Applications of Sortase a in Disulfide-Rich Peptide Engineering

Applications of sortase A in disulfide-rich peptide engineering

Soohyun Kwon

BSc, MSc

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2016 Institute for Molecular Bioscience

Abstract

While the global peptide drug market has been constantly growing, the intrinsic limitations of peptides, such as chemical and physical instability and poor oral bioavailability, continue to require innovative peptide engineering to facilitate the uptake of peptides by the pharmaceutical industry. In recent years, plant-derived cyclotides have been investigated for this purpose because cyclotides are remarkably stable against chemical, enzymatic or thermal attack. The stability of cyclotides is derived from their unique structure, which comprises a head-to-tail cyclized backbone and three disulfide bonds forming a knotted core. Although the overall structure of cyclotides is rigid, certain loops are ‘plastic’ enough to accommodate their synthetic replacement with a foreign peptide epitope. More than a dozen published studies have already demonstrated that bioactive peptide sequences can be grafted onto cyclotide scaffolds and thereby stabilized, while maintaining activity.

Following lessons learnt from studies of cyclotides, other naturally occurring disulfide-rich cyclic peptides, including sunflower trypsin inhibitor I (SFTI-1) and θ-defensins, have been studied for similar purposes. Furthermore, cyclization has been applied to disulfide-rich neurotoxin peptides, derived from cone snails or scorpions, and has resulted in improved therapeutic potential. This thesis investigates a range of these natural peptides for their role as drug templates and explores novel approaches to produce them.

The benefits of engineering cyclotides and other disulfide-rich peptides for therapeutic purposes can be maximized by efficient production systems. In this thesis, an efficient strategy to produce disulfide-rich cyclic peptides is investigated using the enzyme sortase A. Sortase A, originally discovered in Staphylococcus aureus, recognizes a penta-amino acid motif, LPXTG, and cleaves the peptide bond between Thr and Gly to form a thioacyl intermediate. This intermediate undergoes nucleophilic attack by an N-terminal poly-Gly sequence, forming a new peptide bond between the Thr and N-terminal Gly. This site-specific ligation (transpeptidation) activity of sortase A allowed the successful cyclization of a synthetic cyclotide precursor in a straightforward and safe manner compared to traditional chemical methods.

The semi-enzymatic method was also readily applied to produce other disulfide-rich cyclic peptides, including SFTI-1, cVc1.1, and cyclic chlorotoxin. Furthermore, by considering cyclotides in a structural model, cyclization of the κ-conotoxin PVIIA originating from the venom of Conus

purpurascens was also explored. Interestingly, κ-PVIIA and other conopeptides, which belong to the O1 gene superfamily of conotoxins, are topologically very similar to cyclotides except that their backbones are linear. In this thesis, cyclization of κ-PVIIA was studied as a proof-of-concept for other conopeptides in the O1 superfamily. Furthermore, this study provides insights into successful cyclization by suggesting factors to be considered to obtain the desired peptide activity.

Even with enzymatic cyclization, chemical synthesis of disulfide-rich peptides is not as cost- effective as recombinant expression in large scale manufacturing. Therefore, recombinant expression of cyclotides, which are 28-37 amino acid long, is desirable for them to be attractive as potential therapeutics. In this thesis, bacterial periplasmic expression of kalata B1 precursor was investigated. To aid solubility of hydrophobic kalata B1 in the bacterial periplasm, maltose binding protein (MBP) was used as a fusion protein and sorting motifs were placed at both N and C terminus of kalata B1 sequence for subsequent cleavage from MBP and head-to-tail cyclization of kalata B1.

Lastly, sortase A-mediated engineering was used to enable targeted delivery of cyclotides via conjugation to a single-domain antibody fragment (VHH). A non-natural amino acid, azidolysine, was incorporated during cyclotide synthesis, and bacterially expressed anti MHC Class II VHH was equipped with cyclooctyne using sortase A to enable a click reaction between them. The conjugation resulted in the improvement in immunogenicity of cyclotides by targeting antigen presenting cells. Overall, this strategy facilitates the application of a targeted delivery of cyclotides for therapeutic purposes.

iii

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

Publications during candidature

Peer-reviewed papers (1) Kwon S, Bosmans F, Kaas Q, Cheneval O, Conibear AC, Rosengren KJ, Wang CK, Schroeder CI, Craik DJ. (2016) Efficient enzymatic cyclization of an inhibitory cystine knot-containing peptide. Accepted on the 19th April 2016 by Biotechnol Bioeng

(2) Jia X*, Kwon S*, Wang CA, Huang YH, Chan L, Tan C, Rosengren KJ, Mulvenna JP, Schroeder CI, Craik DJ. (2014) Semienzymatic cyclization of disulfide-rich peptides using sortase A. J Biol Chem 289:6627-38 *co-first authors

(3) Cemazar M, Kwon S, Mahatmanto T, Ravipati AS, Craik DJ. (2012) Discovery and applications of disulfide-rich cyclic peptides. Curr Top Med Chem 12:1534-45 Review

Conference abstracts (1) 3rd International Conference on Circular Proteins ‒ Moreton island, Australia (November 2015) Oral presentation: Application of Sortase A in disulfide-rich peptide engineering

(2) Peptide Therapeutics Symposium, Salk Institute, USA (October 2014) Poster presentation: Semienzymatic formation of cyclic cystine knot

(3) Australian Society for Medical Research (ASMR) Postgraduate Student Conference ‒ Brisbane, Australia (May 2013) Poster presentation: Cyclization of kalata B1 using sortase A

(4) 2nd International Conference on Circular Proteins, Heron Island, Australia (October 2012) Poster presentation: Cyclization of kalata B1 using sortase A

Publications included in this thesis

(1) Cemazar M, Kwon S, Mahatmanto T, Ravipati AS, Craik DJ. (2012) Discovery and applications of disulfide-rich cyclic peptides. Curr Top Med Chem 12(14):1534-45 Review ‒ partially incorporated in Chapter 1 (conotoxin and chlorotoxin) Contributor Statement of contribution Cemazar M Revised and edited the paper (20%) Kwon S (Candidate) Wrote and edited paper (20%) Prepared and edited the figures (25%) Mahatmanto T Wrote and edited paper (20%) Prepared and edited the figures (25%) Ravipati AS Wrote and edited paper (20%) Prepared and edited the figures (25%) Craik DJ Wrote and edited paper (20%) Edited the figures (25%)

Contributor Statement of contribution Jia X Designed experiments (40%) Performed experiments for kalata B1 (30%) Kalata B1 structure calculations (45%) Sortase A expression (50%) Wrote paper (40%) Prepared and edited the figures (10%) Kwon S (Candidate) Designed experiments (20%) Performed experiments for kalata B1 (70%) Performed experiments for SFTI-1 (100%) Peptide preparation (70%) Sortase A expression (50%) Hemolytic activity and stability assay (50%) Wrote and edited paper (30%) Prepared and edited the figures (60%) Wang CA Performed experiments for Vc1.1 (100%) Revised paper (5%) Huang YH Hemolytic activity and stability assay (50%) Chan L Peptide preparation (10%) Tan C Peptide preparation (20%) Rosengren KJ Kalata B1 structure calculations (10%) Mulvenna JP Revised the paper (5%) Schroeder CI Designed experiments (40%) Kalata B1 structure calculations (45%) Prepared and edited the figures (20%) Revised the paper (40%) Craik DJ Revised the paper (50%) Prepared and edited the figures (10%)

vii

(3) Kwon S, Bosmans F, Kaas Q, Cheneval O, Conibear AC, Rosengren KJ, Wang CK, Schroeder CI, Craik DJ. (2016) Efficient enzymatic cyclization of an inhibitory cystine knot-containing peptide. Submission on 13/Jan/2016 to Biotechnol. Bioeng. (incorporated as Chapter 4) Contributor Statement of contribution Kwon S. (Candidate) Designed experiments (30%) Peptide preparation (80%) Oxidation, cyclization (100%) Sortase A expression (100%) NMR experiments and structure calculation (100%) Stability assay (100%) Wrote the paper (70%) Prepared figures (55%) Bosmans F. Electrophysiology experiments (100%) Wrote the paper (20%) Prepared figures (15%) Kaas Q. Molecular modelling study (100%) Wrote the paper (10%) Prepared figures (30%) Cheneval O. Peptide preparation (20%) Conibear AC. Structure calculation technical advice (25%) Rosengren KJ Structure calculation technical advice (25%) Wang CK. Structure calculation technical advice (25%) Schroeder CI Designed experiments (70%) Structure calculation technical advice (25%) Edited the figures (50%) Revised the paper (50%) Craik DJ Revised the paper (50%) Edited the figures (50%)

viii

Contributions by others to the thesis

Phillip Walsh, Philip Sunderland, and Olivier Cheneval carried out peptide synthesis using an automatic peptide synthesizer (Symphony®, Protein Technologies, Inc. USA). Figure 2 in Chapter 2 was prepared by Olivier Cheneval. Dr. Xinying Jia designed the kalata B1 expression strategy in Chapter 5.

Statement of parts of the thesis submitted to qualify for the award of another degree

None

Acknowledgements

Immeasurable appreciation and deepest gratitude are extended to the following persons who have supported and guided me throughout this study.

My principal supervisor, Prof. David Craik, for giving me an opportunity to be a part of his group, supervision with patience, passing his knowledge in both science and management skills. Co-supervisor and a mentor, Dr. Christina Schroeder, for encouragement and support through frequent meetings and personal advice. Dr. Xinying Jia, for providing me with inspiring ideas and teaching me protein expression and purification. Phill Walsh, Phil Sunderland, Olivier Cheneval for peptide synthesis and HF cleavage. All other Craik group members, for day-to-day help, stimulating feedback during lab meetings, and friendship. Dr. Anderson Wang, for helpful discussion on recombinant protein expression. Dr. Peta Harvey, for lots of help with NMR, Dr. Alun Jones, for training on MALDI-TOF. Thesis committee, Prof. Kirill Alexandrov and Prof. Norelle Daly, for providing valuable advice. Dr. Amanda Carrozzi, for her encouragement and help with scholarship and travel funding.

I also would like to thank Prof. Hidde Ploegh for giving me the opportunity to visit his laboratory at the Whitehead Institute for six months, and for his support and helpful discussion on the project. Robert Miller for his support for me adapt to the new environment quickly. Ella Li, Joao Duarte, Dr.Tao Fang, Jingjing ling, Dr. Masatoshi Hagiwara for help so that I could achieve the work included in Chapter 6, and all Ploegh lab members, for welcoming me and their kindness.

I also thank Prof. David Liu at Harvard University for providing mutant sortase expression vectors. Dr. In-Jung Kim, my previous supervisor, for her teaching and giving me confidence to start PhD. And special thanks to my family and friends, for their love and support.

Finally, I would like to acknowledge the financial support of Australian Postgraduate Award, graduate school international travel award (GSITA) from the University of Queensland, and student travel grant from IMB.

Keywords peptides, cyclotides, conotoxins, peptide structure, sortase, peptide conjugation

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 030301, Analytical Chemistry, 20% ANZSRC code: 030406, Proteins and Peptides, 60% ANZSRC code: 060112, Structural Biology, 20%

Fields of Research (FoR) Classification

FoR code: 0304, Medicinal and Biomolecular Chemistry, 70% FoR code: 0601, Biochemistry and Cell Biology, 30%

Table of Contents

Abstract ...... ii

Declaration by author ...... iv

Publications during candidature ...... v

Peer-reviewed papers ...... v

Conference abstracts ...... v

Publications included in this thesis ...... vi

Contributions by others to the thesis ...... ix

Statement of parts of the thesis submitted to qualify for the award of another degree ...... ix

Acknowledgements ...... x

Keywords ...... xi

Australian and New Zealand Standard Research Classifications (ANZSRC) ...... xi

Fields of Research (FoR) Classification ...... xi

Table of contents ...... xii

Chapter 1. Introduction ...... 1

1.1. Overview and declaration of authorship ...... 2

1.2. Cyclotides...... 4

1.2.1. Discovery, structure and biological activities of cyclotides ...... 4

1.2.2. Biosynthesis of cyclotides ...... 4

1.2.3. Chemical biology and applications of cyclotides ...... 5

1.3. Conotoxins ...... 7

1.4. Chlorotoxins ...... 8

1.5. Sortase A ...... 10

1.5.1. Discovery, recombinant expression and structure of sortase A (SrtA) ...... 10

1.5.2. Application of SrtA ...... 11

1.5.3. Mutagenesis of SrtA and improvement of SrtA reaction ...... 15

xii

1.6. Summary and outlook ...... 16

1.7. Scope of this thesis ...... 17

1.8. References ...... 18

Chapter 2. Peptide synthesis and NMR spectroscopy ...... 28

2.1. Overview ...... 29

2.2. Solid-phase peptide synthesis ...... 29

2.2.1. Synthesis procedure ...... 29

2.2.2. Peptide cleavage ...... 29

2.2.3. Fmoc-based synthesis of cyclotides ...... 30

2.3. NMR spectroscopy ...... 31

2.3.1. TOCSY ...... 32

2.3.2. NOESY ...... 33

2.3.3. Dihedral angle restraints ...... 33

2.3.4. Hydrogen bonds ...... 35

2.3.5. Structure calculation ...... 36

2.4. References ...... 37

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A ...... 40

3.1. Overview and Declaration of authorship ...... 41

3.2. Semienzymatic cyclization of disulfide-rich peptides using sortase A...... 42

3.3. Abstract ...... 43

3.4. Introduction ...... 43

3.3. Experimental procedure ...... 46

3.3. Results ...... 50

3.5. Discussion ...... 61

3.6. References ...... 65

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot-containing peptide . 71

4.1. Overview and Declaration of authorship ...... 72

xiii

4.2. Efficient enzymatic cyclization of an inhibitory cysteine knot-containing peptide ...... 73

4.3. Abstract ...... 74

4.4. Introduction ...... 75

4.3. Materials and methods ...... 78

4.3. Results and discussion ...... 82

4.3. References ...... 96

Chapter 5. Bacterial expression of kalata B1 ...... 102

5.1. Overview ...... 103

5.2. Experimental procedure ...... 105

5.2.1. Design of expression vector ...... 105

5.2.2. Bacterial culture ...... 105

5.2.3. Bacterial cell lysis ...... 106

5.2.4. Purification of fusion protein ...... 106

5.2.5. SrtA cleavage and peptide purification ...... 106

5.2.6. NMR analysis ...... 106

5.3. Results ...... 106

5.3.1. Expression of MBP-[LPETGG]kalata B1[TGG] ...... 107

5.3.2. Kalata B1 precursor cleavage by SrtA5° ...... 108

5.3.3. Periplasmic expression of [GG]kalata B1[TGG] ...... 109

5.4. Discussion ...... 113

5.4. References ...... 116

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit immune response ...... 119

6.1. Overview ...... 120

6.2. Experimental procedures...... 122

6.2.1. Peptides synthesized ...... 122

6.2.2. Nuclear magnetic resonance (NMR) analysis ...... 123

xiv

6.2.3. Expression of recombinant VHH7 and Enhancer protein ...... 123

6.2.4. Maleimide-thiol reaction on Gly-Gly-Gly-Cys-NH2 ...... 124

6.2.5. Modification of VHH7 using SrtA ...... 124

6.2.6. Conjugation of a cyclotide to VHH7 ...... 124

6.2.7. Immunization and serum collection ...... 124

6.2.8. SDS-PAGE and Western blot ...... 124

6.3. Results ...... 125

6.3.1. Click handle installation on MCoTI-I and a probe for VHH7 ...... 126

6.3.2. Bacterial expression, SrtA-mediated modification of VHH7, and conjugation ...... 126

6.3.3. NMR analysis of MCoTI-I and MCoTI-HA ...... 127

6.3.4. Mouse immunization and western blot analysis ...... 137

6.3.5. Reactivity between MCoTI-HA and monoclonal anti-HA antibody...... 138

6.4. Discussion ...... 140

6.5. References ...... 143

Overall discussion ...... 147

Concluding remarks ...... 152

Appendix ...... 155

List of Figures & Tables Chapter 1 Page Figure 1 Cyclotides from three categories and biosynthesis 5 Figure 2 Applications of cyclotide as scaffolds 6 Figure 3 Linear and cyclic α-conotoxin MII and Vc1.1 8

Figure 4 Schematic illustration of chlorotoxin (PDB 1CHL) cyclization and 10 conjugation with cy5.5 Figure 5 Schematic representation of SrtA-mediated protein engineering 13 Figure 6 Methods to improve efficiency of SrtA reaction 16 Table 1 Proteins and peptides that have been cyclized using SrtA 14 Chapter 2 Figure 1 Overview of Fmoc-solid phase peptide synthesis (SPPS) 30 Figure 2 Fmoc-based cyclization of disulfide-rich peptides 32 Figure 3 An example of TOCSY and COSY 33 Figure 4 The Karplus equation and the dihedral angle ϕ 34

Figure 5 Newman projections of the three possible staggered conformers for the χ1 35 dihedral angle Chapter 3 Figure 1 Overview of semienzymatic synthesis of kalata B1 46 Figure 2 HPLC profile of oxidation and cyclization 52 Figure 3 Hα secondary shift comparison of native kB1 and cyclo-[GGG]kB1[T] 53

Figure 4 NMR spectra of cyclo-[GGG]kB1[T] 54 Figure 5 Superimposition of the 20 lowest energy structures of cyclo-[GGG]kB1[T] 55 Figure 6 Hemolytic and serum-stability assay on cyclo-[GGG]kB1[T] 56 Figure 7 Cyclization of Vc1.1 and SFTI-1 58

Figure 8 NMR spectra of cyclo-[G]Vc1.1[GLPET] 60 Figure 9 NMR spectra of cyclo-[GG]SFTI-1[LPET] 61 Table 1 Energies and structural statistics for the family of 20 lowest energy cyclo- 52 [GGG]kB1[T] structures Chapter 4

Figure 1 Sequence alignment of O1 superfmaily member peptides containing the 78 ICK motif and schematic representation of the cyclization of κ-PVIIA Figure 2 Oxidation and cyclization of [GGG]PVIIA[LPETGG] 78 Figure 3 Structure of cyc-PVIIA 83 Figure 4 Inhibitory toxin activity on Drosophila Shaker and rKV1.6 83

xvi

Figure 5 Model of interactions of κ-PVIIA and cyc-PVIIA with Shaker Kv channel 92 Different hydrogen bond network of Arg2 between native (PDB code: Figure 6 93 1AV3) (Scanlon et al., 1997) and cyc-PVIIA Figure 7 Serum stability assay on native and cyc-PVIIA 94

Table 1 Energies and structural statistics for the family of 20 lowest energy cyc- 87 PVIIA structures Chapter 5 Figure 1 Strategy of cyclization of MBP-LPETGG-kalata B1-TGG 105 Figure 2 Expression of kalata B1 precursor and SrtA-based cyclisation 108 Figure 3 Purification of cyclised [GG]kalata B1[T] 109 Figure 4 Comparison between recombinant cyclic [GG]kalata B1[T] and cyclo- 110 [GGG]kalata B1[T] at pH 3.5 Figure 5 Expression of MBP-kalata B1 precursor and cleavage by SrtA 111 Figure 6 One-dimensional NMR spectra of synthetic [GGG]kalata B1[T], two 111 isomers of [GG]kalataB1[T] derived from bacterial expression. Figure 7 RP-HPLC profile of oxidation of cyclic-[GG]kalata B1[T]. 112 Figure 8 Comparison of one-dimensional spectra between the synthetic 113 [GGG]kalata B1[T] and refolded [GGG]kalata B1[T] after expression. Chapter 6

Figure 1 Sequence of MCoTI-I and MCoTI-HA and schematic illustration of 122 conjugation strategy Figure 2 Schematic representation of other enzymatic modification strategies 123 Figure 3 Schematic representation of peptide modification strategies 126 Figure 4 DBCO NHSester modification on MCoTI-HA 126 Figure 5 VHH7 and its conjugation to cyclotides 127 Figure 6 Structure comparison of MCoTI-I and MCoTI-HA 129 Figure 7 Conjugation to VHH7 made MCoTIs more immunogenic 138 Figure 8 Antisera reactivity against MCoTI-I and linear HAtag 139

Figure 9 Reactivity between MCoTI-HA and commercially available anti HA 140 antibody Table 1 Chemical shift of MCoTI-I 130 Table 2 Chemical shift of MCoTI-HA 131 Table 3 Temperature coefficient of MCoTI-I 132 Table 4 Temperature coefficient of MCoTI-HA 133 Table 5 Dihedral angle restraints for MCoTI-I and MCoTI-HA 134

Table 6 Energies and structural statistics for ensembles of 20 structures with the 136 lowest energy for MCoTI-I and MCoTI-HA Table 7 RMSD values of each loop of MCoTI-I and MCoTI-HA 137

xvii

List of Abbreviations used in the thesis

Acm: acetamidomethyl ACN: acetonitrile APCs: antigen presenting cells Boc: tert-butyloxycarbonyl CCK: cyclic cystine knot COSY: correlation spectroscopy CTR: C-terminal region cVc1.1: cyclic Vc1.1 DBCO: dibenzylcyclooctyne DCM: dichloromethane DIPEA: N,N-diisopropylethylamine DMF: Dimethylformamide DQF-COSY: double quantum-filtered COSY DTT: dithiothreitol

D2O: deuterium oxide E.coli: Escherichia coli E.COSY: exclusive COSY ER: endoplasmic reticulum Fmoc: fluorenylmethyloxycarbonyl HATU: N,N,N’,N’-tetramethyl-O-(7-azabenzotriazol-1-yl)uronium hexafluorophosphate HSQC: heteronuclear single quantum coherence spectroscopy ICK: inhibitor cystine knot IPTG: β-D-thiogalactopyranoside Kv: voltage-gated potassium Oak1: Oldenlandia affinis kalata B1 precursor poly I:C: polyinosinic:polycytidylic acid RMSD: root mean square deviations RMSF: root mean square fluctuation RP-HPLC: reversed-phase high-pressure liquid chromatography SDS-PAGE: sodium dodecyl sufate polyacrylamide gel electrophoresis Sec: secretory

xviii

SFTI-1: sunflower trypsin inhibitor 1 SPPS: solid-phase peptide synthesis SRP: signal recognition particle SrtA: sortase A SrtA5º: mutated sortase A (P94R/D160N/D165A/K190E/196T) LC-MS: liquid chromatography-mass spectrometry MALDI: matrix-assisted laser desorption/ionization MBP: maltose-binding protein MHC: major histocompatibility complex MMP-2: matrix metalloproteinase-2 MBHA: Rink amide 4-methylbenzhydrylamine Ni-NTA: nickel nitrilotriacetic acid NCL: native chemical ligation NHS: N-hydroxysuccinimide NMR: nuclear magnetic resonance NTR: N-terminal repeat Tat: twin-arginine translocation TFA: Trifluoroacetic acid TIPS: Triisopropylsilane TOCSY: total correlation spectroscopy TOF: time-of-flight TCEP: tris(2-carboxyethyl)phosphine 2-CTC: 2-chlorotrityl chloride

xix

Chapter 1 Introduction

Chapter 1. Introduction

1.1. Overview Peptides play critical roles as hormones, neurotransmitters or ligands targeting a wide range of receptors. For example, peptide neurotoxins have attracted attention as potential drug candidates because of their inhibitory activity, coupled with high specificity and potency, against ion channels. However, peptides are generally susceptible to proteolysis, and are typically unstable in vivo and have poor bioavailability. Another problem for peptides as therapeutics is that they go through renal clearance rapidly due to their small size (<5 kDa). To overcome such obstacles, which limit the potential of peptides in drug development, a number of strategies have been introduced to chemically engineer them, including PEGylation (1,2), conjugation (3,4) or incorporation of unnatural amino acids (5-7). In addition, cyclization of peptides has been showing promising results of improved stability, compared to their linear counterparts. It is believed that protecting the N- or C- terminus of a peptide from exposure to proteases makes peptides resistant (8-10). Although there are many approaches to cyclize a peptide utilizing side chains of amino acids, this thesis will focus on head-to-tail backbone cyclization.

The concept for an engineering strategy is often inspired by nature and cyclization is no exception. Naturally occurring cyclic peptides are found in bacteria, fungi, plants and animals and are typically very stable due to their unique conformation. Of the cyclic peptides present in nature, ribosomally synthesized cyclic peptides are expressed as gene-encoded precursors and post- translationally cyclized. (11,12). They belong to the cyclic ribosomally synthesized and post- translationally modified peptides (RiPPs) family of proteins (13), which exist in all kingdoms of life (13) and exist in all kingdoms of life. Bacterial-derived circularin A and gassericin A (14), animal origin of θ-defensins (15,16) and plant-derived cyclotides are well-known examples. In general, they are significantly more resistant to proteases than their linear counterparts (17). There are a variety of biosynthetic mechanisms to generate the cyclic peptides, but cyclic backbones are typically formed via head-to-tail cyclization (12), thioether crosslinks (18), or cysteine sulfur to α- carbon crosslinks (19). Nonribosomal cyclic peptides are also synthesized via several strategies (20). For example, gramicidin S is cyclized via oligomerization, bacitracin is cyclized between an amino acid branch and the C-terminal peptide end, and cyclosporine is head-to-tail cyclized (20). The backbone of cyclosporine is also N-methylated via post-translational modification involving

Chapter 1. Introduction

complex enzyme cascades during non-ribosomal synthesis. Notably, cyclosporine, which is a currently marketed immunosuppressant, is orally bioavailable (21).

This chapter reviews disulfide-rich cyclic peptides focusing on cyclotides in particular. Cyclotides are plant-derived and remarkably stable due to a complex topology known as a cyclic cystine knot (CCK). This chapter also describes the use of cyclotides in drug design as well as how their unique topology has been inspiring chemical, structural engineering of peptide nerotoxins via cyclization. Among various studies published in the last decade on the usefulness of cyclization, cyclization of conopeptides and scorpion toxin is specifically addressed. Finally, this chapter also introduces sortase A which has been employed as an alternative tool to engineer peptides throughout research described in this thesis.

Declaration of authorship

In accordance with the guidelines of the University of Queensland, this chapter includes a part of a published, peer-reviewed review article.

"Discovery and applications of disulfide-rich cyclic peptides" Cemazar M, Kwon S, Mahatmanto T, Ravipati AS, Craik DJ. (2012) Current Topics in Medicinal Chemistry 12:1534-45

This review describes the research that has been published on disulfide-rich peptides including cyclotides, conotoxins, and chlorotoxin. The manuscript was written and figures were prepared by Soohyun Kwon, Tunjung Mahatmanto, Anjaneya Swamy Ravipati and revised by Masa Cemazar and David Craik. This chapter includes the part describing conotoxins and chlorotoxin.

Chapter 1. Introduction

1.2. Cyclotides 1.2.1. Discovery, structure and biological activities of cyclotides Cyclotides were initially discovered by a Norwegian doctor, Lorents Gran, who identified kalata B1 in a traditional medicine taken by Congolese women to facilitate childbirth during labor (22). Kalata B1 consists of ~30 amino acids forming a unique structural motif called a cyclic cystine knot (CCK) (23) which is characterized by a head-to-tail backbone and three disulfide bonds in a knotted topology, i.e. two disulfide bonds forming a ring and penetrated by a third disulfide bond (23,24). Since the discovery of kalata B1, a number of related plant-derived cyclic peptides were independently discovered during the 1990s and categorized with the family name ‘cyclotide’(23).

Cyclotides are divided into three sub-groups, Möbius, bracelet and trypsin inhibitors (Figure 1A). The Möbius group, including kalata B1, possess a conceptual twist formed by a cis-Pro peptide bond in loop 5 of the cyclic backbone while the bracelet group, typified by cycloviolacin O1, do not have such a twist. The third small sub-group, which is consisted of trypsin inhibitors such as MCoTIs identified from Momordica cochinchinensis. Cyclotides within a sub-family share a common structural characteristic as well as their sequences. On average, Möbius possess a 40% hydrophobic surface whereas bracelet cyclotides possess a 57% hydrophobic molecular surface (25).

The native role of cyclotides in plants is believed to be as a defense against insects (26,27) and other pests (28). However, cyclotides have also been reported to posssess pharmaceutically beneficial activities including antiviral (29), antibacterial (30), antifungal (31), antifouling (32), and cytotoxic (33) properties. These activities are thought to be related to the binding and disruption of lipid membranes in the case of kalata B1 (34-36).

1.2.2. Biosynthesis of cyclotides Ribosomally synthesized cyclotides originate from their precursor proteins, which consist of an endoplasmic reticulum (ER)-targeting sequence, a pro-region, a highly conserved N-terminal repeat (NTR) region, a cyclotide domain, and a C-terminal tail (Figure 1B). The cyclotide domain may include multiple copies of sequences that are the same or different, but multiple copies are usually separated by additional NTR sequences. Cyclization is known to take place during posttranslational processing, but many details of the process have not yet been completely elucidated. However, it has been recently confirmed that a conserved Asn or Asp at the C-terminal cleavage 4

Chapter 1. Introduction

site is subjected to catalysis of asparaginyl endoproteinase (AEP) or butelase (37-40). AEP was originally reported to be involved in post-translational peptide bond formation to process concanavalin A from jackbeans (41), and plants transformed to express a cyclotide precursor produced a decreased amount of cyclotides when AEP was inhibited (37).

Figure 1. Cyclotides from three categories and biosynthesis (A) kalata B1 (PDB ID: 1NB1) from Möbius (42), cycloviolacin O1 (PDB ID: 1NBJ) from bracelet (42), and MCoTI-II (PDB ID: 1IB9) from trypsin inhibitors (43), (B) Schematic representation of AEP-catalyzed cyclization of cyclotide precursor. This figure is adapted from Sancheti et al (44).

1.2.3. Chemical biology and applications of cyclotides Cyclotides comprise six loops divided by cysteines in the sequence (Figure 2). Certain loops are more flexible than others implying they can accommodate addition or modification of amino acids without perturbation of the overall structure of the cyclotide scaffold. Therefore, cyclotides have been re-engineered to gain a selected biological function or remove an unwanted natural function (45). The engineering can be as simple as a single residue deletion/insertion, or grafting of a bioactive peptide epitope into a loop (46,47). One motivation for doing this is that the cyclotide scaffold shields the linear peptide epitope from proteolysis. Activities conferred to a cyclotide via epitope grafting include anti-angiogenic (48), pro-angiogenic (49), antiviral (50), and cytomodulatory activities (51). The cyclotide scaffold and selected loops for grafting are shown in Figure 2. 5

Chapter 1. Introduction

One of the breakthroughs recently achieved by an epitope grafted cyclotide was accessing an intracellular target, as such targets are typically difficult to address with peptide or protein therapeutics (52). Ji et al. chose MCoTI-I as a scaffold to accommodate a 15 amino acid, α-helical peptide which can modulate protein-protein interaction to activate p53 tumor suppressor pathway. To maintain the α-helical conformation of the epitope they placed a linker derived from apamin, a bee-venom neurotoxin, which had been previously used to correctly display p53-derived helical peptides (52,53). Overall, these proof-of-concept experiments provide support for generalization of the concept that cyclotides have high potential to be a drug scaffold.

Figure 2. Applications of cyclotide as scaffolds. (A) Kalata B1 was used as a template for grafting a vascular endothelial growth factor-A (VEGF-A) antagonist sequence in place of a Pro residue in loop 3 (highlighted in circle). The epitope was stabilized and active upon grafting without perturbing the structure of kalata B1, thus illustrating the tolerance of the CCK motif to size modification (48), (B) MCoTI-II has been used as a scaffold for several grafting applications. The trypsin inhibitory site residue lysine (K, highlighted in circle) was replaced with Ala or Val and shown to redirect the activity of the native scaffold towards human leukocyte elastase (HLE) (50); and when substituted to Glu, the resultant MCoTI-II analogue was active against foot and mouth disease virus 3C protease (FMDV 3Cpro) (50). Another study demonstrated that grafting extracellular matrix protein epitopes onto loop 6 6

Chapter 1. Introduction

imparted MCoTI-II with angiogenic activity (49). Together, these studies demonstrate the use of cyclotides as carriers for delivering a desired activity of a short peptide sequence. The large gray arrows indicate the grafting sites; the small black arrow indicates the cyclization site in native cyclotide biosynthesis. The 3D structures of kalata B1 (PDB 1NB1) and MCoTI-II (PDB 1IB9) were generated using YASARA v11.9.18. Figure taken from Cemazar et al (47).

1.3. Conotoxins Conotoxins are peptides of 12–35 amino acids produced in the venoms of marine cone snails from the Conus genus (54). Conoserver (http://www.conoserver.org) provides extensive information on more than 1000 of these peptides (55). Their classification is based on gene superfamily, cysteine frameworks and pharmacological activities (56), and they are named using acronyms that include a Greek letter referring to the pharmacological family, a capital letter from the name of the source Conus species, and a Roman number indicating the cysteine framework, followed by a capital letter to show the order of discovery. Conotoxin κ-PVIIA, for example, is a conotoxin targeting voltage- gated K+ channels that originated from Conus purpurascens (57). Another example, conotoxin ω- MVIIA (PRIALT®), was approved by the US Food and Drug Administration in 2004 for the treatment of chronic pain (58). Despite its highly selective mechanism against N-type voltage- sensitive calcium channels, it is administered by direct infusion into the spinal cord, which is invasive for patients, costly, and has a risk of infection. Therefore, approaches were sought for the oral delivery of bioactive conotoxins. In this work, structural studies of a wide range of conotoxins have provided a basis for the design of re-engineered drug leads (59-62).

First, backbone cyclization was proven to improve stability against proteases using the α-conotoxin MII (8). Cyclization by joining the ends of the peptide, with a linker consisting of six alanine and glycine residues, resulted in higher stability with no loss of bioactivity. The structures of linear and cyclized MII are shown in (Figure 3A). This work was subsequently extended the cyclization approach to another α-conotoxin, Vc1.1. Cyclizing Vc1.1 with a 6-amino acid linker, without perturbation of the original structure (Figure 3B), resulted in more potent selective inhibition of calcium channels via a GABAB receptor dependent pathway. Furthermore, its stability in simulated intestinal fluid and human serum was increased relative to the linear counterpart. CyclicVc1.1 was tested in a chronic constriction injury rat model of neuropathic pain and found to be more than two orders of magnitude more potent than the clinically used gold standard for neuropathic pain, gabapentin (9).

Chapter 1. Introduction

Figure 3. Linear and cyclic α-conotoxin MII and Vc1.1. (A) Linear MII (PDB 1MII) and cyclized MII (PDB 2AKO) by 7 amino acid linker (8), (B) Linear Vc1.1 (PDB 2H8S) and its cyclized form by 6 amino acid linker (9). Figure taken from Cemazar et al (47).

1.4. Chlorotoxin Chlorotoxin is a 36 amino acid peptide isolated from the venom of the Israeli scorpion Leiurus quinquestriatus quinquestriatus. Its disulfide-rich knotted topology contributes to its high stability (63). Its name originated from the fact that it inhibits chloride channels in embryonic rat brain and epithelia (64), and interest in the molecule was particularly sparked by the discovery that it binds to glioma cells (65-67). Chlorotoxin has subsequently been shown to interact with a number of other tumor cell types, including medulloblastomas, neuroblastomas, ganglioneuromas, melanomas, pheochromocytomas and lung carcinomas (68). The tumor binding of chlorotoxin is supported by its interaction with matrix metalloproteinase-2 (MMP-2) isoforms, which are upregulated in gliomas and other cancers (69).

Chlorotoxin uptake by glioma cells is different to its entry into normal cells. It enters glioma cells through a clathrin-mediated pathway and localizes near the trans-Golgi apparatus, whereas in normal cells it enters via micropinocytosis and localizes in the cytoplasm (70). The difference in uptake is advantageous for the development of glioma therapeutics (70) and synthetic chlorotoxin

Chapter 1. Introduction

(TM-601) was conjugated to iodine-131 (131I) and administered intracavitarily to patients having recurrent glioma in recent clinical trials. 131I-TM-601 showed an anti-tumor effect and was well tolerated without dose-limiting toxicity (71).

Based on its specific binding ability to brain tumors, chlorotoxin has been conjugated to the near- infrared fluorescent dye Cy5.5 to visualize tumors during surgery (72,73). The dye conjugates to lysine, which chlorotoxin has three of (Lys15, 23, 27) resulting in 75-85% mono-labeled peptide at Lys27 along with di- and tri-labeled side products. Ideally, monolabeled compound is preferred for commercialization and in a recent study, linear chlorotoxin was successfully mono-labeled by substitution of Lys15 and Lys23 with alanine or arginine (74). The study also demonstrated that cyclized chlorotoxin, using a 7-amino acid linker without any lysine substitution, produced only mono-conjugation of the dye, providing an unexpected advantage of cyclisation. A schematic illustration of the cyclization and conjugation of chlorotoxin to Cy5.5 is shown in (Figure 4). Cyclization resulted in enhanced serum stability compared to the linear form of chlorotoxin when conjugated with Cy5.5.

Recently, other targets of chlorotoxin have been reported, including proliferating vascular endothelial cells, where chlorotoxin can play a role in inhibiting angiogenesis. The anti-angiogenic activity was not obstructed in the presence of angiogenesis stimulating factors such as VEGF (vascular endothelial growth factor), whereas the activity of the current anti-VEGF antibody therapeutic, bevacizumab, was reduced (75). Co-administration of chlorotoxin and bevacizumab significantly decreased blood vessel formation in tumors (75). Annexin A2 was recently proposed to be another potential target for chlorotoxin on tumor cell surfaces (76). Annexin A2 is expressed in a wide range of tumor cells and plays a role in tumor recurrence (77). It is also expressed in vascular endothelial cells and a part of regulation of angiogenesis (78). Taking its multi-target activity into consideration, chlorotoxin has great potential to be used in a variety of ways as a therapy for cancer. The cyclic derivatives are currently being developed in our laboratory offer potential for enhancing these applications.

Chapter 1. Introduction

Figure 4. Schematic illustration of chlorotoxin (PDB 1CHL) cyclization and conjugation with cy5.5 (A) Sequence of chlorotoxin including 7-amino acid linker and its disulfide bond connectivity (B) Natural linear chlorotoxin results in mono-, di-, and tri-labeled products when conjugated with Cy5.5 whereas cyclic chlorotoxin is only mono-labeled (74). Figure taken from Cemazar et al (47).

1.5. Sortase A 1.5.1. Discovery, recombinant expression, and structure of sortase A Sortase A (SrtA) is a transpeptidase consisting of 206 amino acids. Its mechanism of action was found in the process of understanding how Staphylococcus aureus displays cell surface proteins.

These proteins present a NH2-terminal domain on the cell surface, while anchoring the COOH- terminus including conserved LPXTG motif to the bacterial cell wall (79). The LPXTG motif has an crucial role in the anchoring mechanism of the staphylococcal cell surface protein as it is a specific site to be recognized by SrtA. Transpeptidase SrtA cleaves the peptide bond between Thr and Gly in the LPXTG motif so that Thr can form a new peptide bond with pentaglycine crossbridge in the cell wall (80,81). In the process, a thioacyl intermediate is found as a single cysteine (Cys184) on the active site of SrtA forms a thioester with the threonyl carboxyl end. (80- 82). The intermediate subsequently is resolved by the amino group of pentaglycine crossbridge in the cell wall via a new peptide bond.

Chapter 1. Introduction

Understanding the rationale of activity and sequence of SrtA enabled its recombinant expression which is subsequently led to its structure determination. Recombinant expression was possible by removing 59 amino acids on the N-terminus including the transmembrane domain of native SrtA, and a fully functional recombinant SrtA was achieved via bacterial cell culture with high yield (82,83). The structure of SrtA, resolved using nuclear magnetic resonance and X-ray crystallography (84,85) revealed eight stranded β-barrel folds. Histidine in the active site of SrtA was initially reported to form a thiolate-imidazolium ion pair with cysteine, and participate in nucleophilic attack by protonating the amide atom of the substrate peptide for Cys/Ser proteases to act (84). Although Cys184 and His120 are conserved in all sortases, later studies found the histidine in the active site of SrtA was uncharged, and thus, does not function as a protonator/deprotonator (85,86). Instead Arg197 is positioned in a catalytic dyad with Cys184. In the same study, calcium binding sites were also determined based on other cysteine proteases (85).

1.5.2. Application of sortase A The first application of SrtA was anchoring heterologous proteins to the cell wall of Gram-positive bacteria via a genetic fusion of the sorting motif to the protein. The heterologous protein decorates the bacterial cell surface with help from endogenous SrtA and this led to an understanding of the protein sorting and anchoring mechanism in bacteria and the application of bacteria in vaccination (79,87,88).

Due to its site-specific activity and ready recombinant expression, SrtA has been widely used in protein engineering over the last decade. It has been especially useful in labeling proteins at their N- or C- terminus (83). In the case of C-terminal labeling of a protein, the LPXTG motif is placed at the C-terminus and a probe containing a NH2-Glyn is required (Figure 5A). Similarly, for N- terminal labeling, a protein to be labeled requires Glyn at its N-terminus and the LPXTG motif needs to be added to the probe. Such SrtA-mediated N- or C-terminal labeling can overcome limitations of traditional genetic approaches. While these genetic approaches allow two protein fusion in a head-to-tail fashion, the site-specific SrtA reaction allows unusual protein fusion techniques such as N-to-N and C-to-C fusion in combination with click chemistry by incorporating cyclooctyne and azide for conjugation (89). There are two advantages when using this tool. Two of the same or two different proteins can be folded prior to conjugation, whereas recombinant 11

Chapter 1. Introduction

expression following genetic fusion may result in incorrectly folded side products requiring a refolding step. In addition, fusion proteins achieved via this approach can possess bispecific functions (89).

Finally, SrtA can be used in head-to-tail backbone cyclization in protein or peptide engineering (90- 93) (Figure 5B). Backbone cyclization is known to improve the in vivo half-life of proteins (12). To date, this cyclization has mostly been achieved by native chemical ligation (NCL) (94,95), or by split-intein methods performed in conjunction with NCL (96-98). Whereas NCL and split-intein methods require chemical modifications to generate a thioester linker, SrtA-mediated cyclization requires a modest modification with an N-terminal oligoglycine nucleophile and a C-terminal LPXTG motif. The high efficiency and general applicability of SrtA-mediated cyclization was confirmed in initial studies on a variety of proteins, including Cre recombinase (~41 kDa), eGFP (~30 kDa), the loop of the ubiquitin C-terminal hydrolase UCHL3 (~26 kDa), and p97, which is a hexameric AAA-ATPase (~97 kDa) (90). In addition, Popp et al. showed that a cyclized cytokine was more resistant to thermal denaturation than its linear counterpart (93), and that SrtA-mediated cyclization could improve the therapeutic properties of some proteins (93). Indeed, the method was subsequently applied to the 38-mer peptide, histatin, with the cyclic version showing significantly improved wound closure activity (91). The study also pointed out factors to be considered when cyclizing peptides. First, active domains should not be located at N- or C-termini to maintain biological activities of peptides. Second, peptide ligands should adopt their bioactive conformations, i.e., the conformation that occurs after initial binding to the receptor (91,99). Third, the minimum length for linear SrtA substrates is 12 amino acids, excluding the sorting motif and glycine nucleophile (100). Finally, the length and region of the protein fragment used for cyclization can affect its activity (101). In addition, the minimum length for linear SrtA substrates is 12 amino acids excluding the sorting motif and glycine nucleophile (100), optimization of the length and region of protein fragment for cyclization resulted in improved or reduced activity (101). The summary of peptides and proteins cyclized using SrtA in recent years is shown in Table 1.

Chapter 1. Introduction

Figure 5. Schematic representation of SrtA-mediated protein engineering (A) C-terminal labeling of a protein, (B) SrtA-mediated cyclization requires both sorting motif and nucleophilic Gly in the sequence of a protein.

Chapter 1. Introduction

Table 1. Proteins and peptides that have been cyclized using SrtA (adapted from van ‘t Hof et al (102)).

Substrate Description Structure Size Ref.

G-Cre-LPETG- human recombinase Two distinct domains ~41 kDa (90,103) His6 forming a C-shaped clamp G5-eGFP- Enhanced green the N and C termini were ~30 kDa (90) LPETG-His6 fluorescent protein positioned on the same end of the β-barrel G-His6-p97- human hexameric N and C termini of ~97 kDa (90) LPETG-XX AAA-ATPase adjacent p97 units were

sufficiently close to permit covalent crosslinking G-His6-UCHL3- human ubiquitin C- sortase recognition site ~26 kDa (90) LPETG-XX terminal hydrolase installed in the crossover loop GG-GCSF-3- human granulocyte A four-helix bundle ~20 kDa (93)

recombinant protein LPETG-His6 colony-stimulating structure factor GG-IFNα2- human interferon-α A four-helix bundle ~20 kDa (93) LPETG-His6 structure GG-EPO- human ∼40% of its mass is ~20 kDa (93) LPETG-His erythropoietin 6 composed of bulky, charged N-linked glycans GG-histatin 1- human salivary Lack of order in solution ~ 5 kDa (91) LPETGG peptide structure GG- human MUC1 N/A ~ 2 kDa (100) VTGalNacSGalNac glycopeptide APDTGalNac- LPKTGGS GGG-kalata B1- plant cyclotide Three disulfide bonds, ~3 kDa (92) TGG cyclic cystine knot G-αVc1.1- marine cone shell Two disulfide bonds, ~2.5 kDa (92) GLPETGGS neurotoxin globular conformation GG-SFTI-1- sunflower trypsin One disulfide bond ~2.2 kDa (92) LPETGG inhibitor cyclic backbone GG-LFcinB17-41- bovine lactoferricin N/A ~ 4 kDa (101)

synthetic or recombinant peptide or recombinantsynthetic LPETGG GGG-MCoTI-II- plant cyclotide Three disulfide bonds, ~3.5 kDa (104) LPETGG cyclic cystine knot

Chapter 1. Introduction

1.5.3. Mutagenesis of SrtA and improvement of SrtA reaction SrtA requires Ca2+ to be activite (84). In the study of Ilangovan et al., recombinant SrtA required about 2 mM of Ca2+ for activity whereas other divalent ions including Fe2+, Zn2+, Cd2+, and Co2+ or monovalent K+ cation resulted in SrtA not being able to cleave peptide bonds. The Ca2+ presumably binds to the β3-β4 loop and β6-β7 loop where glutamic acids with negative electrostatic surface potential exists in a structurally ordered site (84). However, how Ca2+ activates SrtA in vivo is not fully understood. Since the Ca2+-dependency can be an obstacle to using SrtA in an environment with low Ca2+ concentration such as the cytoplasm, or in the presence of Ca2+ binding compounds such as phosphate, carbonate, and ethylendiaminetetraactic acid (EDTA), Ca2+-independent SrtA was produced by making mutations on Ca2+ binding glutamic acids located on the β3-β4 loop (E105K/E108A or E105K/E108Q) (105).

Another potential drawback of SrtA reaction is its poor catalytic efficiency. A strategy using a yeast-display system for evolving catalytic activity of the enzyme was applied to SrtA (106). This strategy isolated SrtA mutants with significantly improved Km values from 7.6 mM to 170 μM for the LPXTG motif (106), implying improved applicability of SrtA in protein engineering.

Another characteristic of SrtA is that it functions as a hydrolase when there is no nucleophilic peptidoglycan or its mimics (107). Given that the product of the SrtA transpeptidation incorporates the LPXTG motif which can be cleaved by SrtA, the hydrolysed side-product with a LPXT sequence has been a drawback in utilizing SrtA in protein engineering. Furthermore, the release of the Gly amine nucleophile from the original LPXTG substrate cleavage makes the SrtA reaction reversible (Figure 6A). Therefore, several studies have investigated how to overcome this reversibility of the SrtA reaction. Firstly, Li et al. (108) used a new nucleophile probe instead of Gly so that the SrtA reaction would not result in re-formation of the LPXTG motif in the product. Using hydrazine as a probe in the same SrtA reaction buffer, substrate proteins become hydrazinolysed thus avoiding recognition by SrtA (Figure 6B). Secondly, Row et al. found a way to improve efficiency of the SrtA reaction via removal of cleaved Gly from the LPXTG motif (109). Based on the fact that GGH has a high affinity for Ni2+ when present on the N-terminus of proteins or peptides (110), they used protein substrates equipped with a LPETGGH motif for SrtA in the presence of NiSO4 (Figure 6C) and observed the yield of final products was increased. Other approaches to overcome reversibility of the SrtA reaction include employing depsipeptides (Figure

Chapter 1. Introduction

6D) (111,112), a β-hairpin formation through the introduction of a Trp zipper around the LPXTG motif (Figure 6E) (113).

Figure 6. Methods to improve the efficiency of the SrtA reaction.

1.6. Summary and outlook Cyclotides are topologically very interesting as they are the only class of proteins known to contain the CCK motif. This motif confers them with exceptional stability and thus potentiates their applications as drug scaffolds. Their ability to tolerate sequence substitutions and modification has led to a range of cyclotide grafting studies, from which potential peptide drugs have been designed and demonstrated to have high stability and potency. Although none of the cyclotides has yet reached the clinic, new approaches to produce or engineer them using an enzyme like sortase A could add more favorable biopharmaceutical properties to cyclotides and inspire additional applications to other classes of disulfide-rich peptides.

Chapter 1. Introduction

1.7. Scope of this thesis The broad aim of this thesis is to explore the application of sortase A in disulfide-rich peptide engineering, and particularly cyclotides and conotoxins were studied because of their properties such as exceptional stability and specific activities that are interesting from pharmaceutical perspectives.

Chapter 1 of this thesis provides the background information to define the molecules studied and to highlight their importance. Chapter 2 describes the main experimental techniques that were used in this thesis, including solid-phase peptide synthesis (SPPS) and nuclear magnetic resonance spectroscopy.

The synthesis of the cyclic peptides has been mostly achieved via native chemical ligation (NCL), which requires anhydrous hydrogen fluoride (HF) to release the peptide from the resin that was assembled based on Boc-SPPS and to remove their side chain protecting groups. Despite the advantages of Boc-SPPS, Fmoc-SPPS became more widespread than Boc-SPPS as it is generally safer by excluding handling HF in the process. To find an alternative way to produce cyclic peptides, Chapter 3 reports semi-enzymatic cyclization of disulfide-rich peptides. Following successful Fmoc-SPPS of several peptides that are not related and with different number of disulfide bonds, cyclization was achieved using srtA. Each peptide had LPXTG motif at their C- terminus that SrtA can recognize and catalyze transpeptidation with oligoglycine equipped at the N- terminus of the peptide.

Encouraged by the successful sortase A-mediated cyclization, Chapter 4 reports on effective conversion of an ICK motif to a CCK motif using sortase A. κ-conotoxin PVIIA, which blocks Shaker K+ channel of Drosophila melanogaster and possesses an ICK structural motif was cyclized and its structure and activity relationships were analyzed. Cyclization of chlorotoxin, which comprises an ICK motif with an extra disulfide bond, was also studied as another example in this application.

Chapter 5 explores the bacterial expression of a cyclotide to overcome a drawback found in the method reported in Chapter 3 and 4. As the semi-enzymatic method required separate steps for synthesis, folding and cyclization in which purification is necessary in between, a molecular biological approach was employed to express folded kalata B1 linear precursor in the bacterial periplasm which is an oxidative environment.

Chapter 1. Introduction

In Chapter 6, sortase A was applied for delivery of cyclotides to dendritic cells to examine the potential of cyclotides as vaccination material. Although peptides are not generally immunogenic, a cyclotide, McoTI-I and its variant could elicit antibody response in mouse when they were conjugated to a dendritic cell specific antibody. In the conjugation process, sortase A was used to modify the antibody and the cyclotides were conjugated via click chemistry. The immunization study and its western blot analysis shows potential applications of antibodies elicited by cyclotides.

In summary, cyclization of disulfide-rich peptides and ligation of a cyclotide to another entity using SrtA have been investigated in this thesis. The work described in each chapter shows feasibility of the application of SrtA in engineering disulfide-rich peptides and a potential for bacterial expression of cyclotides.

1.8. References

1. Lee, S. H., Lee, S., Youn, Y. S., Na, D. H., Chae, S. Y., Byun, Y., and Lee, K. C. (2005) Synthesis, characterization, and pharmacokinetic studies of PEGylated glucagon-like peptide-1. Bioconjug. Chem. 16, 377-382

2. Na, D. H., Lee, K. C., and DeLuca, P. P. (2005) PEGylation of octreotide: II. Effect of N- terminal mono-PEGylation on biological activity and pharmacokinetics. Pharm. Res. 22, 743-749

3. Abraham, S., Guo, F., Li, L. S., Rader, C., Liu, C., Barbas, C. F., 3rd, Lerner, R. A., and Sinha, S. C. (2007) Synthesis of the next-generation therapeutic antibodies that combine cell targeting and antibody-catalyzed prodrug activation. Proc. Natl. Acad. Sci. U. S. A. 104, 5584-5589

4. Swee, L. K., Guimaraes, C. P., Sehrawat, S., Spooner, E., Barrasa, M. I., and Ploegh, H. L. (2013) Sortase-mediated modification of alphaDEC205 affords optimization of antigen presentation and immunization against a set of viral epitopes. Proc. Natl. Acad. Sci. U. S. A. 110, 1428-1433

5. Cheloha, R. W., Maeda, A., Dean, T., Gardella, T. J., and Gellman, S. H. (2014) Backbone modification of a polypeptide drug alters duration of action in vivo. Nat. Biotechnol. 32, 653-655

6. Hong, S. Y., Oh, J. E., and Lee, K. H. (1999) Effect of D-amino acid substitution on the stability, the secondary structure, and the activity of membrane-active peptide. Biochem. Pharmacol. 58, 1775-1780

7. Tugyi, R., Uray, K., Ivan, D., Fellinger, E., Perkins, A., and Hudecz, F. (2005) Partial D- amino acid substitution: Improved enzymatic stability and preserved Ab recognition of a MUC2 epitope peptide. Proc. Natl. Acad. Sci. U. S. A. 102, 413-418 18

Chapter 1. Introduction

8. Clark, R. J., Fischer, H., Dempster, L., Daly, N. L., Rosengren, K. J., Nevin, S. T., Meunier, F. A., Adams, D. J., and Craik, D. J. (2005) Engineering stable peptide toxins by means of backbone cyclization: stabilization of the alpha-conotoxin MII. Proc. Natl. Acad. Sci. U. S. A. 102, 13767-13772

9. Clark, R. J., Jensen, J., Nevin, S. T., Callaghan, B. P., Adams, D. J., and Craik, D. J. (2010) The engineering of an orally active conotoxin for the treatment of neuropathic pain. Angew. Chem. Int. Ed. Engl. 49, 6545-6548

10. Halai, R., Callaghan, B., Daly, N. L., Clark, R. J., Adams, D. J., and Craik, D. J. (2011) Effects of cyclization on stability, structure, and activity of alpha-conotoxin RgIA at the alpha9alpha10 nicotinic acetylcholine receptor and GABA(B) receptor. J. Med. Chem. 54, 6984-6992

11. Trabi, M., and Craik, D. J. (2002) Circular proteins - no end in sight. Trends Biochem. Sci. 27, 132-138

12. Craik, D. J. (2006) Chemistry - Seamless proteins tie up their loose ends. Science 311, 1563- 1564

13. Arnison, P. G., Bibb, M. J., Bierbaum, G., Bowers, A. A., Bugni, T. S., Bulaj, G., Camarero, J. A., Campopiano, D. J., Challis, G. L., Clardy, J., Cotter, P. D., Craik, D. J., Dawson, M., Dittmann, E., Donadio, S., Dorrestein, P. C., Entian, K. D., Fischbach, M. A., Garavelli, J. S., Goransson, U., Gruber, C. W., Haft, D. H., Hemscheidt, T. K., Hertweck, C., Hill, C., Horswill, A. R., Jaspars, M., Kelly, W. L., Klinman, J. P., Kuipers, O. P., Link, A. J., Liu, W., Marahiel, M. A., Mitchell, D. A., Moll, G. N., Moore, B. S., Muller, R., Nair, S. K., Nes, I. F., Norris, G. E., Olivera, B. M., Onaka, H., Patchett, M. L., Piel, J., Reaney, M. J., Rebuffat, S., Ross, R. P., Sahl, H. G., Schmidt, E. W., Selsted, M. E., Severinov, K., Shen, B., Sivonen, K., Smith, L., Stein, T., Sussmuth, R. D., Tagg, J. R., Tang, G. L., Truman, A. W., Vederas, J. C., Walsh, C. T., Walton, J. D., Wenzel, S. C., Willey, J. M., and van der Donk, W. A. (2013) Ribosomally synthesized and post-translationally modified peptide natural products: overview and recommendations for a universal nomenclature. Nat. Prod. Rep. 30, 108-160

14. Kawai, Y., Kemperman, R., Kok, J., and Saito, T. (2004) The circular bacteriocins gassericin A and circularin A. Curr. Protein Pept. Sci. 5, 393-398

15. Conibear, A. C., and Craik, D. J. (2014) The chemistry and biology of theta defensins. Angew. Chem. Int. Ed. Engl. 53, 10612-10623

16. Tang, Y. Q., Yuan, J., Osapay, G., Osapay, K., Tran, D., Miller, C. J., Ouellette, A. J., and Selsted, M. E. (1999) A cyclic antimicrobial peptide produced in primate leukocytes by the ligation of two truncated alpha-defensins. Science 286, 498-502

17. Cascales, L., and Craik, D. J. (2010) Naturally occurring circular proteins: distribution, biosynthesis and evolution. Org. Biomol. Chem. 8, 5035-5047

18. Knerr, P. J., and van der Donk, W. A. (2012) Discovery, biosynthesis, and engineering of lantipeptides. Annu. Rev. Biochem. 81, 479-505

Chapter 1. Introduction

19. Kawulka, K., Sprules, T., McKay, R. T., Mercier, P., Diaper, C. M., Zuber, P., and Vederas, J. C. (2003) Structure of subtilosin A, an antimicrobial peptide from Bacillus subtilis with unusual posttranslational modifications linking cysteine sulfurs to alpha-carbons of phenylalanine and threonine. J. Am. Chem. Soc. 125, 4726-4727

20. Sieber, S. A., and Marahiel, M. A. (2003) Learning from nature's drug factories: nonribosomal synthesis of macrocyclic peptides. J. Bacteriol. 185, 7036-7043

21. Veber, D. F., Johnson, S. R., Cheng, H. Y., Smith, B. R., Ward, K. W., and Kopple, K. D. (2002) Molecular properties that influence the oral bioavailability of drug candidates. J. Med. Chem. 45, 2615-2623

22. Gran, L. (1970) An oxytocic principle found in Oldenlandia affinis DC. Medd. Nor. Farm. Selsk. 12, 173-180

23. Craik, D. J., Daly, N. L., Bond, T., and Waine, C. (1999) Plant cyclotides: A unique family of cyclic and knotted proteins that defines the cyclic cystine knot structural motif. J. Mol. Biol. 294, 1327-1336

24. Saether, O., Craik, D. J., Campbell, I. D., Sletten, K., Juul, J., and Norman, D. G. (1995) Elucidation of the primary and three-dimensional structure of the uterotonic polypeptide Kalata B1. Biochemistry 34, 4147-4158

25. Pelegrini, P. B., Quirino, B. F., and Franco, O. L. (2007) Plant cyclotides: An unusual class of defense compounds. Peptides 28, 1475-1481

26. Gruber, C. W., Cemazar, M., Anderson, M. A., and Craik, D. J. (2007) Insecticidal plant cyclotides and related cystine knot toxins. Toxicon 49, 561-575

27. Jennings, C., West, J., Waine, C., Craik, D., and Anderson, M. (2001) Biosynthesis and insecticidal properties of plant cyclotides: The cyclic knotted proteins from Oldenlandia affinis. Proc. Natl. Acad. Sci. U. S. A. 98, 10614-10619

28. Colgrave, M. L., Kotze, A. C., Kopp, S., McCarthy, J. S., Coleman, G. T., and Craik, D. J. (2009) Anthelmintic activity of cyclotides: In vitro studies with canine and human hookworms. Acta Trop. 109, 163-166

29. Gustafson, K. R., Sowder, R. C., Henderson, L. E., Parsons, I. C., Kashman, Y., Cardellina, J. H., McMahon, J. B., Buckheit Jr, R. W., Pannell, L. K., and Boyd, M. R. (1994) Circulins A and B. Novel human immunodeficiency virus (HIV)-inhibitory macrocyclic peptides from the tropical tree Chassalia parvifolia. J. Am. Chem. Soc. 116, 9337-9338

30. Pranting, M., Loov, C., Burman, R., Goransson, U., and Andersson, D. I. (2010) The cyclotide cycloviolacin O2 from Viola odorata has potent bactericidal activity against Gram-negative bacteria. J. Antimicrob. Chemother. 65, 1964-1971

31. Tam, J. P., Lu, Y. A., Yang, J. L., and Chiu, K. W. (1999) An unusual structural motif of antimicrobial peptides containing end-to-end macrocycle and cystine-knot disulfides. Proc. Natl. Acad. Sci. U. S. A. 96, 8913-8918

Chapter 1. Introduction

32. Goransson, U., Sjogren, M., Svangard, E., Claeson, P., and Bohlin, L. (2004) Reversible antifouling effect of the cyclotide cycloviolacin O2 against barnacles. J. Nat. Prod. 67, 1287-1290

33. Svangard, E., Goransson, U., Hocaoglu, Z., Gullbo, J., Larsson, R., Claeson, P., and Bohlin, L. (2004) Cytotoxic cyclotides from Viola tricolor. J. Nat. Prod. 67, 144-147

34. Henriques, S. T., Huang, Y. H., Rosengren, K. J., Franquelim, H. G., Carvalho, F. A., Johnson, A., Sonza, S., Tachedjian, G., Castanho, M. A., Daly, N. L., and Craik, D. J. (2011) Decoding the membrane activity of the cyclotide kalata B1: the importance of phosphatidylethanolamine phospholipids and lipid organization on hemolytic and anti-HIV activities. J. Biol. Chem. 286, 24231-24241

35. Kamimori, H., Hall, K., Craik, D. J., and Aguilar, M. I. (2005) Studies on the membrane interactions of the cyclotides kalata B1 and kalata B6 on model membrane systems by surface plasmon resonance. Anal. Biochem. 337, 149-153

36. Hall, K., Lee, T. H., Daly, N. L., Craik, D. J., and Aguilar, M. I. (2012) Gly(6) of kalata B1 is critical for the selective binding to phosphatidylethanolamine membranes. Biochim. Biophys. Acta 1818, 2354-2361

37. Gillon, A. D., Saska, I., Jennings, C. V., Guarino, R. F., Craik, D. J., and Anderson, M. A. (2008) Biosynthesis of circular proteins in plants. Plant J. 53, 505-515

38. Harris, K. S., Durek, T., Kaas, Q., Poth, A. G., Gilding, E. K., Conlan, B. F., Saska, I., Daly, N. L., van der Weerden, N. L., Craik, D. J., and Anderson, M. A. (2015) Efficient backbone cyclization of linear peptides by a recombinant asparaginyl endopeptidase. Nat. Commun. 6, 10199

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

40. Saska, I., Gillon, A. D., Hatsugai, N., Dietzgen, R. G., Hara-Nishimura, I., Anderson, M. A., and Craik, D. J. (2007) An asparaginyl endopeptidase mediates in vivo protein backbone cyclization. J. Biol. Chem. 282, 29721-29728

41. Sheldon, P. S., Keen, J. N., and Bowles, D. J. (1996) Post-translational peptide bond formation during concanavalin A processing in vitro. Biochem. J. 320, 865-870

42. Rosengren, K. J., Daly, N. L., Plan, M. R., Waine, C., and Craik, D. J. (2003) Twists, knots, and rings in proteins - Structural definition of the cyclotide framework. J. Biol. Chem. 278, 8606-8616

43. Felizmenio-Quimio, M. E., Daly, N. L., and Craik, D. J. (2001) Circular proteins in plants - Solution structure of a novel macrocyclic trypsin inhibitor from Momordica cochinchinensis. J. Biol. Chem. 276, 22875-22882

44. Sancheti, H., and Camarero, J. A. (2009) "Splicing up" drug discovery. Cell-based expression and screening of genetically-encoded libraries of backbone-cyclized polypeptides. Adv. Drug. Deliv. Rev. 61, 908-917 21

Chapter 1. Introduction

45. Craik, D. J., Clark, R. J., and Daly, N. L. (2007) Potential therapeutic applications of the cyclotides and related cystine knot mini-proteins. Expert Opin. Investig. Drugs 16, 595-604

46. Barry, D. G., Daly, N. L., Clark, R. J., Sando, L., and Craik, D. J. (2003) Linearization of a naturally occurring circular protein maintains structure but eliminates hemolytic activity. Biochemistry 42, 6688-6695

47. Cemazar, M., Kwon, S., Mahatmanto, T., Ravipati, A. S., and Craik, D. J. (2012) Discovery and applications of disulfide-rich cyclic peptides. Curr. Top. Med. Chem. 12, 1534-1545

48. Gunasekera, S., Foley, F. M., Clark, R. J., Sando, L., Fabri, L. J., Craik, D. J., and Daly, N. L. (2008) Engineering stabilized vascular endothelial growth factor-A antagonists: Synthesis, structural characterization, and bioactivity of grafted analogues of cyclotides. J. Med. Chem. 51, 7697-7704

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

50. Thongyoo, P., Roqué-Rosell, N., Leatherbarrow, R. J., and Tate, E. W. (2008) Chemical and biomimetic total syntheses of natural and engineered MCoTI cyclotides. Org. Biomol. Chem. 6, 1462-1470

51. Agyei, D., and Danquah, M. K. (2011) Industrial-scale manufacturing of pharmaceutical- grade bioactive peptides. Biotechnol Adv 29, 272-277

52. Ji, Y., Majumder, S., Millard, M., Borra, R., Bi, T., Elnagar, A. Y., Neamati, N., Shekhtman, A., and Camarero, J. A. (2013) In vivo activation of the p53 tumor suppressor pathway by an engineered cyclotide. J. Am. Chem. Soc. 135, 11623-11633

53. Li, C., Pazgier, M., Liu, M., Lu, W. Y., and Lu, W. (2009) Apamin as a template for structure-based rational design of potent peptide activators of p53. Angew. Chem. Int. Ed. Engl. 48, 8712-8715

54. Olivera, B. M., Gray, W. R., Zeikus, R., McIntosh, J. M., Varga, J., Rivier, J., de Santos, V., and Cruz, L. J. (1985) Peptide neurotoxins from fish-hunting cone snails. Science 230, 1338-1343

55. Kaas, Q., Yu, R., Jin, A. H., Dutertre, S., and Craik, D. J. (2012) ConoServer: Updated content, knowledge, and discovery tools in the conopeptide database. Nucleic Acids Res. 40, D325-330

56. Kaas, Q., and Craik, D. J. (2010) Analysis and classification of circular proteins in CyBase. Biopolymers 94, 584-591

57. Scanlon, M. J., Naranjo, D., Thomas, L., Alewood, P. F., Lewis, R. J., and Craik, D. J. (1997) Solution structure and proposed binding mechanism of a novel potassium channel toxin kappa-conotoxin PVIIA. Structure 5, 1585-1597

58. Miljanich, G. P. (2004) Ziconotide: Neuronal calcium channel blocker for treating severe chronic pain. Curr. Med. Chem. 11, 3029-3040 22

Chapter 1. Introduction

59. Armishaw, C. J., Daly, N. L., Nevin, S. T., Adams, D. J., Craik, D. J., and Alewood, P. F. (2006) Alpha-selenoconotoxins, a new class of potent alpha7 neuronal nicotinic receptor antagonists. J. Biol. Chem. 281, 14136-14143

60. Marx, U. C., Daly, N. L., and Craik, D. J. (2006) NMR of conotoxins: structural features and an analysis of chemical shifts of post-translationally modified amino acids. Magn. Reson. Chem. 44 Spec No, S41-50

61. Nicke, A., Loughnan, M. L., Millard, E. L., Alewood, P. F., Adams, D. J., Daly, N. L., Craik, D. J., and Lewis, R. J. (2003) Isolation, structure, and activity of GID, a novel alpha 4/7- conotoxin with an extended N-terminal sequence. J. Biol. Chem. 278, 3137-3144

62. Skjaerbaek, N., Nielsen, K. J., Lewis, R. J., Alewood, P., and Craik, D. J. (1997) Determination of the solution structures of conantokin-G and conantokin-T by CD and NMR spectroscopy. J. Biol. Chem. 272, 2291-2299

63. Gelly, J. C., Gracy, J., Kaas, Q., Le-Nguyen, D., Heitz, A., and Chiche, L. (2004) The KNOTTIN website and database: A new information system dedicated to the knottin scaffold. Nucleic Acids Res. 32, D156-D159

64. DeBin, J. A., Maggio, J. E., and Strichartz, G. R. (1993) Purification and characterization of chlorotoxin, a chloride channel ligand from the venom of the scorpion. Am. J. Physiol. 264, C361-369

65. Soroceanu, L., Gillespie, Y., Khazaeli, M. B., and Sontheimer, H. (1998) Use of chlorotoxin for targeting of primary brain tumors. Cancer Res. 58, 4871-4879

66. Ullrich, N., Gillespie, G. Y., and Sontheimer, H. (1996) Human astrocytoma cells express a unique chloride current. Neuroreport 7, 1020-1024

67. Ullrich, N., and Sontheimer, H. (1996) Biophysical and pharmacological characterization of chloride currents in human astrocytoma cells. Am. J. Physiol. 270, C1511-1521

68. Lyons, S. A., O'Neal, J., and Sontheimer, H. (2002) Chlorotoxin, a scorpion-derived peptide, specifically binds to gliomas and tumors of neuroectodermal origin. Glia 39, 162-173

69. Deshane, J., Garner, C. C., and Sontheimer, H. (2003) Chlorotoxin inhibits glioma cell invasion via matrix metalloproteinase-2. J. Biol. Chem. 278, 4135-4144

70. Wiranowska, M., Colina, L. O., and Johnson, J. O. (2011) Clathrin-mediated entry and cellular localization of chlorotoxin in human glioma. Cancer Cell Int. 11, 27

71. Mamelak, A. N., Rosenfeld, S., Bucholz, R., Raubitschek, A., Nabors, L. B., Fiveash, J. B., Shen, S., Khazaeli, M. B., Colcher, D., Liu, A., Osman, M., Guthrie, B., Schade-Bijur, S., Hablitz, D. M., Alvarez, V. L., and Gonda, M. A. (2006) Phase I single-dose study of intracavitary-administered iodine-131-TM-601 in adults with recurrent high-grade glioma. J. Clin. Oncol. 24, 3644-3650

72. Veiseh, M., Gabikian, P., Bahrami, S. B., Veiseh, O., Zhang, M., Hackman, R. C., Ravanpay, A. C., Stroud, M. R., Kusuma, Y., Hansen, S. J., Kwok, D., Munoz, N. M., Sze, R. W., Grady, W. M., Greenberg, N. M., Ellenbogen, R. G., and Olson, J. M. (2007) Tumor 23

Chapter 1. Introduction

paint: a chlorotoxin:Cy5.5 bioconjugate for intraoperative visualization of cancer foci. Cancer Res. 67, 6882-6888

73. Veiseh, O., Sun, C., Gunn, J., Kohler, N., Gabikian, P., Lee, D., Bhattarai, N., Ellenbogen, R., Sze, R., Hallahan, A., Olson, J., and Zhang, M. (2005) Optical and MRI multifunctional nanoprobe for targeting gliomas. Nano Lett. 5, 1003-1008

74. Akcan, M., Stroud, M. R., Hansen, S. J., Clark, R. J., Daly, N. L., Craik, D. J., and Olson, J. M. (2011) Chemical re-engineering of chlorotoxin improves bioconjugation properties for tumor imaging and targeted therapy. J. Med. Chem. 54, 782-787

75. Jacoby, D. B., Dyskin, E., Yalcin, M., Kesavan, K., Dahlberg, W., Ratliff, J., Johnson, E. W., and Mousa, S. A. (2010) Potent pleiotropic anti-angiogenic effects of TM601, a synthetic chlorotoxin peptide. Anticancer Res. 30, 39-46

76. Kesavan, K., Ratliff, J., Johnson, E. W., Dahlberg, W., Asara, J. M., Misra, P., Frangioni, J. V., and Jacoby, D. B. (2010) Annexin A2 is a molecular target for TM601, a peptide with tumor-targeting and anti-angiogenic effects. J. Biol. Chem. 285, 4366-4374

77. Takano, S., Togawa, A., Yoshitomi, H., Shida, T., Kimura, F., Shimizu, H., Yoshidome, H., Ohtsuka, M., Kato, A., Tomonaga, T., Nomura, F., and Miyazaki, M. (2008) Annexin II overexpression predicts rapid recurrence after surgery in pancreatic cancer patients undergoing gemcitabine-adjuvant chemotherapy. Ann. Surg. Oncol. 15, 3157-3168

78. Ling, Q., Jacovina, A. T., Deora, A., Febbraio, M., Simantov, R., Silverstein, R. L., Hempstead, B., Mark, W. H., and Hajjar, K. A. (2004) Annexin II regulates fibrin homeostasis and neoangiogenesis in vivo. J. Clin. Invest. 113, 38-48

79. Schneewind, O., Model, P., and Fischetti, V. A. (1992) Sorting of protein A to the staphylococcal cell wall. Cell 70, 267-281

80. Navarre, W. W., and Schneewind, O. (1994) Proteolytic cleavage and cell wall anchoring at the LPXTG motif of surface proteins in gram-positive bacteria. Mol. Microbiol. 14, 115-121

81. Schneewind, O., Fowler, A., and Faull, K. F. (1995) Structure of the cell wall anchor of surface proteins in Staphylococcus aureus. Science 268, 103-106

82. Ton-That, H., Liu, G., Mazmanian, S. K., Faull, K. F., and Schneewind, O. (1999) Purification and characterization of sortase, the transpeptidase that cleaves surface proteins of Staphylococcus aureus at the LPXTG motif. Proc. Natl. Acad. Sci. U. S. A. 96, 12424- 12429

83. Popp, M. W., Antos, J. M., Grotenbreg, G. M., Spooner, E., and Ploegh, H. L. (2007) Sortagging: A versatile method for protein labeling. Nat. Chem. Biol. 3, 707-708

84. Ilangovan, U., Ton-That, H., Iwahara, J., Schneewind, O., and Clubb, R. T. (2001) Structure of sortase, the transpeptidase that anchors proteins to the cell wall of Staphylococcus aureus. Proc. Natl. Acad. Sci. U. S. A. 98, 6056-6061

Chapter 1. Introduction

85. Zong, Y. N., Bice, T. W., Ton-That, H., Schneewind, O., and Narayana, S. V. L. (2004) Crystal structures of Staphylococcus aureus sortase A and its substrate complex. J. Biol. Chem. 279, 31383-31389

86. Connolly, K. M., Smith, B. T., Pilpa, R., Ilangovan, U., Jung, M. E., and Clubb, R. T. (2003) Sortase from Staphylococcus aureus does not contain a thiolate-imidazolium ion pair in its active site. J. Biol. Chem. 278, 34061-34065

87. Nguyen, H. D., and Schumann, W. (2006) Establishment of an experimental system allowing immobilization of proteins on the surface of Bacillus subtilis cells. J. Biotechnol. 122, 473-482

88. Pozzi, G., Contorni, M., Oggioni, M. R., Manganelli, R., Tommasino, M., Cavalieri, F., and Fischetti, V. A. (1992) Delivery and expression of a heterologous antigen on the surface of streptococci. Infect. Immun. 60, 1902-1907

89. Witte, M. D., Cragnolini, J. J., Dougan, S. K., Yoder, N. C., Popp, M. W., and Ploegh, H. L. (2012) Preparation of unnatural N-to-N and C-to-C protein fusions. Proc. Natl. Acad. Sci. U. S. A. 109, 11993-11998

90. Antos, J. M., Popp, M. W., Ernst, R., Chew, G. L., Spooner, E., and Ploegh, H. L. (2009) A straight path to circular proteins. J. Biol. Chem. 284, 16028-16036

91. Bolscher, J. G., Oudhoff, M. J., Nazmi, K., Antos, J. M., Guimaraes, C. P., Spooner, E., Haney, E. F., Garcia Vallejo, J. J., Vogel, H. J., van't Hof, W., Ploegh, H. L., and Veerman, E. C. (2011) Sortase A as a tool for high-yield histatin cyclization. FASEB J. 25, 2650-2658

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

93. Popp, M. W., Dougan, S. K., Chuang, T. Y., Spooner, E., and Ploegh, H. L. (2011) Sortase- catalyzed transformations that improve the properties of cytokines. Proc. Natl. Acad. Sci. U. S. A. 108, 3169-3174

94. Iwai, H., Lingel, A., and Pluckthun, A. (2001) Cyclic green fluorescent protein produced in vivo using an artificially split PI-PfuI intein from Pyrococcus furiosus. J. Biol. Chem. 276, 16548-16554

95. Scott, C. P., Abel-Santos, E., Wall, M., Wahnon, D. C., and Benkovic, S. J. (1999) Production of cyclic peptides and proteins in vivo. Proc. Natl. Acad. Sci. U. S. A. 96, 13638- 13643

96. Camarero, J. A., Fushman, D., Sato, S., Giriat, I., Cowburn, D., Raleigh, D. P., and Muir, T. W. (2001) Rescuing a destabilized protein fold through backbone cyclization. J. Mol. Biol. 308, 1045-1062

97. Camarero, J. A., Kimura, R. H., Woo, Y. H., Shekhtman, A., and Cantor, J. (2007) Biosynthesis of a fully functional cyclotide inside living bacterial cells. Chembiochem 8, 1363-1366

Chapter 1. Introduction

98. Evans, T. C., Jr., Martin, D., Kolly, R., Panne, D., Sun, L., Ghosh, I., Chen, L., Benner, J., Liu, X. Q., and Xu, M. Q. (2000) Protein trans-splicing and cyclization by a naturally split intein from the dnaE gene of Synechocystis species PCC6803. J. Biol. Chem. 275, 9091- 9094

99. Zhou, H. X. (2010) From induced fit to conformational selection: a continuum of binding mechanism controlled by the timescale of conformational transitions. Biophys. J. 98, L15-17

100. Wu, Z., Guo, X., and Guo, Z. (2011) Sortase A-catalyzed peptide cyclization for the synthesis of macrocyclic peptides and glycopeptides. Chem. Commun. 47, 9218-9220

101. Arias, M., McDonald, L. J., Haney, E. F., Nazmi, K., Bolscher, J. G., and Vogel, H. J. (2014) Bovine and human lactoferricin peptides: chimeras and new cyclic analogs. Biometals 27, 935-948

102. van 't Hof, W., Hansenova Manaskova, S., Veerman, E. C., and Bolscher, J. G. (2015) Sortase-mediated backbone cyclization of proteins and peptides. Biol. Chem. 396, 283-293

103. Van Duyne, G. D. (2001) A structural view of cre-loxp site-specific recombination. Annu. Rev. Biophys. Biomol. Struct. 30, 87-104

104. Stanger, K., Maurer, T., Kaluarachchi, H., Coons, M., Franke, Y., and Hannoush, R. N. (2014) Backbone cyclization of a recombinant cystine-knot peptide by engineered Sortase A. FEBS Lett. 588, 4487-4496

105. Hirakawa, H., Ishikawa, S., and Nagamune, T. (2012) Design of Ca2+-independent Staphylococcus aureus sortase A mutants. Biotechnol. Bioeng. 109, 2955-2961

106. Chen, I., Dorr, B. M., and Liu, D. R. (2011) A general strategy for the evolution of bond- forming enzymes using yeast display. Proc. Natl. Acad. Sci. U. S. A. 108, 11399-11404

107. Ton-That, H., Mazmanian, S. K., Faull, K. F., and Schneewind, O. (2000) Anchoring of surface proteins to the cell wall of Staphylococcus aureus - Sortase catalyzed in vitro transpeptidation reaction using LPXTG peptide and NH2-Gly(3) substrates. J. Biol. Chem. 275, 9876-9881

108. Li, Y. M., Li, Y. T., Pan, M., Kong, X. Q., Huang, Y. C., Hong, Z. Y., and Liu, L. (2014) Irreversible site-specific hydrazinolysis of proteins by use of sortase. Angew. Chem. Int. Ed. Engl. 53, 2198-2202

109. Row, R. D., Roark, T. J., Philip, M. C., Perkins, L. L., and Antos, J. M. (2015) Enhancing the efficiency of sortase-mediated ligations through nickel-peptide complex formation. Chem. Commun. 51, 12548-12551

110. Person, M. D., Brown, K. C., Mahrus, S., Craik, C. S., and Burlingame, A. L. (2001) Novel inter-protein cross-link identified in the GGH-ecotin D137Y dimer. Protein Sci. 10, 1549- 1562

111. Liu, F., Luo, E. Y., Flora, D. B., and Mezo, A. R. (2014) Irreversible sortase A-mediated ligation driven by diketopiperazine formation. J. Org. Chem. 79, 487-492

Chapter 1. Introduction

112. Williamson, D. J., Fascione, M. A., Webb, M. E., and Turnbull, W. B. (2012) Efficient N- terminal labeling of proteins by use of sortase. Angew. Chem. Int. Ed. Engl. 51, 9377-9380

113. Yamamura, Y., Hirakawa, H., Yamaguchi, S., and Nagamune, T. (2011) Enhancement of sortase A-mediated protein ligation by inducing a beta-hairpin structure around the ligation site. Chem. Commun. (Camb.) 47, 4742-4744

Chapter 2 Peptide synthesis and NMR spectroscopy

Chapter 2. Peptide synthesis and NMR spectroscopy

2. Peptide synthesis and nuclear magnetic resonance (NMR) spectroscopy 2.1. Overview This chapter describes the main general experimental techniques that were used in this thesis, i.e., solid-phase peptide synthesis performed either manually or using an automated peptide synthesizer and nuclear magnetic resonance (NMR) spectroscopy for peptide structure determination. Experimental techniques specific for individual projects can be found in the corresponding chapters.

2.2. Solid-phase peptide synthesis 2.2.1. Synthesis procedure All peptides described in this thesis were synthesized using fluorenylmethyloxycarbonyl (Fmoc) chemistry. Rink amide 4-methylbenzhydrylamine (MBHA) was used when a C-terminal amide was required after cleavage of peptide from the resin otherwise 2-chlorotrityl chloride (CTC) resin was used. The resin was swelled in DMF for 30 min prior to synthesis and the Fmoc protecting group was removed with 20% piperidine in DMF (2 x 5 min). Amino acid (4 eq) were activated by dissolving in 0.5 M 2-(1Hbenzotriazol-1-yl)-1,1,3,3,-tetramethyluronium hexafluorophosphate (HBTU) in DMF and N,N-diisopropylethylamine (DIPEA) (4 eq) was added. Each coupling (except proline) was subjected to a ninhydrin test which is based on the fact that primary amines react with ninhydrin generating a blue-colored product, thus allowing a check of coupling efficiency to ensure that no free primary amines remained (1,2). Briefly, the procedure included adding 2 drops of 76% w/w phenol in ethanol, 4 drops of 0.2 mM potassium cyanide (KCN) in pyridine, and 2 drops of 0.28 M ninhydrin in ethanol to a small amount of the resin followed by incubation at 95 °C for 5 min. When coupling efficiency was low even after repeating the coupling step with the same amino acid, the N-terminus of the peptide was acetylated by treating the resin in 6% acetic anhydride/0.2M DIPEA/DMF for 2 x 15 min Following coupling of the last amino acid (i.e. first residue at the N-terminus) and final deprotection by 20% piperidine, the resin with peptide assembled on it was thoroughly washed with DMF, followed by DCM. The resin was dried by nitrogen (1 h) to prevent air oxidation of thiol groups. Figure 1 shows a schematic illustration of the peptide synthesis procedure based on Fmoc-chemistry.

2.2.2. Peptide cleavage The advantage of the Fmoc-solid phase peptide synthesis (SPPS over BOC (tert-butyloxycarbonyl)- SPPS is that peptides can be cleaved using mild acidic conditions while BOC-SPPS requires the use of hydrogen fluoride (HF). Fmoc cleavage conditions typically requires 95% TFA/2.5% TIPS 29

Chapter 2. Peptide synthesis and NMR spectroscopy

(scavenger)/2.5% water in which the resin is incubated for at least 2 h. The TFA is evaporated using a rotary evaporator and the released peptide is precipitated with ice-cold ether.

Figure 1. Overview of Fmoc-solid phase peptide synthesis (SPPS). Fmoc is a Nα-protecting group that is base-labile and acid-stable. Amino acid building blocks with the Fmoc protecting group usually have a side-chain protecting group that is also acid-labile and cleaved after completion of synthesis unless it requires selective deprotection (eg. Acm for Cys). The synthesis proceeds from the C- to the N-terminus and it can be fully automated using peptide synthesizer.

2.2.3. Fmoc-based synthesis of cyclotides Cyclotides studied in our laboratory have, until recently, been produced by thioester-mediated native chemical ligaion (NCL). As the thioester-linker is base-labile, Boc-SPPS has been used (3-5) instead of Fmoc-SPPS, which requires piperidine for deprotection of each amino acid. Boc-SPPS allows attachment of a thioester linker to a peptide on resin, which subsequently can make a peptide bond with a thiol group of N-terminal Cys. In spite of the usefulness of NCL, alternative methods for cyclization/ligation are desirable as Boc-SPPS requires repetitive use of highly corrosive TFA to remove the Boc-protecting group of each amino acid, which complicates the use of automated peptide synthesizers (6). The use of highly toxic HF to release peptides from resins is another drawback of Boc-SPPS. Furthermore, as fluorophores, click handles (such as azides), and diazirine- based photoaffinity labels are HF-sensitive, the subsequent use of these peptides for in vitro or in vivo assays is restricted (7). Therefore, many ligation methods compatible with Fmoc chemistry have been developed in recent years using various thioester surrogates (6-13). Following Fmoc-

Chapter 2. Peptide synthesis and NMR spectroscopy

based synthesis of cyclotides via thioesterification and NCL (13,14), a straightforward, in-solution cyclization method was reported (7). The procedure includes peptide chain assembly, peptide cleavage from the 2CTC resin, head-to-tail cyclization with the side-chain protecting groups remaining on the amino acids, deprotection of side chains, and oxidation of cysteines. This method is particularly advantageous because it does not require an N-terminal cysteine or a thioester linker. Therefore, peptides that did not require enzymatic cyclization in this thesis were made using Fmoc- based cyclization (7).

Firstly, peptides were synthesized by Fmoc SPPS with a Symphony automated peptide synthesizer (Protein Technology Inc, USA) using standard protocols. After final deprotection of the terminal Fmoc group, the dried resin was treated with 5 mL of 1% TFA in DCM (v/v) for 5 min in a reaction vessel and this step was repeated 10 times. The eluate contained peptides released from the resin but with the side-chain protecting group still attached. This was collected in a round-bottomed flask containing 150 mL of 50% ACN, 0.05% TFA in water (v/v) since DCM is highly volatile. Prior to lyophilizing the peptide, residual DCM and TFA were removed using a rotary evaporator. For cyclization, the lyophilized peptide was dissolved in DMF at a concentration of 2 mM and N,N,N’,N’-tetramethyl-O-(7-azabenzotriazol-1-yl)uronium hexafluorophosphate (HATU) was added to a final concentration of 5 mM. In addition, 10 mM DIPEA was added after HATU was completely dissolved. The reaction was stirred for 3 h at room temperature, diluted with twice the volume of 50% ACN (0.05% TFA) and lyophilized overnight. Side-chain deprotection was carried out as described in the section on peptide cleavage. Figure 2 summarizes the procedure with schematic illustrations.

2.3. NMR spectrometry All peptides were analysed using a Bruker 500 MHz or Bruker 600 Avance MHz spectrometer equipped with a cryoprobe. Following acquisition, spectra were processed using Topspin (Bruker, Germany). For each experiment, a one-dimensional spectrum was initially run as it gives an indication of peptide folding status. However, for a peptide containing ~30 amino acids, a one- dimensional spectrum often displays overlapped peaks. Therefore two-dimensional experiments such as TOCSY and NOESY were usually run to resolve peak overlap. Further details of two- dimensional experiments required for structure calculation of peptides are discussed in the following sections.

Chapter 2. Peptide synthesis and NMR spectroscopy

Figure 2. Fmoc-based cyclization of disulfide-rich peptides. The procedure includes peptide chain assembly, peptide cleavage from the 2CTC resin, head-to-tail cyclization with side-chain protecting group remained on amino acids, deprotection of sidechains, and oxidation of cysteines.

2.3.1. TOCSY Total correlation spectroscopy (TOCSY) (11), also known as Hartmann-Hahn spectroscopy (HOHAHA) (15), shows all protons of one spin system while correlated spectroscopy (COSY) (16) shows protons directly connected (Figure 3). The characteristic pattern from each amino acid on F1 axis of TOCSY spectra allows identification of an amino acid. Amino acids such as Ala, Gly, Val, Glu, Gln and Pro display particularly distinctive chemical shifts and are good starting points for sequential assignment. Therefore, TOCSY is commonly used for structural analysis, D2O exchange, temperature and pH titration experiments (17).

Chapter 2. Peptide synthesis and NMR spectroscopy

Figure 3. An example of TOCSY and COSY. Signals of threonine on the diagonal divide the spectrum in two equal halves. Symmetrical to this diagonal (another half not shown), cross signals are generated by TOCSY (red) or COSY (green). TOCSY shows all protons of one spin system while COSY shows protons directly connected.

2.3.2. NOESY Nuclear Overhauser effect spectroscopy (NOESY) allows the identification of protons at a distance of less than 5 Å apart (18). In other words, NOESY experiments produce peaks based on spatial couplings by using the dipolar interaction of spins, nuclear Overhauser effects (NOE), for protons. The intensity of the NOE is proportional to 1/r6, with r representing the distance between protons. The NOESY experiment correlates not only protons that are in close proximity, but also protons of amino acids that are distant in the sequence but spatially close due to tertiary structure. Therefore, NOESY is the most crucial spectral tool to determine protein structures. Access to TOCSY and NOESY spectra will allow for the sequential assignment of a peptide, which is typically followed by the assignment of non-sequential peaks.

2.3.3. Dihedral angle restraints The backbone conformation of a peptide is determined by the dihedral angles, i.e., ϕ, ψ, and ω, of each amino acid. Because peptide bonds are usually in the trans conformation and the ω angle is thus assumed to be 180°, mainly the values of ϕ and ψ need to be determined from experiments. 3 The ϕ angle restraints of a peptide backbone can be determined from the JHα-HN couplings derived from a double-quantum filtered correlation spectroscopy (DQF-COSY), or a one dimensional spectrum if the signals in the region of HN are not overlapped. The correlation between the observed coupling constant (3J) and dihedral angle, ϕ, (Figure 4A) is based on the Karplus equation (Figure 4B) (19). In general, the value of ϕ ranges from -30° to -180° (20) and it is common to set

Chapter 2. Peptide synthesis and NMR spectroscopy

3 3 the value of ϕ at -120 ± 30° for JHα-HN >8 Hz and -60 ± 30° for JHα-HN <6 Hz in structure calculation. In addition, secondary structures of a peptide tend to have characteristic values of 3 3 coupling constant such that JHα-HN >8 Hz for β strands, and JHα-HN <6 Hz for α-helix (Figure 4C) (21).

Figure 4. (A) Newman projection showing the concept of dihedral angle, ϕ (B) The Karplus equation. A, B, and C are experimentally determined parameters whose values are variable depending on the atoms and substituents involved. The original paper determines them as A=4.22, 3 B=-0.5 and C=4.5 (19), (C) Relationship of coupling constant JHα-HN and backbone dihedral angle ϕ (19,22). Approximate values of ϕ associated with secondary structures (α helix and β sheet) are 3 3 highlighted in grey and these two elements determine values of JHα-HN <6 Hz, JHα-HN >8 Hz respectively (21,23).

The ψ backbone dihedral angle is not as readily experimentally measurable as ϕ because the relative coupling constants are usually small (23). However, it is possible to estimate both ϕ and ψ via the TALOS program (24,25), which is based on the fact that conformation affects secondary chemical shifts of the 1Hα, 13C, 13Cα, 13Cβ, and amide nitrogen nuclei (26-28). Therefore, the TALOS program searches the best match in a protein database to predict both dihedral angles using chemical shifts given as input (29).

In addition to the backbone dihedral angles, the sidechain dihedral angle χ1 provides important information for structure calculation. For amino acids with a β-methylene group χ1 can be 3 3 determined by measurement of JHαHβ2 and JHαHβ3 and stereospecific assignment of β-methylene protons can be achieved in combination with certain NOEs patterns. The coupling constants usually correspond to one of the three possible staggered rotamers, in which each atom attached to Cβ is either in one of the two gauche positions relative to Hα or in the trans position (Figure 5). To 3 measure JHαHβ experimentally, exclusive correlation spectroscopy (ECOSY) (30) is used, which is a

Chapter 2. Peptide synthesis and NMR spectroscopy

modified COSY experiment but simplifies the multiplet structure of the Hα-Hβ crosspeaks, however spectral overlap is often found for large proteins.

Figure 5. Newman projections of the three possible staggered conformers for the χ1 dihedral angle. 3 3 χ1 can be defined from information combining JHαHβ2 and JHαHβ3 coupling constants and NOE intensity.

2.3.4. Hydrogen bonds

There are two main methods to identify protons that are shielded from the solvent and therefore potentially involved in hydrogen bonding. These two methods are D2O exchange experiments and temperature coefficient experiments.

D2O exchange experiment

Hydrogen bond restraints can be identified by D2O exchange, in which amide protons will exchange rapidly if they are not involved in a hydrogen bond. The exchange rates are usually measured by recording a series of 1D and/or TOCSY spectra immediately after dissolving the peptide sample in 100% D2O whereby slow-exchanging amide protons can be traced over time. Protons buried within the peptide structure or involved in a hydrogen bond will exchange with deuterium at a slower rate compared to those exposed to the surface (31). In such deuterium exchange experiments, it is only the hydrogen bond donor that is detected, so preliminary structures are required to identify H-bond acceptors. Protons still present after >4 h are considered shielded and thus possibly involved in hydrogen bonding.

Chapter 2. Peptide synthesis and NMR spectroscopy

Temperature coefficient experiment Temperature coefficients measure the rate of change of amide proton chemical shifts with temperature. A slow rate of the change indicates protection of amide protons from solvent by being involved in hydrogen bonds. The most common interpretation is that a temperature coefficient value less negative than -4.5 ppb/K is indicative of HN solvent protection (32).

2.3.5. Structure calculation After obtaining chemical shifts, dihedral angles, and hydrogen bonds, the structure of a peptide can be calculated using programs including CYANA, which produces a set of preliminary structures (33). Those structures need to be refined using programs such as CNS (34) which use more rigorous molecular dynamics and includes solvents. In general, the angles and distance restraints that violate the restraints set within the program are corrected or removed, and final structures with the lowest overall energies are chosen. These structures are analyzed in MolMol (35) to calculate the overall root-mean-square deviations (RMSD), which is a good indicator of structural precision. Finally, the quality of the structures is judged by Molprobity (this can be accessed online at http://molprobity.biochem.duke.edu) (36), which generates a report including a Ramachandran plot with bond angles, dihedral angles and unfavorable sidechain rotamers.

Chapter 2. Peptide synthesis and NMR spectroscopy

2.4. References

1. Kaiser, E., Colescott, R. L., Bossinger, C. D., and Cook, P. I. (1970) Color test for detection of free terminal amino groups in the solid-phase synthesis of peptides. Anal. Biochem. 34, 595-598

2. Sarin, V. K., Kent, S. B., Tam, J. P., and Merrifield, R. B. (1981) Quantitative monitoring of solid-phase peptide synthesis by the ninhydrin reaction. Anal. Biochem. 117, 147-157

3. Dawson, P. E., Muir, T. W., Clark-Lewis, I., and Kent, S. B. (1994) Synthesis of proteins by native chemical ligation. Science 266, 776-779

4. Daly, N. L., Love, S., Alewood, P. F., and Craik, D. J. (1999) Chemical synthesis and folding pathways of large cyclic polypeptide: Studies of the cystine knot polypeptide kalata B1. Biochemistry 38, 10606-10614

5. Tam, J. P., Lu, Y. A., Yang, J. L., and Chiu, K. W. (1999) An unusual structural motif of antimicrobial peptides containing end-to-end macrocycle and cystine-knot disulfides. Proc. Natl. Acad. Sci. U. S. A. 96, 8913-8918

6. Mende, F., and Seitz, O. (2011) 9-Fluorenylmethoxycarbonyl-based solid-phase synthesis of peptide alpha-thioesters. Angew. Chem. Int. Ed. Engl. 50, 1232-1240

7. Cheneval, O., Schroeder, C. I., Durek, T., Walsh, P., Huang, Y. H., Liras, S., Price, D. A., and Craik, D. J. (2014) Fmoc-based synthesis of disulfide-rich cyclic peptides. J. Org. Chem. 79, 5538-5544

8. Blanco-Canosa, J. B., Nardone, B., Albericio, F., and Dawson, P. E. (2015) Chemical protein synthesis using a second-generation N-acylurea linker for the preparation of peptide- thioester precursors. J. Am. Chem. Soc. 137, 7197-7209

9. Zheng, J. S., Tang, S., Guo, Y., Chang, H. N., and Liu, L. (2012) Synthesis of cyclic peptides and cyclic proteins via ligation of peptide hydrazides. Chembiochem 13, 542-546

10. Lelievre, D., Terrier, V. P., Delmas, A. F., and Aucagne, V. (2016) Native chemical ligation strategy to overcome side reactions during Fmoc-based synthesis of C-terminal cysteine- containing peptides. Org. Lett. 18, 920-923

11. Pusterla, I., and Bode, J. W. (2015) An oxazetidine amino acid for chemical protein synthesis by rapid, serine-forming ligations. Nat. Chem. 7, 668-672

12. Raz, R., and Rademann, J. (2011) Fmoc-based synthesis of peptide thioesters for native chemical ligation employing a tert-butyl thiol linker. Org. Lett. 13, 1606-1609

Chapter 2. Peptide synthesis and NMR spectroscopy

13. Aboye, T. L., Clark, R. J., Craik, D. J., and Goransson, U. (2008) Ultra-stable peptide scaffolds for protein engineering - Synthesis and folding of the circular cystine knotted cyclotide cycloviolacin O2. Chembiochem 9, 103-113

14. Gunasekera, S., Aboye, T. L., Madian, W. A., El-Seedi, H. R., and Goransson, U. (2013) Making Ends Meet: Microwave-Accelerated Synthesis of Cyclic and Disulfide Rich Proteins Via In Situ Thioesterification and Native Chemical Ligation. Int J Pept Res Ther 19, 43-54

15. Bax, A., Davis, D. G., and Sarkar, S. K. (1985) An improved method for two-dimensional heteronuclear relayed-coherence-transfer NMR-spectroscopy. J. Magn. Reson. 63, 230-234

16. Aue, W. P., Bartholdi, E., and Ernst, R. R. (1976) 2-Dimensional spectroscopy - application to nuclear magnetic-resonance. J. Chem. Phys. 64, 2229-2246

17. Jacobsen, N. E. (2007) NMR spectroscopy explained: simplified theory, applications and examples for organic chemistry and structural biology. John Wiley and Sons, New jersey, 1- 668

18. Overhauser, A. W. (1953) Polarization of nuclei in metals. Phys Rev 92, 411-415

19. Karplus, M. (1963) Vicinal proton coupling in nuclear magnetic resonance. J. Am. Chem. Soc. 85, 2870-2871

20. Richardson, J. S. (1981) The anatomy and taxonomy of protein structure. Adv. Protein Chem. 34, 167-339

21. Fletcher, J. I., Smith, R., O'Donoghue, S. I., Nilges, M., Connor, M., Howden, M. E., Christie, M. J., and King, G. F. (1997) The structure of a novel insecticidal neurotoxin, omega-atracotoxin-HV1, from the venom of an Australian funnel web spider. Nat. Struct. Biol. 4, 559-566

22. Billeter, M., Neri, D., Otting, G., Qian, Y. Q., and Wuthrich, K. (1992) Precise vicinal coupling constants 3JHN alpha in proteins from nonlinear fits of J-modulated [15N,1H]- COSY experiments. J. Biomol. NMR 2, 257-274

23. King, G. F., and Mobli, M. (2010) Determination of peptide and protein structures using NMR Spectroscopy Comprehensive Natural Products Ii: Chemistry and Biology 9, 1-769

24. Shen, Y., and Bax, A. (2013) Protein backbone and sidechain torsion angles predicted from NMR chemical shifts using artificial neural networks. J. Biomol. NMR 56, 227-241

25. Shen, Y., Delaglio, F., Cornilescu, G., and Bax, A. (2009) TALOS+: a hybrid method for predicting protein backbone torsion angles from NMR chemical shifts. J. Biomol. NMR 44, 213-223

Chapter 2. Peptide synthesis and NMR spectroscopy

26. Spera, S., and Bax, A. (1991) Empirical correlation between protein backbone conformation and C-alpha and C-beta C-13 nuclear-magnetic-resonance chemical-shifts. J. Am. Chem. Soc. 113, 5490-5492

27. Wishart, D. S., and Sykes, B. D. (1994) The C-13 chemical-shift index - a simple method for the identification of protein secondary structure using C-13 chemical-shift data. J. Biomol. NMR 4, 171-180

28. Wishart, D. S., Sykes, B. D., and Richards, F. M. (1991) Relationship between nuclear- magnetic-resonance chemical-shift and protein secondary structure. J. Mol. Biol. 222, 311- 333

29. Cornilescu, G., Delaglio, F., and Bax, A. (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13, 289- 302

30. Griesinger, C., Sørensen, O. W., and Ernst, R. R. (1987) Practical aspects of the E.COSY technique. Measurement of scalar spin-spin coupling constants in peptides. J. Magn. Reson. 75, 474-492

31. Wagner, G., Braun, W., Havel, T. F., Schaumann, T., Go, N., and Wuthrich, K. (1987) Protein structures in solution by nuclear magnetic resonance and distance geometry. The polypeptide fold of the basic pancreatic trypsin inhibitor determined using two different algorithms, DISGEO and DISMAN. J. Mol. Biol. 196, 611-639

32. Baxter, N. J., and Williamson, M. P. (1997) Temperature dependence of 1H chemical shifts in proteins. J. Biomol. NMR 9, 359-369

33. Guntert, P. (2004) Automated NMR structure calculation with CYANA. Methods Mol. Biol. 278, 353-378

34. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905-921

35. Koradi, R., Billeter, M., and Wuthrich, K. (1996) MOLMOL: A program for display and analysis of macromolecular structures. J. Mol. Graph. 14, 51-55

36. Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S., and Richardson, D. C. (2010) MolProbity: All-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12-21

Chapter 3 Semienzymatic cyclization of disulfide-rich peptides using sortase A

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

3.1. Overview and declaration of authorship

Overview

This chapter includes a publication describing the enzymatic cyclization of three disulfide-rich peptides (SFTI-1, Vc1.1, kalata B1) possessing one to three disulfide bonds using sortase A.

Declaration of authorship

In accordance with the guidelines of the University of Queensland, this chapter is presented as a published, peer-reviewed research article.

“Semienzymatic cyclization of disulfide-rich peptides using sortase A”

Journal of Biological Chemistry, 2014, 289:6627-38 originally published online January 14, 2014 doi: 10.1074/jbc.M113.539262

The experiments were conceived and designed by Xinying Jia, Soohyun Kwon under the supervision of Christina Schroeder and David Craik. Soohyun Kwon did peptide purification, oxidative folding and cyclization for kalata B1 and SFTI-I, NMR analysis on SFTI-1. Xinying Jia and Christina Schroeder did the structure calculation on kalata B1. Purification, oxidative folding and cyclization, NMR analysis on Vc1.1 was done by Ching-I Anderson Wang. Soohyun Kwon and Xinying Jia expressed and purified sortase A. Yen-Hua Huang helped with hemolytic activity assay and serum stability assay analysis, Lai Y.Chan prepared native SFTI-I. K. Johan Rosengren participated in the analysis of structure of kalata B1. All the peptides were synthesized by Fmoc chemistry on an automated synthesizer by Chia Chia Tan, Phillip Walsh. The manuscript and figures were prepared by Xinying Jia and Soohyun Kwon and revised with guidance from Christina Schroeder and David Craik.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

3.2. (Cover page of the published manuscript)

Semienzymatic cyclization of disulfide-rich peptides using sortase A

Xinying Jia‡§1, Soohyun Kwon§1, Ching-I Anderson Wang§, Yen-Hua Huang§, Lai Y. Chan§, Chia Chia Tan§, K. Johan Rosengren¶2, Jason P. Mulvenna‡2, Christina I. Schroeder§3, and David J. Craik§4

1QIMR Berghofer Medical Research, Brisbane 4000, Qld, Australia. 2The University of Queensland, Institute for Molecular Bioscience, Brisbane 4072, Qld, Australia 3The University of Queensland, School of Biomedical Sciences, Brisbane 4072, Qld, Australia

*Running title: Sortase A as a tool for cyclizing disulfide-rich peptides

1 Both authors contributed equally to this work. 2 Supported by Career Development Awards from the Australian NHMRC. 3 To whom correspondence may be addressed. Tel.: 61-7-33462019; Fax: 61-7-33462101; E-mail: [email protected]. 4 An Australian NHMRC Professorial Fellow (Grant APP1026501). To whom correspondence may be addressed. Tel.: 61-7-33462019; Fax: 61-7-33462101; E-mail: [email protected].

Background: Sortase A (SrtA) is a transpeptidase capable of catalysing the formation of amide bonds. Results: SrtA was used to backbone cyclize disulfide-rich peptides including kalata B1, α-conotoxin Vc1.1 and SFTI-1. Conclusion: SrtA-mediated cyclization is applicable to small disulfide-rich peptides. Significance: SrtA-mediated cyclization is an alternative to native chemical ligation for the cyclization of small peptides of therapeutic interest.

Keywords: cyclic peptides, cyclotides, enzymatic cyclization, sortase A, kalata B1

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

3.3. Abstract

Disulfide-rich cyclic peptides have generated great interest in the development of peptide-based therapeutics due to their exceptional stability toward chemical, enzymatic, or thermal attack. In particular, they have been used as scaffolds onto which bioactive epitopes can be grafted to take advantage of the favorable biophysical properties of disulfide-rich cyclic peptides. To date, the most commonly used method for the head-to-tail cyclization of peptides has been native chemical ligation. In recent years, however, enzyme-mediated cyclization has become a promising new technology due to its efficiency, safety, and costeffectiveness. Sortase A (SrtA) is a bacterial enzyme with transpeptidase activity. It recognizes a C-terminal penta-amino acid motif, LPXTG, and cleaves the amide bond between Thr and Gly to form a thioacyl-linked intermediate. This intermediate undergoes nucleophilic attack by an N-terminal poly-Gly sequence to form an amide bond between the Thr and N-terminal Gly. Here, we demonstrate that sortase A can successfully be used to cyclize a variety of small disulfide-rich peptides, including the cyclotide kalata B1, α-conotoxin Vc1.1, and sunflower trypsin inhibitor 1. These peptides range in size from 14 to 29 amino acids and contain three, two, or one disulfide bond, respectively, within their head-to-tail cyclic backbones. Our findings provide proof of concept for the potential broad applicability of enzymatic cyclization of disulfide-rich peptides with therapeutic potential.

3.4. Introduction Disulfide-rich cyclic peptides are generating great interest in the field of drug design because of their remarkable stability toward enzymatic, thermal, or chemical attack (1). Additionally, anecdotal evidence of some of them being orally bioavailable in indigenous medicine applications (2) has recently received support from studies demonstrating oral activity of engineered cyclic peptides in animal pain models (3, 4). Several classes of cyclic peptides, including the sunflower trypsin inhibitors and cyclotides, have now been discovered in a variety of plant species (5). In addition to the plant-derived cyclic peptides, disulfide-rich peptides isolated from cone snails, including α-conotoxins, have successfully been cyclized (3, 6) and are showing promise as potential treatments for neuropathic pain (3). Sunflower trypsin inhibitor 1 (SFTI-1) comprises 14 amino acids with a head-to-tail cyclized backbone. It was isolated from sunflower seeds and specifically inhibits trypsin activity at 0.1 nM (7), which makes it the smallest and most potent serine proteinase inhibitor known. It was also reported to have inhibition activity against the epithelial serine protease matriptase at subnanomolar concentration (8). Its structure consists of two antiparallel β-strands connected by an extended loop and a sharp hairpin turn containing a cis-proline residue. The β-strands are constrained by a single disulfide bond, which stabilizes the molecule and divides it into two regions: the active site loop and the cyclization loop (7) (Fig. 1A).

Cyclotides are characterized by the combination of head-totail backbone cyclization and six conserved Cys residues forming a distinctive disulfide linkage pattern in which one disulfide bond passes through a ring 43

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

formed by two other disulfide bonds (9, 10) (Fig. 1B). The combination of a cyclic backbone and a knotted arrangement of three disulfide bonds is known as the cyclic cystine knot motif, and this motif is believed to be responsible for the exceptional stability of cyclotides toward chemical, enzymatic, or thermal degradation (1). Although thought to have a natural function as defense molecules in plants via their pesticidal activities (11–13), cyclotides possess a range of pharmaceutically interesting biological activities, including uterotonic (2), anti-HIV (14), antimicrobial (15), and insecticidal (11,16) activities, and are cytotoxic to tumor cells (17, 18). Thousands of cyclotides are thought to exist in nature, and more than 250 sequences have been described so far (see the CyBase Web site) (19), making them one of the most abundant plant protein families discovered.

Cyclotides are able to accommodate the synthetic introduction of a range of biologically active sequence motifs in the backbone loops between the conserved Cys residues while retaining thermal, chemical, and enzymatic stability. This toleranceto sequence substitutions makes them ideal scaffolds for the development of stable peptide-based drugs incorporating epitopes of therapeutic value. For example, a kB1 hybrid, with an anti-angiogenic epitope incorporated into loop 3, combined anti-angiogenic activity with the native stability of kB1 (20). Likewise, molecules possessing activity against foot-and mouth disease (21), inhibitory activity against β-tryptase and human leukocyte elastase (22), angiogenic activity (23), and the ability to antagonize intracellular p53 degradation (24) have also been achieved by grafting relevant bioactive epitopes into the cyclotide MCoTI-I or MCoTI-II. The capacity to target intracellular sites is consistent with a range of recent biophysical studies demonstrating the ability of disulfide-rich cyclic peptides to internalize into cells (24–28). As well as being valuable scaffolds, naturally occurring cyclic peptides, such as cyclotides and SFTI-1, have inspired the concept of using artificial head-to-tail backbone cyclization as a stabilizing strategy for peptides and proteins of therapeutic interest (3, 6). For example, Clark et al. (3) showed that cyclization of α-conotoxin Vc1.1, a potential neuropathic pain treatment, endowed it with oral activity. Thus, great interest has been shown in the development of cost-effective and efficient methods for the N- to C- terminal backbone cyclization of synthesized or expressed proteins (29, 30). In the case of cyclic Vc1.1 (cVc1.1), the termini of the naturally occurring 16-residue peptide were covalently linked using a 6-amino acid linker, as illustrated in Fig. 1C (3).

Currently, the most widely used strategy for backbone cyclization of synthetic peptides is native chemical ligation (NCL) (31, 32), which utilizes an N-terminal cysteine and a functionalized C terminus to form a thioester-linked intermediate that undergoes an S,N-acyl migration to form a native peptide bond (32). In general, peptides are synthesized using solid phase peptide synthesis, using either t-butoxycarbonyl or Fmoc protection. Despite the usefulness of NCL for splicing or cyclizing protein, there is still a demand for alternative methods of amide bond formation that are chemoselective, rapid, cheap, safe, and waste-free (30). Moreover, the requirement for an N-terminal cysteine and a C-terminal thioester in NCL is not always 44

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

compatible with Fmoc solid phase peptide synthesis or recombinant expression. Accordingly, a number of different approaches to backbone cyclization have been developed (29), including expressed protein ligation (33, 34), intein-mediated protein trans-splicing (35), and genetic code reprogramming (36).

Recently, a family of thiol-containing transpeptidases, known as the sortases, was found to successfully cyclize linear proteins (37–40). Sortases are found in most Gram-positive bacteria, where their primary function is to attach surface proteins to the bacterial cell wall (41). This ligation reaction has been exploited for a variety of protein engineering purposes, in most cases utilizing sortase A (SrtA) from Staphylococcus aureus (42). SrtA recognizes a 5-residue sequence motif, LPXTG, and cleaves the peptide bond between the Thr and Gly residues, forming an acyl enzyme intermediate between a cysteine at the active site of SrtA and the carboxylate at the truncatedCterminus of the substrate (43, 44). Although in a natural system, this covalent intermediate is resolved by nucleophilic attack from a pentaglycine side chain in a peptidoglycan precursor, a variety of glycine-based nucleophiles can activate SrtA-catalyzed transpeptidation (38). The result is the efficient attachment of various moieties to the protein substrates, and this process has been applied to site-specific protein labeling (45), PEGylation (46), protein thioester generation (47), and protein- protein fusion (48). When such glycine-based nucleophiles originate from the N terminus of the substrate, intramolecular transpeptidation reactions occur to yield covalently closed (circular) polypeptides by amide bond formation between the two termini (38–40). So far, this approach has been examined only for a limited number of possible substrates and has not been explored for disulfide-rich peptides.

In the current study, we demonstrate the efficient enzymatic cyclization of disulfide-rich peptides, including SFTI-1, cVc1.1, and kalata B1 (kB1), containing one, two, or three disulfide bonds, respectively, by SrtA without the need for a thioester linker and using Fmoc chemistry. This study provides a proof of concept that SrtA-mediated ligation can be utilized to cyclize disulfide-rich peptides ranging from the simple one- disulfidecontaining SFTI-1 to the complex cyclic cystine knot-containing cyclotide structure of kB1 while retaining the native disulfide bond connectivity that is vital for the stability of these proteins. Fig. 1D summarizes the cyclization strategy for kB1.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 1. (A) Representation of the sequence and structure (PDB code 1JBL) of native SFTI-1, (B) native kB1 (PDB code 1NB1), which is the prototypic cyclotide, (C) native linear Vc1.1 (PDB code 2H8S), which has been cyclized by a 6-amino acid linker using native chemical ligation (3, 32). The disulfide bonding network is shown as sticks, with the cysteine residues numbered. Amino acids are represented by their one- letter code. (D) The linear oxidized [GGG]kB1[TGG] contains a SrtA recognition sequence (LPVTG) and undergoes an intramolecular transpeptidation reaction catalyzed by SrtA to form cyclo-[GGG]kB1[T].

3.5. EXPERIMENTAL PROCEDURES 3.5.1. Peptide Synthesis The linear peptides [GGG]kB1[TGG], [G]Vc1.1[GLPETGGS], and [GG]SFTI-1[LPETGG] were synthesized by solid-phase peptide synthesis using an automatic peptide synthesizer (Symphony®, Protein Technologies, Inc.) following standard Fmoc chemistry on a 2-chlorotrityl chloride resin. Acetamidomethyl (Acm) protecting groups were used on Cys4 and Cys17 of [G]Vc1.1[GLPETGGS] for selective disulfide bond (CysII–CysIV) formation. Cleavage of the peptide from the resin was achieved using trifluoroacetic acid 46

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

(TFA)/triisopropylsilane/water (95:2.5:2.5) at room temperature for 2 h. The TFA was evaporated, and the peptide was precipitated with ice-cold diethyl ether and dissolved in 50% acetonitrile (ACN) containing 0.05% TFA and lyophilized. Wild type Vc1.1 and SFTI-1 were prepared as described previously (3, 49). SrtA Expression—The plasmid of and recombinant expression protocol for S. aureus SrtA, which comprises the catalytic domain (residues 60–206 with an N-terminal hexahistidine tag), was as described by Popp et al. (50). Briefly, an overnight culture of Escherichia coli BL21 (DE3) transformed with SrtA plasmid was cultured in the presence of kanamycin (50 μg/ml) until A600 reached ~0.7. The protein expression was induced by adding isopropyl β-D-thiogalactopyranoside to a final concentration of 1mM for 3 h at 37 °C. The cells were then harvested by centrifugation at 7000 x g at 4 °C for 10 min. Cells were then resuspended in ice-cold lysis buffer (50 mM Tris-HCl, pH 8.0, 150mM NaCl, 25mM imidazole, and 10% glycerol) and lysed by passing through a cell disrupter (TS Series Bench Top) operating at 26,000 p.s.i. The cell debris was removed by centrifugation at 12,000 x g at 4 °C for 30 min. The lysate was subjected to immobilized metal ion affinity chromatography using a 5-ml nickel-nitrilotriacetic acid FF column (GE Healthcare). The column was washed extensively with lysis buffer after sample loading, and SrtA was eluted with a linear gradient over 10 column volumes using elution buffer (50mM Tris-HCl, pH 8.0, 150 mM NaCl, and 500 mM immidazole). Fractions containing SrtA were buffer-exchanged (50 mM Tris-HCl, 150 mM NaCl, and 10% glycerol) and concentrated by using a centrifugal concentrator (<10 kDa). The SrtA was then stored at -80 °C until further use. The protein concentration was calculated by its extinction coefficient at 280 nm.

3.5.2. Oxidative Folding

The linear precursor [GGG]kB1[TGG] was oxidized in 50% isopropyl alcohol (v/v), 0.1 M NH4HCO3 with 1 mM reduced glutathione at pH 8 for 20 h at room temperature (51). The progress of oxidation was monitored by analytical reversed-phase high performance liquid chromatography (RP-HPLC) with a gradient from 0 to 50% ACN (0.05% TFA) in 50 min at a flow rate of 0.3 ml/min using an analytical C18 column (Agilent ZORBAX 300SB-C18, 5 μm, 2.1 x 150 mm). The oxidation yield was calculated based on HPLC profile. The mixture was then purified by RP-HPLC using a 0.5% gradient, and the molecular weight was confirmed by mass spectrometry. Native kB1 was isolated from Oldenlandia affinis as described previously (9, 52). Both linear precursor [G]Vc1.1[GLPETGGS] and cyclic reduced [G]Vc1.1[GLPET] were oxidized in 0.1 M I III NH4HCO3 at a 0.1 mg/ml concentration for 24 h to form disulfide bond Cys –Cys . Purified Vc1.1 (both linear and cyclic) with one disulfide bond was then dissolved in 50% acetic acid to a final concentration of

0.5 mg/ml under N2. Excess iodine dissolved (10 eq with respect to the Acm-protected Cys) in 50% acetic acid was added until the color of the reaction became yellow, and the reaction was left at room temperature for 24 h. The reaction was quenched by the addition of ascorbic acid until the solution became colorless, followed by RP-HPLC purification using a gradient of 0–60% acetonitrile in 60 min.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

3.5.3. Cyclization The concentration of modified kB1 was determined by its molar extinction coefficient at 280 nm using NanoDrop2000c (Thermo Scientific). Linear oxidized [GGG]kB1[TGG] and linear precursor [GG]SFTI- 1[LPETGG] (150 μM) were incubated with SrtA (50 μM) in the reaction buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 10 mM CaCl2) at 37 °C until the reaction was completed. The cyclization progress was monitored using RP-HPLC with a gradient from 0 to 50% ACN (0.05% TFA) in 50 min at a flow rate of 1 ml/min using an analytical C18 column (GraceSmartTM RP18, 5 μm, 4.6 x 150 mm). After completion of cyclization, the reaction was loaded to a semipreparative C18 column (Phenomenex, Jupiter 5UC18, 300 Å, 250 x 10.0 nm, 5 μm), eluted with a gradient from 36 to 45% ACN (0.05% TFA) in 30 min at a flow rate of 3 ml/min. The molecular mass of purified cyclo-[GGG]kB1[T] was confirmed by matrix-assisted laser desorption/ionization time of flight (MALDI-TOF). The fractions showing the correct mass were lyophilized for further assays.

The concentration of Vc1.1 was also determined as described above. 10 μM linear precursor [G]Vc1.1[GLPETGGS] was incubated with 0.2 μM SrtA in the reaction buffer containing 10mM TCEP for 24 h at 4 °C. SrtA was then removed from the sample using a 5-ml nickel-nitrilotriacetic acid FF column (GE Healthcare), and the flow-through containing cyclic reduced [G]Vc1.1[GLPET] was collected and purified by RP-HPLC for oxidation.

Because there were no aromatic residues present, the concentration of SFTI-1 was determined at 214 nm using the NanoDrop2000c (Thermo Scientific) UV-visible function with the extinction coefficient calculated based on the equation, ε214 nm = (nAA ̶ 1 + nN + nQ)2846 + nF7200 + nH6309 + nW22735+nγ5755 (53). 150μM [GG]SFTI-1[LPETGG] and 50μM SrtA were incubated in the reaction buffer at 37 °C with pH adjusted to 8.5 to form a disulfide bond at the same time. The progress of cyclization was monitored using analytical RP-HPLC with a gradient from 0 to 50% ACN (0.05% TFA) in 50 min at a flow rate of 1 ml/min using an analytical C18 column (GraceSmartTM RP18, 5 μm, 4.6 x 150 mm).

3.5.4. Nuclear Magnetic Resonance Analysis Cyclo-[GGG]kB1[T] was dissolved in 90% H2O, 10% D2O (v/v) or 100% D2O to a final concentration of 1 mM and pH 6.1 (54). Cyclo-[G]Vc1.1[GLPET] and cyclo-[GG]SFTI-1[LPET] were dissolved in 90% H2O,

10% D2O (v/v) to a final concentration of ~1mM and pH 3.1. All two-dimensional nuclear magnetic resonance (NMR) spectra were acquired at 298 K. Spectra recorded for assignments and Hα secondary chemical shift analysis included TOCSY with an 80-ms mixing time and NOESY with a 200-ms mixing time. All spectra were acquiredon a Bruker Avance 600-MHz NMR spectrometer equipped with a cryogenically cooled probe (Bruker, Karlsruhe, Germany).

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Water suppression for two-dimensional TOCSY and two-dimensional NOESY spectra were achieved by using excitation sculpting (55). Spectra were recorded with 2048 data points in the direct F2 dimension and 600 increments in the indirect F1 dimension. Additional spectra recorded for cyclo-[GGG]kB1[T] to be used in full structure calculations were NOESY with a 100-ms mixing time as well as E.COSY (56), DQF-COSY, and 13C HSQC to produce the carbon chemical shifts. All spectra were processed using Topspin (Bruker) and analyzed by CCPNMR (57) and Xeasy (58). The temperature coefficients were determined from recording TOCSY spectra from 283 to 303 K in steps of 10 K. Chemical shifts were referenced to internal 2,2- dimethyl-2-silapentane-5-sulfonate (DSS) at 0.0 ppm (59). The peaks in the NOESY spectra with mixing times of 100 and 200 ms were manually picked, intraresidual and sequential NOEs were assigned, and a full list of interproton distances was generated using the AUTO function in CYANA (tolerances used for CYANA were set to 0.02 ppm for both indirect and direct 1H dimension) (60). Several rounds of AUTO, including additional restraints, such as disulfide bonds, hydrogen bonds, and dihedral angle restraints derived from TALOS_, were run to ensure correct peak assignment. Hα, HN, Cα, and Cβ chemical shifts derived from 1H-1H NOESY and 1H-13CHSQCwere used to generate φ and ψ backbone dihedral angles using TALOS+(61). The ANNEAL function in CYANA was used to perform 10,000-step torsion angle dynamics to generate an ensemble from which 20 with the lowest penalty function were chosen for further analysis. Several rounds of ANNEAL were run to resolve distance and dihedral constraint violations. Using protocols from the RECOORD database (62), 50 structures were calculated using CNS (63) with force fields distributed with Haddock version 2.0 (64) and refined in a water shell (65) as described previously (66). A set of 20 structures with no NOE violations of greater than 0.3 Å and dihedral violations greater than 3.0º was chosen for MolProbity analysis (67). MOLMOL (68) was used for generating figures and final superimposition.

3.5.5. Hemolytic Activity Assay on [GGG]kB1[T] The hemolytic assay was conducted as described previously (69). Briefly, erythrocytes were isolated from human blood and washed in phosphate- buffered saline (PBS, pH 7.4) with repeated centrifugation at 1500 x g. The hemolytic activities of cyclo-[GGG]kB1[T], linear oxidized [GGG]kB1[TGG], and kB1 were determined with eight concentration points from 300 to 1.17μM by serial dilution with a 2-fold interval with a final volume of 20 μL/well in a 96-well plate. Triton X-100 solution (1% (v/v), 20 μL) was used as a positive control to achieve the 100% lysis, and PBS solution (20 μL) was used as a negative control. 100 μL of erythrocyte stock solution was added to each well and incubated at 37 °C for 1 h. After centrifugation of intact cells from the 96-well plate, the supernatant of each well was measured by visual absorption spectroscopy at 415 nm. Each experiment was performed in triplicate.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

3.5.6. Serum Stability Assay on [GGG]kB1[T] The stability of cyclo-[GGG]kB1[T] in human serum was evaluated using a method described previously (23). Briefly, human serum from maleABplasma (Sigma-Aldrich) was centrifuged at 17,000xg for 10 min to remove the lipid component. The resultant supernatant was collected and incubated for 15 min at 37 °C before the assay. Cyclo-[GGG]kB1[T], linear oxidized [GGG]kB1[TGG], and kB1 were tested at a final concentration of 20 μM. Stock solutions of peptides were diluted 1:10 with human serum, whereas the dilution with PBS was a negative control and a linear 12-mer peptide tested in parallel was a positive control of the peptidase activity of the serum. Controls and test peptides were incubated at 37 °C, and 40 μl of samples was taken for measurements at time points of 0, 1, 2, 3, 5, 8, 11, and 24 h. Serum proteins in samples were denatured with 40 μl of 6M urea at 4 °C for 10 min and precipitated with an additional 40 μl of 20% trichloroacetic acid at 4 °C for 10 min. The supernatant containing peptides was recovered by centrifugation at 17,000xg for 10 min. 100 μl of supernatant from each time point of the serum-treated and PBS-treated (control) peptides was injected and analyzed in triplicate on an analytical RP-HPLC column (Phenomenex) at a flow rate of 0.3 ml/min with a gradient of 0–50% ACN over 50 min. The retention time for each peptide was determined by PBS control at the zero time point. The stability of each peptide at each time point was calculated as the percentage of the integrated peak area of the serum-treated peptide over that of the 0-h serum-treated peptide on RP-HPLC at 215 nm.

3.6. RESULTS 3.6.1. Protein Production and Purification Linear precursors, with a SrtA recognition motif (LPXTG) at the C terminus and an oligoglycine motif at the N terminus (full sequences as follows:[GGG]kB1[TGG], GGGCGETCVGGTCNTPGCTCSWPVCTRNGLPVTGG;[G]Vc1.1[GLPETGGS], GGCCSDPRCNYDHPEICGLPETGGS;[GG]SFTI-1[LPETGG], GGGRCTKSIPPICFPDLPETGG (with the native sequences in boldface type)), were assembled using automated Fmoc chemistry, cleaved from the resin with TFA, and purified using RPHPLC. No oxidation of cysteine residues was observed during the purification process, as determined using mass spectrometry. The molecular mass and purity were confirmed using mass spectrometry and analytical RP-HPLC, respectively.

3.6.2. Oxidation and Cyclization of [GGG]kB1[TGG] In preliminary experiments two approaches were used to obtain oxidized, cyclic kB1 using SrtA-mediated cyclization. First, linear reduced [GGG]kB1[TGG] was cyclized using the SrtA reaction buffer but with the addition of a 2 mM concentration of the reducing agent TCEP. The peptide was then oxidized after completion of the cyclization reaction. Second, the linear peptide [GGG]kB1[TGG] was oxidized prior to cyclization with SrtA, which proceeded without the addition of reducing agent to the reaction buffer. 50

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Cyclization prior to oxidation resulted in higher yields of cyclic reduced peptide as determined by RPHPLC (~80%; data not shown), but subsequent oxidative refolding yielded several isomers (<10% yield for each isomer; data not shown), presumably due to misfolding during the oxidation reaction. Oxidation prior to cyclization resulted in a predominance of correctly folded peptide, and this approach was therefore used for subsequent experiments.

The oxidative folding of the linear precursor [GGG]kB1[TGG]was monitored by RP-HPLC (Fig. 2A), and the yield of correctly folded product was ~27% based on RP-HPLC peak integration. Following oxidative folding, linear [GGG]kB1[TGG] was purified by RP-HPLC, and the correct folding was confirmed by NMR analysis (data not shown).

Cyclization of 150 μM linear oxidized [GGG]kB1[TGG] with 50 μM SrtA was monitored over 24 h using RP-HPLC (Fig. 2B). The reaction was not extended beyond 24 h due to the possibility of a competing hydrolysis reaction in which SrtA irreversibly hydrolyzes the LPVTG motif (70). After 24 h, ~49% of linear, oxidized [GGG]kB1[TGG] was converted to the cyclic product (highlighted in Fig. 2B). MALDI-TOF analysis of the highlighted peak fraction showed a monoisotopic mass (3162.92 Da) corresponding to that of cyclo-[GGG]kB1[T] (Fig. 2C), which was purified to _95% purity. Interestingly, the elution time of this peak (40.8 min) was similar to that of native kB1(Fig. 2D) despite the additional amino acids in the sequence. Apart from this peak, two other peaks, at ~37 and 40 min, appeared to be intermediate species in the cyclization reaction because their abundance over time was negatively correlated with the abundance of correctly folded cyclo-[GGG]kB1[T] (Fig. 2B). A small peak appearing at ~42.8 min contained a peptide with the mass corresponding to that of cyclo-[GGG]kB1[T]. This peak also increased during the course of the experiment and is most probably an alternatively folded form of the peptide (Fig. 2B).

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 2. (A) HPLC profile of [GGG]kB1[TGG] oxidation at 0 and 20 h. A major peak of linear precursor peptide and linear oxidised product is marked with schematic representation of the peptide. (B) Time course analysis of the SrtA reaction progress monitored by RP-HPLC. Aliquots were taken from the reaction mixture at 0, 4, 8, 12 and 24 h, and loaded onto an analytical RP-HPLC. MALDI-TOF analysis of cyclo- [GGG]kB1[T] revealed a mass corresponding to the linear oxidised [GGG]kB1[TGG] minus two glycines and a water molecule, indicative of the formation of a amide bond between the N-and the C-terminal. (C) Analytical RP-HPLC profile of purified cyclo-[GGG]kB1[T].

3.6.3. Characterization of the cyclo-[GGG]kB1[T] by NMR The NMR signals of cyclo-[GGG]kB1[T] were well dispersed in the amide proton region, indicating that the peptide is folded and has a well defined structure. The individual amino acid spin systems were readily assigned using the sequential assignment procedure based on TOCSY and NOESY spectra (71). Cyclization was confirmed based on observation of a sequential αHi-HNi+1 NOE between Thr30 and Gly31, indicative of the amide linkage of the C-terminal LPVT and N-terminal GGG motif (Fig. 3,4). Chemical shifts are extremely sensitive to the local chemical environment, and H_ chemical shifts were thus used to compare the structure of cyclo-[GGG]kB1[T] with that of native kB1. Overall, the chemical shift differences between the two molecules are minimal (Fig. 3). With the exception of residues 1 and 27–33 (numbering system in Fig. 3), which are in loop 6 and either very close to or part of the SrtA sorting motif, no other cyclo-[GGG]kB1[T] 52

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

residue exhibited significant deviations from the corresponding residue in native kB1. The very small shifts from random coil values for residues 1 and 27–33 of cyclo-[GGG]kB1[T] indicate that this region is disordered, and this was confirmed in the three-dimensional structure of cyclo-[GGG]kB1[T] (Fig. 5A). The solution structure was calculated using CYANA (60) followed by refinement in a water shell using CNS (63) based on 388 NOE distance restraints; 53 dihedral angle restraints, including 26 φ, 20 ψ, and 6 χ1 dihedral angle restraints; and five hydrogen bonds derived from temperature coefficient data (Table 1). The 20 conformers with the lowest energy superimposed with a backbone root mean square deviation of 0.53 ±0.14 Å across resides 1–24 (Fig. 5A). Comparison of the cyclo-[GGG]kB1[T] structure with that of native kB1 (PDB code 1NB1) shows them to be very similar except for the more flexible loop 6 of cyclo-[GGG]kB1[T] (Fig. 5B).

Figure 3. Hα secondary shift comparison of native kB1 (empty squares with dashed line) and cyclo- [GGG]kB1[T] (filled squares with solid line). The Hα secondary shifts were calculated by subtracting random coil shifts from the experimental Hα shifts. The white box highlights loop 6, where the extra residues TGGG were introduced by the SrtA-mediated cyclization.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 4. TOCSY spectrum (upper panel) and NOESY spectrum (lower panel, with enlarged HA region) of cyclo-[GGG]kB1[T] recorded at 298 K, 90% H2O/10% D2O. The upper panel shows the spin systems by vertical lines, except Trp19 which was not observable in the TOCSY, while the lower panel highlights the sequential walk, except for Pro13, Pro20 and Pro28.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 5. (A) Superimposition of the 20 lowest energy structures of cyclo-[GGG]kB1[T] shown in blue with disulfide bonds shown in orange. (B) Comparison of cyclo-[GGG]kB1[T] in blue with native kB1 in red (PDB ID:1NB1) by superimposition of residues 1-24 of cyclo-[GGG]kB1[T] against residues 1-24 of the native kB1 structure. Disulfide bonds shown in orange. Highlighting the similarities between the two structures across loop 1-5 and the increased flexibility of loop 6 by insertion of the sSrtA sorting-motif. RMSD across heavy backbone atoms for residues 1-24 cyclo-[GGG]kB1[T] / 1-24 kB1 is 0.461.

3.6.4. Hemolytic Activity and Serum Stability Assay on Cyclo- [GGG]kB1[T] Native kB1 possesses mild hemolytic activity (15, 51), which is undesired in a pharmaceutical context. Thus, to assess the effects of SrtA cyclization on the novel kB1 analog, the hemolytic activity of cyclo- [GGG]kB1[T] and linear oxidized [GGG]kB1[TGG] was determined and compared with that of the native peptide. Compared with native kB1, both peptides exhibited significantly less hemolytic activity (Fig.6A). After 24h of incubation, 50μM native kB1 lysed 64% of human erythrocytes, whereas in the same time period, only 6 or2%of erythrocytes were lysed by the same concentration of cyclo-[GGG]kB1[T] or linear oxidized [GGG]kB1[TGG], respectively. The reduced hemolytic activity is consistent with previous mutagenesis studies on kB1, which showed that linearization or strategic replacement of certain residues can abrogate hemolytic activity (72, 73).

To determine the effects of SrtA-mediated cyclization on the stability of the peptide, cyclo-[GGG]kB1[T] was incubated in human serum. Encouragingly, its stability was comparable with that of native kB1, with only a slight reduction in peptide survival observed. After a 24-h incubation, 77.4 and 95.9% of the starting concentration of cyclo-[GGG]kB1[T] and kB1, respectively, was still present (Fig. 6B). Interestingly, 71% of linear oxidized [GGG]kB1[TGG] also remained in human serum after 24 h. In contrast, a linear 12-mer control peptide, without any disulfide bond, was completely degraded within the first 1 h.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 6. Hemolytic and serum-stability assay on cyclo-[GGG]kB1[T]. (A) Percentage of hemolysis by cyclo-[GGG]kB1[T] compared to native kB1 included as a positive control. (B) Percentage of peptide remaining after 0, 1, 2, 3, 5, 8, 11 and 24 h incubation in human serum. The data are presented as means ± standard deviation (SD).

3.6.5. Oxidation and Cyclization of [G]Vc1.1[GLPETGGS] α-Conotoxin Vc1.1 was successfully synthesized with a SrtA sorting motif at the C terminus and an additional Gly residue at the N terminus. Acm protecting groups on Cys4 and Cys17 were used for selective disulfide bond formation of CysII–CysIV (Fig. 7A). As for cyclo-[GGG]kB1[T], two approaches, namely oxidation prior to cyclization and cyclization prior to oxidation, were attempted to obtain the cyclo- [G]Vc1.1[GLPET]. The former approach gave rise to only a low yield (~1%) of oxidized product following Acm removal. In contrast, isomerization during oxidative folding was averted by cyclizing the linear precursor [G]Vc1.1[GLPETGGS] prior to oxidation. The cyclization performed in the presence of 10 mM

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

TCEP resulted in one major product with ~50% yield. The two disulfide bonds in the cyclic reduced [G]Vc1.1[GLPET] were then successfully formed via a two-step oxidation process with_15% yield. The linear precursor [G]Vc1.1[GLPETGGS], cyclic reduced [G]Vc1.1[GLPET], and cyclic oxidized [G]Vc1.1[GLPET] (hereafter named cyclo-[G]Vc1.1[GLPET]) showed a change in retention time after each reaction (Fig. 6B). The molecular mass and purity of the desired products for each step were confirmed by MALDI-TOF and RP-HPLC (Fig. 7B).

3.6.6. NMR Analysis of Cyclo-[G]Vc1.1[GLPET] The structure of cyclo-[G]Vc1.1[GLPET] was deduced using secondary Hα chemical shifts determined from TOCSY and NOESY spectra. A comparison of the secondary shifts of cyclo-[G]Vc1.1[GLPET] with both native and chemically cyclized Vc1.1 (3) suggested no significant difference in the overall structure.

Cyclization of cyclo-[G]Vc1.1[GLPET] was confirmed by observation of a sequential Hα-NHi+1 NOE between the C and N termini (Fig. 8).

3.6.7. SrtA Reaction for SFTI-1 For this single disulfide-bonded peptide, the SrtA reaction was performed at pH 8.5 to achieve cyclization and oxidation in a one-pot reaction. After 9 h, ~70% of the linear precursor had been converted to the cyclic oxidized form based on the RP-HPLC profile (peak at 20 min; Fig.7C). The mass of cyclic oxidized [GG]SFTI-1[LPET] was confirmed by MALDI-TOF (Fig. 7C). After a 24-h incubation, the peak at 20 min decreased, and a peak at 22.8 min increased in height and was found to contain both cyclic oxidized and hydrolyzed linear [GG]SFTI-1[LPET].

3.6.8. NMR Analysis of Cyclo-[GG]SFTI-1[LPET] Purified cyclized and oxidized cyclo-[GG]SFTI-1[LPET] and wild type SFTI-1 (both peptides_95% pure) were analyzed using TOCSY and NOESY experiments (Fig. 9). Two distinct conformations, in an approximate 9:1 ratio, were observed for cyclo-[GG]SFTI-1[LPET]. The secondary shifts of the major conformation were almost identical with those of wild type SFTI-1 (Fig. 7D). The minor conformation appears to be due to cis-trans isomerization of a proline residue. Pro8 in wild type STFI-1 is in the cis- conformation, with strong NOEs observed between Ile7 Hα and Pro8 Hi+1. By contrast, in cyclo-[GG]SFTI- 1[LPET], two sets of NOEs interacting with either δ-proton or α-proton of Pro8 were observed, suggesting that Pro8 exists in both cis and trans conformations (49).

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 7. (A) Cyclization strategy of [G]Vc1.1[GLPETGGS] and [GG]SFTI-1[LPETGG] by SrtA. (B) Retention time and MALDI-TOF analysis of linear precursor [G]Vc1.1[GLPETGGS], cyclic reduced and cyclic oxidized [G]Vc1.1[LPET]. The peptide status is shown with schematic representations. (C) One-pot reaction of oxidation and cyclization of [GG]SFTI-1[LPETGG]. Time course analysis of SrtA reaction by RP-HPLC (left), MALDI-TOF data (right) of linear precursor [GG]SFTI-1[LPETGG] and cyclic oxidised(cyclo)-[GG]SFTI-1[LPET]. (D) Hα secondary shift comparison of cyclo-[GG]SFTI-1[LPET] and wild type SFTI-1. (E) Hα secondary shift comparison of cyclo-[G]Vc1.1[GLPET], wild type Vc1.1, and previously published data of orally active cVc1.1. Sorting motif is shown in a dashed box.

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Table 1: Energies and structural statistics for the family of 20 lowest energy cyclo-[GGG]kB1[T] structuresa.

Energies (kcal/mol) Overall -975.7 ± 38.5 Bonds 20.5 ± 1.4 Angles 52.5 ± 4.3 Improper 17.4 ± 2.3 Van der Waals -96.1 ± 4.7 NOE 0.3 ± 0.03 cDih 0.3 ± 0.2 Dihedral 129.7 ± 1.6 Electrostatic -1100.3 ± 36.4 MolProbity Statistics Clashes (>0.4 Å / 1000 atoms) 12.8 ± 4.1 Poor rotamers 0.3 ± 0.4 Ramachandran Outliers (%) 0.0 ± 0.0 Ramachandran Favoured (%) 87.6 ± 2.7 MolProbity score 2.3 ± 0.2 MolProbity score percentile 56.2 ± 13.4b Atomic RMSDc (Å) Mean global backbone (1-24) 0.53 ± 0.14 Mean global heavy (1-24) 1.11 ± 0.25 Mean global backbone (1-33) 1.53 ± 0.58 Mean global heavy (1-33) 1.90 ± 0.53 Distance Restraints Intraresidue (i-j = 0) 59 Sequential (/i-j/ = 1) 169 Medium range (/i-j/ < 5) 53 Long range (/i-j/ > 5) 107 Hydrogen bonds 10 Total 398 Dihedral angle restraints f 26 Ψ 20 c1 6 Total 53 Violations from experimental restraints Total NOE violations exceeding 0.3 Å 0 Total Dihedral violations exceeding 3.0° 0

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 8. NMR spectra of cyclo-[G]Vc1.1[GLPET]. TOCSY spectrum (upper panel) and NOESY spectrum

(lower panel) recorded at 298 K, 90% H2O/10% D2O. There was a subset of extra residues found to be from Gly1 to Ser4 and Glu14 to Gly22 with overlapping or similar chemical shifts to the major conformer. This is caused by cis-trans isomerization of Pro19 in the linker (GLPETG), but it did not affect the overall structure of cyclo-[G]Vc1.1[GLPET].

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

Figure 9. NMR spectra of cyclo-[GG]SFTI-1[LPET]. TOCSY spectrum (upper panel) and NOESY spectrum (lower panel) recorded at 298 K, 90% H2O/10% D2O.

3.7. DISCUSSION Disulfide-rich cyclic peptides offer an attractive framework for the development of stable bioactive peptides with pharmaceutical potential. They display a remarkable degree of functional plasticity and high stability. Accordingly, interest has been shown both in the cyclization of linear disulfide-rich peptides (especially those isolated from venoms) and the development of native cyclotides as therapeutics and in their use as scaffolds for the grafting of pharmaceutically relevant epitopes onto the native cyclotide framework (74, 75). A bottleneck in the development of these peptides as potential therapeutics, however, is the production of 61

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

linear precursors and their subsequent cyclization and folding. NCL (using t-butoxycarbonyl chemistry) has been the preferred strategy for the production of disulfide-rich cyclic peptides, but this approach requires reaction conditions that cannot always be achieved using Fmoc chemistry or recombinantly expressed proteins. We therefore investigated SrtA-mediated cyclization as a mechanism for the increased efficiency of one of the steps in cyclic peptide production, namely head-to-tail cyclization of synthetically produced peptides. Specifically, we explored the possibility of using SrtA to enzymatically cyclize SFTI-1, Vc1.1 (with a linker), and kB1 and showed that SrtA offers an efficient, safe, and cost-effective method for the cyclization of disulfide-rich peptides ranging in size from 14 to 29 amino acids with 1–3 disulfide bonds.

The presence of a native Leu-Pro-Val in loop 6 of kB1 made this peptide an ideal model for cyclization using SrtA. These three residues fortuitously form a part of the penta-amino acid sorting motif, LPXTG, and cyclization was achieved with minimal disruption to the native sequence of kB1. In NCL-based synthesis of kB1, the cyclization and oxidation is carried out in a one-pot fashion. However, the reaction requires a two- step procedure when using SrtA for cyclization, with oxidation occurring either after or before SrtA- mediated cyclization. In our hands, the order of the reaction (i.e. oxidation/cyclization versus cyclization/oxidation) had a significant effect on the yield of correctly folded peptide. Despite the minimal changes introduced in kB1 by SrtA cyclization, no correctly folded peptide was produced when oxidation occurred after cyclization. Small changes in the amino acid composition of cyclotides can have profound effects on the refolding of synthetic forms (6,20), and it is possible that the introduction of the TGGG amino acid sequence in loop 6 of kB1 had a negative effect on the oxidative refolding of the cyclic form. The proximity of the cyclization point to CysI (Fig. 1) might also have adversely affected oxidative folding. The only region of significant structural perturbation in cyclo-[GGG]kB1[T] was for the Leu-Pro-Val residues immediately preceding CysI, and, when combined with restraints imposed by cyclization, conformational variations in this region of the reduced, cyclized peptide might have prevented correct disulfide formation.

Production of correctly folded cyclo-[GGG]kB1[T] was achieved by oxidative refolding prior to cyclization with SrtA. Monitoring of the cyclization reaction showed what appeared to be intermediates at various stages of the reaction (Fig. 2B), but by 24 h, the reaction had efficiently converted the linear peptides to the cyclic form in _49% yield. Several pieces of NMR evidence suggest that cyclo- [GGG]kB1[T] adopts the fold of native kB1: (i) Pro20 adopts a cis-conformation as in native kB1; (ii) The HN of Trp19 is not observed in TOCSY spectra but is observed in NOESY spectra, as characteristically applies to kB1; (iii) The HB of Pro20 is shifted upfield to -0.26 ppm; (iv) the protonation/deprotonation of Glu3 significantly affects the chemical shifts of HN of Asn11 and Thr12; (v) the χ1 dihedral angles of all six cysteine residues agree with those for kB1, providing strong evidence for an identical disulfide connectivity. Consistent with this connectivity, NOEs between Cys1 Hβ2/Hβ3 to Cys15 Hβ3 and Cys10 Hβ2/Hβ3 to Cys22 Hβ2 were observed (NOEs between Cys5Hβ and Cys17Hβ could not be unambiguously identified due to extensive 62

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

spectral overlap). Interestingly, NOEs between Hβ protons of non-connected cysteine residues, including Cys10–Cys15, Cys1–Cys22, and Cys15–Cys22, were also observed, but this is in complete accordance with a study on native kB1 by Rosengren et al. (54), where it was suggested that for cyclic cystine knot peptides, H_ cross-peaks are not the preferred diagnostic tool for determining disulfide connectivity due to the close proximity of all disulfide bonds. From the extensive evidence noted above, especially the fact that the two peptides share identical χ1 dihedral angles for all of the Cys residues, we were confident that we produced cyclo-[GGG]kB1[T] with the native disulfide connectivity. Thus, we calculated the structure of cyclo- [GGG]kB1[T] using the cysteine connectivities as native kB1 (PDB code 1NB1).

The hemolytic activity of cyclo-[GGG]kB1[T] was significantly reduced compared with native kB1. Because structural perturbation of cyclo-[GGG]kB1[T] was minimal (Fig. 5), the loss of hemolytic activity is presumably due to the effects of the residues introduced into loop 6 upon cyclization. Previous grafting experiments utilizing kB1 have also shown the abrogation of hemolytic activity after relatively minor changes to the primary sequence; for example, the grafting of the vascular endothelial growth factor sequence (RRKRRR) into loop 2, 3, 5, or 6 significantly reduced the hemolytic activity of kB1 (20). The effects of single point mutations on the hemolytic activityof kB1 have also been extensively studied (73). The hemolytic activity of kB1 was dramatically reduced by Ala substitution on a specific set of residues, including the last three residues of loop 6 (Leu27, Pro28, and Val29). Similar results were observed during Lys substitution at the same positions (69). Barry et al.(72) also noted the complete loss of hemolytic activity of linearized kB1 with the RNGLP sequence deletion. All of the studies suggest that the hemolytic activity of kB1 can be diminished when a less hydrophobic residue is introduced into the last three positions of loop 6. In kalata B2, a close homolog of kB1, loop 6 forms part of a hydrophobic interface that mediates interactions between the cyclotide and the plasma membrane (76). In this context, the loss of hemolytic activity after the introduction of additional residues to loop 6 suggests the interruption of critical membrane/peptide interactions in this region of the peptide.

The residues introduced into loop 6 might have also contributed to the slight reduction in stability of cyclo- [GGG]kB1[T] relative to the native peptide. Chemical shifts for these residues differed little from random coil values, suggesting that structural disorder at the cyclization point made the peptide slightly more susceptible to plasma proteases than native kB1. This disorder was also confirmed in the NMR solution structure of [GGG]kB1[T], with loop 6 displaying markedly increased disorder compared with native kB1. The serum stability results show that the cystine knot formation is crucial for stability because linear oxidized [GGG]kB1[TGG] was largely (~71%) intact after 24 h.

The sortase-catalyzed process requires modest modifications in order to render peptides amenable to cyclization. In this work, the cyclization point was chosen on the basis of a pre-existing Leu-Pro-Val motif in 63

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

loop 6, which overlapped with the SrtA recognition motif. This fortuitous coincidence resulted in the introduction of three fewer non-native amino acids than would be required using a different cyclization point. Analysis of chemical shifts showed that apart from differences of ~0.5 ppm in the Leu-Pro-Val residues of the cyclization point, the remainder of the SrtA cyclized peptide exhibited very little structural perturbation. Loop 6 of cyclotides exhibits the greatest diversity in structure, sequence, and length (77), and the lack of structural perturbation in this case reflects the ability of this loop to accommodate a range of structural motifs without affecting the overall fold of the peptide. Accordingly, the use of this loop for cyclization should provide a convenient and simple mechanism for production of synthetic cyclotides for a variety of applications.

To test the generality of SrtA-mediated cyclization, we investigated if it could be applied to other small disulfide-rich peptides, such as_-conotoxin Vc1.1 and SFTI-1. MALDI-TOF and NMR analysis confirmed that their cyclized and oxidized structures were similar to the native structures. In contrast to kB1, 9 h of SrtA reaction appeared to be sufficient to obtain the maximum yield for cyclo-[GG]SFTI-1[LPET], presumably due to its smaller size. Beyond this incubation time, some hydrolysis of the peptide was found to occur, resulting in linearization because the final product still contains a LPETG motif. Interestingly, introduction of the sortase motif in SFTI-1 (cyclo-[GG]SFTI-1[LPET]) gave rise to cis-trans isomerization of Pro8, as highlighted by the observation of strong NOEs between both Ile7 Hαi to Pro8 Hαi+1 (cis) and Ile7

Hαi to Pro8 Hδi+1 (trans). Although the LPXTG motif was introduced to the cyclization loop of SFTI-1, presumably placing the motif next to the Asp14 made the structure of the peptide less constrained than the wild type. This is consistent with a previous study showing that Ile7, Pro8, Pro9, and Asp14 are important for structural integrity (49). In that study, when Asp14 was replaced with Ala, the turn region was destabilized, resulting in cis-trans isomerization of Pro13 (49). Although the isomerization site is different, introduction of the SrtA sorting motif appears to have destabilized the turn region, giving rise to the minor conformation observed by NMR.

The work described here demonstrates the applicability of the SrtA method for the cyclization of different classes of disulfide-rich peptides without significant structural perturbation to the overall peptide fold. We have shown that this method is applicable whether the peptides are cyclic in their native form, such as SFTI- 1 containing one disulfide bond or kB1 with a three-disulfide bond cyclic cystine knot motif, or whether the peptide is linear in its native form and the cyclization site is introduced by the addition of a linker, such as for cVc1.1. SrtA mediated head-to-tail cyclization works optimally with peptides larger than 12-mers, excluding the SrtA sorting signal (40). Cyclotides are typically 28–37 amino acids in size, making all characterized cyclotides suitable for SrtA cyclization. In addition, other disulfide-rich peptides (e.g. conotoxins (78) or defensins (79, 80)) are also within this size range, making SrtA mediated cyclization a possibility for a wide range of peptides of therapeutic interest. Importantly, the ability to place the SrtAtags 64

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

at any position within a peptide sequence makes it possible to introduce cyclization points anywhere in the peptide of interest and thus minimize disruption of biologically or structurally important motifs. Thus, peptide bond closure of linear precursors through sortase-catalyzed circularization represents a simple, safe, and effective method not only for the production of synthetic cyclotides but also for the introduction of increased stability and bioavailability to other disulfide-rich peptides with therapeutic value.

3.8. REFERENCES 1. Colgrave, M. L., and Craik, D. J. (2004) Thermal, chemical, and enzymatic stability of the cyclotide kalata B1. The importance of the cyclic cystine knot. Biochemistry 43, 5965–5975

2. Gran, L. (1970) An oxytocic principle found in Oldenlandia affinis DC. An indigenous, Congolese drug “Kalata-Kalata” used to accelerate delivery. Medd. Nor. Farm. Selsk. 32, 173–180

3. Clark, R. J., Jensen, J., Nevin, S. T., Callaghan, B. P., Adams, D. J., and Craik, D. J. (2010) The engineering of an orally active conotoxin for the treatment of neuropathic pain. Angew. Chem. Int. Ed. Engl. 49, 6545–6548

4. Wong, C. T., Rowlands, D. K., Wong, C. H., Lo, T. W., Nguyen, G. K., Li, H. Y., and Tam, J. P. (2012) Orally active peptidic bradykinin B1 receptor antagonists engineered from a cyclotide scaffold for inflammatory pain treatment. Angew. Chem. Int. Ed. Engl. 51, 5620–5624

5. Cemazar, M., Kwon, S., Mahatmanto, T., Ravipati, A. S., and Craik, D. J. (2012) Discovery and applications of disulfide-rich cyclic peptides. Curr. Top. Med. Chem. 12, 1534–1545

6. Clark, R. J., Fischer, H., Dempster, L., Daly, N. L., Rosengren, K. J., Nevin, S. T., Meunier, F. A., Adams, D. J., and Craik, D. J. (2005) Engineering stable peptide toxins by means of backbone cyclization. Stabilization of the α-conotoxin MII. Proc. Natl. Acad. Sci. U.S.A. 102, 13767–13772

7. Luckett, S., Garcia, R. S., Barker, J. J., Konarev, A. V., Shewry, P. R., Clarke, A. R., and Brady, R. L. (1999) High-resolution structure of a potent, cyclic proteinase inhibitor from sunflower seeds. J. Mol. Biol. 290, 525–533

8. Long, Y. Q., Lee, S. L., Lin, C. Y., Enyedy, I. J., Wang, S., Li, P., Dickson, R. B., and Roller, P. P. (2001) Synthesis and evaluation of the sunflower derived trypsin inhibitor as a potent inhibitor of the type II transmembrane serine protease, matriptase. Bioorg. Med. Chem. Lett. 11, 2515–2519

9. Craik, D. J., Daly, N. L., Bond, T., and Waine, C. (1999) Plant cyclotides. A unique family of cyclic and knotted proteins that defines the cyclic cystine knot structural motif. J. Mol. Biol. 294, 1327–1336

10. Trabi, M., and Craik, D. J. (2002) Circular proteins. No end in sight. Trends Biochem. Sci. 27, 132–138

11. Jennings, C., West, J., Waine, C., Craik, D., and Anderson, M. (2001) Biosynthesis and insecticidal properties of plant cyclotides. The cyclic knotted proteins from Oldenlandia affinis. Proc. Natl. Acad. Sci. U.S.A. 98,10614–10619

12. Jennings, C. V., Rosengren, K. J., Daly, N. L., Plan, M., Stevens, J., Scanlon, M. J., Waine, C., Norman, D. G., Anderson, M. A., and Craik, D. J. (2005) Isolation, solution structure, and insecticidal activity of kalata B2, a circular protein with a twist. Do Mobius strips exist in nature? Biochemistry 44,851–860

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

13. Plan, M. R., Saska, I., Cagauan, A. G., and Craik, D. J. (2008) Backbone cyclised peptides from plants show molluscicidal activity against the rice pest Pomacea canaliculata (golden apple snail). J. Agric. Food Chem. 56, 5237–5241

14. Gustafson, K. R., Sowder, R. C., Henderson, L. E., Parsons, I. C., Kashman, Y., Cardellina, J. H., McMahon, J. B., Buckheit, R. W., Jr., Pannell, L. K., and Boyd, M. R. (1994) Circulins A and B. Novel human immunodeficiency virus (HIV)-inhibitory macrocyclic peptides from the tropical tree Chassalia parvifolia. J. Am. Chem. Soc. 116, 9337–9338

15. Tam, J. P., Lu, Y. A., Yang, J. L., and Chiu, K. W. (1999) An unusual structural motif of antimicrobial peptides containing end-to-end macrocycle and cystine-knot disulfides. Proc. Natl. Acad. Sci. U.S.A. 96, 8913–8918

16. Gruber, C. W., Cemazar, M., Anderson, M. A., and Craik, D. J. (2007) Insecticidal plant cyclotides and related cystine knot toxins. Toxicon 49, 561–575

17. Lindholm, P., Go¨ransson, U., Johansson, S., Claeson, P., Gullbo, J., Larsson, R., Bohlin, L., and Backlund, A. (2002) Cyclotides. A novel type of cytotoxic agents. Mol. Cancer Ther. 1, 365–369

18. Svangård, E., Go¨ransson, U., Hocaoglu, Z., Gullbo, J., Larsson, R., Claeson, P., and Bohlin, L. (2004) Cytotoxic cyclotides from Viola tricolor. J. Nat. Prod. 67, 144–147

19. Kaas, Q., and Craik, D. J. (2010) Analysis and classification of circular proteins in CyBase. Biopolymers 94, 584–591

20. Gunasekera, S., Foley, F. M., Clark, R. J., Sando, L., Fabri, L. J., Craik, D. J., and Daly, N. L. (2008) Engineering stabilized vascular endothelial growth factor-A antagonists. Synthesis, structural characterization, and bioactivity of grafted analogues of cyclotides. J. Med. Chem. 51, 7697–7704

21. Thongyoo, P., Roque´-Rosell, N., Leatherbarrow, R. J., and Tate, E. W. (2008) Chemical and biomimetic total syntheses of natural and engineered MCoTI cyclotides. Org. Biomol. Chem. 6, 1462– 1470

22. Thongyoo, P., Bonomelli, C., Leatherbarrow, R. J., and Tate, E. W. (2009) Potent inhibitors of_-tryptase and human leukocyte elastase based on the MCoTI-II scaffold. J. Med. Chem. 52, 6197–6200

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

24. Ji, Y., Majumder, S., Millard, M., Borra, R., Bi, T., Elnagar, A. Y., Neamati, N., Shekhtman, A., and Camarero, J. A. (2013) In vivo activation of the p53 tumor suppressor pathway by an engineered cyclotide. J. Am. Chem. Soc. 135, 11623–11633

25. Contreras, J., Elnagar, A. Y., Hamm-Alvarez, S. F., and Camarero, J. A. (2011) Cellular uptake of cyclotide MCoTI-I follows multiple endocytic pathways. J. Control. Release 155, 134–143

26. Greenwood, K. P., Daly, N. L., Brown, D. L., Stow, J. L., and Craik, D. J. (2007) The cyclic cystine knot miniprotein MCoTI-II is internalized into cells by macropinocytosis. Int. J. Biochem. Cell Biol. 39, 2252–2264

27. Cascales, L., Henriques, S. T., Kerr, M. C., Huang, Y. H., Sweet, M. J., Daly, N. L., and Craik, D. J. (2011) Identification and characterization of a new family of cell-penetrating peptides. J. Biol. Chem. 286, 36932–36943 66

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

28. Henriques, S. T., Huang, Y. H., Rosengren, K. J., Franquelim, H. G., Carvalho, F. A., Johnson, A., Sonza, S., Tachedjian, G., Castanho, M. A., Daly, N. L., and Craik, D. J. (2011) Decoding the membrane activity of the cyclotide kalata B1. The importance of phosphatidylethanolamine phospholipids and lipid organization on hemolytic and anti-HIV activities. J. Biol. Chem. 286, 24231–24241

29. Aboye, T. L., and Camarero, J. A. (2012) Biological synthesis of circular polypeptides. J. Biol. Chem. 287, 27026–27032

30. Pattabiraman, V. R., and Bode, J. W. (2011) Rethinking amide bond synthesis. Nature 480, 471–479

31. Clark, R. J., and Craik, D. J. (2010) Native chemical ligation applied to the synthesis and bioengineering of circular peptides and proteins. Biopolymers 94, 414–422

32. Dawson, P. E., Muir, T. W., Clark-Lewis, I., and Kent, S. B. (1994) Synthesis of proteins by native chemical ligation. Science 266, 776–779

33. Kimura, R., and Camarero, J. A. (2005) Expressed protein ligation. A new tool for the biosynthesis of cyclic polypeptides. Protein Pept. Lett. 12,789–794

34. Muir, T. W. (2003) Semisynthesis of proteins by expressed protein ligation. Annu. Rev. Biochem. 72, 249–289

35. Tavassoli, A., and Benkovic, S. J. (2007) Split-intein mediated circular ligation used in the synthesis of cyclic peptide libraries in E. coli. Nat. Protoc. 2, 1126–1133

36. Kawakami, T., Ohta, A., Ohuchi, M., Ashigai, H., Murakami, H., and Suga, H. (2009) Diverse backbone-cyclized peptides via codon reprogramming. Nat. Chem. Biol. 5, 888–890

37. Parthasarathy, R., Subramanian, S., and Boder, E. T. (2007) Sortase A as a novel molecular “stapler” for sequence-specific protein conjugation. Bioconjug. Chem. 18, 469–476

38. Antos, J. M., Popp, M. W., Ernst, R., Chew, G. L., Spooner, E., and Ploegh, H. L. (2009) A straight path to circular proteins. J. Biol. Chem. 284,16028–16036

39. Bolscher, J. G., Oudhoff, M. J., Nazmi, K., Antos, J. M., Guimaraes, C. P., Spooner, E., Haney, E. F., Garcia Vallejo, J. J., Vogel, H. J., van’t Hof, W.,Ploegh, H. L., and Veerman, E. C. (2011) Sortase A as a tool for high-yield histatin cyclization. FASEB J. 25, 2650–2658

40. Wu, Z., Guo, X., and Guo, Z. (2011) Sortase A-catalyzed peptide cyclization for the synthesis of macrocyclic peptides and glycopeptides. Chem. Commun. 47, 9218–9220

41. Mazmanian, S. K., Liu, G., Ton-That, H., and Schneewind, O. (1999) Staphylococcus aureus sortase, an enzyme that anchors surface proteins to the cell wall. Science 285, 760–763

42. Popp, M. W., and Ploegh, H. L. (2011) Making and breaking peptide bonds. Protein engineering using sortase. Angew. Chem. Int. Ed. Engl. 50, 5024–5032

43. Marraffini, L. A., Dedent, A. C., and Schneewind, O. (2006) Sortases and the art of anchoring proteins to the envelopes of Gram-positive bacteria. Microbiol. Mol. Biol. Rev. 70, 192–221

44. Ton-That, H., Liu, G., Mazmanian, S. K., Faull, K. F., and Schneewind, O. (1999) Purification and characterization of sortase, the transpeptidase that cleaves surface proteins of Staphylococcus aureus at the LPXTG motif. Proc. Natl. Acad. Sci. U.S.A. 96, 12424–12429

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

45. Antos, J. M., Chew, G. L., Guimaraes, C. P., Yoder, N. C., Grotenbreg, G. M., Popp, M. W., and Ploegh, H. L. (2009) Site-specific N- and C-terminal labeling of a single polypeptide using sortases of different specificity. J. Am. Chem. Soc. 131, 10800–10801

46. Popp, M. W., Dougan, S. K., Chuang, T. Y., Spooner, E., and Ploegh, H. L. (2011) Sortase-catalyzed transformations that improve the properties of cytokines. Proc. Natl. Acad. Sci. U.S.A. 108, 3169–3174

47. Ling, J. J., Policarpo, R. L., Rabideau, A. E., Liao, X., and Pentelute, B. L. (2012) Protein thioester synthesis enabled by sortase. J. Am. Chem. Soc.134, 10749–10752

48. Witte, M. D., Cragnolini, J. J., Dougan, S. K., Yoder, N. C., Popp, M. W., and Ploegh, H. L. (2012) Preparation of unnatural N-to-N and C-to-C protein fusions. Proc. Natl. Acad. Sci. U.S.A. 109, 11993– 11998

49. Daly, N. L., Chen, Y. K., Foley, F. M., Bansal, P. S., Bharathi, R., Clark, R. J.,Sommerhoff, C. P., and Craik, D. J. (2006) The absolute structural requirement for a proline in the P3_-position of Bowman-Birk protease inhibitors is surmounted in the minimized SFTI-1 scaffold. J. Biol. Chem. 281, 23668–23675

50. Popp, M. W., Antos, J. M., and Ploegh, H. L. (2009) Site-specific protein labeling via sortase-mediated transpeptidation. Curr. Protoc. Protein Sci.Chapter 15, Unit 15.13

51. Daly, N. L., Love, S., Alewood, P. F., and Craik, D. J. (1999) Chemical synthesis and folding pathways of large cyclic polypeptide. Studies of the cystine knot polypeptide kalata B1. Biochemistry 38, 10606– 10614

52. Huang, Y. H., Colgrave, M. L., Daly, N. L., Keleshian, A., Martinac, B., and Craik, D. J. (2009) The biological activity of the prototypic cyclotide kalata B1 is modulated by the formation of multimeric pores. J. Biol. Chem. 284, 20699–20707

53. Moffatt, F., Senkans, P., and Ricketts, D. (2000) Approaches towards the quantitative analysis of peptides and proteins by reversed-phase high performance liquid chromatography in the absence of a pure reference sample. J. Chromatogr. A 891, 235–242

54. Rosengren, K. J., Daly, N. L., Plan, M. R., Waine, C., and Craik, D. J. (2003) Twists, knots, and rings in proteins. Structural definition of the cyclotide framework. J. Biol. Chem. 278, 8606–8616

55. Hwang, T. L., and Shaka, A. J. (1995) Water suppression that works. Excitation sculpting using arbitrary wave-forms and pulsed-field gradients. J. Magn. Reson. 112, 275–279

56. Griesinger, C., Sørensen, O. W., and Ernst, R. R. (1987) Practical aspects of the E.COSY technique. Measurement of scalar spin-spin coupling constants in peptides. J. Magn. Reson. 75, 474–492

57. Vranken, W. F., Boucher, W., Stevens, T. J., Fogh, R. H., Pajon, A., Llinas, M., Ulrich, E. L., Markley, J. L., Ionides, J., and Laue, E. D. (2005) The CCPN data model for NMR spectroscopy. Development of a software pipeline. Proteins 59, 687–696

58. Bartels, C., Xia, T. H., Billeter, M., Gu¨ntert, P., and Wu¨thrich, K. (1995) The program XEASY for computer-supported NMR spectral analysis of biological macromolecules. J. Biomol. NMR 6, 1–10

59. Wishart, D. S., Bigam, C. G., Yao, J., Abildgaard, F., Dyson, H. J., Oldfield, E., Markley, J. L., and Sykes, B. D. (1995) 1H, 13C and 15N chemical shift referencing in biomolecular NMR. J. Biomol. NMR 6, 135–140

60. Gu¨ntert, P. (2004) Automated NMR structure calculation with CYANA. Methods Mol. Biol. 278, 353– 68

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

378

61. Shen, Y., Delaglio, F., Cornilescu, G., and Bax, A. (2009) TALOS_. A hybrid method for predicting protein backbone torsion angles from NMR chemical shifts. J. Biomol. NMR 44, 213–223

62. Nederveen, A. J., Doreleijers, J. F., Vranken, W., Miller, Z., Spronk, C. A., Nabuurs, S. B., Gu¨ntert, P., Livny, M., Markley, J. L., Nilges, M., Ulrich, E. L., Kaptein, R., and Bonvin, A. M. (2005) RECOORD. A recalculated coordinate database of 500_proteins from the PDB using restraints from the BioMagResBank. Proteins 59, 662–672

63. Bru¨nger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Crystallography&NMRsystem. A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921

64. Dominguez, C., Boelens, R., and Bonvin, A. M. (2003) HADDOCK. A protein-protein docking approach based on biochemical or biophysical information. J. Am. Chem. Soc. 125, 1731–1737

65. Linge, J. P., and Nilges, M. (1999) Influence of non-bonded parameters on the quality of NMR structures. A new force field for NMR structure calculation. J. Biomol. NMR 13, 51–59

66. Conibear, A. C., Rosengren, K. J., Harvey, P. J., and Craik, D. J. (2012) Structural characterization of the cyclic cystine ladder motif of θ-defensins. Biochemistry 51, 9718–9726

67. Chen, V. B., Arendall, W. B., 3rd, Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S., and Richardson, D. C. (2010) MolProbity. All-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12–21

68. Koradi, R., Billeter, M., and Wu¨thrich, K. (1996) MOLMOL. A program for display and analysis of macromolecular structures. J. Mol. Graph. 14,51–55, 29–32

69. Huang, Y. H., Colgrave, M. L., Clark, R. J., Kotze, A. C., and Craik, D. J. (2010) Lysine-scanning mutagenesis reveals an amendable face of the cyclotide kalata B1 for the optimization of nematocidal activity. J. Biol. Chem. 285, 10797–10805

70. Levary, D. A., Parthasarathy, R., Boder, E. T., and Ackerman, M. E. (2011) Protein-protein fusion catalyzed by sortase A. PLoS One 6, e18342

71. Wüthrich, K. (1986) NMR of Proteins and Nucleic Acids, pp. 16–157, Wiley-Interscience, New York

72. Barry, D. G., Daly, N. L., Clark, R. J., Sando, L., and Craik, D. J. (2003) Linearization of a naturally occurring circular protein maintains structure but eliminates hemolytic activity. Biochemistry 42, 6688– 6695

73. Simonsen, S. M., Sando, L., Rosengren, K. J., Wang, C. K., Colgrave, M. L., Daly, N. L., and Craik, D. J. (2008) Alanine scanning mutagenesis of the prototypic cyclotide reveals a cluster of residues essential for bioactivity. J. Biol. Chem. 283, 9805–9813

74. Craik, D. J., Swedberg, J. E., Mylne, J. S., and Cemazar, M. (2012) Cyclotides as a basis for drug design. Expert Opin. Drug. Discov. 7, 179–194

75. Gould, A., Ji, Y., Aboye, T. L., and Camarero, J. A. (2011) Cyclotides, a novel ultrastable polypeptide scaffold for drug discovery. Curr. Pharm. Des. 17, 4294–4307

Chapter 3. Semienzymatic cyclization of disulfide-rich peptides using sortase A

76. Wang, C. K., Colgrave, M. L., Ireland, D. C., Kaas, Q., and Craik, D. J. (2009) Despite a conserved cystine knot motif, different cyclotides have different membrane binding modes. Biophys. J. 97, 1471– 1481

77. Mulvenna, J. P., Sando, L., and Craik, D. J. (2005) Processing of a 22 kDa precursor protein to produce the circular protein tricyclon A. Structure 13, 691–701

78. Terlau, H., and Olivera, B. M. (2004) Conus venoms.Arich source of novel ion channel-targeted peptides. Physiol. Rev. 84, 41–68

79. Ganz, T. (2003) Defensins. Antimicrobial peptides of innate immunity. Nat. Rev. Immunol. 3, 710–720

80. Wong, J. H., Xia, L., and Ng, T. B. (2007) A review of defensins of diverse origins. Curr. Protein Pept. Sci. 8, 446–459

81. Wishart, D. S., Bigam, C. G., Holm, A., Hodges, R. S., and Sykes, B. D. (1995) 1H, 13C and 15N random coil NMR chemical shifts of the common amino acids. I. Investigations of nearest-neighbor effects. J. Biomol. NMR 5, 67–81

82. Davis, I. W., Leaver-Fay, A., Chen, V. B., Block, J. N., Kapral, G. J., Wang, X., Murray, L. W., Arendall, W. B., 3rd, Snoeyink, J., Richardson, J. S., and Richardson, D. C. (2007) MolProbity. All-atom contacts and structure validation for proteins and nucleic acids. Nucleic Acids Res. 35, W375–W383

Chapter 4 Efficient enzymatic cyclization of an inhibitory cystine knot- containing peptide

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.1. Overview and Declaration of authorship Overview This chapter includes an accepted manuscript describing enzymatic cyclization of an inhibitory cystine knot containing peptide.

Declaration of authorship

In accordance with the guidelines of the University of Queensland, this chapter is presented as an accepted manuscript for publication of a research article.

“Efficient enzymatic cyclization of an inhibitory cystine knot-containing peptide”

Biotechnology and Bioengineering, accepted on April 19, 2016

The experiments were conceived and designed by Soohyun Kwon under the supervision of Christina Schroeder and David Craik. Soohyun Kwon did peptide purification, oxidative folding, cyclization, NMR analysis, stability assay, and sortase A expression. Frank Bosmans carried out electrophysiology experiments. Modelling study was performed by Quentin Kaas. Anne Conibear and Christina Schroeder guided NMR structure calculation, K. Johan Rosengren and Conan Wang assisted with structure refinement using CNS. Peptides were synthesized by Fmoc chemistry on an automated synthesizer by Olivier Cheneval and Phill Walsh. The manuscript and figures were prepared by Soohyun Kwon, Quentin Kaas, and Christina Schroeder, and revised by Frank Bosmans, Christina Schroeder and David Craik.

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.2. (Cover page of the published manuscript)

Efficient enzymatic cyclization of an inhibitory cystine knot-containing peptide

Running title: Enzymatic cyclization of ICK peptide κ-PVIIA

Soohyun Kwon1, Frank Bosmans2, Quentin Kaas1, Olivier Cheneval1, Anne C. Conibear1, K. Johan Rosengren3, Conan K. Wang1, Christina I. Schroeder1*, David J. Craik1*

1 The University of Queensland, Institute for Molecular Bioscience, Brisbane 4072, Qld, Australia

2 Department of Physiology and Solomon H Snyder Department of Neuroscience, Johns Hopkins University School of Medicine, Baltimore, Maryland, 21205, USA

3 The University of Queensland, School of Biomedical Sciences, Brisbane 4072, Qld, Australia

Note: The structure of cyclized PVIIA has been deposited to the Protein Data Bank (code 2N8E) and the Biological Magnetic Resonance Bank (accession number 25847).

*To whom correspondence should be addressed:

Prof. David J. Craik

Tel: +61 7 3346 2019

Email: [email protected] or

Dr. Christina I. Schroeder

Tel: +61 7 3346 2021

Email: [email protected]

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.3. Abstract

Disulfide-rich peptides isolated from cone snails are of great interest as drug leads due to their high specificity and potency toward therapeutically relevant ion channels and receptors. They commonly contain the inhibitor cystine knot (ICK) motif comprising three disulfide bonds forming a knotted core. Here we report the first successful enzymatic backbone cyclization of a linear ICK-containing peptide κ-PVIIA, a 27-amino acid conopeptide from Conus purpurascens, using a mutated version of the bacterial transpeptidase, sortase A. Although a slight loss of activity was observed compared to native κ-PVIIA, cyclic κ-PVIIA is a functional peptide that inhibits the Shaker voltage-gated potassium (Kv) channel. Molecular modeling suggested that the decrease in potency may be related to the loss of crucial, but previously unidentified electrostatic interactions between the N-terminus of the peptide and the Shaker channel. This hypothesis was confirmed by testing an N-terminally acetylated κ-PVIIA, which showed a similar decrease in activity. We also investigated the conformational dynamics and hydrogen bond network of cyc-PVIIA, both of which are important factors to be considered for successful cyclization of peptides. We found that cyc-PVIIA has the same conformational dynamics, but different hydrogen bond network compared to those of κ-

PVIIA. The ability to efficiently cyclize ICK peptides using sortase A will enable future protein engineering for this class of peptides and may help in the development of novel therapeutic molecules.

Keywords: κ-PVIIA, inhibitory cystine knot, cyclization, sortase A

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.4. Introduction

Peptides are undergoing an exciting revival of interest due to recent technological advances in protein and peptide engineering (1). These advances have led to an increase in interest from the pharmaceutical industry in using peptides in a number of therapeutic applications because, unlike small molecules, peptides tend to be highly selective and typically very potent for a specific target.

However, a drawback often associated with peptides is a short half-life due to a lack of proteolytic stability. One way to improve the stability of a peptide is through backbone or side chain cyclization, as this reduces susceptibility to proteases and, in some cases, improves activity (2) and potentiates orally delivered activity (2-4). One class of peptides that has gained particular interest is venom-derived disulfide-rich peptides due to their inherent stability (brought about by the high number of cysteine residues) and ability to selectively and potently inhibit voltage-gated potassium

(KV), calcium (CaV) and sodium (NaV) channels. This inhibitory activity is particularly relevant given the involvement of these channels in numerous pathologies, including pain, stroke, epilepsy and multiple sclerosis (5-8).

κ-PVIIA is a 27-amino acid peptide isolated from the venom of the purple cone snail, Conus purpurascens (9). It has been reported to block the conductance of the Shaker Kv channel of

Drosophila melanogaster with an IC50 of 60 ± 3 nM (9-11). Three-dimensional structural analysis and a complete alanine scan suggested that the peptide inhibits the channel by physically occluding the pore (11-13). NMR structural analyses showed that κ-PVIIA has an inhibitor cystine knot (ICK)

(12,13) – a unique structural motif characterized by a triple-stranded anti-parallel β-sheet connected by three disulfide bonds, forming a knotted core. The ICK significantly contributes to the stability of peptides (14,15). The disulfide connectivity of the ICK motif is CI-CIV, CII-CV and CIII-CVI, and a common sequence patterns is: CX3-7CX3-6CX0-5CX1-4CX4-13C, (X being any amino acid) (16).

The ICK motif is found in a broad spectrum of phyla, including animals, plants, and fungi, with a particularly large prevalence in the venoms of cone snails, spiders and scorpions (17). 75

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

Interestingly, the ICK is widespread in venoms regardless of the pharmacological target, mechanism of action, or the species from which the peptides are extracted. κ-PVIIA is a member of the O1 conotoxin gene superfamily (9,18,19), and contains an ICK motif with CX6-CX6-CCX3-

CX5-C cysteine spacing. In addition to κ-PVIIA, this family comprises a diverse series of conotoxins with an ICK motif; for example, ω-conotoxin MVIIA (20), μ-conotoxin GS (21), δ- conotoxin PVIA (22), and μO-conotoxin MrVIB (23), which influence CaV and NaV channels

(Figure 1a) (24,25).

One class of peptides that is structurally related to the O1 conotoxin superfamily is the cyclotides (26). This family of plant-derived peptides comprises a cystine knot motif that is embedded within a cyclic backbone, a motif referred to as a cyclic cystine knot (CCK) (15). Six loops protrude from the core and are thought to direct the activity of cyclotides, which have been referred to as a natural combinatorial template (27,28). The combination of the cystine knot coupled with the cyclized backbone makes this family of peptides ultrastable and able to withstand harsh chemical, enzymatic or thermal conditions (29). In addition, their use in African folk medicinal teas suggests that they have a degree of oral bioavailability (30).

Chemical cyclization has been used in a few cases to improve the pharmaceutical properties of ICK peptides by making them more cyclotide-like (31,32). Cyclotides are typically sequentially assembled using standard Boc solid-phase peptide synthesis and cyclized using native chemical ligation (33-35). Recent reports describe the production of cyclotides using Fmoc compatible methods (36,37). We also recently showed that cyclization of disulfide-rich peptides, including cyclotide kalata B1, sunflower trypsin inhibitor-I (SFTI-I) and cyclic Vc1.1, could be achieved using the bacterial transpeptidase, sortase A (SrtA), as an alternative to chemical ligation methods

(38). In addition, another recent study demonstrated successful backbone cyclization of a recombinant cyclotide precursor of MCoTI-II using a mutant SrtA (P94S/D160N/D165A/K196T)

(39). 76

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

SrtA, originally isolated from Staphylococcus aureus, recognizes the penta-amino acid LPXTG motif (X = any amino acid) and cleaves the peptide bond between Thr and Gly, forming a thioacyl enzyme intermediate. The intermediate then undergoes nucleophilic attack by Glyn, which results in the formation of a new peptide bond between Thr and Glyn (40). SrtA has been widely used in protein engineering (41,42), and efforts to improve its catalytic activity were recently successful; specifically, through the use of a directed evolution strategy leading to the modified SrtA

(P94R/D160N/D165A/K190E/K196T), hereafter referred to as SrtA5º, with significantly enhanced ligation activity (43).

Little is known about the subtype selectivity of κ-PVIIA across KV channels and with the increasing interest in the potential treatment of multiple sclerosis by KV1.3 channel inhibition (44), we chose

κ-PVIIA as a model system not only to enzymatically cyclize an ICK peptide, but also to search for potential targets in addition to Shaker. Following the successful cyclization of kalata B1, SFTI-I and cyclic Vc1.1 using SrtA, we investigated whether cyclization of κ-PVIIA could be achieved by a semi-enzymatic approach using SrtA5º and determined how it would affect the activity and the stability of the peptide. This study describes the synthesis, oxidation and cyclization using SrtA5º of linear κ-PVIIA (Figure 1b) and reports its serum stability. Structural analysis using nuclear magnetic resonance (NMR) allowed us to compare the structures of cyclic κ-PVIIA (cyc-PVIIA) with κ-PVIIA; electrophysiological studies and molecular dynamics simulations provided insights into differences in interaction between cyc-PVIIA, κ-PVIIA and Shaker, and allowed us to define the κ-PVIIA KV channel selectivity profile.

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

Figure 1. (a) Sequence alignment of O1 gene superfamily conotoxins displaying an ICK motif. (b) Schematic representation of the experimental procedures used to cyclize κ-PVIIA.

4.5. Materials and Methods

4.5.1. Peptide synthesis

Linear peptides were synthesized by solid-phase peptide synthesis using an automatic peptide synthesizer (Symphony®, Protein Technologies, Inc. USA) following standard Fmoc chemistry protocols. 2-Chlorotrityl chloride resin was used for [GGG]PVIIA[LPETGG], whereas rink amide resin was used for linear κ-PVIIA to achieve C-terminal amidation. κ-PVIIA with N-terminal acetylation was produced by treating the resin with 6% acetic anhydride/0.2 M DIPEA in DMF following completion of elongation of the peptide. Peptides were released from the resins using trifluoroacetic acid (TFA)/triisopropylsilane/water (95:2.5:2.5) at room temperature for 2 h. TFA was removed using a rotary evaporator, and peptides were precipitated with ice-cold diethyl ether, dissolved in 50% v/v acetonitrile (ACN) containing 0.05% v/v TFA and lyophilized. Peptides were purified using reversed-phase high performance liquid chromatography (RP-HPLC) and the mass was confirmed using mass spectrometry.

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.5.2. Oxidative folding of peptides

κ-PVIIA and GGG-PVIIA-LPETGG were oxidized in 50 mM Tris pH 7.5, 1 mM GSH, 0.5 mM

GSSG at 4 °C for 4–5 days. The progress of oxidation was monitored by analytical RP-HPLC with a gradient from 0–50% v/v ACN (0.05% v/v TFA) over 50 min with a flow rate of 0.3 mL/min using an analytical C18 column (Agilent ZORBAX 300SB-C18, 5 μm, 2.1 x 150 mm). The oxidation yield was calculated based on HPLC profile. The mixture was purified by RP-HPLC using a 0.5%/min gradient, and the molecular weight was confirmed using mass spectrometry.

4.5.3. SrtA5º expression

SrtA5º (43) was expressed as previously described (45). Briefly, an overnight culture of E.coli

BL21 (DE3) transformed with SrtA5º plasmid was cultured in the presence of 50 μg/mL kanamycin until OD600=0.7. Protein expression was induced by addition of isopropyl-β-D-1- thiogalactopyranoside (IPTG) to a concentration of 1 mM for 3 h at 37 °C. Cells were harvested by centrifugation and resuspended in lysis buffer (50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 10 mM imidazole, 10% v/v glycerol, 20 μg/mL DNase I). Cells were sonicated, the supernatant was collected and the protein was purified using Ni-Sepharose (HisTrap FF, GE Healthcare). Fractions containing SrtA5º were combined and the buffer was exchanged using a PD-10 desalting column

(GE Healthcare).

4.5.4. Cyclization of peptides using SrtA5°

For cyclization, 50 μM of the peptide was incubated with 50 μM of SrtA5º in reaction buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 10 mM CaCl2) at 37 °C. Cyclization progress was monitored using RP-HPLC with a gradient of 0–50% v/v ACN (0.05% v/v TFA) over 50 min with a flow rate of 1 mL/min using an analytical C18 column (GraceSmartTM RP18, 5 μm, 4.6 x 150 mm). When the majority of linear peptide was deemed to be cyclic, the peptide was separated from SrtA5º using a 79

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

centrifugal concentrator (10 kDa), the solution containing the cyclic peptide was loaded onto a semipreparative C18 column (Phenomenex, Jupiter 5UC18, 300 Å, 250 x 10.0 mm, 5 μm) and eluted with a gradient of 36–45% ACN (0.05% TFA) over 30 min with a flow rate of 3 mL/min.

The molecular mass of purified cyclic-[GGG]PVIIA[LPET] was confirmed by matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry. Fractions showing the correct mass were lyophilized for further assays.

4.5.5. NMR spectroscopy and structure calculations

Cyc-PVIIA was dissolved in 90% H2O, 10% D2O (v/v) or 99.96% D2O at a concentration of 1 mM and ~pH 3.5. Two-dimensional NMR spectra were acquired at 298K on a Bruker Avance 600 MHz

NMR spectrometer equipped with a cryogenically cooled probe (Bruker, Karlsruhe, Germany).

Spectra were recorded with 4096 data points in the F2 and 512 increments in the F1 dimension for

TOCSY, NOESY, E.COSY, DQF-COSY, 2048 data points and 256 increments for 1H-13C HSQC, and 2048 data points and 128 increments for 1H-15N HSQC. Sequential assignment and structure calculations were performed as previously described (46). Briefly, The peaks in the NOESY spectra with 100 ms mixing time were picked manually, intraresidual and sequential NOEs were assigned, and a list of interproton distances was generated using the AUTO and CALC function in CYANA

(47). Following structural refinement within CNS (48), a set of 20 structures with the lowest energies, no NOE violations greater than 0.2 Å and no dihedral violations greater than 2.0º, was chosen for MolProbity analysis (49). PyMOL and MOLMOL (50) were used for generating figures and final superimposition.

4.5.6. Two-electrode voltage-clamp recording from Xenopus oocytes

Shaker-IR (IR for inactivation removed) (51,52), rKV2.1Δ7 (53,54), hKV1.3 (55), rKV1.6 (56,57), rNaV1.4 (58), and hNaV1.6 (59,60) were used in this study. The rKV2.1Δ7 construct contains seven 80

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

point mutations in the outer vestibule, rendering the channel sensitive to agitoxin-2 (61). cRNA of all constructs and mutants was synthesized using T7/SP6 polymerase (Message Machine kit, Life

Technologies, USA) after linearizing the sequenced DNA with appropriate restriction enzymes.

Channel constructs were expressed in Xenopus laevis oocytes and studied following 1–2 days incubation after manual cRNA injection as described before (62). Leak and background conductances, identified by blocking the channel with agitoxin-2, were subtracted for all KV channel currents shown. Voltage-activation relationships were obtained by measuring tail currents for KV channels and a single Boltzmann function was fitted to the data. Voltage-activation relationships were recorded in the absence and presence of different toxin concentrations. The ratio of currents (I/I0) recorded in the presence (I) and absence (I0) of toxin was calculated for various strength depolarizations. The value of I/I0 measured in the linear phase at 0 mV was taken as

Fraction unblocked. The IC50 of the toxin for KV channels was calculated using the Hill equation.

Off-line data analysis was performed using Clampfit 10 (Molecular Devices, USA) and Origin 7.5

(Originlab, USA).

4.5.7. Molecular modeling

The molecular model used for the studies of interaction between the Shaker K+ channel and linear

κ-PVIIA was kindly provided by Mhadavi and Kuyucak (63). A model of the interaction between cyc-PVIIA and the Shaker K+ channel was built by homology with Modeller 9v13 (64) using the model of Mhadavi and Kuyucak and the solution structure of cyc-PVIIA as templates. The homology model was then embedded in a pre-equilibrated 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphocholine (POPC) bilayer: the shaker channel was placed in the bilayer and POPC residues that overlap with the channel were removed. SPC water molecules were added to the system and sodium and chloride ions were added to simulate a physiological concentration of 150 mM sodium

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

chloride. The simulated system contained 100 POPC molecules, ~9,000 water molecules, 23 sodium ions and 26 chloride ions. After 10,000 steps of steepest descent minimization, the system was equilibrated for 10 ns while progressively releasing position restraints on the chanel and toxin.

The system was then simulated for 100 ns without restraint. All simulations used the Gromos 54a7 force field (65), the v-rescale thermostat (66), the Parrinello-Rahman barostat (67), and the reaction field method (68) to simulate the effect of long range dipole-dipole interactions. The electrostatic potential generated by the Shaker K+ channel was computed using APBS software (69). Molecular dynamics simulations of cyc-PVIIA and κ-PVIIA individually in solution were carried out using the

Amber 99SB-ILDN force field (70). Each peptide was embedded in a water box of ~2,600 SPC water molecules, containing sodium and chloride ions at 150 mM concentrations. All molecular dynamics simulations were carried out using the Gromacs 5.1 molecular dynamics engine (71).

4.5.8. Serum stability assay

The stability of native and cyc-PVIIA in human serum was tested at a peptide concentration of 20

μM using a method previously described (72).

4.6. Results and Discussion

4.6.1. κ-PVIIA precursor synthesis, oxidation and cyclization

κ-PVIIA and the linear precursor [GGG]PVIIA[LPETGG] were synthesized by Fmoc chemistry.

Oxidative folding was monitored by analytical HPLC over time, and both peptides showed a similar profile, with one prominent peak after 4 days (Figure 2a). For cyclization, we used a mutated S. aureus SrtA which has improved catalytic activity compared to wild-type SrtA (43). Among the mutant SrtAs described by Chen et al., we tested the two most active (i)

P94S/D160N/D165A/K196T and (ii) P94R/D160N/D165A/K190E/K196T (SrtA5º) using GFP-

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

LPETGG as the substrate. In our hands, SrtA5º showed slightly better catalytic activity compared to the tetramutant, and consequently was used in subsequent experiments (data not shown). The linear oxidized [GGG]PVIIA[LPETGG] was incubated with SrtA5º for cyclization, and progress was monitored by RP-HPLC. Based on calculations of peak area derived from HPLC, 46% and 57% of conversion from linear [GGG]PVIIA[LPETGG] to cyclic-[GGG]PVIIA[LPET] was observed after

3.5 h and 8 h, respectively (Figure 2b, c). The peak eluting before cyc-PVIIA contained a peptide with the same mass of cyclic peptide and was assumed to be an isomer, but did not yield sufficient material for further analysis. An attempt to synthesize κ-PVIIA by Boc-chemistry using an automatic synthesizer was unsuccessful; therefore, cyclization by native chemical ligation was not achieved.

Figure 2. Oxidation and cyclization of [GGG]PVIIA[LPETGG]. (a) RP-HPLC profile of [GGG]PVIIA[LPETGG] oxidation. The asterisk represents the major peak of linear oxidized product. (b) Cyclization progress of [GGG]PVIIA[LPETGG] over time using SrtA5º and (c)

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

MALDI-TOF analysis of linear oxidized (calculated monoisotopic mass: 3992.7 Da) and cyclic oxidized PVIIA (calculated monoisotopic mass: 3860.6 Da).

4.6.2. NMR solution structure of cyc-PVIIA

Linear oxidized [GGG]PVIIA[LPETGG] and cyc-PVIIA were analysed by two-dimensional 1H

NMR. The peaks for both peptides in the amide proton region of the NMR spectra were well dispersed and all protons could be assigned by the sequential assignment procedure using TOCSY and NOESY spectra (73). Secondary Hα chemical shift comparison confirmed them to be similar to native κ-PVIIA (PDB identifier 1AV3) (12) and the positive secondary shifts for residues Q6-C8,

K19-C20 and K25-V27 indicated the presence of the characteristic β-sheet secondary structure

(Figure 3a). Backbone cyclization was confirmed by the presence of a αHi-NHi+1 NOE observed between Thr31 and Gly32, indicating peptide bond formation of the C-terminal LPET and N- terminal GGG motif of cyc-PVIIA.

The three-dimensional structure of cyc-PVIIA was calculated using 384 distance restraints, as well

as 17 φ, 17 ψ, four χ1 dihedral angle restraints and four hydrogen bonds derived from a D2O exchange experiment. The structure of cyc-PVIIA calculated using CYANA (47) was refined further in a water shell using CNS (48). The 20 conformers with the lowest energy were chosen for structural statistical analysis and showed a root mean square deviation of 0.91 ± 0.25 Å across the backbone, excluding the flexible loop 2 (residues F9-D14) and the disordered linker (residues L28-

G34) (Table I). Cyc-PVIIA showed high similarity to κ-PVIIA (PDB identifier 1AV3) (12) across the overall structure, except in the linker region (Figure 3b). As seen in other members of the O1 conotoxin superfamily, including MVIIA (20) and MrVIB (74), loop 2 of PVIIA is the most flexible region (not including the linker) based on RMSD comparison of each loop of NMR structures. To determine if cyclization introduced any changes in backbone flexibility of the structure, the Cα root-mean-square fluctuation (RMSF) of κ-PVIIA and cyc-PVIIA were measured 84

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

from molecular dynamics simulations. Overall, both peptides showed similar RMSF profiles

(Figure 3c). Loop 2 of cyc-PVIIA had slightly less flexibility than that of native PVIIA, which is consistent with cyclization being suggested to introduce structural rigidity and prevent disulfide bond shuffling (2).

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

Figure 3. (a) Secondary Hα chemical shift comparison of κ-PVIIA and cyc-PVIIA. Residues forming β-sheets are indicated with arrows. (b) The 20 lowest-energy structres of κ-PVIIA (PDB identifier 1AV3, left) (12) and cyc-PVIIA (PDB identifier 2N8E, right) superimposed over the backbone residues 2–7 and 15–26. Disulfide bonds are shown in grey with Roman numerals. (c) Root-mean-square fluctuation (RMSF) highlighting the flexibility of each loop of native and cyc- PVIIA.

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

Table I. Energies and structural statistics for the family of 20 lowest energy cyc-PVIIA structures. Energies (kcal/mol) Overall –1119.8 ± 73.9 Bonds 14.1 ± 1.5 Angles 44.5 ± 5.2 Improper 18.3 ± 4.9 Dihedral 158.5 ± 3.3 van der Waals –125.0 ± 3.3 Electrostatic –1230.3 ± 77.7 NOE 0.1 ± 0.0 cDih 0.1 ± 0.1 MolProbity statistics Clashes (> 0.4Å/1000 atoms) 7.5 ± 2.5 Poor rotamers 1.0 ± 1.0 Ramachandran outliers (%) 0.0 ± 0.0 Ramachandran favored (%)a 93.3 ± 4.1 RMSDb from mean structure (Å) Mean global backbone (residues 3–8, 15–27) 0.91 ± 0.25 Mean global heavy (residues 3–8, 15–27) 1.78 ± 0.26 Mean global backbone (residues 1–34) 1.65 ± 0.31 Mean global heavy (residues 1–34) 2.42 ± 0.30 Number of distance restraints Intraresidue (i-j = 0) 139 Sequential (|i-j| = 1) 149 Medium range (|i-j| < 5) 46 Long range (|i-j| ≥ 5) 43 Hydrogen bonds 4 Disulfide bonds 3 Total 384 Number of dihedral angle restraints φ 17 ψ 17

χ1 4 Total 38 Violations from experimental restraints Total NOE violations exceeding 0.2 Å 0 Total dihedral violations exceeding 2.0° 0 Note: Data are presented as mean ± S.D. a The remaining (6.7%) were found in the allowed regions of the Ramachandran plot. b Mean RMSD was calculated using MOLMOL (50). 87

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.6.3. Native and cyclic κ-PVIIA inhibit Shaker

To evaluate the effect of cyclization on toxin activity, κ-PVIIA and cyc-PVIIA were tested on the

Shaker KV channel expressed in Xenopus oocytes. κ-PVIIA completely blocked Shaker currents at

1 μM over a wide range of voltages, whereas the same concentration of the cyclic peptide only partially blocked the channel (Figure 4a). Fitting the concentration-dependence of Shaker block with the Hill equation revealed an IC50 value for native κ-PVIIA of 80 ± 5 nM, which is comparable to that previously published, 60 ± 3 nM (9,11). In contrast, an IC50 value ~10-fold higher was observed for cyc-PVIIA (824 ± 60 nM). Both κ-PVIIA and cyc-PVIIA displayed a slope of 1.1 ±

0.1, suggesting a toxin:channel ratio of 1:1, which is commonly observed for pore-blocking toxins

(Figure 4b). As the ion-conducting pore of KV channels is known to be conserved (75), both native and cyc-PVIIA were tested on other KV channels, including rKV2.1Δ7, hKV1.3, rKV1.6. Of the tested channels, 1 μM κ-PVIIA only slightly blocked rKV1.6, whereas cyc-PVIIA had no effect

(Figure 4c). Although Shaker has been used as a model for studying other Kv channels due to the high degree of homology, our results suggest that a ligand of Shaker does not necessarily bind to other Kv channels. Similar experimental or modeling results were also obtained previously using

Kv1.1 and Kv1.2, (10,63,76). Finally, neither peptide is active on rNaV1.4 and hNaV1.6.

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

Figure 4. Inhibitory toxin activity on Drosophila Shaker and rKV1.6 (a) 1 μM of κ-PVIIA completely blocks Shaker-IR (inactivation removed) currents over a wide voltage range whereas the same concentration of the cyc-PVIIA and N-terminally acetylated κ-PVIIA only partially block the channel. Representative example traces shown (left) were selected from n = 5 recordings under each condition at a voltage of 20 mV (100 ms) followed by a voltage-step to –70 mV (100 ms) to evoke tail currents (holding potential was –90 mV). Voltage-activation relationships (right) were obtained by plotting tail currents vs applied voltage step. Black lines and markers represent control currents and grey lines and markers show the response on currents following application of toxin. (b) Fitting the concentration-dependence for toxin, block of Shaker-IR with the Hill equation reveals an IC50 value for the κ-PVIIA of 80 ± 5 nM (black squares), cyc-PVIIA is 824 ± 60 nM (grey filled circles) and that of N-terminally acetylated κ- PVIIA is 1616 ± 98 nM (grey empty circles). κ- PVIIA, cyc-PVIIA, and acetylated κ-PVIIA fits display a slope of ~1.1 ± 0.1, suggesting a toxin:channel ratio of 1:1. The data is presented as mean ± s.e.m. (c) 1 μM κ-PVIIA weakly blocks rKV1.6 whereas cyc-PVIIA has no effect (n = 5). Example traces shown on the left were evoked by depolarizing the channel to 0 mV followed by –50 mV (tail) from a holding potential of –90mV. Black lines and markers represent control currents and grey lines and markers show the response on currents following application of toxin.

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.6.4. The N-terminus of κ-PVIIA is required for electrostatic interaction with Shaker.

Previous molecular modeling studies have predicted how κ-PVIIA binds to Shaker based on structural analyses and mutagenesis studies (11,13,63,77). Initially, recognition of κ-PVIIA by

Shaker takes place via electrostatic interactions between positively charged residues on the surface of the peptide and negatively charged residues on the extracellular side of the channel (77). The charged residues play an important role in adjusting the conformation of the peptide in the binding process until Lys7 of κ-PVIIA inserts into the pore of the channel and blocks the flow of potassium ions. In addition to Lys7, several residues were found to have a critical role in binding; for example,

Arg2 of κ-PVIIA interacts strongly with Asp447 of the channel and Phe9 of κ-PVIIA and Phe425 make hydrophobic interactions (63,77).

We developed a model of cyc-PVIIA bound to the Shaker channel using the model of Mahdavi and

Kuyucak (63). In our model, the binding mode of cyc-PVIIA was predicted to be generally unchanged compared with that of κ-PVIIA (Figure 5a-c). The stability of the cyc-PVIIA and Shaker channel interaction was substantiated by a 100 ns molecular dynamics simulation. This simulation was analyzed using the backbone RMSD of Shaker, cyc-PVIIA and the position of cyc-PVIIA on the Shaker (Figure 5d-g). The transmembrane helices of the Shaker channel were very rigid and did not deviate significantly from the starting conformation with a backbone RMSD of ~1 Å during the entire simulation. The backbone RMSD of the entire channel including transmembrane helices and loops was on average 2.2 Å, indicating the loops were more flexible during the simulation. The conformation of the cyc-PVIIA was also very stable as its backbone RMSD was ~1 Å (excluding cyclization linker) (Figure 5g). The change of position of cyc-PVIIA on the Shaker was evaluated by computing the backbone RMSD of the peptide (excluding the linker) after fitting on the Shaker transmembrane helices. The average RMSD of the cyc-PVIIA position (~2.5 Å) is comparable to the backbone RMSD of the Shaker, which is mostly accounted by the flexibility of the loops, thus suggesting that the binding mode of cyc-PVIIA to the channel is preserved. This is further 90

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

supported by an overlay of the starting and final binding modes of the simulation (Figure 5d-f), which shows excellent conservation of the relative position of Lys7 and of the interface.

Despite the unchanged binding mode after cyclization, cyc-PVIIA showed decreased activity

(Figure 4a), and this was assumed to be due to loss of electrostatic interaction between the electronegative potential generated by Glu418 and Glu422 on the turret of Shaker and the N- terminus of κ-PVIIA (Figure 5a-c). To test this hypothesis, N-terminally acetylated PVIIA was synthesized and indeed found to have ~20-fold lower activity (IC50, 1616 ± 98 nM) than κ-PVIIA when tested on Shaker (Figure 4a,b) confirming that the N-terminus of κ-PVIIA is required for full inhibition of the conductivity of Shaker. However, we expect that the loss of electrostatic interaction could possibly be compensated by incorporation of a positively charged residue in the linker.

4.6.5. The conformations of κ-PVIIA and cyc-PVIIA are similar upon binding to Shaker, but the hydrogen bond networks differ.

Although the overall backbone conformations of cyc-PVIIA and native κ-PVIIA in the NMR solution structures are very similar, Cys1 and Arg2, which are adjacent to the cyclization linker, resided in different spatial locations in the cyclized and native peptides (Figure 6b,c). By contrast, molecular models suggest that these two residues have similar conformations and the interactions at the interface with Shaker. We investigate here if the difference of affinity between the cyclized and linear form could be partially explained by difference in the unbound state. An overlay of the unbound and bound peptide conformations suggests that Cys1 is displaced by 3.0 Å for κ-PVIIA and 2.5 Å for cyc-PVIIA. In contrast, the hydrogen bond networks of Arg2 for the native and cyclic peptides are quite different when they are bound and unbound to Shaker (Figure 6). The change of the hydrogen bond networks established by Arg2 side chain was studied using 100 ns molecular dynamics simulations. Arg2 of κ-PVIIA forms stable hydrogen bonds with Lys7 and Cys16 and 91

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

transiently hydrogen bonds with Asp14. On the other hand, Arg2 of cyc-PVIIA lacks hydrogen bonds with Lys7 or Cys16 but also forms a transient hydrogen bond with Asp14. The models suggest that hydrogen bond network of unbound κ-PVIIA is more similar to the bound state than that of cyc-PVIIA. Although it is not certain whether this difference contributes to decreased activity of cyc-PVIIA, it appears likely that the two hydrogen bonds between Arg2 and Lys7, and

Arg2 and Cys16 stabilize the conformation of κ-PVIIA upon binding to Shaker.

Figure 5. Model of interactions of κ-PVIIA and cyc-PVIIA with the Shaker channel, (a) View of the extracellular entrance of the channel in the molecular model of κ-PVIIA in complex with Shaker (63). The solvent-accessible surface of the channel is colored according to the Poisson- Boltzmann electrostatic potential it generates, as computed by the APBS software (69), with a scale ranging from -4 kT/e (red) to +4 kT/e (blue). The extracellular loop of each of the four subunits of the channel creates an electronegative potential with which the positively charged residues R22 and K25 and N-terminus of κ-PVIIA interact on three different sides of the peptide. (b) Side view of the κ-PVIIA/Shaker complex highlighting two glutamic acids located on the Shaker turret, E418 and E422 that contribute to generate the electronegative potential surrounding the N-terminus of κ- PVIIA. The side chain of K7 of κ-PVIIA is also shown deeply buried in the ion channel selectivity filter located. (c) Molecular model of the complex between cyc-PVIIA and Shaker, which suggests that the binding mode is globally similar to that of the linear peptide apart from the absence of electrostatic interactions between the N-terminus and the Shaker channel. (d) Complex between cyc-PVIIA and Shaker at the start of the molecular dynamics simulation carried out in a POPC lipid membrane environment and explicit solvent. Cyc-PVIIA is in orange and Shaker is in purple; K7 is shown using stick representation. (e) Complex between cyc-PVIIA and Shaker at the end of the molecular dynamics simulation. Cyc-PVIIA is in dark green and Shaker is in light green; K7 is 92

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

shown using stick representation; POPC lipids are shown as grey balls. (f) superimposition of first and final frame of the simulation as represented in (d) and (e). (g) Evolution of backbone RMSD from the initial conformation during the 100 ns molecular dynamics simulation of cyc- PVIIA/Shaker embedded in a POPC lipid membrane. The backbone RMSD values were computed using the transmembrane helices of the Shaker only (red), all the Shaker channel residues (green), cyc-PVIIA residues (light blue) and the position of cyc-PVIIA (blue), which is the backbone RMSD of cyc-PVIIA computed after fitting on the structure of the transmembrane helices.

Figure 6. Hydrogen bond network established by Arg2 of κ-PVIIA and cyc-PVIIA. Hydrogen bond network of (a) Arg2 of κ-PVIIA when bound to Shaker in the model (63), (b) Arg2 of κ-PVIIA Arg2 in the NMR solution structure (PDB identifier 1AV3) (12), and (c) Arg2 in the molecular model of unbound cyc-PVIIA. The evolution of the Arg2 hydrogen bond network of κ-PVIIA (d) and cyc-PVIIA (e) was studied using 100 ns molecular dynamics simulations. In these simulations, Arg2 of κ-PVIIA formed stable hydrogen bonds with Lys7 and Cys16 and transiently established a hydrogen bond with Asp14. In contrast to the linear peptide, Arg2 of cyc-PVIIA did not form hydrogen bonds with Lys7 and Cys16 but it established a hydrogen bond more frequently with Asp14 than Arg2 of the linear peptide.

4.6.6. κ-PVIIA and cyc-PVIIA serum stability

To determine the effect of cyclization on stability, native and cyc-PVIIA were incubated in human serum. The majority of native κ-PVIIA was still detectable after 12 h; this was expected given that 93

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

the ICK motif is known to contribute to peptide stability. Indeed, a high percentage of acyclic kalata

B1, which is topologically similar to κ-PVIIA because of its knotted conformation, remains in serum even after 24 h (38). Surprisingly, only ~50% of cyc-PVIIA was detectable in serum after 12 h (Figure 7). One interpretation of this finding is that the additional constraint introduced by cyclization has made it more susceptible to proteolysis, thus highlighting the importance of optimizing the length and sequence of a linker in cyclisation studies.

Figure 7. Serum stability assay on native and cyc-PVIIA. Percentage of remaining peptide following 0, 3, 6, 9 and 12 h incubation in human serum is presented as mean ± s.e.m of triplicates.

4.7. Conclusion

The ICK motif is found in many venom peptides that have therapeutic potential, but chemical and protein engineering is typically required to achieve suitable pharmaceutical properties for this motif.

This study demonstrates that cyclization of an ICK motif-containing conotoxin is achievable via transpeptidation using SrtA. The linear oxidized peptide precursor of κ-PVIIA was efficiently cyclized, producing cyc-PVIIA with a head-to-tail cyclic backbone. The overall backbone structure and conformational dynamics of cyc-PVIIA is similar to native κ-PVIIA; however, following cyclization the peptide has partially reduced activity on the Shaker KV channel. The decrease in activity appears to be due to the removal of previously unrecognized electrostatic interactions

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

between the N-terminus of the peptide and negatively charged residues within the channel. Overall, the results of this study suggests that criteria which need to be considered in the head-to-tail backbone cyclization of ICK peptides include the effect of terminal mutations on activity, as well as the intramolecular hydrogen bond network and the conformational dynamics of the peptides. The study reveals that SrtA is an efficient tool for cyclizing disulfide-rich peptides that are inaccessible to cyclization using traditional synthetic methods.

Acknowledgements

We thank Prof. David Liu for providing the SrtA5° expression plasmid, and K. J. Swartz, C.

Deutsch, O. Pongs, and B. Chanda for sharing Shaker, rKV2.1Δ7, hKV1.3, rKV1.6, and rNaV1.4, respectively, and Dr. C. A. Wang for testing the catalytic activity of SrtA5°. This work was funded by a National Health and Medical Research Council (NHMRC) Project Grant to DJC and CIS

(APP1047857) and by the National Institute of Neurological Disorders And Stroke (NINDS) of the

National Institutes of Health (NIH) under Award Number 1R01NS091352 to FB. SK is the recipient of an Australian Research Council (ARC) Australian Postgraduate Award, ACC was supported by a University of Queensland International PhD scholarship, CIS in an Institute for

Molecular Bioscience Industry Fellow, DJC is an ARC Australian Laureate Fellow (FL150100146) and KJR is an ARC Future Fellow (FT130100890).

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

4.8. References

1. Fosgerau, K., and Hoffmann, T. (2015) Peptide therapeutics: current status and future directions. Drug Discov. Today 20, 122-128

2. Clark, R. J., Jensen, J., Nevin, S. T., Callaghan, B. P., Adams, D. J., and Craik, D. J. (2010) The engineering of an orally active conotoxin for the treatment of neuropathic pain. Angew. Chem. Int. Ed. Engl. 49, 6545-6548

3. Nielsen, D. S., Lohman, R. J., Hoang, H. N., Hill, T. A., Jones, A., Lucke, A. J., and Fairlie, D. P. (2015) Flexibility versus Rigidity for Orally Bioavailable Cyclic Hexapeptides. Chembiochem 16, 2289-2293

4. Wang, C. K., Northfield, S. E., Colless, B., Chaousis, S., Hamernig, I., Lohman, R. J., Nielsen, D. S., Schroeder, C. I., Liras, S., Price, D. A., Fairlie, D. P., and Craik, D. J. (2014) Rational design and synthesis of an orally bioavailable peptide guided by NMR amide temperature coefficients. Proc. Natl. Acad. Sci. U. S. A. 111, 17504-17509

5. Miljanich, G. P. (2004) Ziconotide: Neuronal calcium channel blocker for treating severe chronic pain. Curr. Med. Chem. 11, 3029-3040

6. Cushman, D. W., and Ondetti, M. A. (1999) Design of angiotensin converting enzyme inhibitors. Nat. Med. 5, 1110-1113

7. Barton, M. E., White, H. S., and Wilcox, K. S. (2004) The effect of CGX-1007 and CI-1041, novel NMDA receptor antagonists, on NMDA receptor-mediated EPSCs. Epilepsy Res. 59, 13- 24

8. Koo, G. C., Blake, J. T., Talento, A., Nguyen, M., Lin, S., Sirotina, A., Shah, K., Mulvany, K., Hora, D., Jr., Cunningham, P., Wunderler, D. L., McManus, O. B., Slaughter, R., Bugianesi, R., Felix, J., Garcia, M., Williamson, J., Kaczorowski, G., Sigal, N. H., Springer, M. S., and Feeney, W. (1997) Blockade of the voltage-gated potassium channel Kv1.3 inhibits immune responses in vivo. J. Immunol. 158, 5120-5128

9. Terlau, H., Shon, K. J., Grilley, M., Stocker, M., Stuhmer, W., and Olivera, B. M. (1996) Strategy for rapid immobilization of prey by a fish-hunting marine snail. Nature 381, 148-151

10. Shon, K. J., Stocker, M., Terlau, H., Stuhmer, W., Jacobsen, R., Walker, C., Grilley, M., Watkins, M., Hillyard, D. R., Gray, W. R., and Olivera, B. M. (1998) kappa-Conotoxin PVIIA is a peptide inhibiting the shaker K+ channel. J. Biol. Chem. 273, 33-38

11. Jacobsen, R. B., Koch, E. D., Lange-Malecki, B., Stocker, M., Verhey, J., Van Wagoner, R. M., Vyazovkina, A., Olivera, B. M., and Terlau, H. (2000) Single amino acid substitutions in kappa-conotoxin PVIIA disrupt interaction with the shaker K+ channel. J. Biol. Chem. 275, 24639-24644

12. Scanlon, M. J., Naranjo, D., Thomas, L., Alewood, P. F., Lewis, R. J., and Craik, D. J. (1997) Solution structure and proposed binding mechanism of a novel potassium channel toxin kappa- conotoxin PVIIA. Structure 5, 1585-1597

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

13. Savarin, P., Guenneugues, M., Gilquin, B., Lamthanh, H., Gasparini, S., Zinn-Justin, S., and Menez, A. (1998) Three-dimensional structure of kappa-conotoxin PVIIA, a novel potassium channel-blocking toxin from cone snails. Biochemistry 37, 5407-5416

14. Pallaghy, P. K., Nielsen, K. J., Craik, D. J., and Norton, R. S. (1994) A common structural motif incorporating a cystine knot and a triple-stranded beta-sheet in toxic and inhibitory polypeptides. Protein Sci. 3, 1833-1839

15. Craik, D. J., Daly, N. L., and Waine, C. (2001) The cystine knot motif in toxins and implications for drug design. Toxicon 39, 43-60

16. Norton, R. S., and Pallaghy, P. K. (1998) The cystine knot structure of ion channel toxins and related polypeptides. Toxicon 36, 1573-1583

17. Zhu, S. Y., Darbon, H., Dyason, K., Verdonck, F., and Tytgat, J. (2003) Evolutionary origin of inhibitor cystine knot peptides. FASEB J. 17, 1765-1767

18. Kaas, Q., Westermann, J. C., and Craik, D. J. (2010) Conopeptide characterization and classifications: an analysis using ConoServer. Toxicon 55, 1491-1509

19. Kaas, Q., Yu, R., Jin, A. H., Dutertre, S., and Craik, D. J. (2012) ConoServer: Updated content, knowledge, and discovery tools in the conopeptide database. Nucleic Acids Res. 40, D325-330

20. Nielsen, K. J., Thomas, L., Lewis, R. J., Alewood, P. F., and Craik, D. J. (1996) A consensus structure for omega-conotoxins with different selectivities for voltage-sensitive calcium channel subtypes: Comparison of MVIIA, SVIB and SNX-202. J. Mol. Biol. 263, 297-310

21. Hill, J. M., Alewood, P. F., and Craik, D. J. (1997) Solution structure of the sodium channel antagonist conotoxin GS: A new molecular caliper for probing sodium channel geometry. Structure 5, 571-583

22. Shon, K. J., Grilley, M. M., Marsh, M., Yoshikami, D., Hall, A. R., Kurz, B., Gray, W. R., Imperial, J. S., Hillyard, D. R., and Olivera, B. M. (1995) Purification, characterization, synthesis, and cloning of the lockjaw peptide from Conus purpurascens venom. Biochemistry 34, 4913-4918

23. Fainzilber, M., van der Schors, R., Lodder, J. C., Li, K. W., Geraerts, W. P., and Kits, K. S. (1995) New sodium channel-blocking conotoxins also affect calcium currents in Lymnaea neurons. Biochemistry 34, 5364-5371

24. Heinemann, S. H., and Leipold, E. (2007) Conotoxins of the O-superfamily affecting voltage- gated sodium channels. Cell. Mol. Life Sci. 64, 1329-1340

25. Kalia, J., Milescu, M., Salvatierra, J., Wagner, J., Klint, J. K., King, G. F., Olivera, B. M., and Bosmans, F. (2015) From foe to friend: Using animal toxins to investigate ion channel function. J. Mol. Biol. 427, 158-175

26. Daly, N. L., and Craik, D. J. (2011) Bioactive cystine knot proteins. Curr. Opin. Chem. Biol. 15, 362-368

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

27. Craik, D. J., and Conibear, A. C. (2011) The chemistry of cyclotides. J. Org. Chem. 76, 4805- 4817

28. Ireland, D. C., Colgrave, M. L., and Craik, D. J. (2006) A novel suite of cyclotides from Viola odorata: Sequence variation and the implications for structure, function and stability. Biochem. J. 400, 1-12

29. Colgrave, M. L., and Craik, D. J. (2004) Thermal, chemical, and enzymatic stability of the cyclotide kalata B1: The importance of the cyclic cystine knot. Biochemistry 43, 5965-5975

30. Gran, L. (1970) An oxytocic principle found in Oldenlandia affinis DC. Medd. Nor. Farm. Selsk. 12, 173-180

31. Hemu, X., Taichi, M., Qiu, Y., Liu, D. X., and Tam, J. P. (2013) Biomimetic synthesis of cyclic peptides using novel thioester surrogates. Biopolymers 100, 492-501

32. Akcan, M., Clark, R. J., Daly, N. L., Conibear, A. C., de Faoite, A., Heghinian, M. D., Sahil, T., Adams, D. J., Mari, F., and Craik, D. J. (2015) Transforming conotoxins into cyclotides: Backbone cyclization of P-superfamily conotoxins. Biopolymers 104, 682-692

33. Dawson, P. E., Muir, T. W., Clark-Lewis, I., and Kent, S. B. (1994) Synthesis of proteins by native chemical ligation. Science 266, 776-779

34. Tam, J. P., Lu, Y. A., Yang, J. L., and Chiu, K. W. (1999) An unusual structural motif of antimicrobial peptides containing end-to-end macrocycle and cystine-knot disulfides. Proc. Natl. Acad. Sci. U. S. A. 96, 8913-8918

35. Daly, N. L., Love, S., Alewood, P. F., and Craik, D. J. (1999) Chemical synthesis and folding pathways of large cyclic polypeptide: Studies of the cystine knot polypeptide kalata B1. Biochemistry 38, 10606-10614

36. Cheneval, O., Schroeder, C. I., Durek, T., Walsh, P., Huang, Y. H., Liras, S., Price, D. A., and Craik, D. J. (2014) Fmoc-based synthesis of disulfide-rich cyclic peptides. J. Org. Chem. 79, 5538-5544

37. Blanco-Canosa, J. B., Nardone, B., Albericio, F., and Dawson, P. E. (2015) Chemical protein synthesis using a second-generation N-acylurea linker for the preparation of peptide-thioester precursors. J. Am. Chem. Soc. 137, 7197-7209

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

39. Stanger, K., Maurer, T., Kaluarachchi, H., Coons, M., Franke, Y., and Hannoush, R. N. (2014) Backbone cyclization of a recombinant cystine-knot peptide by engineered Sortase A. FEBS Lett. 588, 4487-4496

40. Ton-That, H., Liu, G., Mazmanian, S. K., Faull, K. F., and Schneewind, O. (1999) Purification and characterization of sortase, the transpeptidase that cleaves surface proteins of Staphylococcus aureus at the LPXTG motif. Proc. Natl. Acad. Sci. U. S. A. 96, 12424-12429

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

41. Witte, M. D., Cragnolini, J. J., Dougan, S. K., Yoder, N. C., Popp, M. W., and Ploegh, H. L. (2012) Preparation of unnatural N-to-N and C-to-C protein fusions. Proc. Natl. Acad. Sci. U. S. A. 109, 11993-11998

42. Popp, M. W., and Ploegh, H. L. (2011) Making and breaking peptide bonds: protein engineering using sortase. Angew. Chem. Int. Ed. Engl. 50, 5024-5032

43. Chen, I., Dorr, B. M., and Liu, D. R. (2011) A general strategy for the evolution of bond- forming enzymes using yeast display. Proc. Natl. Acad. Sci. U. S. A. 108, 11399-11404

44. Rangaraju, S., Chi, V., Pennington, M. W., and Chandy, K. G. (2009) Kv1.3 potassium channels as a therapeutic target in multiple sclerosis. Expert Opin. Ther. Targets 13, 909-924

45. Popp, M. W., Antos, J. M., and Ploegh, H. L. (2009) Site-specific protein labeling via sortase- mediated transpeptidation. Curr. Protoc. Protein. Sci, 56, 15.13.11-15.13.19.

46. Conibear, A. C., Rosengren, K. J., Harvey, P. J., and Craik, D. J. (2012) Structural characterization of the cyclic cystine ladder motif of theta-defensins. Biochemistry 51, 9718- 9726

47. Guntert, P. (2004) Automated NMR structure calculation with CYANA. Methods Mol. Biol. 278, 353-378

48. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905-921

49. Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S., and Richardson, D. C. (2010) MolProbity: All-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12- 21

50. Koradi, R., Billeter, M., and Wuthrich, K. (1996) MOLMOL: A program for display and analysis of macromolecular structures. J. Mol. Graph. 14, 51-55

51. Hoshi, T., Zagotta, W. N., and Aldrich, R. W. (1990) Biophysical and molecular mechanisms of Shaker potassium channel inactivation. Science 250, 533-538

52. Tempel, B. L., Papazian, D. M., Schwarz, T. L., Jan, Y. N., and Jan, L. Y. (1987) Sequence of a probable potassium channel component encoded at Shaker locus of Drosophila. Science 237, 770-775

53. Frech, G. C., VanDongen, A. M., Schuster, G., Brown, A. M., and Joho, R. H. (1989) A novel potassium channel with delayed rectifier properties isolated from rat brain by expression cloning. Nature 340, 642-645

54. Li-Smerin, Y., and Swartz, K. J. (1998) Gating modifier toxins reveal a conserved structural motif in voltage-gated Ca2+ and K+ channels. Proc. Natl. Acad. Sci. U. S. A. 95, 8585-8589

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

55. Stuhmer, W., Ruppersberg, J. P., Schroter, K. H., Sakmann, B., Stocker, M., Giese, K. P., Perschke, A., Baumann, A., and Pongs, O. (1989) Molecular basis of functional diversity of voltage-gated potassium channels in mammalian brain. EMBO J. 8, 3235-3244

56. Grupe, A., Schroter, K. H., Ruppersberg, J. P., Stocker, M., Drewes, T., Beckh, S., and Pongs, O. (1990) Cloning and expression of a human voltage-gated potassium channel. A novel member of the RCK potassium channel family. EMBO J. 9, 1749-1756

57. Swanson, R., Marshall, J., Smith, J. S., Williams, J. B., Boyle, M. B., Folander, K., Luneau, C. J., Antanavage, J., Oliva, C., Buhrow, S. A., Bennet, C., Stein, R. B., and Kaczmarek, L. K. (1990) Cloning and expression of cDNA and genomic clones encoding three delayed rectifier potassium channels in rat brain. Neuron 4, 929-939

58. Trimmer, J. S., Cooperman, S. S., Tomiko, S. A., Zhou, J. Y., Crean, S. M., Boyle, M. B., Kallen, R. G., Sheng, Z. H., Barchi, R. L., Sigworth, F. J., Goodman, R. H., Agnew, W. S., and Mandel, G. (1989) Primary structure and functional expression of a mammalian skeletal muscle sodium channel. Neuron 3, 33-49

59. Burgess, D. L., Kohrman, D. C., Galt, J., Plummer, N. W., Jones, J. M., Spear, B., and Meisler, M. H. (1995) Mutation of a new sodium channel gene, Scn8a, in the mouse mutant 'motor endplate disease'. Nat. Genet. 10, 461-465

60. Schaller, K. L., Krzemien, D. M., Yarowsky, P. J., Krueger, B. K., and Caldwell, J. H. (1995) A novel, abundant sodium channel expressed in neurons and glia. J. Neurosci. 15, 3231-3242

61. Garcia, M. L., Garcia-Calvo, M., Hidalgo, P., Lee, A., and MacKinnon, R. (1994) Purification and characterization of three inhibitors of voltage-dependent K+ channels from Leiurus quinquestriatus var. hebraeus venom. Biochemistry 33, 6834-6839

62. Gilchrist, J., Das, S., Van Petegem, F., and Bosmans, F. (2013) Crystallographic insights into sodium-channel modulation by the beta4 subunit. Proc. Natl. Acad. Sci. U. S. A. 110, E5016- 5024

63. Mahdavi, S., and Kuyucak, S. (2013) Why the Drosophila Shaker K+ channel is not a good model for ligand binding to voltage-gated Kv1 channels. Biochemistry 52, 1631-1640

64. Sali, A., and Blundell, T. L. (1993) Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779-815

65. Schmid, N., Eichenberger, A. P., Choutko, A., Riniker, S., Winger, M., Mark, A. E., and van Gunsteren, W. F. (2011) Definition and testing of the GROMOS force-field versions 54A7 and 54B7. Eur. Biophys. J. 40, 843-856

66. Bussi, G., Donadio, D., and Parrinello, M. (2007) Canonical sampling through velocity rescaling. J. Chem. Phys. 126, 014101

67. Parrinello, M., and Rahman, A. (1981) Polymorphic transitions in single-crystals - a new molecular-dynamics method. J Appl Phys 52, 7182-7190

100

Chapter 4. Efficient enzymatic cyclization of an inhibitory cystine knot containing peptide

68. Heinz, T. N., van Gunsteren, W. F., and Hunenberger, P. H. (2001) Comparison of four methods to compute the dielectric permittivity of liquids from molecular dynamics simulations. J. Chem. Phys. 115, 1125-1136

69. Baker, N. A., Sept, D., Joseph, S., Holst, M. J., and McCammon, J. A. (2001) Electrostatics of nanosystems: Application to microtubules and the ribosome. Proc. Natl. Acad. Sci. U. S. A. 98, 10037-10041

70. Lindorff-Larsen, K., Piana, S., Palmo, K., Maragakis, P., Klepeis, J. L., Dror, R. O., and Shaw, D. E. (2010) Improved side-chain torsion potentials for the Amber ff99SB protein force field. Proteins 78, 1950-1958

71. Pronk, S., Pall, S., Schulz, R., Larsson, P., Bjelkmar, P., Apostolov, R., Shirts, M. R., Smith, J. C., Kasson, P. M., van der Spoel, D., Hess, B., and Lindahl, E. (2013) GROMACS 4.5: a high- throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29, 845-854

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

73. Wüthrich, K. (1986) NMR of proteins and nucleic acids, Wiley-Interscience, New York.

74. Daly, N. L., Ekberg, J. A., Thomas, L., Adams, D. J., Lewis, R. J., and Craik, D. J. (2004) Structures of muO-conotoxins from Conus marmoreus. Inhibitors of tetrodotoxin (TTX)- sensitive and TTX-resistant sodium channels in mammalian sensory neurons. J. Biol. Chem. 279, 25774-25782

75. Lu, Z., Klem, A. M., and Ramu, Y. (2001) Ion conduction pore is conserved among potassium channels. Nature 413, 809-813

76. Boccaccio, A., Conti, F., Olivera, B. M., and Terlau, H. (2004) Binding of kappa-conotoxin PVIIA to Shaker K+ channels reveals different K+ and Rb+ occupancies within the ion channel pore. J. Gen. Physiol. 124, 71-81

77. Huang, X., Dong, F., and Zhou, H. X. (2005) Electrostatic recognition and induced fit in the kappa-PVIIA toxin binding to Shaker potassium channel. J. Am. Chem. Soc. 127, 6836-6849

101

Chapter 5

Bacterial expression of kalata B1

102

Chapter 5. Bacterial expression of kalata B1

5.1. Overview

Chemical synthesis of cyclotides is advantageous in terms of being able to make chemical modifications or substitution of a specific amino acid, or to incorporate unnatural amino acids or fluorescent labels in a peptide. However, producing large amounts of cyclotides would be expensive and for cyclotides to be attractive as future therapeutics, alternative routes of production are necessary. Recombinant expression of cyclotides on the other hand is desirable because it could, in principle, be more cost-effective than solid phase peptide synthesis. In addition to economic benefits, recombinant expression might simplify the process of purification significantly, especially for disulfide-rich peptides if the peptides can be correctly folded in a host cell. Chemical synthesis of disulfide-rich peptides, including cyclotides, typically requires multiple purification steps to ensure removal of scavengers, side chain protecting groups, and impurities including side products and disulfide bond isomers. Therefore, alternative methods to achieve correct folding, such as recombinant expression, have been sought to boost the production of proteins/peptides containing disulfide bonds.

There are three methods most commonly used. Firstly, expression carried out in the bacterial periplasm where disulfide oxidoreductases and isomerases catalyze the formation of disulfide bonds (1). This approach requires the use of a signal peptide to directly export protein from the cytoplasm to the periplasm. Although yield of final product from periplasmic expression may not be as high compared to cytoplasmic expression, the environment in the cytoplasm is not ideal for disulfide bond formation due to the presence of numerous reductases and reducing agents such as glutathione. While translocation of a protein to the periplasm takes place by a signal peptide, which needs to be encoded by DNA in an expression vector, disulfide bond formation is catalyzed by DsbA. For example, MalE which is located at the N-terminus of the gene encoding MBP directs secretion of MBP into the periplasmic space. (In other words, the MalE sequence needs be deleted from the gene when MBP is used as a fusion protein for cytoplasmic expression.)

The second approach to fold proteins in host cells is using fusion protein partners such as bacteria- derived DsbA (2,3), DsbC (4) and protein disulfide bond isomerase (PDI) (5-7) from eukaryotes that can help fusion protein substrates fold even in the cytoplasm. Thirdly, engineered E. coli strains

103

Chapter 5. Bacterial expression of kalata B1

such as OrigamiTM(DE3)(Novagen), Rosetta-gamiTM(Novagen), and SHuffle®(NEB) have all been shown to overcome reductive properties of cytoplasm.

Recently, a protocol describing the recombinant expression of a range of disulfide-rich peptides in E. coli was published (8). The authors systematically described parameters to consider including bacterial culture media, induction conditions, lysis methods, and fusion proteins. It is worth noting that in ~30 disulfide-rich peptides they tried to express, access to the desired product was achieved in 75% of the cases described (8). In this chapter, the possibility of inexpensive production of kalata B1 by expressing a linear kalata B1 precursor in bacterial periplasm was investigated based on the successful semi-enzymatic cyclization of synthetic kalata B1 shown in Chapter 3. In Chapter 3, the introduction of tri-Gly into kalata B1 made loop 6 more flexible than in the native protein. In addition, the SrtA-cyclized kalata B1 was slightly less stable than the native kalata B1 (9). To reduce this flexibility, a di-Gly was installed at the N-terminus of kalata B1. Whereas Klint et al. (8) used TEV protease to cleave a linear disulfide-rich peptide from the fusion protein, linear kalata B1 precursor was designed to be cleaved using SrtA and simultaneously cyclized in vitro. A schematic representation of the map of gene encoding MBP-kalata B1 fusion protein for bacterial periplasmic expression and the cyclization strategy are shown in Figure 1.

104

Chapter 5. Bacterial expression of kalata B1

Figure 1. Strategy for cyclization of MBP-LPETGG-kalata B1-TGG expressed in E. coli periplasm. (A) Schematic representation of the pLic-MBP expression vector for bacterial periplasmic expression of disulfide-rich peptides. The figure was adapted from Klint et al (8). (B) The linear oxidized [GG]kalata B1[TGG] containing the SrtA recognition motif at both the N- and C-termini undergoes an intramolecular transpeptidation reaction catalyzed by SrtA to form cyclic [GG]kalata B1[T].

5.2. Experimental procedures

5.2.1. Design of expression vector

A synthetic gene encoding the kalata B1 precursor, [GG]kalata B1[TGG], was codon optimized for expression in E. coli and cloned into a modified pLic-MBP expression vector which encodes N- terminal His6 tag (10) by Geneart AG (Regenburg, Germany).

5.2.2. Bacterial culture

Escherichia coli BL21 (DE3) transformed with the expression plasmid, pLic-MBP-

[LPETGG]kalataB1[TGG] was cultured in the presence of ampicillin (100 μg/mL) until the OD600 reached ~0.7. Protein expression was induced by adding isopropyl β-D-thiogalactopyranoside (IPTG) to a final concentration of 1 mM, and incubating for 3 h at 30°C or overnight at 18°C. Also, 105

Chapter 5. Bacterial expression of kalata B1

another E. Coli strain, SHuffle® T7 Express lysY (NEB, C3030H), was transformed with the same expression plasmid and cultured at room temperature until the OD600 reached ~0.7. Subsequently, protein expression was induced by adding IPTG to a final concentration of 1 mM, and incubating overnight at 18°C.

5.2.3. Bacterial cell lysis The cells were harvested by centrifugation at 6000 x g for 20 min at 4°C and cell pellets were resuspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 25 mM imidazole, 10% glycerol). Small scale trial lysis was carried out using (i) repeat of freezing and thawing of cells, (ii) sonication, (iii) FastBreakTM, (iv) osmotic shock using sucrose buffer (50 mM Tris pH 7.5, 20% sucrose w/v) with or without proteinase inhibitor to delineate optimal lysis condictions. Cell debris was removed by centrifugation at 12,000 x g for 30 min at 4°C.

5.2.4. Purification of fusion protein The lysate was subjected to metal ion affinity chromatography using a 5 mL Ni2+ sepharose FF column (HisTrap FF, GE Healthcare). The column was washed extensively with lysis buffer (see above) following sample loading. MBP-kalata B1 was eluted with a linear gradient over 10 column volumes using elution buffer (50mM Tris-HCl, pH 7.5, 150 mM NaCl, and 500 mM immidazole). Fractions containing the fusion protein were buffer-exchanged (50 mM Tris-HCl, 150 mM NaCl,

10 mM CaCl2 and 10% glycerol).

5.2.5. SrtA cleavage and peptide purification

Two SrtA: substrate ratios were attempted. First, 50 μM purified His6-MBP-LPETGG-kalata B1- TGG was incubated with the same concentration of wild-type ∆59 SrtA or SrtA5° at pH 7.5 and at room temperature. Second, 150 μM purified His6-MBP-LPETGG-kalata B1-TGG was incubated with 30 μM SrtA5° at pH 6 for 4 h. Both reactions were subjected again to the Ni2+ affinity column to remove SrtA as well as MBP, and the flow-through was collected to purify the cleaved the peptide using RP-HPLC.

5.2.6. Refolding of cyclic-[GG]kalata B1[T]

Cyclic-[GG]kalata B1[T] was dissolved in 0.1 M NH4HCO3 (pH 8) with 3 M urea and 5 mM DTT to a final concentration of 0.5 mg/mL and incubated at 37°C for 30 min to reduce its cystines. After

106

Chapter 5. Bacterial expression of kalata B1

purification using RP-HPLC with a gradient from 20 to 50% ACN (0.05% TFA) over 30 min, the cyclic reduced [GG]kalata B1[T] was re-oxidized in 50% isopropyl alcohol (v/v), 0.1 M NH4HCO3 with 1 mM reduced glutathione at pH 8 for ~24 h at room temperature (11). The oxidized product was purified by RP-HPLC using a gradient from 30 to 50% ACN (0.05% TFA) over 20 min.

5.2.7. NMR analysis

Purified cyclic-[GG]kalata B1[T] was dissolved in 10% D2O/H2O (v/v), pH 3.5 and subjected to one-dimensional and two-dimensional TOCSY experiments to be compared to the structure of cyc- [GGG]kalata B1[LPET] as described in Jia et al. (9) (Also in Chapter 3).

5.3. Results

5.3.1. Expression of MBP-[LPETGG]-kalata B1[TGG]

In the first round of kalata B1 protein expression, E. coli cells transformed with kalata B1 precursor construct, linear kalata B1 fused at the C-terminus of MBP including MalE signal sequence (Figure 1), were lysed by sonication in the buffer excluding proteinase inhibitor. However, upon cleavage of MBP-kalata B1 using wild-type SrtA, no cleaved peptide was detected. In order to determine if it was because kalata B1, which is located at the C-terminus of MBP, was degraded by proteinases released during cell lysis, cell lysis was carried out again in the presence or absence of proteinase inhibitors. Simultaneously, several additional lysis methods were investigated in order to identify the ultimate method to extract MBP-[LPETGG]kalata B1[TGG] (~46 kD). Lysis methods included a gentle method using FastBreakTM, osmotic shock (which extracts out only periplasmic expression), mechanical lysis using sonication and freeze-thaw method.

The SDS-PAGE analysis showed a protein band size for MBP-[LPETGG]kalataB1[TGG] consistent throughout the process from cell lysis to purification regardless of lysis methods or the presence of the proteinase inhibitor in the lysis buffer. In terms of yield, while the osmotic shock method resulted in the lowest amount of the fusion protein, the lysate resulting from this extraction method was found to contain the least impurities of any of the methods. The other three lysis methods showed similar intensity of the protein band from the lysate or lysate supernatant sample except when FastBreakTM was used without the proteinase inhibitor (Figure 2A). The lysate

107

Chapter 5. Bacterial expression of kalata B1

supernatant from each method was purified using affinity chromatography showing one major product (Figure 2B).

Figure 2. Expression of kalata B1 precursor and SrtA-based cyclization (A) SDS-PAGE analysis of four different bacterial cell lysis methods, (B) Fast protein liquid chromatography (FPLC) profile of MBP-kalata B1, (C) SDS-PAGE showing progress of cyclization sampled at different times.

5.3.2. Kalata B1 precursor cleavage by SrtA5º

The purified MBP-kalata B1 from each lysis method was subjected to cleavage and cyclization using SrtA5° which has better catalytic activity than the wild-type SrtA. Surprisingly SrtA5° cleaved the linker (LPETGG) between MBP and kB1 of every sample from the various lysis methods and cyclized the precursor of kB1 very efficiently. Figure 2(C) shows the time course of the SrtA reaction on MBP-[LPETGG]kalata B1[TGG]. The cyclization was complete within 1 h and the final product showed cyclized and oxidized mass of cyc-[GG]kB1[T] after purification (Figure 3). The total yield of peptide showing the correct mass was about 1 mg as a result of 2 L

108

Chapter 5. Bacterial expression of kalata B1

bacterial culture. Other peaks showing the same mass did not yield enough sample for NMR analysis.

Figure 3. Purification of cyclised [GG]kalata B1[T] (A) HPLC profile of cyclic [GG]kalata B1[T] purification after SrtA5°+MBP-kalata B1 reaction. Fraction marked with▼ was subjected to NMR analysis, (B) MALDI-TOF analysis of cyclized oxidized [GG]kB1[T].

5.3.3. Periplasmic expression of [GG]kalata B1[TGG] did not achieve correct folding.

The structure of recombinant cyclic [GG]kalata B1[T] was compared to the synthetically produced but enzymatically cyclized cyclo-[GGG]kalata B1[T] studied in Chapter 3. Although the recombinant cyclic [GG]kalata B1[T] appeared to be folded from the one-dimensional spectrum displaying widely spread peaks in the amide proton region, the chemical shift of each peak was different when superimposed with cyclo-[GGG]kalata B1[T] described in Chapter 3 (Figure 4). Further comparison with TOCSY spectra confirmed that their folding states were indeed different to each other (Figure 4). As it was reported that expressed disulfide-rich venom peptides are generally more soluble when the temperature is lowered to 16°C before induction (8), induction of expression at 18°C (which was the lowest temperature possible with the incubator available) for the same BL21 strain was attempted. However, overexpression of the MBP-kalata B1 fusion protein was not successful (Figure 5A). When the same expression strategy was applied to the SHuffle® strain, overexpression of MBP-kalata B1 was observed (Figure 5A), and the SrtA-mediated cleavage and cyclization of [GG]kalata B1[T] was successful (Figure 5B). However, two major isomers of [GG]kalata B1[T] were found from purification using RP-HPLC (Figure 5C). Unfortunately, both isomers appeared to have different structures to that of synthetically made and correctly folded [GGG]kalata B1[T] (Figure 6). 109

Chapter 5. Bacterial expression of kalata B1

Figure 4. Comparison between recombinant cyclic [GG]kalata B1[T] and cyclo-[GGG]kalata B1[T] (Chapter 3) at pH 3.5 (A) 1D 1H spectra shows difference in the amide region between the two peptides, (B) 2D TOCSY spectrum of recombinant [GG]kalata B1[T] shows that its amide region is not as widely spread as that of cyclo-[GGG]kalata B1[T], which is shown in (C).

110

Chapter 5. Bacterial expression of kalata B1

Figure 5. Expression of MBP-kalata B1 precursor and cleavage by SrtA. (A) SDS-PAGE analysis of MBP-kalata B1 precursor expression in BL21 and SHuffle® T7 express lysY strains,which was induced overnight at 18°C. (B) SrtA5° reaction on MBP-kalata B1 expressed in the SHuffle® strain. (C) HPLC profile of SrtA-cleaved and cyclized kalata B1.

Figure 6. One-dimensional NMR spectra of synthetic [GGG]kalata B1[T] made by native chemical ligation, and the two isomers of [GG]kalataB1[T] derived from bacterial expression.

111

Chapter 5. Bacterial expression of kalata B1

5.3.4. Refolding of cyclic-[GG]kalata B1[T] results in correct disulfide bond connectivities.

Cyclic-[GG]kalata B1[T] with incorrect folding that resulted from expression using SHuffle® was completely reduced and subjected to oxidative folding. The peaks collected from 35% to 40% on the ACN gradient showed masses corresponding to the cyclic oxidized [GG]kalata B1[T] (Figure 7). The peak at ~40% on the ACN gradient (Figure 7; red arrow) was chosen for NMR characterization, as the native kalata B1 is hydrophobic and elutes later than its isomers from in vitro oxidation. The one-dimensional NMR spectrum was highly similar to that of synthetically made [GGG]kalata B1[T] (Figure 8).

mV % Detector A Ch1:215nm 1900 Detector A Ch2:280nm 95.0 1800

90.0 1700

85.0 1600

1500 80.0

1400 75.0

1300 70.0

1200 65.0

1100 60.0

1000 55.0

900 50.0

800 45.0

700 40.0 600 35.0 500 30.0 400 25.0 300 20.0 200 15.0 100 10.0 0

5.0 -100

0.0 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 min

Figure 7. RP-HPLC profile of oxidation of cyclic-[GG]kalata B1[T]. The peak indicated with an arrow was lyophilized and analyzed to NMR experiment.

112

Chapter 5. Bacterial expression of kalata B1

Figure 8. Comparison of one-dimensional NMR spectra between the synthetic [GGG]kalata B1[T] and refolded [GGG]kalata B1[T] after expression.

5.4. Discussion In this chapter, bacterial periplasmic expression of the kalata B1 precursor, MBP- [LPETGG]kalataB1[TGG], was attempted to enable cost-effective and efficient mass production of a cyclotide. The purpose of expression with the fusion protein, MBP, was for enhanced solubility and the His6 tag was for easy purification by affinity chromatography. Among the methods investigated to extract the MBP-[LPETGG]kalataB1[TGG], sonication, using FastBreakTM, and freez-thaw methods showed similarly good yield, but sonication would be more suitable to carry out large volume of bacterial culture than other methods. Although lysis by osmotic shock resulted in the sample with the least impurities as it allows extraction of proteins in the periplasmic space, the yield of the fusion protein was much lower than those from the other methods. This observation suggests two possibilities; (i) the yield of the fusion protein appears to be low from the cell lysate obtained by osmotic shock because the lysis condition was not optimal for complete bacterial cell breakage, or (ii) export from the cytoplasm to periplasm was not effective and thus the majority of the MBP-kalata B1 fusion protein remained in the cytoplasm in a reduced state. There are three

113

Chapter 5. Bacterial expression of kalata B1

major pathways (12) involved in the translocation of polypeptides from the cytoplasmic membrane into the periplasm of Gram-negative bacteria: the Sec pathway (12), the signal recognition particle (SRP) pathway (13,14) and the Tat pathway (15,16). The Sec pathway transports polypeptides post- translationally, whereas the SRP pathway does it cotranslationally. While these two pathways translocate unfolded polypeptides from the cytoplasm to the periplasm via Sec translocon, Tat pathway transports only folded polypeptides via Tat translocon. MalE favors the Sec pathway along with other signal peptides such as OmpA, PelB, Lamb, PhoA and MglB (17). If the translocation of MBP-[LPETGG]kalataB1[TGG] was indeed an issue, it is probably worth trying signal peptides that favor SRP pathway such as DsbA, SfmC, TolB, and TorT(17).

The expressed and purified MBP-[LPETGG]kalataB1[TGG] was cleaved by SrtA5° and cyclization of [GG]kalataB1[TGG] was immediately achieved. Unfortunately, cyclic [GG]kalataB1[T] appeared to have a different structure compared to the cyclo-[GGG]kalataB1[T], which was successfully cyclized by SrtA (9), when analyzed by NMR. This is most likely due to misfolding because of non-native disulfide-connectivity. The proton signals in the TOCSY spectrum of recombinant [GG]kalataB1[T] were not as dispersed as those for cyclo-[GGG]kalataB1[T], and it was not possible to find characteristic spectral ‘markers’ of native kalata B1 such as NOEs reflecting Pro20 in a cis-conformation and the characteristic chemical shifts for HN of ASN11, HB of Pro20.

Assuming the translocation of the fusion protein across the cytoplasmic membrane was not effective, it is possible that the kalata B1 precursor was expressed with reduced cysteines and subsequently formed wrong disulfide bonds in the step of lysis or cyclization in which pH 7~8 of buffer was used. On the other hand, if the fusion protein was misfolded in periplasm, one way to troubleshoot the problem is reducing the temperature of bacterial culture after addition of IPTG because induction temperature is an important factor to limit aggregation of recombinant protein expression (18,19). To test the latter hypothesis, induction at lower temperature was tried using two bacterial strains. While BL21 did not express the desired protein after overnight IPTG induction at 18°C, SHuffle® over-expressed MBP-kalata B1 at the same condition. SHuffle® is an engineered E.coli strain to produce protein with disulfide bond formation within its cytoplasm. This strain expresses a choromosomal copy of the disulfide bond isomerase DsbC, which aids the correction of misfolded proteins into their correct form (20). Given that the SHuffle® strain performs the best at low temperature such as 16°C and at 1 mM IPTG (20), a similar strategy was attempted in this 114

Chapter 5. Bacterial expression of kalata B1

study. Although no correctly folded [GG]kalata B1[T] was found after SrtA cleavage and cyclization, substantial amount of isomers were produced for refolding. Refolded [GG]kalata B1[T] appeared to have a highly similar structure to that of synthetically made [GG]kalata B1[T] when examined using NMR. This result implies that [GG]kalta B1[T] can be more efficiently produced via expression in the cytoplasm in the reduced state followed by SrtA cyclization and in vitro folding.

While this project was being pursued, a study demonstrated the backbone cyclization of a recombinant cystine-knot peptide, MCoTI-II, by SrtA (21). The authors of that study expressed

MCoTI-II with GST fusion protein with a His6 tag and placed the sorting motif, LPETGG, only at the C-terminus of MCoTI-II. At the N-terminus a TEV sequence followed by tri-Gly was used instead of the sorting motif for cleavage of the peptide from the fusion protein. Folding of MCoTI- II was achieved while MCoTI-II was being cleaved by TEV protease because of the oxidative condition of the buffer. The authors also used mutant SrtA (22) with better catalytic activity than the wild-type SrtA to cyclize McoTI-II.

The ultimate goal of recombinant expression of a cyclotide would be to achieving both oxidation and cyclization in cellular environment. Cyclization and secretion of the eGFP, in the cytosol and endoplasmic reticulum (ER) lumen of S. cerevisiae and mammalian HEK293T cells was achieved utilizing SrtA (23). A linear substrate protein was equipped with an N-terminal signal peptide that displays a glycine residue after cleavage and a C-terminal sortase recognition motif followed by a Myc epitope tag and an ER retention signal (Myc-HDEL) which allows observation of substrate recognition by SrtA followed by entering the secretory pathway. S. cerevisiae was transformed with the substrate protein expression plasmid and the pGAL-ER-SrtA plasmid. Sortase (of S.pyogenes) was active in the ER lumen and subsequently cyclic protein was secreted which could be regulated by galactose-inducible ER-SrtA expression (23).

Although the strategy used in this chapter did not achieve a correctly folded kalata B1, there are many factors that can be optimized such as changing the induction temperature, signal peptide sequence, fusion protein partners or E. coli host strains to name a few. Nevertheless, this study highlights the potential to produce cyclic peptides with disulfide bonds formed via recombinant expression. Recently, an enzyme involved in the biosynthesis of cyclotides, asparaginyl endopeptidase (AEP), was successfully isolated from Clitoria ternatea, and a related enzyme was 115

Chapter 5. Bacterial expression of kalata B1

recombinantly expressed (24-26). Hopefully, with tools including SrtA and AEP at hand, recombinant expression of cyclotides (with or without a linker region) will be possible in the near future.

5.5. References

1. Berkmen, M. (2012) Production of disulfide-bonded proteins in Escherichia coli. Protein Expr. Purif. 82, 240-251

2. Collins-Racie, L. A., McColgan, J. M., Grant, K. L., DiBlasio-Smith, E. A., McCoy, J. M., and LaVallie, E. R. (1995) Production of recombinant bovine enterokinase catalytic subunit in Escherichia coli using the novel secretory fusion partner DsbA. Biotechnology. (N. Y.) 13, 982-987

3. Fan, C. F., Zeng, R. H., Sun, C. Y., Mei, X. G., Wang, Y. F., and Liu, Y. (2005) Fusion of DsbA to the N-terminus of CTL chimeric epitope, F/M2:81-95, of respiratory syncytial virus prolongs protein- and virus-specific CTL responses in Balb/c mice. Vaccine 23, 2869- 2875

4. Zhang, Z., Li, Z. H., Wang, F., Fang, M., Yin, C. C., Zhou, Z. Y., Lin, Q., and Huang, H. L. (2002) Overexpression of DsbC and DsbG markedly improves soluble and functional expression of single-chain Fv antibodies in Escherichia coli. Protein Expr. Purif. 26, 218- 228

5. Chun, H., Joo, K., Lee, J., and Shin, H. C. (2011) Design and efficient production of bovine enterokinase light chain with higher specificity in E. coli. Biotechnol Lett 33, 1227-1232

6. Liu, Y., Zhao, T. J., Yan, Y. B., and Zhou, H. M. (2005) Increase of soluble expression in Escherichia coli cytoplasm by a protein disulfide isomerase gene fusion system. Protein Expr. Purif. 44, 155-161

7. Kajino, T., Ohto, C., Muramatsu, M., Obata, S., Udaka, S., Yamada, Y., and Takahashi, H. (2000) A protein disulfide isomerase gene fusion expression system that increases the extracellular productivity of Bacillus brevis. Appl. Environ. Microbiol. 66, 638-642

8. Klint, J. K., Senff, S., Saez, N. J., Seshadri, R., Lau, H. Y., Bende, N. S., Undheim, E. A., Rash, L. D., Mobli, M., and King, G. F. (2013) Production of recombinant disulfide-rich venom peptides for structural and functional analysis via expression in the periplasm of E. coli. PLoS ONE 8, e63865

116

Chapter 5. Bacterial expression of kalata B1

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

10. Cabrita, L. D., Dai, W., and Bottomley, S. P. (2006) A family of E. coli expression vectors for laboratory scale and high throughput soluble protein production. BMC Biotechnol. 6, 12

11. Daly, N. L., Love, S., Alewood, P. F., and Craik, D. J. (1999) Chemical synthesis and folding pathways of large cyclic polypeptide: Studies of the cystine knot polypeptide kalata B1. Biochemistry 38, 10606-10614

12. Fekkes, P., and Driessen, A. J. (1999) Protein targeting to the bacterial cytoplasmic membrane. Microbiol. Mol. Biol. Rev. 63, 161-173

13. Koch, H. G., Moser, M., and Muller, M. (2003) Signal recognition particle-dependent protein targeting, universal to all kingdoms of life. Rev. Physiol. Biochem. Pharmacol. 146, 55-94

14. Luirink, J., and Sinning, I. (2004) SRP-mediated protein targeting: structure and function revisited. Biochim. Biophys. Acta 1694, 17-35

15. Fisher, A. C., and DeLisa, M. P. (2004) A little help from my friends: quality control of presecretory proteins in bacteria. J. Bacteriol. 186, 7467-7473

16. Robinson, C., and Bolhuis, A. (2004) Tat-dependent protein targeting in prokaryotes and chloroplasts. Biochim. Biophys. Acta 1694, 135-147

17. Steiner, D., Forrer, P., Stumpp, M. T., and Pluckthun, A. (2006) Signal sequences directing cotranslational translocation expand the range of proteins amenable to phage display. Nat. Biotechnol. 24, 823-831

18. Kiefhaber, T., Rudolph, R., Kohler, H. H., and Buchner, J. (1991) Protein Aggregation Invitro and Invivo - a Quantitative Model of the Kinetic Competition between Folding and Aggregation. Bio-Technol 9, 825-829

19. Schein, C. H. (1990) Solubility as a Function of Protein-Structure and Solvent Components. Bio-Technol 8, 308-315

20. Lobstein, J., Emrich, C. A., Jeans, C., Faulkner, M., Riggs, P., and Berkmen, M. (2012) SHuffle, a novel Escherichia coli protein expression strain capable of correctly folding disulfide bonded proteins in its cytoplasm. Microb Cell Fact 11, 56

21. Stanger, K., Maurer, T., Kaluarachchi, H., Coons, M., Franke, Y., and Hannoush, R. N. (2014) Backbone cyclization of a recombinant cystine-knot peptide by engineered Sortase A. FEBS Lett. 588, 4487-4496

117

Chapter 5. Bacterial expression of kalata B1

22. Chen, I., Dorr, B. M., and Liu, D. R. (2011) A general strategy for the evolution of bond- forming enzymes using yeast display. Proc. Natl. Acad. Sci. U. S. A. 108, 11399-11404

23. Strijbis, K., Spooner, E., and Ploegh, H. L. (2012) Protein Ligation in Living Cells Using Sortase. Traffic 13, 780-789

24. Nguyen, G. K., Cao, Y., Wang, W., Liu, C. F., and Tam, J. P. (2015) Site-Specific N- Terminal Labeling of Peptides and Proteins using Butelase 1 and Thiodepsipeptide. Angew. Chem. Int. Ed. Engl.

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

26. Harris, K. S., Durek, T., Kaas, Q., Poth, A. G., Gilding, E. K., Conlan, B. F., Saska, I., Daly, N. L., van der Weerden, N. L., Craik, D. J., and Anderson, M. A. (2015) Efficient backbone cyclization of linear peptides by a recombinant asparaginyl endopeptidase. Nat. Commun. 6, 10199

118

Chapter 6 Conjugation of a cyclotide to a single domain antibody to elicit an immune response

119

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

6.1. Overview As described in previous chapters, cyclotides have been used extensively as drug scaffolds to accommodate therapeutically useful linear peptide sequences to confer stability. For example, linear peptides with proangiogenesis (1), regulation of p53 tumor suppressor pathway (2) or immunoregulatory (3) activity have been successfully grafted into cyclotides. As cyclotides are being developed for clinical applications, a more detailed understanding of their immunological properties is desirable. For example, the immunogenicity of cyclotide scaffolds, the immunological effect caused by conformational constraints introduced to the epitope, and how the compact structure of cyclotides affect the process required to generate MHC-restricted epitopes essential for T-cell help in the production of antibodies are all unknown factors because cyclotides have not been systematically explored as immunogens.

In the context of studying the biosynthesis of cyclotides, generating antibodies against cyclotides can be useful e.g. to localize the peptides in plant tissues. Cyclotides originate from their precursor proteins in plants, and the backbone is cyclized by asparaginyl endopeptidase (AEP) during posttranslational processing (4,5). However, more information on the mechanism and distribution of cyclotides and their linear precursor needs to be elucidated. Therefore, antibodies that specifically bind to and distinguish cyclotides from their linear precursors would help characterize the location and timing of the conversion of linear precursors into the corresponding cyclotides. To date, such studies have been achieved using MALDI imaging (6), MALDI-TOF mass spectrometry and an antibody elicited against the precursor protein of kalata B1, Oak 1 (~14 kDa), as an immunogen (7). The Oak 1 antibody was able to detect kalata B1 in an immunoblot analysis, but showed limited sensitivity (7).

In general, peptide-based immunization suffers limitations because of the inefficient delivery of peptides to professional antigen presenting cells (APCs) (8). This is because peptides are often intrinsically weak due to poor chemical and physical stability. This drawback for peptides as immunogens can be overcome by delivering peptides using carriers like nanoparticles or microspheres (8,9). Another strategy improving peptides as immunogens is conjugation to an antibody. For example, the use of an antibody targeting MHC class II can efficiently deliver antigens to dendritic cells (10-12). Hopefully bringing the antibody-antigen complex to the dendritic cells can process the antigen in the context of MHC class II products and subsequently

120

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

lead to the generation of T-helper cells, which is crucial for an effective secondary antibody response (13).

Recently, a camelid (alpaca)-derived single-domain antibody fragment, VHH, was reported to target mouse a MHC class II molecule, namely VHH7 (14). Camelids, whose members include camels, llama, and vicugna, produce not only conventional antibodies, but also heavy-chain-only antibodies in which VHH exists as an antigen-binding domain (i.e., variable domain). VHH can be easily identified from heavy-chain antibodies after immunizing a camelid, via cloning the VHH repertoire of B cells and panning by phage display (15). They bind an antigen with nanomolar affinity (16), highly soluble, and are much smaller than conventional antibodies (~15 kDa vs. ~150 kDa). These characteristics have led to the bacterial recombinant expression of VHHs (14,17) which is more straightforward and cost-effective than general antibody recombinant expression in mammalian cells.

In this thesis, click chemistry was employed to conjugate a cyclotide to an anti-class II MHC VHH (14) to improve its immunogenicity. The purpose of using anti Class II VHHs derives from their ability to deliver attached payloads to the very same cells responsible for antigen presentation in vivo, including dendritic cells.A non-natural amino acid was incorporated during cyclotide synthesis to enable a subsequent click reaction with a DBCO-modified anti MHC Class II VHH (18) (Figure 1). This strategy overcomes the fact that cyclotides lack a free N- and C-terminus, contain only fully oxidized cysteines and in their natural form have no amino acid residue enabling conjugation to a carrier protein. The comparatively small size of cyclotides also makes the presence of an adequate set of T helper epitopes unlikely, a complication confounded by the compact fold and the presence of three disulfide bonds, each of which is presumably required to be reduced for antigen presentation to occur. Whereas the range of chemical modifications available for site- specific attachment of the antigenic cargo was limited in previous conjugation studies (19), here a sortase-catalyzed reaction that affords both excellent yields (20,21) and absolute site specificity (22) was used. There are other enzymes used for antibody conjugation, such as transglutaminase (TGase) (23) (Figure 2A) or formylglycine-generating enzyme (FGE) (24) (Figure 2B). However, TGases tend to have low substrate specificity (25), and reconstituted FGE has poor activity in vitro (26) and requires the addition of a CuI cofactor to improve its activity (27).

121

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Two cyclotides were examined for immunogenicity: MCoTI-I in its native form with the exception of the single azido-lysine substitution, and MCoTI-I with a HA epitope grafted into loop 6, here after referred to as MCoTI-HA (Figure 1A). It was found that conjugating cyclotides to an alpaca- derived single domain antibody VHH targeting MHC class II product, VHH7 (14) dramatically improves the induction of antibodies reacting with cyclotides. Surprisingly, MCoTI-HA does not induce HA-specific antibodies and show less reactivity with a commercially available anti HA antibodies. Our results suggest that constraining the conformation of a particular sequence will hamper its reactivity against antibodies raised against the sequence. The data is also relevant for the design of peptide derivatives with enhanced or reduced immunogenicity.

Figure 1. (A) Sequence of MCoTI-I and MCoTI-HA (HA-tag shown in light blue). The site of azidolysine is shown in bold. (B) Schematic representation of Cu-free click reaction between cyclooctyne and azide, (C) Schematic illustration of conjugation strategy. VHH7 targeting antigen presenting cells (APCs) expressed with sorting motif, LPETGG, gets equipped with dibenzylcyclooctyne (DBCO) by SrtA reaction and subsequently conjugated with MCoTI-I (shown in cartoon representation) possessing azidolysine.

6.2. Experimental procedures 6.2.1. Peptides synthesized

122

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

All the peptides were synthesized following the Fmoc-chemistry protocol outlined in Chapter 2. 2- Chlorotritylchloride resin and an unnatural amino acid, Fmoc-Lys(N3)-OH (Chem-Impex International) were used for the synthesis of MCoTI-I and MCoTI-HA, for conjugation to VHH7. Cyclization was achieved following the strategy described by Cheneval et al. (28). The cyclic reduced peptides were oxidized in 0.1 M NH4HCO3 at pH 8. The probe peptides used in SrtA reaction, Gly-Gly-Gly-Cys-NH2 and Gly-Gly-Gly-HA-tag (YPYDVPDYA)-NH2 and linear HA-tag peptides were assembled on rink amide resin.

Figure 2. Schematic representation of other enzymatic modification strategies (A) Transglutaminase catalyzes transamination of Gln in the LLQGA recognition sequence installing diverse amines. (B) Formylglycine generating enzyme (FGE) converts thiol of the cysteine to an aldehyde group, which subsequently goes through reactions with aminooxy or hydrazide probes.

6.2.2. Nuclear magnetic resonance (NMR) analysis MCoTI-I and MCoTI-HA were subjected to structure calculation as described in Chapter 4. Briefly, peptides were dissolved in 90% H2O/10% D2O (v/v) or 100% D2O to a final concentration of 0.5 mM at pH 3. All two-dimensional NMR spectra were acquired at 298 K, except for temperature coefficient experiments (283-308 K), on a Bruker Avance 600-MHz NMR spectrometer equipped with a cryogenically cooled probe (Bruker, Karlsruhe, Germany).

6.2.3. Expression of recombinant VHH7 and Enhancer protein (VHH raised against green fluorescent protein) 123

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

E. coli glycerol stocks for expression of recombinant VHH7 and Enhancer VHH were obtained from Prof. Hidde Ploegh’s laboratory at the Whitehead Institute for Biomedical Research, Cambridge, MA, USA, and expressed as described previously (29). Briefly, the gene encoding a

His6 tagged VHH and LPXTG motif near the C-terminus was cloned in pHEN6 expression vector for periplasmic expression. Protein expression was induced by addition of 1 mM IPTG at

OD600=0.7 and cells were grown overnight at 30ºC. The periplasmic expression was harvested by osmotic shock and VHHs purified by Ni-NTA affinity and PD-10 desalting columns (GE Healthcare Life Sciences).

6.2.4. Maleimide-thiol reaction on Gly-Gly-Gly-Cys-NH2 Purified GGGC peptide (~80 μmol) was dissolved in PBS and dibenzylcyclooctyne-maleimide (DBCO-maleimide) (40 μmol, Click Chemistry Tools) dissolved in DMF was added. The reaction was incubated at room temperature overnight and subjected to purification using RP-HPLC with a gradient of 0‒40% ACN (0.05% TFA) over 40 min to remove unreacted GGGC peptide. Fractions showing correct mass were lyophilized (18).

6.2.5. Modification of VHH7 using SrtA

SrtA heptamutant (30) of S. aureus (20 μM final concentration) and GGGC(DBCO)-NH2 (1 mM final concentration) were added to the VHH (200 μM final concentration) in sortase buffer (50 mM

Tris, pH 7.5, 150 mM NaCl). The reaction was incubated at room temperature for ~4 h. His6-tagged SrtA and unreacted VHH were removed by Ni-NTA agarose and VHH-GGGC(DBCO) was purified using Superdex 75 size exclusion column and purity was analysed by SDS-PAGE.

6.2.6. Conjugation of a cyclotide to VHH7 80 μM of VHH equipped with DBCO and 160 μM MCoTI-I or MCoTI-HA containing azidolysine were conjugated in PBS at 4 ºC overnight. Conjugation progress was monitored by SDS-PAGE and LC-MS. Additionally, GGG-HA-tag was ligated to the C-terminus of VHH by SrtA reaction.

6.2.7. Immunization and serum collection BALB/C mice (6 weeks old) were immunized with 2.8 nmol of peptides (MCoTI-I, MCoTI-HA, HA) or VHH-peptide conjugates together with 50 μg of polyinosinic:polycytidylic acid (poly I:C) and 25 μg αCD40. Mouse blood was collected by cheek bleeding and serum was separated using

124

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

BD Microtainer™ tube. All described procedures were approved by the Massachusetts Institute of Technology Committee on Animal Care.

6.2.8. SDS-PAGE and Western blot Enhancer-peptide conjugates were subjected to SDS-PAGE at 80 V for 100 min and transferred to nitrocellulose membrane at 60 V for 120 min. The presence of protein samples on the nitrocellulose membrane was determined by Ponceau staining. Anti-sera collected from mice was used as primary antibody after 1:1000 dilution and anti-mouse IgG, horseradish peroxidase conjugated antibody was used for detection. For commercially available antibodies, monoclonal anti-HA antibody produced in mouse (Sigma-Aldrich, MO, #H3663 and Cell Signaling Technology, MA, #2367), anti-HA rabbit polyclonal antibody (Y-11, Santa Cruz Biotechnology, TX, #sc-805) were used.

6.3. Results 6.3.1. Click handle installation on MCoTI-I and a probe for VHH7 To conjugate MCoTI peptides to VHH7 using a strain-promoted click reaction, click handles such as azide or cyclooctyne needed to be attached to MCoTIs and two strategies were tried: (i) an NHS (N-hydroxysuccinimide) ester reaction for chemical conjugation to the primary amine of lysine (Figure 3A), (ii) incorporation of azidolysine during the synthesis of MCoTIs. Given that MCoTI-I has only one lysine, modification using NHS ester activated DBCO was attempted first, but yielded less than 50% modification of the peptide in the reaction and required additional purification (Figure 4). Therefore, a second strategy was chosen to produce MCoTIs with a click handle. GGGC was used as a probe in the SrtA reaction to modify VHH7 and was equipped with DBCO via a maleimide-thiol reaction (Figure 3B).

125

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Figure 3. Schematic representation of peptide modification strategies (A) NHS (N- hydroxysuccinimide) ester reaction for chemical conjugation to a primary amine of lysine, (B) Maleimide-thiol reaction for chemical conjugation to a sulfhydryl. R represents a labeling reagent (DBCO).

Figure 4. DBCO NHSester modification on MCoTI-HA. (A) LC profile of the reaction. The peak rising at 0.7 min contained unmodified MCoTI-HA and the peak rising at 0.9 min contained DBCO-modified MCoTI-HA. (B) Mass data corresponding to 0.7 min (upper) and 0.9 min (lower).

6.3.2. Bacterial expression, SrtA-mediated modification of VHH7, and conjugation to MCoTI- I / MCoTI-HA VHH7 with the SrtA motif near the C-terminus was expressed in the bacterial periplasm in order to have a disulfide bond formed. After purification by affinity and size exclusion chromatography,

126

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

purity and yield was analysed by SDS-PAGE (Figure 5A) and GGGC(DBCO) was coupled at the C-terminus via SrtA reaction. Subsequently, the click reaction between the azido-lysine of the MCoTIs and cyclo-octyne (DBCO) was performed in aqueous buffer at neutral pH, thus avoiding any unwanted side reactions, and yielded VHH7 modified at its C-terminus with either McoTI-I or McoTI-HA. Identity of the VHH7-McoTI-I and VHH7-McoTI-HA adducts after 21 h was confirmed by SDS-PAGE (Figure 5B), which yielded polypeptides of the expected sizes, and further verified by LC-MS (Figure 5C).

Figure 5. VHH7 and its conjugation to cyclotides (A) Commassie stained gel showing expression of VHH7 lane1: E.coli culture media collected after IPTG induction, 2: E.coli lysate, 3: flow- through sample collected after loading the lysate to Ni-NTA agarose column, 4-12: fractions collected after size-exclusion chromatography, (B) Instant blue stained gel showing the progress of click reaction between VHH7-GGGC(DBCO) and MCoTI-I / MCoTI-HA, (C) Mass of modified VHH7 after SrtA reaction and conjugates after click reaction.

6.3.3. NMR analysis of MCoTI-I and MCoTI-HA The amide region of both peptides was well dispersed in the TOCSY and NOESY spectra and the sequential connectivity could be readily confirmed. Comparison of the secondary Hα chemical shifts between MCoTI-I and MCoTI-HA showed that grafting of the HA-tag did not affect the 127

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

overall structure integrity of MCoTI-I (Figure 6A). Additional two-dimensional experiments were subsequently performed to determine three-dimensional structures of both peptides. Chemical shifts obtained for 1H, 13C, 15N are shown in Tables 1 and 2. Potential residues involved in hydrogen bond network were also found from temperature coefficient (Table 3, 4) and D2O exchange experiments. Dihedral angle restraints for the backbone ϕ, ψ angles were predicted by TALOS using the Hα, Cα, 3 Cβ, HN, and N chemical shifts in addition to JHN-Hα coupling constants measured from antiphase cross-peak splitting in the DQF-COSY spectrum (Table 5).

For each of MCoTI-I and MCoTI-HA, an ensemble of 20 minimum-energy structures was selected from the structures calculated and refined in a water shell using CNS. Structural statistics are summarized in Table 6. Initially, grafting of the HA-tag was anticipated to make the structure of MCoTI-I more rigid because the HA-tag sequence includes two prolines. Proline is conformationally rigid because of its fixed ϕ torsion angle of -60°. Therefore, flexibility of each loop was compared between MCoTI-I and MCoTI-HA based on Hβ chemical shifts of Cys as well as RMSD values. In general, a large βH chemical shift separation for an individual methylene pair of protons indicates that the pair is located in a highly structured environment (31). A similar degree of Hβ chemical shift separation was observed for the cysteines of both peptides, suggesting that the Pro-rich HA-tag did not dramatically affect the flexibility of MCoTI-I, although slightly larger Hβ separation of CysIV was found for MCoTI-HA (0.15 ppm versus 0.08 ppm). RMSD values for each loop measured using MolMol (32) agreed with this result (Table 7).

128

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Figure 6. Structural comparison of MCoTI-I and MCoTI-HA. (A) Hα secondary-shift comparison of MCoTI-I and MCoTI-HA. (B) Superimposition of the 20 lowest energy structures of MCoTI-I (left) and MCoTI-HA (right). Disulfide bonds are shown in grey in both structures.

129

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 1. Chemical shifts (ppm) of MCoTI-I

H N NE HA HB HG HD HE CA CB CG CD Others 1 Gly 8.21 108.4 - 4.13,3.85 - - - - 47.87 - - - 2 Gly 8.11 108.1 - 3.96,3.95 - - - - 47.05 - - - 3 Val 8.47 121.1 - 4.06 1.99 0.95,0.85 - - 65.05 35.37 23.34,23.60 - 4 Cys 8.69 126.6 - 5.14 2.91,2.90 - - - 54.88 43.71 - - 5 Pro - - - 4.45 2.32,2.33 1.94,2.04 3.89,3.83 - 65.85 35.19 29.85 54.11 7.55:HZ, 3.01, 6 Lys - - - 4.42 1.88,1.65 1.36,1.48 1.70,1.70 61.61 33.99 27.47 31.49 126.79:NZ, 3.02 44.61:CE 1.35,0.88, 20.42, 7 Ile 7.65 120.3 - 4.42 1.89 0.86 - 62.9 42.99 16.45 1.13 29.33 26.41, 8 Leu 8.67 126.2 - 4.49 1.56,1.72 1.49 0.76,0.77 - 57.06 44.17 29.82 27.36 9 Gln 8.7 - 112.57 4.57 2.00,1.74 2.37,2.47 - 6.99,7.51 57.95 34.39 36.38 - 10 Arg 8.74 127.6 119.69 4.48 1.69,1.61 1.21,1.21 2.88,2.97 6.96 58.4 33.49 30.49 45.42 11 Cys 8.4 115.8 - 4.87 3.03,3.21 - - - 56.75 50.66 - - 12 Arg 9.45 117.9 116.64 4.47 1.75,1.90 1.65,1.65 3.21,3.20 7.31 58.85 34.92 29.76 45.62 13 Arg 8.09 117 119.47 4.75 2.03,1.81 1.58,1.47 3.15 7.16 56.43 35.54 28.35 46.04 14 Asp 9.31 123.5 - 4.13 2.85,3.01 - - - 60.97 40.83 - - 15 Ser 8.19 111.1 - 4.3 4.12,3.84 - - - 62.69 64.86 - - 16 Asp 7.72 120.4 - 4.64 3.04,3.06 - - - 58.29 44.94 - - 17 Cys 8.07 117.5 - 4.99 2.74,2.85 - - - 54.81 43.26 - - 18 Pro - - - 4.64 1.95,2.32 2.03,2.11 3.42,3.87 - 64.87 35.13 30.11 52.61 19 Gly 8.48 106.5 - 3.79,3.84 - - - - 49.93 - - - 20 Ala 8.41 125.1 - 4.5 1.41 - - - 54.36 21.03 - - 21 Cys 8.23 117.3 - 4.7 3.23,3.31 - - - 58.74 48.24 - - 1.09,0.98, 22.09, 22 Ile 9.04 113.2 - 4.43 1.93 0.83 - 58.32 43.97 17.01 0.87 28.66 23 Cys 8.94 120.4 - 4.98 2.51,2.81 - - - 57.63 40.79 - - 24 Arg 8.14 128.6 118.19 4.35 2.07,2.53 1.76,1.63 3.26,3.27 6.91 59.07 33.8 30.51 45.94 25 Gly 8.9 108.2 - 3.94,3.95 - - - - 49.47 - - - 26 Asn 7.81 115.7 - 4.72 3.34,2.88 - 7.56,6.60 - 54.84 39.63 - - 108.55: ND 27 Gly 8.41 107.3 - 3.72,4.01 - - - - 48.99 - - - 28 Tyr 7.31 119.8 - 5.28 3.06,2.66 - 6.86 6.71 59.77 44.32 - - 29 Cys 8.81 120.7 - 5.39 3.17,2.83 - - - 58.26 43.57 - - 30 Gly 9.79 109.6 - 3.94,4.54 - - - - 48.32 - - - 31 Ser 8.84 116 - 4.54 3.91,3.98 - - - 62.3 66.66 - - 32 Gly 9.1 111.4 - 4.42,3.90 - - - - 47.29 - - - 33 Ser 8.67 115.9 - 4.46 3.91,3.91 - - - 61.76 66.2 - - 34 Asp 8.6 119.6 - 4.82 2.90,2.95 - - - 55.7 41.19 - -

130

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 2. Chemical shifts (ppm) of MCoTI-HA

H N NE HA HB HG HD HE CA CB CG CD Others 1 Gly 7.62 106.91 - 3.89,3.89 - - - - 48.01 - - - - 2 Gly 8.05 108.13 - 3.99,3.99 - - - - 47.59 - - - - 3 Val 8.2 118.19 - 4.22 2.04 0.94,0.89 - - 65.06 35.95 23.77,24.02 - - 4 Cys 8 122.58 - 5.04 2.23,2.28 - - - 55.29 42.17 - - - 5 Pro - - - 4.4 1.93,2.30 2.01,1.96 3.71,3.76 - 65.84 35.04 30.19 53.64 - 6 Lys 8.06 121.23 - 4.32 1.72,1.87 1.46,1.40 1.62 3.33 58.98 34.38 27.76 32.86 56.32:CE 0.86,1.13, 7 Ile 7.56 120.36 - 4.26 1.85 0.85 - 62.94 42.48 29.58,20.21 16.14 - 1.37 8 Leu 8.47 125.76 - 4.35 1.58 1.51 0.73,0.71 - 57.6 44.05 29.72 26.20,27.40 - 9 Gln 8.69 124.93 112.15 4.57 1.98,1.78 2.42,2.35 - 6.92,7.48 57.68 34.02 36.36 - - 10 Arg 8.71 127 119.68 4.45 1.68,1.63 1.26 2.99,2.91 6.97 58.8 33.48 30.49 45.48 - 11 Cys 8.39 115.99 - 4.87 3.18,3.00 - - - 56.55 50.49 - - - 12 Arg 9.37 117.89 119.74 4.45 1.74,1.89 1.63 3.18 7.3 58.54 34.79 29.73 45.65 - 13 Arg 8.06 116.75 119.45 4.72 1.81,2.04 1.45,1.56 3.14,3.15 7.14 56.45 35.47 28.36 45.96 - 14 Asp 9.41 122.97 - 4.13 2.92,3.10 - - - 60.57 40.06 - - - 15 Ser 8.21 111.46 - 4.3 4.12,3.83 - - - 62.71 64.81 - - - 16 Asp 7.68 120.31 - 4.63 3.02,3.05 - - - 58.13 44.76 - - - 17 Cys 8.03 117.58 - 5.01 2.77,2.64 - - - 54.57 42.8 - - - 18 Pro - - - 4.59 2.10,2.31 1.93,2.02 3.86,3.35 - 64.96 34.6 29.97 52.82 - 19 Gly 8.44 106.35 - 3.84,3.84 - - - - 49.95 - - - - 20 Ala 8.24 125.25 - 4.46 1.32 - - - 54.21 21.45 - - - 21 Cys 8.13 115.45 - 4.55 3.16,3.01 - - - 57.51 45.15 - - - 0.84,0.99, 22 Ile 8.83 112.12 - 4.42 1.84 0.7 - 61.58 44.1 28.88,22.13 17.22 - 0.80 23 Cys 8.6 119.86 - 4.94 2.50,2.77 - - - 57.59 40.66 - - - 24 Arg 8.17 128.68 117.96 4.31 2.48,2.05 1.66,1.47 3.20,3.19 6.76 59.35 34.04 31.07 45.65 - 25 Gly 8.79 108.06 - 3.95,3.95 - - - - 49.49 - - - - 26 Asn 7.82 115.79 - 4.7 3.32,2.88 - 7.58,6.65 - 54.91 39.66 - - 108.67:ND 27 Gly 8.38 107.22 - 3.98,3.69 - - - - 48.91 - - - - 28 Tyr 7.27 116.5 - 5.28 2.62,3.01 - 6.83 6.7 59.57 44.37 - - - 29 Cys 8.82 120.58 - 5.38 3.10,2.77 - - - 58.04 43.2 - - - 30 Gly 9.66 109.07 - 4.36,3.85 - - - - 48.35 - - - - 31 Tyr 8.59 119.63 - 4.63 2.76,3.02 - 7.16 6.91 60.5 40.7 - - - 32 Pro - - - 4.28 1.81,2.22 1.89,1.94 3.44,3.82 - 65.64 34.39 30.32 53.33 - 33 Tyr 7.88 118.77 - 4.45 3.12,2.83 - 7.15 6.84 61.21 41.25 - - - 34 Asp 8.42 116.8 - 4.61 2.94,2.84 - - - 55.73 40.19 - - - 35 Val 7.65 118.86 - 4.48 2.07 0.88,0.93 - - 62.23 35.7 23.20,22.86 - - 36 Pro - - - 4.28 1.67,2.15 1.94 3.76,3.66 - 66.01 34.6 30 53.53 - 37 Asp 8.26 117.26 - 4.48 2.75,2.83 - - - 55.51 40.23 - - - 38 Tyr 7.77 119.57 - 4.44 2.99,3.00 - 7.06 6.79 60.58 41.01 - - - 39 Ala 8.16 126.04 - 4.29 1.31 - - - 55.17 21.13 - - -

131

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 3. Temperature coefficients of MCoTI-I. Amino acid residues involved in hydrogen bonds are highlighted in yellow (i.e. temperature coefficients less negative than -4.7 ppb/K) Temperature (K) No. a.a. 283 288 293 298 303 308 ppb/K 1 G 8.281 8.255 8.235 8.213 8.183 8.161 -4.8 2 G 8.216 8.174 8.145 8.113 8.08 8.049 -6.6 3 V 8.592 8.549 8.512 8.469 8.415 8.376 -8.7 4 C 8.816 8.772 8.728 8.687 8.638 8.594 -8.9 5 P 6 K 7 I 7.708 7.687 7.669 7.653 7.624 7.608 -4 8 L 8.785 8.746 8.709 8.67 8.625 8.59 -7.9 9 Q 10 R 8.841 8.804 8.769 8.739 8.694 8.663 -7.1 11 C 8.422 8.412 8.406 8.4 8.384 8.376 -1.8 12 R 9.486 9.472 9.462 9.444 9.433 9.421 -2.6 13 R 8.118 8.108 8.102 8.084 8.075 8.069 -2.1 14 D 9.416 9.38 9.342 9.309 9.277 9.241 -7 15 S 8.318 8.272 8.232 8.194 8.147 8.108 -8.4 16 D 7.755 7.746 7.735 7.723 7.706 7.697 -2.4 17 C 8.127 8.108 8.091 8.072 8.052 8.033 -3.8 18 P 19 G 8.586 8.552 8.512 8.484 8.439 8.411 -7.1 20 A 8.504 8.47 8.44 8.413 8.377 8.347 -6.2 21 C 8.271 8.258 8.244 8.231 8.205 8.19 -3.3 22 I 9.069 9.074 9.058 9.042 9.014 8.988 -3.4 23 C 8.989 8.974 8.964 8.942 8.928 8.909 -3.2 24 R 8.188 8.173 8.16 8.143 8.129 8.119 -2.8 25 G 9.007 8.964 8.932 8.898 8.858 8.821 -7.3 26 N 7.837 7.828 7.821 7.811 7.8 7.792 -1.8 27 G 8.436 8.429 8.417 8.411 8.398 8.387 -2 28 Y 7.328 7.319 7.312 7.306 7.292 7.288 -1.6 29 C 8.856 8.834 8.825 8.81 8.79 8.778 -3.1 30 G 9.827 9.815 9.803 9.79 9.77 9.756 -2.9 31 S 8.96 8.915 8.873 8.835 8.787 8.751 -8.4 32 G 9.215 9.17 9.139 9.102 9.057 9.022 -7.7 33 S 8.784 8.75 8.706 8.669 8.625 8.586 -8 34 D 8.692 8.658 8.628 8.597 8.56 8.528 -6.5

132

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 4. Temperature coefficients of MCoTI-HA. Amino acid residues involved in hydrogen bonds are highlighted in yellow (i.e. temperature coefficients less negative than -4.7 ppb/K) Temperature (K) No. a.a. 283 288 293 298 303 308 ppb/K 1 G 7.641 7.628 7.632 7.625 7.616 7.605 -1.3 2 G 8.141 8.108 8.085 8.052 8.029 8.002 -5.5 3 V 8.34 8.292 8.254 8.2 8.163 8.116 -8.9 4 C 8.025 8.018 8.017 8 7.992 7.971 -2.1 5 P 6 K 7 I 7.638 7.606 7.593 7.567 7.549 7.528 -4.3 8 L 8.545 8.517 8.501 8.469 8.444 8.411 -5.3 9 Q 8.797 8.762 8.733 8.691 8.656 8.614 -7.3 10 R 8.833 8.791 8.759 8.714 8.68 8.635 -7.8 11 C 8.423 8.411 8.409 8.398 8.387 8.377 -1.8 12 R 9.422 9.408 9.399 9.374 9.365 9.343 -3.1 13 R 8.094 8.081 8.077 8.06 8.06 8.045 -1.9 14 D 9.516 9.477 9.45 9.415 9.382 9.346 -6.7 15 S 8.342 8.298 8.262 8.214 8.176 8.131 -8.4 16 D 7.725 7.71 7.707 7.688 7.678 7.662 -2.5 17 C 8.091 8.07 8.06 8.036 8.02 7.998 -3.7 18 P 19 G 8.55 8.512 8.484 8.445 8.416 8.377 -6.8 20 A 8.306 8.282 8.269 8.238 8.222 8.197 -4.3 21 C 8.182 8.166 8.151 8.13 8.118 8.095 -3.4 22 I 8.886 8.865 8.852 8.827 8.812 8.79 -3.8 23 C 8.657 8.637 8.625 8.603 8.59 8.565 -3.6 24 R 8.213 8.202 8.194 8.179 8.172 8.156 -2.2 25 G 8.89 8.855 8.828 8.792 8.763 8.726 -6.5 26 N 7.854 7.842 7.838 7.826 7.819 7.81 -1.7 27 G 8.412 8.401 8.396 8.38 8.374 8.358 -2.1 28 Y 7.304 7.295 7.287 7.274 7.267 7.256 -1.9 29 C 8.862 8.844 8.839 8.823 8.813 8.795 -2.5 30 G 9.699 9.685 9.677 9.661 9.647 9.629 -2.7 31 Y 8.692 8.656 8.63 8.593 8.561 8.524 -6.6 32 P 33 Y 8.044 7.991 7.942 7.884 7.837 7.779 -10.5 34 D 8.506 8.474 8.457 8.424 8.404 8.376 -5.1 35 V 7.744 7.71 7.688 7.656 7.631 7.602 -5.6 36 P 37 D 8.388 8.341 8.305 8.258 8.224 8.18 -8.2 38 Y 7.833 7.815 7.8 7.779 7.761 7.735 -3.8 39 A 8.267 8.248 8.212 8.161 8.122 8.078 -7.9

133

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 5. Dihedral angle restraints for MCoTI-I and MCoTI-HA

MCoTI-I MCoTI-HA

amino dihedral min. max. 3 amino dihedral min. max. 3 No. J No. J acid angles value value HN-Hα acid angles value value HN-Hα

4 CYS PSI 111.5 158.7 2 GLY PHI -84.3 -44.3 5 PRO PSI -54 -14 3 VAL PHI -150 -90 ~8Hz 6 LYS PHI -88.1 -48.1 4 CYS PSI 113.7 157.1 9 GLN PHI -190.6 -50.6 5 PRO PSI -51.5 -11.5 9 GLN PSI 116.5 171.8 6 LYS PHI -90.3 -50.3 10 ARG PHI -222.2 -16.2 7 ILE PHI -150 -90 ~9Hz 10 ARG PSI 86.2 184.3 9 GLN PHI -166.8 -71.5 11 CYS PHI -169.3 -126 9 GLN PSI 111.7 172.7 11 CYS PSI 136.1 188.8 10 ARG PHI -157.1 -16.1 11 CYS CHI1 30 90 10 ARG PSI 87.6 186.8 12 ARG PHI -131.5 -45.1 11 CYS PHI -166.2 -122.2 12 ARG PSI -62.9 15.4 11 CYS PSI 131.7 178.5 13 ARG PHI -182.3 -82.7 11 CYS CHI1 30 90 13 ARG PSI 113.2 185.2 12 ARG PHI -134 -47.2 13 ARG CHI1 30 90 12 ARG PSI -62.1 17.4 14 ASP PHI -81 -41 13 ARG PHI -180.3 -88.4 14 ASP PSI -57.6 -17.6 13 ARG PSI 115.9 184.2 15 SER PHI -84.8 -44.8 13 ARG CHI1 30 90 15 SER PSI -53.9 -13.9 14 ASP PHI -80.7 -40.7 15 SER CHI1 30 90 14 ASP PSI -57.5 -17.5 16 ASP PHI -95.5 -53 15 SER PHI -84.6 -44.6 16 ASP PSI -53.3 -13.3 15 SER PSI -53.1 -13.1 17 CYS PHI -138 -31.5 16 ASP PHI -95.7 -53.5 17 CYS PSI 65.2 168.7 16 ASP PSI -53.7 -13.7 20 ALA PHI -90 -30 ≤ 5Hz 17 CYS PHI -147.8 -29.2 21 CYS PHI -90 -30 ≤ 5Hz 17 CYS PSI 54.1 171.2 21 CYS CHI1 -90 -30 20 ALA PHI -90 -30 ≤ 5Hz 22 ILE PHI -152.5 -112.5 21 CYS PHI -90 -30 ≤ 5Hz 22 ILE PSI 135.2 175.2 22 ILE PHI -153.8 -113.8 23 CYS PHI -88.6 -48.6 23 CYS PHI -88.9 -48.9 23 CYS PSI 118.7 169.8 23 CYS CHI1 150 210 23 CYS CHI1 150 210 25 GLY PHI -82.9 -42.9 25 GLY PHI -82.2 -42.2 25 GLY PSI -61.5 -21.5 25 GLY PSI -60.5 -20.5 26 ASN PHI -86.8 -46.8 134

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

26 ASN PHI -87.3 -47.3 26 ASN PSI -55.4 -15.4 26 ASN PSI -51.8 -11.8 26 ASN CHI1 30 90 26 ASN CHI1 30 90 28 TYR CHI1 -90 -30 28 TYR CHI1 -90 -30 29 CYSS PHI -100.1 -34.7 29 CYSS CHI1 -90 -30 29 CYSS CHI1 -90 -30 30 GLY PHI -80.9 -40.9 30 GLY PHI -84.4 -44.4 30 GLY PSI -62.2 -22.2 31 TYR PHI -82.1 -42.1 31 SER PHI -86.4 -46.4 31 TYR CHI1 -90 -30 31 SER PSI -62 -22 33 TYR PHI -88.6 -48.6 32 GLY PHI -83.5 -43.5 33 TYR PSI -55 -15 32 GLY PSI -60.5 -20.5 34 ASP PHI -94 -54 34 ASP PSI -47.6 2.3 34 ASP CHI1 30 90 35 VAL PHI -115.6 -39.8 36 PRO PSI -51.9 -11.9 37 ASP PHI -91.2 -51.2 37 ASP PSI -39.1 1.6 37 ASP CHI1 30 90 38 TYR PHI -150 -90 13.8Hz

3 *Note: Backbone dihedral restraints for residues written in red were inferred from JHN-Hα coupling 3 3 constants, with ϕ restraind to 120 ± 30° for a JHN-Hα greater than 8 Hz, and -60 ± 30° for a JHN-Hα less than 5 Hz.

135

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 6. Energies and structural statistics for MCoTI-I and MCoTI-HA

Energies (kcal/mol) MCoTI-I MCoTI-HA Overall -1015.10 ± 28.33 -1195.31 ± 27.15 Bonds 13.69 ± 1.60 13.88 ± 1.37 Angles 41.28 ± 2.98 41.94 ± 4.64 Improper 16.94 ± 2.76 19.31 ± 2.41 Dihedral 140.60 ± 1.51 160.63 ± 2.00 van der Waals -118.47 ± 4.79 -150.84 ± 7.56 Electrostatic -1110.04 ± 32.34 -1281.23 ± 33.17 NOE 0.04 ± 0.02 0.01 ± 0.01 cDih 0.85 ± 0.46 0.99 ± 0.34

MolProbity statistics Clashes (>0.4Å/1000 atoms) 14.67 ± 4.55 10.73 ± 5.14 Poor rotamers 0.33 ± 0.62 0.06 ± 0.26 Ramachandran outliers (%) 0.00 ± 0.00 0.00 ± 0.00 Ramachandran favored (%)a 88.96 ± 3.89 88.71 ± 3.88

RMSDb from mean structure (Å) Mean global backbone (residues 8-30) 0.91 ± 0.27 0.74 ± 0.21 Mean global heavy (residues 8-30) 1.98 ± 0.30 1.68 ± 0.30

Distance restraints Intraresidue ( i-j=0 ) 67 42 Sequential ( |i-j|=1 ) 81 109 Medium range ( |i-j|<5 ) 35 55 Long range ( |i-j|≥5 ) 58 76 Hydrogen bonds 5 5 Disulfide bonds 3 3 Total 249 290

Dihedral angle restraints φ 23 28 ψ 16 20 χ1 7 11 Total 46 59

Violations from experimental restraints Total NOE violations exceeding 0.2 Å 0 0 Total dihedral violations exceeding 2.0° 0 0 a The rest (11.04%, 11.29%) were found in the allowed regions of the Ramachandran plot. b Mean RMSD was calculated using MOLMOL.

136

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Table 7. RMSD values of each loop of MCoTI-I and MCoTI-HA

Residue RMSD* MCoTI-I (Å) MCoTI-HA (Å)

bb 1.10 ± 0.39 1.05 ± 0.51 Loop 1 4–11 heavy 2.11 ± 0.65 2.24 ± 0.72 bb 0.42 ± 0.13 0.44 ± 0.17 Loop 2 11–17 heavy 1.97 ± 0.47 1.64 ± 0.35 bb 0.44 ± 0.22 0.58 ± 0.47 Loop 3 17–21 heavy 0.71 ± 0.26 0.93 ± 0.77 bb 0.16 ± 0.06 0.14 ± 0.05 Loop 4 21–23 heavy 0.33 ± 0.12 0.26 ± 0.09 bb 0.20 ± 0.07 0.22 ± 0.09 Loop 5 23–29 heavy 1.27 ± 0.49 0.90 ± 0.34 1–4, bb 2.67 ± 0.80 2.51 ± 0.82 Loop 6 29–34/39# heavy 3.25 ± 0.88 3.39 ± 1.01

*Measured using MolMol. #34/39 refers to MCoTI-I and MCoTI-HA numbering, respectively.

6.3.4. Mouse immunization and western blot analysis BALB/C mouse received five injections 2.8 nmol (~50 μg) of the VHH7_MCoTI-I, VHH7_MCoTI-HA or VHH7-HA adducts at 14-day intervals. Equimolar amount of peptide samples, MCoTI-I and MCoTI-HA were also used for immunization. Following primary injection, mouse anti-sera were collected after the second boost injection for assessment of antigen-specific immunoglobulins by immunoblots. To distinguish between antibodies specific for HA, cyclotide or a combination of the two target molecules based on an unrelated VHH was produced, the llama anti-GFP referred to as enhancer (33). We also assessed reactivity on unmodified VHH7, on enhancer modified with the HA epitope and on unmodified enhancer. Because the enhancer is also equipped with an LPETG sortase recognition motif near its C-terminus, the identical click conjugation strategy employed for the production of VHH7-MCoTI-I and VHH7-MCoTI-HA was used. The HA epitope was attached to the enhancer in a conventional sortase-catalyzed reaction, using GGG-HA as the incoming nucleophile.

First, reactivity of the antiserum of MCoTI-I and VHH7_MCoTI-I was compared with MCoTI-I. Whereas anti-MCoTI-I did not show any reactivity to the samples on the immunoblot (Figure 7A), anti-VHH7_MCoTI-I showed strong reactivity to 50 pmol of cyclic MCoTI-I (Figure 7B). 137

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Interestingly, the anti-VHH_MCoTI-I showed much higher reactivity towards the cyclic MCoTI-I over the linear counterpart (Figure 7B). From the comparison between anti-VHH_MCoTI-I and anti-VHH_MCoTI-HA, similar reactivity on adducts formed between enhancer and MCoTI-I or MCoTI-HA was observed. Ten picomole of enhancer_McoTI-HA was readily detectable using antisera of either VHH7_MCoTI-I or VHH7_MCoTI-HA. The serum produced after the forth immunization failed to react with unmodified VHH7, or with HA-modified enhancer. Interestingly, both anti VHH7_MCoTI-I and VHH7_MCoTI-HA reacted with a partial linear peptide (16-mer) derived from MCoTI-I. Of 16 the amino acid residues, 12 amino acids are in common between MCoTI-I and MCoTI-HA (Figure 8C,D). Using the same set of samples, the reactivity of anti-sera of VHH7-HA was also tested, but did not show any band implying the HA-tag alone was not antigenic enough under this experimental condition. Mice immunized with equimolar amounts of peptides also did not produce anti-sera to react with MCoTI-I or HA-tag. Anti-sera of VHH7_MCoTI-I and VHH7_MCoTI-HA showed reactivity after the second boost injection (data not shown), although the result shown below (Figure 8) was obtained using serum samples collected after the fourth boost injection.

6.3.5. Reactivity between MCoTI-HA and monoclonal anti HA antibody The next question asked was whether an epitope inserted into a constrained cyclotide scaffold can react with commercially available HA-tag antibody given that anti VHH7-MCoTI-HA did not react with linear HA. Interestingly, monoclonal HA-tag antibody showed no or less activity to HA-tag in the MCoTI-I scaffold compared to the reactivity to the linear HA-tag (Figure 9). Similarly, polyclonal HA-tag antibody also showed slightly less reactivity to MCoTI-HA than the linear HA.

Figure 7. Conjugation with VHH7 increases the immunogenicity of MCoTI-I. (A) Antiserum of MCoTI-I showed no reactivity with any samples on the immunoblot. (B) antiserum of MCoTI-I conjugated with VHH7 showed different reactivity to linear and cyclic MCoTI-I. Both antisera collected from mice were used as the primary antibody after 1:2000 dilution; anti- mouse IgG horseradish peroxidase-conjugated antibody was used for detection.

138

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Figure 8. Antisera reactivity against MCoTI-I and linear HA-tag (A) Mouse immunization and serum collection schedule. The antisera used for western blot in this figure was collected at the point shown in red, (B) Sequence of 16-mer highlighted in pink within MCoTI-I. Sequence of MCoTI-HA is also shown for comparison, (C) Western blot was performed using antisera of VHH7_MCoTI-HA and (D) VHH7-MCoTI-I. Both sera showed reactivity against cyclic and linear MCoTIs. Whereas no reactivity against linear HA-tag was found, anti VHH7_MCoTI-HA reacted with 16mer derived from MCoTI-I indicating HA-tag was not recognized because of its small size (9 amino acid residues). Enhancer protein used for conjugation of peptides for comparative transfer from SDS gel to nitrocellulose membrane was also used as a negative control.

139

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

Figure 9. Reactivity between MCoTI-HA and commercially available anti HA antibody (A) HA-tag constrained in the MCoTI-I scaffold was either not detected by monoclonal HA antibody, or (B) showed less reactivity than linear HA-tag by another monoclonal HA antibody and (C) polyclonal HA antibody.

6.4. Discussion Plant-derived cyclotides are remarkably stable due to their unique structure comprising a head-totail cyclic backbone and three disulfide bonds in their knotted core. They have been intensively studied for therapeutic application as well as elucidation of their biosynthesis. However, the immunogenicity of cyclotides has not been systematically studied for either purpose, even though such knowledge would be highly desirable: (i) to have better a better understanding of cyclotides in a clinical setting or (ii) to generate cyclotide specific antibodies to detect them in plant tissues, which might provide information on the location and timing of events in the biosynthesis of cyclotides.

Peptide-based immunization is usually inefficient because the delivery of peptides to professional APCs is often hampered by their weak chemical and physical stability. To overcome such drawbacks, peptide antigens are delivered using nanoparticles (8), or using antibodies targeting receptors on APCs such as DNGR-1 (34), DEC205 (35,36), Dectin-1 (37), DCIR2 (36), CD11c (38), Langerin (39) FcγR (40) and MHC class II products (10-12). Targeting APCs enables efficient

140

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

cross-presentation of an antigen and thus has been drawing attention for immunotherapy (11). In particular, targeting MHC class II products, which are expressed on the number of myeloid cells including dendritic cells, neutrophils, and activated macrophages, can potentially be applied to imaging immune response taking place during anticancer therapy as myeloid cells surround the tumor (14).

In this study, a cyclotide MCoTI-I was conjugated to the VHH7, an alpaca-derived single domain antibody fragment specific for MHC Class II, to attempt generation of a MCoTI-I specific antibody. Given that antigen delivery in a stable manner is an issue in peptide-based vaccine, the possibility for MCoTI-I to elicit epitope-specific antibodies was also tested using HA-tag peptide as a proof- of-concept. MCoTI-I and MCoTI-HA were synthesized, cyclized and oxidized as described, with one modification, namely the inclusion of an azido-lysine residue (Figure 1). The cyclic nature of both peptides obviously precludes their expression as genetic fusions with antibodies or antibody fragments, and the absence of readily modifiable functional groups hampers covalent attachment to carrier proteins more generally. We thus resorted to a click reaction to enable the conjugation of MCoTI-I and MCoTI-HA bearing the single azido-modifications to a suitably modified protein carrier. While click reaction has been typically achieved by involving copper as a catalyst, copper- free and strain-promoted click reaction used in this study has comparable kinetics to the traditional method but with non-cytotoxic effect (41), and has been used in other conjugation studies (18,30,42).

VHH7, isolated by phage display from a library constructed using peripheral blood lymphocytes from an alpaca immunized with total mouse splenocytes, was conveniently expressed in bacterial cells unlike full-length antibodies that require more cumbersome production in mammalian cells. VHH7 reacts with murine class II MHC products, as inferred from a lack of reactivity on cells from MHC Class II -/- mice (14,42) and capturing MHC Class II positive cells from mouse peripheral blood using immobilized VHH7 (43). VHH7 has been previously used to generate bispecific constructs by fusion with an anti-GFP VHH, using a strategy extended here to the production of VHH7-McoTI-I and VHH7-McoTI-HA adducts (42).

An immunization strategy that avoids the use of noxious complete Freund’s adjuvant and instead used a combination of anti-CD40 and the TLR3 agonist poly I:C was employed (44). The former activates antigen presenting cells such as dendritic cells (45), whereas the latter up regulates co- 141

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

stimulatory molecules essential for an optimal adaptive immune response (46). Although the serum produced after the forth immunization failed to react with unmodified VHH7, which is consistent with the poor immunogenicity of soluble monomeric VHHs, strong reactivity was observed with MCoTI-I or MCoTI-HA. Of particular interest is the complete absence of immunoreactivity with HA-modified enhancer. The HA epitope tag is frequently used in all manner of genetically modified proteins to track their presence and distribution. The HA epitope contained in MCoTI-HA is identical in sequence to that recognized by the commonly used anti-HA monoclonal antibodies. The failure of MCoTI-HA to elicit antibodies that recognize enhancer-HA can be attributed to conformational constraints imposed by grafting the HA epitope into loop 6 of MCoTI-HA (Figure 1). The HA epitope, when placed at the C-terminus of enhancer or VHH7 is unlikely to assume the constrained conformation present in MCoTI-HA. In the course of immunization, antigen processing may or may not have occurred, as the experiments did not directly address this issue. Should opening of the cyclotide have occurred in concert with reduction of the three disulfides, a linear HA epitope could have been produced, the presence of which would have been expected to elicit antibodies reactive with enhancer-HA. Given that antiserum raised against MCoTI-HA showed reactivity with the linear 16-mer peptide derived from loop 1 and 6 of MCoTI-I (Figure 6), it can be assumed that certain loops of a cyclotide have a better chance for antigen processing because of rigidity. In fact, the structure determination showed that loop 6 of MCoTI-I is the most flexible loop because of high contents of Gly, and replacing the amino acid composition with HA-tag did not change the flexibility (Table 7). This issue could be elucidated by additional immunoblot experiments to check reactivity between each loop and the antiserum, and further investigation on how MCoTI-I is processed by antigen presenting cells. However, commercially available polyclonal sera or a certain monoclonal antibody raised against the HA epitope do show some measure of reactivity with MCoTI-HA, demonstrating that the production of antibodies that recognize constrained epitopes is possible, as corroborated by the existence of many antibodies that show a conformation-sensitive pattern of recognition.

In conclusion, antibody generation against MCoTI-I was examined in this study. The conjugation of MCoTI-I to a VHH targeting MHC class II product significantly improved immunogenicity of the cyclotide, but antisera raised against the cyclotide reacted with both linear and cyclic version of MCoTI-I, thus feasibility of cyclotide conformation-specific antibody generation still remains to be investigated. In addition, no reactivity between the antiserum raised against MCoTI-HA and linear

142

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

HA-tag epitope implies that most of antibodies present in the serum are capable of recognizing the highly constrained peptide.

6.5. References

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

2. Ji, Y., Majumder, S., Millard, M., Borra, R., Bi, T., Elnagar, A. Y., Neamati, N., Shekhtman, A., and Camarero, J. A. (2013) In vivo activation of the p53 tumor suppressor pathway by an engineered cyclotide. J. Am. Chem. Soc. 135, 11623-11633

3. Wang, C. K., Gruber, C. W., Cemazar, M., Siatskas, C., Tagore, P., Payne, N., Sun, G., Wang, S., Bernard, C. C., and Craik, D. J. (2014) Molecular grafting onto a stable framework yields novel cyclic peptides for the treatment of multiple sclerosis. ACS Chem. Biol. 9, 156-163

4. Saska, I., Gillon, A. D., Hatsugai, N., Dietzgen, R. G., Hara-Nishimura, I., Anderson, M. A., and Craik, D. J. (2007) An asparaginyl endopeptidase mediates in vivo protein backbone cyclization. J. Biol. Chem. 282, 29721-29728

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

6. Poth, A. G., Mylne, J. S., Grassl, J., Lyons, R. E., Millar, A. H., Colgrave, M. L., and Craik, D. J. (2012) Cyclotides associate with leaf vasculature and are the products of a novel precursor in petunia (Solanaceae). J. Biol. Chem. 287, 27033-27046

7. Gillon, A. D., Saska, I., Jennings, C. V., Guarino, R. F., Craik, D. J., and Anderson, M. A. (2008) Biosynthesis of circular proteins in plants. Plant J. 53, 505-515

8. Petrizzo, A., Conte, C., Tagliamonte, M., Napolitano, M., Bifulco, K., Carriero, V., De Stradis, A., Tornesello, M. L., Buonaguro, F. M., Quaglia, F., and Buonaguro, L. (2015) Functional characterization of biodegradable nanoparticles as antigen delivery system. J. Exp. Clin. Cancer Res. 34, 114

9. Newman, K. D., Elamanchili, P., Kwon, G. S., and Samuel, J. (2002) Uptake of poly(D,L- lactic-co-glycolic acid) microspheres by antigen-presenting cells in vivo. J. Biomed. Mater. Res. 60, 480-486

10. Carayanniotis, G., and Barber, B. H. (1987) Adjuvant-free IgG responses induced with antigen coupled to antibodies against class II MHC. Nature 327, 59-61

11. Dickgreber, N., Stoitzner, P., Bai, Y., Price, K. M., Farrand, K. J., Manning, K., Angel, C. E., Dunbar, P. R., Ronchese, F., Fraser, J. D., Backstrom, B. T., and Hermans, I. F. (2009) Targeting antigen to MHC class II molecules promotes efficient cross-presentation and enhances immunotherapy. J. Immunol. 182, 1260-1269 143

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

12. Estrada, A., McDermott, M. R., Underdown, B. J., and Snider, D. P. (1995) Intestinal immunization of mice with antigen conjugated to anti-MHC class II antibodies. Vaccine 13, 901-907

13. Schwartz, R. H. (1985) T-lymphocyte recognition of antigen in association with gene products of the major histocompatibility complex. Annu. Rev. Immunol. 3, 237-261

14. Rashidian, M., Keliher, E. J., Bilate, A. M., Duarte, J. N., Wojtkiewicz, G. R., Jacobsen, J. T., Cragnolini, J., Swee, L. K., Victora, G. D., Weissleder, R., and Ploegh, H. L. (2015) Noninvasive imaging of immune responses. Proc. Natl. Acad. Sci. U. S. A. 112, 6146-6151

15. Nguyen, V. K., Desmyter, A., and Muyldermans, S. (2001) Functional heavy-chain antibodies in Camelidae. Adv. Immunol. 79, 261-296

16. Ghahroudi, M. A., Desmyter, A., Wyns, L., Hamers, R., and Muyldermans, S. (1997) Selection and identification of single domain antibody fragments from camel heavy-chain antibodies. FEBS Lett. 414, 521-526

17. Conrath, K. E., Lauwereys, M., Galleni, M., Matagne, A., Frere, J. M., Kinne, J., Wyns, L., and Muyldermans, S. (2001) beta-lactamase inhibitors derived from single-domain antibody fragments elicited in the Camelidae. Antimicrob. Agents Chemother. 45, 2807-2812

18. Witte, M. D., Theile, C. S., Wu, T., Guimaraes, C. P., Blom, A. E., and Ploegh, H. L. (2013) Production of unnaturally linked chimeric proteins using a combination of sortase-catalyzed transpeptidation and click chemistry. Nat. Protoc. 8, 1808-1819

19. Agarwal, P., and Bertozzi, C. R. (2015) Site-Specific Antibody-Drug Conjugates: The Nexus of Biciorthogonal Chemistry, Protein Engineering, and Drug Development. Bioconjug. Chem. 26, 176-192

20. Antos, J. M., Popp, M. W., Ernst, R., Chew, G. L., Spooner, E., and Ploegh, H. L. (2009) A straight path to circular proteins. J. Biol. Chem. 284, 16028-16036

21. Antos, J. M., Chew, G. L., Guimaraes, C. P., Yoder, N. C., Grotenbreg, G. M., Popp, M. W., and Ploegh, H. L. (2009) Site-specific N- and C-terminal labeling of a single polypeptide using sortases of different specificity. J. Am. Chem. Soc. 131, 10800-10801

22. Popp, M. W., Antos, J. M., Grotenbreg, G. M., Spooner, E., and Ploegh, H. L. (2007) Sortagging: A versatile method for protein labeling. Nat. Chem. Biol. 3, 707-708

23. Yokoyama, K., Nio, N., and Kikuchi, Y. (2004) Properties and applications of microbial transglutaminase. Appl. Microbiol. Biotechnol. 64, 447-454

24. Agarwal, P., van der Weijden, J., Sletten, E. M., Rabuka, D., and Bertozzi, C. R. (2013) A Pictet-Spengler ligation for protein chemical modification. Proc. Natl. Acad. Sci. U. S. A. 110, 46-51

25. Sunbul, M., and Yin, J. (2009) Site specific protein labeling by enzymatic posttranslational modification. Org. Biomol. Chem. 7, 3361-3371

144

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

26. Smith, E. L., Giddens, J. P., Iavarone, A. T., Godula, K., Wang, L. X., and Bertozzi, C. R. (2014) Chemoenzymatic Fc glycosylation via engineered aldehyde tags. Bioconjug. Chem. 25, 788-795

27. Knop, M., Engi, P., Lemnaru, R., and Seebeck, F. P. (2015) In Vitro Reconstitution of Formylglycine-Generating Enzymes Requires Copper(I). Chembiochem 16, 2147-2150

28. Cheneval, O., Schroeder, C. I., Durek, T., Walsh, P., Huang, Y. H., Liras, S., Price, D. A., and Craik, D. J. (2014) Fmoc-based synthesis of disulfide-rich cyclic peptides. J. Org. Chem. 79, 5538-5544

29. Ashour, J., Schmidt, F. I., Hanke, L., Cragnolini, J., Cavallari, M., Altenburg, A., Brewer, R., Ingram, J., Shoemaker, C., and Ploegh, H. L. (2015) Intracellular expression of camelid single-domain antibodies specific for influenza virus nucleoprotein uncovers distinct features of its nuclear localization. J. Virol. 89, 2792-2800

30. Li, Z., Theile, C. S., Chen, G. Y., Bilate, A. M., Duarte, J. N., Avalos, A. M., Fang, T., Barberena, R., Sato, S., and Ploegh, H. L. (2015) Fluorophore-Conjugated Holliday Junctions for Generating Super-Bright Antibodies and Antibody Fragments. Angew. Chem. Int. Ed. Engl. 54, 11706-11710

31. Daly, N. L., and Craik, D. J. (2000) Acyclic permutants of naturally occurring cyclic proteins. Characterization of cystine knot and beta-sheet formation in the macrocyclic polypeptide kalata B1. J. Biol. Chem. 275, 19068-19075

32. Koradi, R., Billeter, M., and Wuthrich, K. (1996) MOLMOL: A program for display and analysis of macromolecular structures. J. Mol. Graph. 14, 51-55

33. Kirchhofer, A., Helma, J., Schmidthals, K., Frauer, C., Cui, S., Karcher, A., Pellis, M., Muyldermans, S., Casas-Delucchi, C. S., Cardoso, M. C., Leonhardt, H., Hopfner, K. P., and Rothbauer, U. (2010) Modulation of protein properties in living cells using nanobodies. Nat. Struct. Mol. Biol. 17, 133-U162

34. Sancho, D., Mourao-Sa, D., Joffre, O. P., Schulz, O., Rogers, N. C., Pennington, D. J., Carlyle, J. R., and Reis e Sousa, C. (2008) Tumor therapy in mice via antigen targeting to a novel, DC-restricted C-type lectin. J. Clin. Invest. 118, 2098-2110

35. Bonifaz, L. C., Bonnyay, D. P., Charalambous, A., Darguste, D. I., Fujii, S. I., Soares, H., Brimnes, M. K., Moltedo, B., Moran, T. M., and Steinman, R. M. (2004) In vivo targeting of antigens to maturing dendritic cells via the DEC-205 receptor improves T cell vaccination. J. Exp. Med. 199, 815-824

36. Dudziak, D., Kamphorst, A. O., Heidkamp, G. F., Buchholz, V. R., Trumpfheller, C., Yamazaki, S., Cheong, C., Liu, K., Lee, H. W., Park, C. G., Steinman, R. M., and Nussenzweig, M. C. (2007) Differential antigen processing by dendritic cell subsets in vivo. Science 315, 107-111

37. Carter, R. W., Thompson, C., Reid, D. M., Wong, S. Y. C., and Tough, D. F. (2006) Preferential induction of CD4(+) T cell responses through in vivo targeting of antigen to dendritic cell-associated C-type lectin-1. J. Immunol. 177, 2276-2284

145

Chapter 6. Conjugation of a cyclotide to a single domain antibody to elicit an immune response

38. Castro, F. V. V., Tutt, A. L., White, A. L., Teeling, J. L., James, S., French, R. R., and Glennie, M. J. (2008) CD11c provides an effective immunotarget for the generation of both CD4 and CD8 T cell responses. Eur. J. Immunol. 38, 2263-2273

39. Idoyaga, J., Cheong, C., Suda, K., Suda, N., Kim, J. Y., Lee, H., and Park, C. G. (2008) Cutting edge: Langerin/CD207 receptor on dendritic cells mediates efficient antigen presentation on MHC I and II products in vivo. J. Immunol. 180, 3647-3650

40. Soleimanpour, S., Farsiani, H., Mosavat, A., Ghazvini, K., Eydgahi, M. R., Sankian, M., Sadeghian, H., Meshkat, Z., and Rezaee, S. A. (2015) APC targeting enhances immunogenicity of a novel multistage Fc-fusion tuberculosis vaccine in mice. Appl. Microbiol. Biotechnol.

41. Baskin, J. M., Prescher, J. A., Laughlin, S. T., Agard, N. J., Chang, P. V., Miller, I. A., Lo, A., Codelli, J. A., and Bertozzi, C. R. (2007) Copper-free click chemistry for dynamic in vivo imaging. Proc. Natl. Acad. Sci. U. S. A. 104, 16793-16797

42. Witte, M. D., Cragnolini, J. J., Dougan, S. K., Yoder, N. C., Popp, M. W., and Ploegh, H. L. (2012) Preparation of unnatural N-to-N and C-to-C protein fusions. Proc. Natl. Acad. Sci. U. S. A. 109, 11993-11998

43. Chen, G. Y., Li, Z., Theile, C. S., Bardhan, N. M., Kumar, P. V., Duarte, J. N., Maruyama, T., Rashidfarrokh, A., Belcher, A. M., and Ploegh, H. L. (2015) Graphene Oxide Nanosheets Modified with Single-Domain Antibodies for Rapid and Efficient Capture of Cells. Chemistry 21, 17178-17183

44. Fortier, M. E., Kent, S., Ashdown, H., Poole, S., Boksa, P., and Luheshi, G. N. (2004) The viral mimic, polyinosinic:polycytidylic acid, induces fever in rats via an interleukin-1- dependent mechanism. Am. J. Physiol. Regul. Integr. Comp. Physiol. 287, R759-766

45. Demangel, C., Palendira, U., Feng, C. G., Heath, A. W., Bean, A. G., and Britton, W. J. (2001) Stimulation of dendritic cells via CD40 enhances immune responses to Mycobacterium tuberculosis infection. Infect. Immun. 69, 2456-2461

46. Meylan, E., and Tschopp, J. (2006) Toll-like receptors and RNA helicases: two parallel ways to trigger antiviral responses. Mol. Cell 22, 561-569

146

Overall discussion

147

Overall discussion

In this thesis, disulfide-rich peptides were engineered using sortase A (SrtA) with the main focus on two peptide families, namely cyclotides and neurotoxins. A summary of the cyclized peptides examined in this thesis is shown in Table 1. The major goal to achieve using SrtA was to efficiently cyclize linear peptides, thus providing an alternative option to native chemical ligation (NCL). Although NCL has been the most commonly used method for cyclization to date, it favours Boc- chemistry thus needing the use of dangerous reagents such as hydrogen fluoride. Methods have been developed for NCL to be Fmoc-chemistry compatible or enable ligation based on the use of coupling reagents (1-3), demonstrating the importance and active interest in this field. As another alternative option, enzyme-based ligations have also been drawing attention, with SrtA from Staphylococcus aureus becoming a leading enzyme in the field due to its efficient catalytic activity and site-specificity.

Table 1. Disulfide-rich peptides cyclized by sortase A in this thesis Peptide No. of No. of Structural SrtA Order of disulfide amino description used cyclization and bonds acids oxidation [GG]SFTI- 1 (I-II) 22 cyclic backbone wild- one-pot reaction 1[LPETGG] type [G]α-Vc1.1 2 (I-III,II- 22 globular wild- cycliation [GLPETGG] IV) conformation type → oxidation [GGG]kalata B1 3 (I-IV, II- 35 cyclic cystine wild- oxidation [TGG] V, III-VI) knot type → cyclization [GG]kalata B1 3 (I-IV, II- 34 cyclic cystine mutant oxidation in [TGG] V, III-VI) knot vivo(misfolding) [GGG]κ-PVIIA 3 (I-IV, II- 36 inhibitor cystine mutant oxidation [LPETGG] V, III-VI) knot → cyclization

This thesis has provided a strategy for the SrtA-catalyzed cyclization of disulfide-rich peptides. Its general applicability was proven by cyclizing three different peptides, kalata B1, α-conotoxin Vc1.1, and SFTI-1 in Chapter 3. All three peptides, which have a different number of disulfide bonds and different structures, were efficiently cyclized via site-specific transpeptidation by SrtA. This proof- of-concept study became a basis for SrtA-catalyzed cyclization to convert an ICK peptide to CCK (as shown in Chapter 4) and for the recombinant expression of kalata B1 (Chapter 5). While using a one-pot reaction to achieve both oxidation and cyclization was possible for SFTI-1, which has one disulfide bond, separate steps for cyclization and folding were required for kalata B1 and α-Vc1.1. Optimization was also required to decide the order of cyclization and oxidation due to the 148

Overall discussion

complexity introduced by the presence of multiple disulfide bonds. While hydrolysis of cyclic peptides was prevented by monitoring the SrtA reaction in this thesis, methods to overcome the reversibility of the SrtA-catalyzed reaction (2,4-6) should be applied in future studies. Reducing the number of isomers caused by disulfide bond shuffling should also be considered to improve the yield of cyclic disulfide-rich peptides during the SrtA reaction. Disulfide bond shuffling observed during the cyclization is presumably caused by the cysteine protease nature of SrtA. Stepwise formation of disulfide bonds using different thiol protecting groups or selenocysteines can potentially improve the yield and should be considered in future study.

In an important result, the standard SrtA reaction was successfully applied to κ-conotoxin PVIIA and demonstrated the possibility of cyclization of conopeptides from the O1 gene superfamily, none of which had ever previously been cyclized. However, decreased activity was observed after cyclization due to the loss of electrostatic interaction between the N-terminus of native κ-PVIIA and the shaker channel. The result shows the importance of identifying active sites of a peptide so that peptide modification can be made without losing activity. A molecular modeling study would help to understand the interaction between the peptide and its target. In Chapter 4, the native and cyclic PVIIA were compared using molecular modeling, and a difference in the electrostatic interaction with the shaker channel was identified between the two peptides. They also had different hydrogen bonding networks, which are presumably caused by the linker used for cyclization. Therefore, optimization of the linker residues should be included in future research to achieve identical or improved molecular dynamics after cyclization. Unfortunately, limited therapeutic potential was found for cyclic-PVIIA based on activity screening on multiple ion channels. However, the success in cyclization is encouraging, as the strategy can be readily extended to other topologically similar peptides in the O1 superfamily, including the currently marketed drug ω- MVIIA (ziconotide, Prialt®).

In conjunction with the cyclization of synthetic peptides, the simplicity of the SrtA reaction mechanism can mediate the recombinant expression of cyclic disulfide-rich peptides. In this thesis, bacterial expression of the kalata B1 precursor was attempted and was coupled with backbone cyclization using SrtA. Periplasmic expression facilitated disulfide bond formation within bacterial cells, and maltose binding protein (MBP) used as a fusion protein aided solubility of expressed kalata B1. The SrtA recognition motif was introduced at both the N- and C-termini of the kalata B1 sequence (LPXTGG-kalata B1-LPVTGG) so that SrtA can not only cleave kalata B1 from MBP but 149

Overall discussion

also cyclize it. Although the mutant SrtA, Srt5°, efficiently cleaved and cyclized expressed kalata

B1, the resultant cyclic peptide was misfolded. However, refolding of cyclic-[GG]kalata B1[T] resulted in correct disulfide bond formation. Therefore, a new expression strategy for kalata B1 may be to exclude the signal peptide from the DNA sequence to have kalata B1 in a reduced state in the cytoplasm of E. coli and then to fold it in vitro. Based on the results obtained by using different bacterial strains, as well as changing the induction temperature, the lysis method, and parameters of the SrtA reaction, it can be concluded that MBP-kalata B1 can be expressed in the cytoplasm of

BL21 cell at 30°C and subjected to the sortase A (Srt5°) reaction for less than 2 h to prevent hydrolysis. Furthermore, co-expression of SrtA and peptide precursors in the cytoplasm should be considered in future to achieve cyclic kalata B1 within the bacterial cells, thereby minimizing the expression costs.

The site-specific ligation activity of SrtA not only enabled efficient cyclization of disulfide-rich peptides, but was also used for conjugation of a cyclotide to another protein. Specifically, the cyclotide MCoTI-I was conjugated to a single-domain antibody (sdAb, VHH), which targets MHC Class II products expressed on antigen presenting cells. The immunogenicity of MCoTI-I was significantly improved and mouse serum sampled after immunization showed high reactivity with MCoTI-I. Cyclotide specific antibodies can be useful as an immunohistochemistry tool, for example, to visualize their locations in plant tissues, thus providing more information on the biogenesis of cyclotides. Similarly, cyclotide antibodies could be potentially be applied to animal tissue sections to track how cyclotides are processed in vivo after administration for therapeutic purposes.

Although advanced techniques such as MALDI imaging are currently available, it is desirable to perform histological staining using antibodies after mass measurements to allow a histology- directed analysis of the mass spectra. Such immunohistochemistry requires antibodies with high selectivity and sensitivity. Conformation specific antibodies for cyclotides, which were not isolated in this thesis, would help to detect cyclotides in vivo. By doing so, we can find out under what circumstances cyclotide scaffolds lose their rigidity given that various gastrointestinal systems have varying pH and proteases. In a similar context, cyclotide antibodies would also allow us to study cyclotides via biological approaches at the cellular level, thus revealing intracellular trafficking or antigen processing upon endocytosis of cyclotides into antigen presenting cells. However, the most 150

Overall discussion

obvious potential of the conjugation strategy, as shown in Chapter 6, is targeted delivery of cyclotides. This means an epitope grafted onto a cyclotide scaffold can be delivered to the target in a stable and efficient manner.

References

1. Cheneval, O., Schroeder, C. I., Durek, T., Walsh, P., Huang, Y. H., Liras, S., Price, D. A., and Craik, D. J. (2014) Fmoc-based synthesis of disulfide-rich cyclic peptides. J. Org. Chem. 79, 5538-5544

2. Li, Y. M., Li, Y. T., Pan, M., Kong, X. Q., Huang, Y. C., Hong, Z. Y., and Liu, L. (2014) Irreversible site-specific hydrazinolysis of proteins by use of sortase. Angew. Chem. Int. Ed. Engl. 53, 2198-2202

3. Blanco-Canosa, J. B., Nardone, B., Albericio, F., and Dawson, P. E. (2015) Chemical protein synthesis using a second-generation N-acylurea linker for the preparation of peptide- thioester precursors. J. Am. Chem. Soc. 137, 7197-7209

4. Liu, F., Luo, E. Y., Flora, D. B., and Mezo, A. R. (2014) Irreversible sortase A-mediated ligation driven by diketopiperazine formation. J. Org. Chem. 79, 487-492

5. Williamson, D. J., Fascione, M. A., Webb, M. E., and Turnbull, W. B. (2012) Efficient N- terminal labeling of proteins by use of sortase. Angew. Chem. Int. Ed. Engl. 51, 9377-9380

6. Row, R. D., Roark, T. J., Philip, M. C., Perkins, L. L., and Antos, J. M. (2015) Enhancing the efficiency of sortase-mediated ligations through nickel-peptide complex formation. Chem. Commun. 51, 12548-12551

151

Concluding remarks

152

Concluding remarks

As noted throughout this thesis, peptides are great drug candidates because they have high specificity and potency towards their targets. Therefore, peptide drugs potentially require lower dose administration than conventional drugs, with therefore a lower chance of side effects. However, peptides tend to have short plasma half-lives and poor oral bioavailability, and thus require structure-based design or engineering approaches to improve their stability and bioavailability. Among a number of technologies developed to address these issues, cyclization of linear peptides and grafting a bioactive peptide onto a cyclotide framework have been shown to make therapeutically relevant peptides orally active (1,2).

Overall, this thesis has shown how sortase A (SrtA) can be used to engineer naturally occurring disulfide-rich peptides. A cyclotide kalata B1, α-conotoxin Vc1.1, and sunflower trypsin inhibitor-1 (SFTI-1) were successfully synthesized with the C-terminal SrtA recognition motif (LPXTG) and the N-terminal oligoglycine nucleophile, and subsequently cyclized by SrtA. The strategy to produce kalata B1, which has a unique topology called a cyclic cystine knot (CCK), using SrtA was extended to two additional studies. First, κ-conotoxin PVIIA, which has a linear cystine knot conformation, was efficiently converted to a cyclic cystine knot by SrtA. Second, SrtA was applied to kalata B1 precursor (maltose binding protein-LPETGG-kalata B1-LPVTGG) expressed in the periplasm of E. coli. Although periplasmic expression followed by SrtA-mediated cyclization resulted in misfolded kalata B1, subsequent refolding in vitro was successful. Therefore, cyclization prior to oxidation of expressed kalata B1 precursor can potentially reduce the cost of production significantly compared to that of chemical synthesis.

In addition to head-to-tail cyclization, SrtA-catalyzed reaction coupled with strain-promoted click chemistry was attempted to conjugate a cyclotide to a nanobody, VHH7. A non-natural amino acid was incorporated during cyclotide synthesis to enable a subsequent click reaction with a DBCO- modified VHH7. This strategy has overcome the fact that cyclotides lack a free N- and C-terminus, contain only fully oxidized cysteines, and in their natural form have no amino acid residue enabling conjugation to a carrier protein. The conjugation facilitated targeted delivery of cyclotides to antigen presenting cells in an attempt to examine cyclotides as immunogens. Mouse immunization and immunoblot analysis revealed that the VHH7-MCoTI-I conjugate elicits antibodies against cyclic or linear MCoTI-I, whereas MCoTI-I alone did not show any obvious antibody response.

153

Concluding remarks

The SrtA-catalyzed reaction used in this thesis achieved both absolute site specificity and excellent yields of engineered peptides. However, the necessity of a LPXTG motif hinders the production of cyclotides with native amino acid sequences. Recently, asparaginyl endopeptidase (AEP), a transpeptidase involved in the biosynthesis of cyclotides with site-specific activity (3), was successfully expressed in bacterial cells (4). Therefore, it is now possible to semi-enzymatically produce cyclotides without a linker sequence. As SrtA and AEP share a similar mechanism of transpeptidation, strategies and results shown in this thesis can be used as a basis for AEP-mediated peptide engineering. By utilizing these novel engineering strategies, we can enrich the current pool of peptide-based drug leads available for therapeutic development.

References

1. Clark, R. J., Jensen, J., Nevin, S. T., Callaghan, B. P., Adams, D. J., and Craik, D. J. (2010) The engineering of an orally active conotoxin for the treatment of neuropathic pain. Angew. Chem. Int. Ed. Engl. 49, 6545-6548

2. Wong, C. T., Rowlands, D. K., Wong, C. H., Lo, T. W., Nguyen, G. K., Li, H. Y., and Tam, J. P. (2012) Orally active peptidic bradykinin B1 receptor antagonists engineered from a cyclotide scaffold for inflammatory pain treatment. Angew. Chem. Int. Ed. Engl. 51, 5620- 5624

3. Gillon, A. D., Saska, I., Jennings, C. V., Guarino, R. F., Craik, D. J., and Anderson, M. A. (2008) Biosynthesis of circular proteins in plants. Plant J. 53, 505-515

4. Harris, K. S., Durek, T., Kaas, Q., Poth, A. G., Gilding, E. K., Conlan, B. F., Saska, I., Daly, N. L., van der Weerden, N. L., Craik, D. J., and Anderson, M. A. (2015) Efficient backbone cyclization of linear peptides by a recombinant asparaginyl endopeptidase. Nat. Commun. 6, 10199

154

Appendix

155

Concluding remarks

DNA sequence of [LPETGG]kalataB1[TGG] and restriction enzyme sites available in pLIC expression plasmid. The amino acid translation of [LPETGG]kalataB1[TGG] corresponding to the DNS sequence is shown in bold.

156