Mechanisms and consequences of bacterial resistance to antimicrobial peptides. Drug Resist Updat, 26, 43-57.

Arnison, P. G., M. J. Bibb, G. Bierbaum, A. A. Bowers, T. S. Bugni, G. Bulaj, J. A. Camarero, D. J. Campopiano, G. L. Challis, J. Clardy, P. D. Cotter, D. J. Craik, M. Dawson, E. Dittmann, S. Donadio, P. C. Dorrestein, K. D. Entian, M. A. Fischbach, J. S. Garavelli, U. Goransson, C. W. Gruber, D. H. Haft, T. K. Hemscheidt, C. Hertweck, C. Hill, A. R. Horswill, M. Jaspars, W. L. Kelly, J. P. Klinman, O. P. Kuipers, A. J. Link, W. Liu, M. A. Marahiel, D. A. Mitchell, G. N. Moll, B. S. Moore, R. Muller, S. K. Nair, I. F. Nes, G. E. Norris, B. M. Olivera, H. Onaka, M. L. Patchett, J. Piel, M. J. T. Reaney, S. Rebuffat, R. P. Ross, H. G. Sahl, E. W. Schmidt, M. E. Selsted, K. Severinov, B. Shen, K. Sivonen, L. Smith, T. Stein, R. D. Sussmuth, J. R. Tagg, G. L. Tang, A. W. Truman, J. C. Vederas, C. T. Walsh, J. D. Walton, S. C. Wenzel, J. M. Willey & W. A. van der Donk (2013) Ribosomally synthesized and post-translationally modified peptide natural products: overview and recommendations for a universal nomenclature. Nat Pro Repo30, 108-160.

Bals, R., X. Wang, Z. Wu, T. Freeman, V. Bafna, M. Zasloff & J. M. Wilson (1998a) Human beta-defensin 2 is a salt-sensitive peptide antibiotic expressed in human lung. J Clin Invest., 102, 874-80.

Bals, R., X. Wang, M. Zasloff & J. M. Wilson (1998b) The peptide antibiotic LL-37/hCAP-18 is expressed in epithelia of the human lung where it has broad antimicrobial activity at the airway surface. PNAS, 95, 9541-9546.

58

Barber, C. J., P. T. Pujara, D. W. Reed, S. Chiwocha, H. Zhang & P. S. Covello (2013) The two-step biosynthesis of cyclic peptides from linear precursors in a member of the plant family Caryophyllaceae involves cyclization by a serine protease-like enzyme. J Biol Chem., 288, 12500-10.

Bensch, K. W., M. Raida, H. J. Mägert, P. Schulz-Knappe & W. G. Forssmann (1995) hBD-1: a novel β-defensin from human plasma. FEBS Lett, 368, 331-335.

Borowski, R. G. V., S. C. B. Gnoatto, A. J. Macedo & R. Gillet (2018) Promising antibiofilm activity of peptidomimetics. Front Microbiol.

Burman, R., S. Gunasekera, A. A. Strömstedt & U. Göransson (2014) Chemistry and biology of cyclotides: Circular plant peptides outside the box. J Nat Prod, 77, 724-736.

Cao, Y., G. K. T. Nguyen, J. P. Tam & C. F. Liu (2015) Butelase-mediated synthesis of protein thioesters and its application for tandem chemoenzymatic ligation. Chem Comm, 51, 17289-17292.

Chan, L. Y., S. Gunasekera, S. T. Henriques, N. F. Worth, S. J. Le, R. J. Clark, J. H. Campbell, D. J. Craik & N. L. Daly (2011) Engineering pro-angiogenic peptides using stable, disulfide-rich cyclic scaffolds. Blood, 118, 6709-6717.

Ciornei, C. D., T. Sigurdardottir, A. Schmidtchen & M. Bodelsson (2005) Antimicrobial and chemoattractant activity, lipopolysaccharide neutralization, cytotoxicity, and inhibition by serum of analogs of human cathelicidin LL-37. Antimicrob Agents Chemother, 49, 2845-50.

Craik, D. J. & J. Du (2017) Cyclotides as drug design scaffolds. Curr Opin in Chem Biol, 38, 8-16.

Dathe, M., H. Nikolenko, J. Klose & M. Bienert (2004) Cyclization increases the antimicrobial activity and selectivity of arginine- and tryptophan-containing hexapeptides. Biochemistry, 43, 9140-50.

Durr, U. H., U. S. Sudheendra & A. Ramamoorthy (2006) LL-37, the only human member of the cathelicidin family of antimicrobial peptides. Biochim Biophys Acta, 1758, 1408-25.

Epand, R. M. & R. F. Epand (2009) Lipid domains in bacterial membranes and the action of antimicrobial agents. BBA, 1788, 289-294.

Falagas, M. E. & S. K. Kasiakou (2006) Toxicity of polymyxins: a systematic review of the evidence from old and recent studies. Crit Care, 10, R27.

Falagas, M. E., S. K. Kasiakou & L. D. Saravolatz (2005) Colistin: The Revival of Polymyxins for the Management of Multidrug-Resistant Gram-Negative Bacterial Infections. Clinical Infectious Diseases 40, 1333-1341.

Fernandez-Lopez, S., H. S. Kim, E. C. Choi, M. Delgado, J. R. Granja, A. Khasanov, K. Kraehenbuehl, G. Long, D. A. Weinberger, K. M. Wilcoxen & M. R. Ghadiri (2001) Antibacterial agents based on the cyclic D,L-alpha-peptide architecture. Nature, 412, 452-455.

Fleming, A. (1922) On a remarkable bacteriolytic element found in tissues and secretions. Proc Ro Soc Lond 93, 306-317.

Franke, B., J. S. Mylne & K. J. Rosengren (2018) Buried treasure: Biosynthesis, structures and applications of cyclic peptides hidden in seed storage albumins. Nat Prod Rep, 35, 137-146.

Fritsche, T. R., P. R. Rhomberg, H. S. Sader & R. N. Jones (2008) Antimicrobial activity of omiganan pentahydrochloride tested against contemporary bacterial pathogens commonly responsible for catheter-associated infections. J Antimicrob Chemother., 61, 1092-8.

59

Goldman, M. J., G. M. Anderson, E. D. Stolzenberg, U. P. Kari, M. Zasloff & J. M. Wilson (1997) Human beta-defensin-1 is a salt-sensitive antibiotic in lung that is inactivated in cystic fibrosis. Cell, 88, 553-60.

Gombart, A. F. (2009) The vitamin D–antimicrobial peptide pathway and its role in protection against infection. Future Microbiol., 4, 1151-65.

Grönberg, A., M. Mahlapuu, ., M. Ståhle, C. Whately-Smith & O. Rollman (2014) Treatment with LL-37 is safe and effective in enhancing healing of hard-to-heal venous leg ulcers: a randomized, placebo-controlled clinical trial. Wound Repair Regen., 22, 613-21.

Gronberg, A., M. Mahlapuu, M. Stahle, C. Whately-Smith & O. Rollman (2014) Treatment with LL-37 is safe and effective in enhancing healing of hard-to-heal venous leg ulcers: a randomized, placebo-controlled clinical trial. Wound Repair Regen, 22, 613-21.

Gudmundsson, G. H., B. Agerberth, J. Odeberg, T. Bergman, B. Olsson & R. Salcedo (1996) The human gene FALL39 and processing of the cathelin precursor to the antibacterial peptide LL-37 in granulocytes. Eur J Biochem 238, 325-32.

Gunasekera, S., T. L. Aboye, W. A. Madian, H. R. El-Seedi & U. Göransson (2013) Making ends meet: Microwave-accelerated synthesis of cyclic and disulfide rich proteins via in situ thioesterification and native chemical ligation. International Journal of Peptide Research and Therapeutics, 19, 43-54.

Hancock, R. E. & D. S. Chapple (1999) Peptide antibiotics. Antimicrob Agents Chemother, 43, 1317-23.

Hancock, R. E. & H. G. Sahl (2006) Antimicrobial and host-defense peptides as new anti-infective therapeutic strategies. Nature Biotechnol 24, 1551-7.

Hancock, R. E. W. & R. Lehrer (1998) Cationic peptides: a new source of antibiotics. Trends in Biotechnol 16, 82-88.

Harris, K. S., T. Durek, Q. Kaas, A. G. Poth, E. K. Gilding, B. F. Conlan, I. Saska, N. L. Daly, N. L. Van Der Weerden, D. J. Craik & M. A. Anderson (2015) Efficient backbone cyclization of linear peptides by a recombinant asparaginyl endopeptidase. Nature Comm, 6, 1-10.

Huang, W. H. (2000) Action of antimicrobial peptides: Two-State Mode. Biochemistry, 39, 8347-8352.

Hultmark, D., A. Engstrom, H. Bennich, R. Kapur & H. G. Boman (1982) Isolation and structure of cecropin D and four minor antibacterial components from cecropia pupae. Eur J Biochem , 127, 207-17.

James, A. M., J. Haywood, J. Leroux, K. Lgnasiak, A. G. Elliott, J. W. Schmidberger, M. F. Fisher, S. G. Nonis, R. Fenske, S. C. Bond & J. S. Mylne (2019) The macrocyclizing protease butelase 1 remains autocatalytic and reveals the structural basis for ligase activity. Plant J, 98, 988-999.

Jeffcoate, W. J., B. A. Lipsky, A. R. Berendt, P. R. Cavanagh, S. A. Bus, E. J. G. Peters, W. H. Van Houtum, G. D. Valk & K. Bakker (2008) Unresolved issues in the management of ulcers of the foot in diabetes. Diabetic Medicine 25, 1380-1389.

Jia, X., S. Kwon, C. I. A. Wang, Y. H. Huang, L. Y. Chan, C. C. Tan, K. J. Rosengren, J. P. Mulvenna, C. I. Schroeder & D. J. Craik (2014) Semienzymatic cyclization of disulfide-rich peptides using sortase A. J Bio Chem, 289, 6627-6638.

60

Knerr, J. P., A. Tzekou, D. Ricklin, H. Qu, H. Chen, W. A. Van der Donk & J. D. Lambris (2011) Synthesis and Activity of Thioether-Containing Analogues of the Complement Inhibitor Compstatin. ACS Chem Biol., 6, 753-760.

Koehnke, J., A. Bent, W. E. Houssen, D. Zollman, F. Morawitz, S. Shirran, J. Vendome, A. F. Nneoyiegbe, L. Trembleau, C. H. Botting, M. C. Smith, M. Jaspars & J. H. Naismith (2012) The mechanism of patellamide macrocyclization revealed by the characterization of the PatG macrocyclase domain. Nat Struct Mol Biol, 19, 767-72.

Koehnke, J., Bent A. F., W. E. Houssen, G. Mann, M. Jaspars & J. H. Naismith (2014) The structural biology of patellamide biosynthesis. Curr Opin in Structural Bio, 29, 112-121.

Koo, B. H. & J. Seo (2019) Antimicrobial peptides under clinical investigation. Peptide Science, 111.

Lamb, H. M., Wiseman, L.R., (1998) Pexiganan acetate. Drugs, 56, 1047-1052.Larrick, J. W., M. Hirata, R. F. Balint, J. Lee, J. Zhong & S. C. Wright (1995) Human CAP18: a novel antimicrobial lipopolysaccharide-binding protein. Infect Immun, 63, 1291-1297.

Li, X., Y. Li, H. Han, D. W. Miller & G. Wang (2006) Solution structures of humanLL-37 fragments and NMR-based identification of a minimal membrane-targeting antimicrobial and anticancer region. J Am Chem Soc, 128, 5776-85.

Lofton, H., M. Pränting, E. Thulin & D. I. Andersson (2013a) Mechanisms and fitness costs of resistance to antimicrobial peptides LL-37, CNY100HL and wheat germ histones. PLoS One, 8, e68875. .

Lofton, H., M. Pränting, E. Thulin & D. I. Andersson (2013b) Mechanisms and Fitness Costs of Resistance to Antimicrobial Peptides LL-37, CNY100HL and Wheat Germ Histones. PLoS One, 8, e68875.

Luo, H., S. Y. Hong, R. M. Sgambelluri, E. Angelos, X. Li & J. D. Walton (2014) Peptide macrocyclization catalyzed by a Prolyl Oligopeptidase involved in α-Amanitin biosynthesis. Chemistry & Biology 21, 1610-1617.

Mahlapuu, M., J. Håkansson, L. Ringstad & B. Björn (2016) Antimicrobial peptides: An emerging category of therapeutic agents. Frontiers in Cellular and Infection Microbiology 6, 194.

Malanovic, N., R. Leber, M. Schmuck, M. Kriechbaum, R. A. Cordfunke, J. W. Drijfhout, A. de Breij, P. H. Nibbering, D. Kolb & K. Lohner (2015) Phospholipid-driven differences determine the action of the synthetic antimicrobial peptide OP-145 on Gram-positive bacterial and mammalian membrane model systems. Biochim. Biophys. Acta 1848, 2437-2447.

Malik, S. Z. 2017. Interaction of cyclotides and bacteria: A study of the cyclotide action and the bacterial reaction. In Pharmacognosy, Doctoral Dissertation 48. Uppsala, Acta Universitatis Upsaliensis: Uppsala University

Malik, S. Z., M. Linkevicius, U. Goransson & D. I. Andersson (2017) Resistance to the Cyclotide Cycloviolacin O2 in Salmonella enterica Caused by Different Mutations That Often Confer Cross-Resistance or Collateral Sensitivity to Other Antimicrobial Peptides. Antimicrob Agents Chemother, 61.

Marx, U. C., M. L. J. Korsinczky, H. J. Schirra, A. Jones, B. Condie, L. Otvos & D. J. Craik (2003) Enzymatic cyclization of a potent Bowman-Birk protease inhibitor, sunflower trypsin inhibitor-1, and solution structure of an acyclic precursor peptide. Journal of Biological Chemistry, 278, 21782-21789.

61

Migoń, D., D. Neubauer & W. Kamysz (2018) Hydrocarbon stapled sntimicrobial peptides. Protein J, 37, 2-12.

Mishra, B., R. F. Epand, R. M. Epand & G. S. Wang (2013) Structural location determines functional roles of the basic amino acids of KR-12, the smallest antimicrobial peptide from human cathelicidin LL-37. RSC Advances, 3, 19560-19571.

Mishra, B., R. M. Golla, K. Lau, T. Lushnikova & G. Wang (2016) Anti-Staphylococcal Biofilm Effects of Human Cathelicidin Peptides. ACS Med Chem Lett, 7, 117-21.

Mishra, B. & G. Wang (2017) Titanium surfaces immobilized with the major antimicrobial fragment FK-16 of human cathelicidin LL-37 are potent against multiple antibiotic-resistant bacteria. Biofouling, 33, 544-555.

Moiola, M., M. G. Memeo & P. Quadrelli (2019) Stapled Peptides-A Useful Improvement for Peptide-Based Drugs. Molecules, 24.

Moll, G. N., A. Kuipers, L. De Vries, T. Bosma & R. Rink (2009) A biological stabilization technology for peptide drugs: Enzymatic introduction of thioether-bridges. Drug Discov Today: Technologies, 6, e13-e18.

Mourtada, R., H. D. Herce, D. J. Yin, J. A. Moroco, T. E. Wales, J. R. Engen & L. D. Walensky (2019) Design of stapled antimicrobial peptides that are stable, nontoxic and kill antibiotic-resistant bacteria in mice. Nature Biotech 37, 1186-1197.

Münzker, L., A. Oddo & P. R. Hansen (2017) Chemical synthesis of antimicrobial peptides. Methods Mol Biol, 1548, 35-49.

Murai, N. & S. Sugai (1974) Side-chain effect on conformation of ionizable polypeptides in aqueous solution. Biopolymers, 13, 1161-71.

Nell, M. J., G. S. Tjabringa, A. R. Wafelman, R. Verrijk, P. S. Hiemstra, J. W. Drijfhout & J. J. Grote (2005) Development of novel LL-37 derived antimicrobial peptides with LPS and LTA neutralizing and antimicrobial activities for therapeutic application. Peptides, 27, 649-660.

Nguyen, G. K., S. Wang, Y. Qiu, X. Hemu, Y. Lian & J. P. Tam (2014) Butelase 1 is an Asx-specific ligase enabling peptide macrocyclization and synthesis. Nat Chem Biol, 10, 732-738.

Nguyen, G. K. T., X. Hemu, J. P. Quek & J. P. Tam (2016) Butelase-Mediated Macrocyclization of d-Amino-Acid-Containing Peptides. Angewandte Chemie International Edition, 55, 12802-12806.

Nguyen, G. K. T., A. Kam, S. Loo, A. E. Jansson, L. X. Pan & J. P. Tam (2015) Butelase 1: A Versatile Ligase for Peptide and Protein Macrocyclization. J Am Chem Soc, 137, 15398-15401.

Nilsson, A. C., H. Janson, H. Wold, A. Fugelli, K. Andersson, C. Hakangard, P. Olsson & W. M. Olsen (2015) LTX-109 is a novel agent for nasal decolonization of methicillin-resistant and -sensitive Staphylococcus aureus. Antimicrob. Agents Chemother., 59, 145-151.

Oueis, E., H. Stevenson, M. Jaspars, J. N. Westwood & H. J. Naismith (2017) Bypassing the proline/thiazoline requirement of the macrocyclase PatG. Chem Comm

Parachin, N. S., K. C. Mulder, A. A. BarbosaViana, S. C. Dias & O. L. Franco (2012) Expression systems for heterologous production of antimicrobial peptides. Peptides, 38, 446-456.

62

Pasupuleti, M., A. Schmidtchen, A. Chalupka, L. Ringstad & M. Malmsten (2009) End-Tagging of Ultra-Short Antimicrobial Peptides by W/F Stretches to Facilitate Bacterial Killing. Plos One, 4, -.

Peek, F., M. J. Nell, R. Brand, T. Jansen-Werkhoven, E. Van Hoogdalem & J. Frijns (2009) Double-blind placebo-controlled study of the novel peptide drug P60.4Ac in cronic middle ear infection. ICAAC, L1-L337.

Pokorny, A., T. H. Birkbeck & P. F. F. Almeida (2002) Mechanism and kinetics of delta-lysin interaction with phospholipid vesicles. Biochemistry, 41, 11044-11056.

Powers, J. S., A. Rozek & R. E. W. Hancock (2004) Structure–activity relationships for the β-hairpin cationic antimicrobial peptide polyphemusin I. . Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics, 1698, 239-250.

Pränting, M., C. Loov, R. Burman, U. Goransson & D. I. Andersson (2010) The cyclotide cycloviolacin O2 from Viola odorata has potent bactericidal activity against Gram-negative bacteria. J Antimicrob Chemother., 65, 1964-71.

Rapala-Kozik, M., O. Bochenska, M. Zawrotniak, N. Wolak, G. Trebacz, M. Gogol, D. Ostrowska, W. Aoki, M. Ueda & A. Kozik (2015) Inactivation of the antifungal and immunomodulatory properties of human cathelicidin LL-37 by aspartic proteases produced by the pathogenic yeast Candida albicans. Infect Immun, 83, 2518-30.

Riedl, S., D. Zweytick & K. Lohner (2011) Membrane-active host defense peptides--challenges and perspectives for the development of novel anticancer drugs. Chem Phys Lipids, 164, 766-81.

Rogers, L. A. & E. O. Whittier (1928) Limiting factores in the lactic fermentation. J Bacteriol., 16, 211.

Rosengren, K. J., U. Göransson, L. J. Otvos & D. Craik, J., (2004) Cyclization of pyrrhocoricin retains structural elements crucial for the antimicrobial activity of the native peptide. Biopolymers, 76, 446-58.

Rubinchik, E., D. Dugourd, T. Algara, C. Pasetka & H. D. Friedland (2009) Antimicrobial and antifungal activities of a novel cationic antimicrobial peptide, omiganan, in experimental skin colonisation models. Int J Antimicrob Agents., 34, 457-61.

Ryan, K. J., L. I. Schainuck, R. O. Hickman & G. Striker (1969) Colistimethate Toxicity: Report of a Fatal Case in a Previously Healthy Child. JAMA, 207, 2099-101.

Schmidt, M., T. Ana, J. L. M. Q. Peter, I. Hans, J. J. R. Gaston, M. H. Tilman, H. v. M. Jan & N. Timo (2017) Omniligase-1: A Powerful Tool for Peptide Head-to-Tail Cyclization. Adv.Synth.Catal, 359, 2050-2055.

Scott, M. G., D. J. Davidson, M. R. Gold, D. Bowdish & R. E. Hancock (2002) The human antimicrobial peptide LL-37 is a multifunctional modulator of innate immune responses. J Immunol, 169, 3883-91.

Selsted, M. E., D. M. Brown, R. J. DeLange, S. S. Harwig & R. I. Lehrer (1985) Primary structures of six antimicrobial peptides of rabbit peritoneal neutrophils. J Biol Chem, 260, 4579-84.

Shin, J. M., J. W. Gwak, P. Kamarajan, J. C. Fenno, A. H. Rickard & Y. Y. Kapila (2016) Biomedical applications of Nisin. J Appl Microbiol. , 120, 1449-1465.

Sieprawska-Lupa, M., P. Mydel, K. Krawczyk, K. Wójcik, M. Puklo, B. Lupa, P. Suder, J. Silberring, M. Reed, J. Pohl, W. Shafer, F. McAleese, T. Foster, J.

63

Travis & J. Potempa (2004) Degradation of human antimicrobial peptide LL-37 by Staphylococcus aureus-derived proteinases. Antimicrob. Agents Chemother., 48, 4673-4679.

Sitaram, N., C. Subbalakshmi & R. Nagaraj (2003) Indolicidin, a 13-residue basic antimicrobial peptide rich in tryptophan and proline, interacts with Ca(2+)-calmodulin. Biochem Biophys Res Commun, 309, 879-84.

Skogman, M. E., P. M. Vuorela & A. Fallarero (2012) Combining biofilm matrix measurements with biomass and viability assays in susceptibility assessments of antimicrobials against Staphylococcus aureus biofilms. J Antibiot (Tokyo), 65, 453-9.

Sohlenkamp, C. & O. Geiger (2016) Bacterial membrane lipids: diversity in structures and pathways. FEMS Mico Rev, 40, 133-159.

Sørensen, O., K. Arnljots, J. B. Cowland, D. F. Bainton & N. Borregaard (1997) The human antibacterial cathelicidin, hCAP-18, is synthesized in myelocytes and metamyelocytes and localized to specific granules in neutrophils. Blood, 90, 2796-803.

Stanger, K., T. Maurer, H. Kaluarachchi, M. Coons, Y. Franke & R. N. Hannoush (2014) Backbone cyclization of a recombinant cystine-knot peptide by engineered Sortase A. FEBS Lett, 588, 4487-96.

Strömstedt, A. A., S. Park, R. Burman & U. Goransson (2017) Bactericidal activity of cyclotides where phosphatidylethanolamine-lipid selectivity determines antimicrobial spectra. Biochim Biophys Acta Biomembr, 1859, 1986-2000.

Strömstedt, A. A., M. Pasupuleti, A. Schmidtchen & M. Malmsten (2009) Evaluation of strategies for improving proteolytic resistance of antimicrobial peptides by using variants of EFK17, an internal segment of LL-37. Antimicrob Agents Chemother, 53, 593-602.

Tam, J. P., Y. Lu & J. Yang (2002) Correlations of cationic charges with salt sensitivity and microbial specificity of cystine-stabilized β-Strand antimicrobial peptides. J Biol Chem, 277, 50450-50456.

Thongyoo, P., E. W. Tate & R. J. Leatherbarrow (2006) Total synthesis of the macrocyclic cysteine knot microprotein MCoTI-II. Chemical Comm, 2848-2850.

Tjabringa, G. S., J. Aarbiou, D. K. Ninaber, J. W. Drijfhout, O. E. Sorensen, N. Borregaard, K. F. Rabe & P. S. Hiemstra (2003) The antimicrobial peptide LL-37 activates innate immunity at the airway epithelial surface by transactivation of the epidermal growth factor receptor. J Immunol, 171, 6690-6.

Turner, J., Y. Cho, N. N. Dinh, A. J. Waring & R. I. Lehrer (1998) Activities of LL-37, a cathelin-associated antimicrobial peptide of human neutrophils. Antimicrob Agents Chemother, 42, 2206-14.

Utsugi, T., A. J. Schroit, J. Connor, C. D. Bucana & I. J. Fidler (1991) Elevated expression of phosphatidylserine in the outer membrane leaflet of human tumor cells and recognition by activated human blood monocytes. Cancer Res, 51, 3062-6.

Vandamme, D., B. Landuyt, W. Luyten & L. Schoofs (2012) A comprehensive summary of LL-37, the factotum human cathelicidin peptide. Cell Immunol, 280, 22-35.

64

Velden, W. J., T. M. van Iersel, N. M. Blijlevens & J. P. Donnelly (2009) Safety and tolerability of the antimicrobial peptide human lactoferrin 1–11 (hLF1–11). BMC Medicine, 7, 44.

Vlieghe, P., V. Lisowski, J. Martinez & M. Khrestchatisky (2010) Synthetic therapeutic peptides: science and market. Drug Disco Today, 15, 40-56.

Wan, Q. & S. J. Danishefsky (2007) Free-radical-based, specific desulfurization of cysteine: a powerful advance in the synthesis of polypeptides and glycopolypeptides. Angew Chem Int Ed Engl, 46, 9248-52.

Wang, G. (2008) Structures of human host defense cathelicidin LL-37 and its smallest antimicrobial peptide KR-12 in lipid micelles. J Biol Chem, 283, 32637-43.

--- (2012) Post-translational modifications of natural antimicrobial peptides and strategies for peptide engineering. Curr Biotechnol. 1, 72-79.

Wang, G., X. Li & Z. Wang (2009) APD2: the updated antimicrobial peptide database and its application in peptide design. . Nucleic Acids Res, 37, D933-D937.

Wang, G., X. Li & Z. Wang (2016) APD3: the antimicrobial peptide database as a tool for research and education. Nucleic Acids Res, 44, D1087-D1093.

Wang, G., K. M. Watson & R. W. Buckheit, Jr. (2008) Anti-human immunodeficiency virus type 1 activities of antimicrobial peptides derived from human and bovine cathelicidins. Antimicrob Agents Chemother, 52, 3438-40.

Wang, Y., B. Agerberth, A. Löthgren, A. Almstedt & J. Johansson (1998) Apolipoprotein A-I Binds and Inhibits the Human Antibacterial/Cytotoxic Peptide LL-37. J Bio Chem, 273, 33115-8.

Wessolowski, A., M. Bienert & M. Dathe (2004) Antimicrobial activity of arginine- and tryptophan-rich hexapeptides: the effects of aromatic clusters, D-amino acid substitution and cyclization. J Pept Res., 64, 159-69.

Wimley, W. C. (2010) Describing the mechanism of antimicrobial peptide action with the interfacial activity model. ACS Chem Biol, 5, 905-917.

Wu, Z., X. Guo & Z. Guo (2011) Sortase A-catalyzed peptide cyclization for the synthesis of macrocyclic peptides and glycopeptides. Chem Comm, 47, 9218-20.

Xie, Y. & L. Yang (2016) Calcium and Magnesium Ions Are Membrane-Active against Stationary-Phase Staphylococcus aureus with High Specificity. Sci Rep., 6.

Yeaman, M. R. & N. Y. Yount (2003) Mechanisms of antimicrobial peptide action and resistance. Pharmacol Rev, 55, 27-55.

Yeung, A. T., S. L. Gellatly & R. E. Hancock (2011) Multifunctional cationic host defence peptides and their clinical applications. Cell Mol Life Sci., 68, 2161-76.

Yilong Li, Tao Bi & J. A. Camarero (2015) Chemical and biological production of cyclotides. Adv Bot Res 76.

Yu, Q., R. I. Lehrer & J. P. Tam (2000) Engineered salt-insensitive alpha-defensins with end-to-end circularized structures. J Biol Chem., 275, 3943-9.

Zaffiri, L., J. Gardner & L. H. Toledo-Pereyra (2012) History of antibiotics. From salvarsan to cephalosporins. . J Invest Surg, 25, 67-77.

Zaman, S. B., M. A. Hussain, R. Nye, V. Mehta, K. T. Mamun & N. Hossain (2017) A Review on Antibiotic Resistance: Aaarm bells are ringing. Cureus, 9, e1403.

Zanetti, M. (2004) Cathelicidins, multifunctional peptides of the innate immunity. J Leukoc Biol, 75, 39-48.

65

Zanetti, M., R. Gennaro & D. Romeo (1995) Cathelicidins-A novel protein family with a common proregion and a variable C-terminal antimicrobial domain. FEBS Letters, 374, 1-5.

Zasloff, M. (1987) Magainins, a class of antimicrobial peptides from Xenopus skin: isolation, characterization of two active forms, and partial cDNA sequence of a precursor. Proc Natl Acad Sci U S A, 84, 5449-5453.

--- (2002) Antimicrobial peptides of multicellular organisms. Nature, 415, 389-95. Zeya, H. I. & J. K. Spitznagel (1963) Antibacterial and enzymic basic proteins from

Leukocyte Lysosomes: separation and identification. Science 142, 1085-1087.

Zhao, C., L. Liu & R. I. Lehrer (1994) Identification of a new member of the protegrin family by cdna cloning. FEBS Lett, 346, 285-288.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 281

Editor: The Dean of the Faculty of Pharmacy

A doctoral dissertation from the Faculty of Pharmacy, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofPharmacy. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Pharmacy”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-397191

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

ll-37-derived cyclic antimicrobial drug...

Documents