Molecular Insights Into Amyloid Toxicity Using Model Lipid Membranes

Molecular insights into Amyloid toxicity using model lipid membranes

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Varun Prasath Babu Reddiar

BSc (Biotechnology) Monash University & MSc (Biotechnology) Deakin University

School of Science

College of Science, Engineering and Health

RMIT University

March 2019

Declaration I certify that except where due to acknowledgement has been made, the work is that of the author alone; the work has not been submitted previously, in whole or in part, to qualify for any other academic award; the content of the thesis is the result of work which has been carried out since the official commencement date of the approved research program; any editorial work, paid or unpaid, carried out by a third party is acknowledged; and, ethics procedures and guidelines have been followed.

I acknowledge the support I have received for my research through the provision of an RMIT Research Stipend Scholarship.

B. Varun Prasath

1st March 2019

Acknowledgements This thesis covers the research I have conducted during the last four years during my PhD education in Chemistry at RMIT University, in Melbourne, Australia. The location on which this research conducted and thesis written was traditionally owned by the Wurundjeri people of the Kulin nation, and I respect the elder’s past, present and emerging. I want to take this opportunity to mention a few people who have assisted and supported me to achieve this milestone.

Drummond Supervisory Team and collaborators

This PhD program was fully funded through Higher Degree by Research professional scholarship and along with a stipend for the duration of my PhD program. This research project was supervised by both Professor Calum J Drummond who is also the Deputy Vice-Chancellor of Research and Innovation and Associate Professor Charlotte E Conn from the first day to submission. I want to thank both my supervisors, for their patience, guidance, encouragement and advice they have provided throughout my time as their student. Charlotte and Calum were critical to keeping me on task and maintaining direction for this project through regular meetings which were important and invaluable for me. I have been extremely lucky to have them as my supervisors, who cared so much and above all I appreciate their humility and honesty, which allowed me to grow and develop myself over the years.

I honestly have no words to describe my appreciation for Charlotte. She is a very caring person, also extremely passionate about research which is reflected by her accomplishments which are remarkable. Her undivided attention and support got me this far in my PhD, and she is an ideal role model for me.

I would like to thank Dr. Nicholas (Nick) Reynolds at Swinburne University, Melbourne and Dr. Celine Valery at RMIT (Bundoora campus) and Associate Professor Helmut Hugel at RMIT (City Campus), for their supervision and support. Especially, Nick for imaging the samples I had prepared, and I am indebted for his contribution.

Also, I would like to thank SAXS/WAXS beamline scientists (Dr Stephen Mudie and Dr Nigel Kirby) for their assistance with SAXS experiments at the Australian Synchrotron. Also, I’d like to thank Karin Vittrup, Soren Pape Moller and Nykola (Nyk) Jones at ISA, Centre for

III

Storage Ring Facilities in Aarhus (Denmark), especially, Nyk for assisting me with SRCD experiments.

I would like to thank my colleagues mainly, Asoka Weerawardena, Dr. Maggie Zhai, Dr. Thomas Meikle, Dr. Brendan Dyett, Dr. Leonie Van ’t Hag, Jamie Strachan, Radhika Arunkumar, Dilek Yalcin Tuncali, Hayden Broomhall and Sachini Kadaoluwa, who have assisted me with my project during my supervisor’s absence. The Technical staff were also critical to the completion of this project, therefore, thank you, Shane Edmonds, Rebecca Flower, Erica McDonald, Dilara Kaymakci, Tanya Nguyen-Milgate, Karl Lang, Bebeto Lay, Vivian Basile, Ruth Cepriano-Hall, Howard Anderson and Fiona De-Mendonca.

A huge thanks to all the admins and security personnel’s at RMIT, for creating such a dynamic, fun and above all, a safe environment for all of us. I have experienced a tonne of support, culture and extracurricular activities that made my four years a lot more bearable.

Family and close friends

Lastly, though certainly not least, to my close friends in research and outside academia, provided a much-needed form of escape from my studies, also deserve thanks for helping me keep things in perspective. Completing this work would have been all the more difficult were it not for the support and friendship of Dr. Andrew Coley, Dr. William Sullivan, Julia Liang, Dr. Mandeep Singh, Durga Dharmadana, Dr. Thomas McGrath, Timothy Coggan, Damien Moodie, Patience Shoko, Dr. Firoozeh Pourjavahe, Sana Subzwari, Buddima Edi, Sutha Satgunarajah, Buraq Tirli and Stuart Brown, thank you all very much.

To my very close family, I must express my gratitude to my beautiful wife Veena Vibhushini, for her continued support and encouragement. I was continually amazed by her willingness to proofread countless pages of my thesis and listen to several mock presentations. Also, thanks to the unconditional love and support received from my parents (Thilaga Rani and Babu Reddiar) and in-laws (Maheshwari and Balakrishnan). I would like to thank my sisters, Sanjeevini and Gayathiri along with my brother in law Sooriya and niece Shanaya for their love, support and encouragement. Lastly, my adorable pets, Tony and Cooper for keeping me active and healthy and above all they have helped me get through the tough times.

Conference presentations Experimental work from chapter 3, the effects of lipid composition on fibrillisation of the functional amyloid-forming peptide Somatostatin-14, was presented at two conferences where the details are listed below.

1. Australian Colloid and Surface science student conference, February 16 – 18, 2016, Sydney, Australia

2. Enabling Capability Platform conference, February 11 – 13, 2017, Melbourne, Australia

Abstract Functional amyloid-forming peptides can be found in nature, including stored in secretory granules as hormones. In contrast to toxic amyloid-forming peptides, the fibrillisation process for functional amyloids is typically “reversible”, a mechanism that may be related to the lack of toxicity for these amyloids. The mechanism of fibrillisation for functional amyloids has been reported to be dependent on changes to environmental conditions such as pH, temperature and solvent conditions. Some molecules are known to act as aggregation enhancers for functional amyloids, including, metal ions, heparin and lipid bilayer while others inhibit fibrillisation.

Lipid membranes have been shown to significantly affect fibrillisation, and membrane- mediated fibrillisation has been well characterised for toxic amyloid species such as the amyloid-beta peptide involved in Alzheimer’s disease. Depending on the lipid composition and physicochemical properties of the membrane, lipid membranes can both promote or inhibit fibril growth. The following parameters, including a surface charge on the membrane, membrane fluidity, chain packing, and lipid solubility have been shown to drive changes in the electrostatic and hydrophobic interactions between the peptide and the membrane, and hence the rate of fibrillisation. However, essentially nothing is known about the membrane-mediated assembly of functional amyloid-forming peptides.

In this thesis, the effect of the lipid membrane on fibrillisation of five different peptide hormones (Somatostatin, Oxytocin, Vasopressin, Substance P and Deslorelin) was investigated. In general, liposomes were used as model membranes. In Chapter 1, the effect of liposomes of different lipid composition on the fibrillisation of SST was investigated. SST reliably self-assembles into amyloid fibrils above a critical concentration (>5% w/w). Interactions of SST with the lipid bilayer were characterised using synchrotron radiation circular dichroism and intrinsic tryptophan fluorescence. The kinetics of fibrillisation was determined using a Thioflavin T assay, while the morphology of fibrils formed was directly visualised using atomic force microscopy (AFM). Reciprocal effects of the peptide and the growing fibril on the lipid bilayer were investigated using synchrotron small-angle x-ray scattering (SAXS).

In Chapter 4 a similar suite of techniques was used to determine the effect of the lipid membrane on the fibrillisation of two additional cyclic peptides, the neurohypophysial hormone-like peptides, vasopressin (AVP-9) and oxytocin (OT-9). These two nine-amino acid peptides are likely to form aggregates due to mutations and are known to form functional

amyloids in secretory granules. They are part of a family of structurally and functionally related peptide hormones with a conserved intramolecular ring, allowing us to determine whether substitutions in amino acids can impact the interactions of these peptides with the membrane and subsequent fibrillisation.

In Chapter 5 the effect of linear vs cyclic peptides was determined using two linear peptides, substance P (SP), and Deslorelin (Des). Substance P has been reported to form fibrils and has also been observed to self-assemble into nanotubes. Deslorelin is an analogue of LHRH, which has also been shown to adopt nanotubes in solution.

Finally, in Chapter 6, the anti-amyloidogenic effect of the polyphenol resveratrol, and its derivatives, Trimethoxy stilbene (TMS) and trans-stilbene (TS) was investigated using Somatostatin. Several studies have reported the anti-amyloidogenic effect of resveratrol on toxic amyloids such as Aβ and HIAPP. To our knowledge, this is the first study conducted on the effect of stilbenes on functional amyloids (somatostatin). All three stilbenes were shown to inhibit fibril formation in solution. When doped in a bilayer, this effect changes, particularly for TS, which actually promoted fibril growth.

This body of work has significantly added to our understanding of the effect of the lipid bilayer on the fibrillisation of functional amyloids, about which almost nothing is known to date. The trends observed not only assist in our fundamental understanding of how such peptide hormones are stored in vivo, but also adds to our understanding of fundamental peptide-lipid interactions, and the differences in toxicity between toxic and functional amyloids.

VII

Table of Contents

Declaration ...... II Acknowledgements ...... III Conference presentations ...... V Abstract ...... VI Table of Contents ...... VIII List of figures and tables ...... XII Abbreviations ...... XXVIII Explanatory notes ...... XXXII 1. The role of the lipid bilayer and molecular additives on fibrillisation of functional amyloid-forming peptides ...... 1 1.1 Introduction ...... 1 1.2 Common structural features and mechanism of amyloid formation ...... 2 1.2.1 Toxic Amyloids ...... 2 1.2.2 Functional Amyloids ...... 5 1.3 Molecular aggregation modifiers ...... 8 1.3.1 Impact of Metal Ions ...... 8 1.3.2 Impact of Heparin/Glycosaminoglycans (GAGs) ...... 8 1.3.3 Effects of Resveratrol (RSV) ...... 9 1.4 Membrane-based aggregation modifiers ...... 10 1.4.1 Ganglioside clusters, Cholesterol and Sphingomyelin ...... 12 1.4.2 Charged lipids ...... 13 1.4.3 Neutral lipid ...... 15 1.4.4 Physicochemical properties of the lipid bilayer ...... 16 1.5 Effect of amyloid-forming peptides on the structure of the lipid bilayer ...... 18 1.6 Model Lipid Bilayers ...... 20 1.6.1 Bulk Lamellar Phase ...... 21 1.6.2 Liposomes ...... 22 1.7 Effect of lipid bilayer composition on the fibrillisation of functional amyloid-forming peptides .. 23 2. Materials and Methods ...... 29 2.1 Materials ...... 29 2.1.1 Lipids ...... 29 2.1.2 Non-lipid materials ...... 29 2.1.3 Peptides ...... 29

VIII

2.2 Methods...... 30 2.2.1 Liposome Preparation ...... 30 2.2.2 Thioflavin T (ThT) Binding Assay ...... 30 2.2.3 Tryptophan Fluorescence ...... 31 2.2.4 Atomic Force Microscopy (AFM) ...... 31 2.2.5 Circular Dichroism (CD) ...... 32 2.2.6 Synchrotron Circular Dichroism (SR-CD) (ASTRID II facility in Aarhus Denmark) sample measurements and characterisation ...... 33 3. Effect of lipid composition on fibrillisation of the functional amyloid-forming peptide Somatostatin-14 (SST-14) ...... 35 3.1 Introduction ...... 35 3.2 Materials and Methods ...... 38 3.2.1 Liposome preparation and analysis ...... 38 3.2.2 Small Angle X-ray Diffraction (SAXD) on bulk lamellar sample, preparation and data analysis ...... 38 3.2.3 Small Angle X-ray Scattering (SAXS) ...... 39 3.2.4 Preparation of Functional amyloid-forming peptides (FAFP’s) ...... 40 3.2.5 Thioflavin T (ThT) Binding Assay ...... 40 3.2.6 Time-resolved kinetic studies using Dynamic Light Scattering (DLS) ...... 40 3.2.7 Atomic Force Microscopy (AFM) ...... 40 3.2.8 Tryptophan Fluorescence ...... 40 3.2.9 Synchrotron Circular Dichroism (SR-CD) (ASTRID II facility in Aarhus Denmark) sample measurements and characterisation ...... 40 3.3 Results ...... 41 3.3.1 Interaction with the lipid bilayer changes the secondary structure of SST-14 and the kinetics of fibril growth ...... 41 3.3.2 The peptide and bilayer interaction reveals no significant changes to bilayer size and shape . 48 3.3.3 The interaction between the peptide and the lipid bilayer do not significantly alter the bilayer structure ...... 49 3.3.4 Clear morphological differences are observed between fibrils formed in the buffer and those incubated with liposomes ...... 54 3.4 Discussions ...... 58 4. The interaction between the cyclic neurohypophysial hormone-like peptides Vasopressin (AVP-9) and Oxytocin (OT-9) and model membranes ...... 62 4.1 Introduction ...... 62 4.2 Materials and Methods ...... 64 4.2.1 Liposome preparation ...... 64 4.2.2 Preparation of Functional amyloid-forming peptides (FAFP’s) ...... 64

4.2.3 Thioflavin T (ThT) Binding Assay ...... 64 4.2.4 Atomic Force Microscopy (AFM) ...... 64 4.2.5 Circular Dichroism (CD) ...... 64 4.3 Results ...... 65 4.3.1 Effect of the lipid bilayer on the secondary structure of cyclic neurohypophysial hormones . 65 4.3.2 Effect of the lipid bilayer on the kinetics of fibril growth of OT-9 ...... 69 4.3.3 Effect of the lipid bilayer on the kinetics of fibril growth for AVP ...... 74 4.3.4 AFM a surface technique used to study aggregate deposited from solution both in the absence and presence of liposomes ...... 76 4.4 Discussion ...... 78 5. Effect of lipid composition on fibrillisation in linear functional amyloid-forming peptides, Substance P-11 (SP-11) and Deslorelin-10 (Des-10) ...... 83 5.1 Introduction ...... 83 5.2 Materials and Methods ...... 85 5.2.1 Liposome preparation ...... 85 5.2.2 Preparation of Functional amyloid-forming peptides (FAFP’s) ...... 85 5.2.3 Thioflavin T (ThT) Binding Assay ...... 86 5.2.4 Atomic Force Microscopy (AFM) ...... 86 5.2.5 Circular Dichroism (CD) ...... 86 5.2.6 Synchrotron Circular Dichroism (SR-CD) (ASTRID II facility in Aarhus Denmark) sample measurements and characterisation ...... 86 5.3 Results ...... 86 5.3.1 Interaction with the lipid bilayer changes the secondary structure of SP-11 and Des-10 ...... 86 5.3.2 Effect of the lipid bilayer on the kinetics of aggregation for SP-11 and ...... 91 Des-10 ...... 91 5.3.3 Clear morphological differences are observed between aggregates formed in the buffer and those incubated with liposomes ...... 97 5.4 Discussion ...... 99 6. Effects of Stilbenes on functional amyloid-forming peptide Somatostatin-14 (SST-14)...... 103 6.1 Introduction ...... 103 6.2 Materials and Methods ...... 106 6.2.1 Liposome preparation ...... 106 6.2.2 Preparation of stilbenes and liposome doped with stilbenes (RSV, TMS and TS) ...... 106 6.2.3 Preparation of Functional amyloid-forming peptides (FAFP’s) ...... 106 6.2.4 Thioflavin T (ThT) Binding Assay ...... 106 6.2.5 Tryptophan Fluorescence ...... 107 6.2.6 Circular Dichroism (CD) ...... 107

6.2.7 Atomic Force Microscopy (AFM) ...... 107 6.3 Results ...... 107 6.3.1 Addition of RSV to liposomes ...... 107 6.3.2 SST-14 interaction with the lipid bilayer doped with RSV ...... 109 6.3.3 Effect of RSV on fibrillisation of SST-14 ...... 113 6.3.4 SST-14 interaction with the lipid bilayer doped with tri-methoxy RSV and trans-stilbene .. 118 6.3.5 Effect of the stilbenes TMS and TS on fibrillisation of SST-14 ...... 121 6.3.6 SST-14 interaction with anionic charged lipid bilayer doped with tri-methoxy RSV and trans- stilbene ...... 127 6.3.7 Clear morphological differences are observed between fibrils formed in the buffer and RSV doped with lipid vesicles ...... 130 6.4 Discussion ...... 132 7. Conclusion and Future work ...... 136 References ...... 139 Appendix ...... 160

List of figures and tables Figure Title Page

1.1 Shows amyloid-forming proteins that form β-sheets are linked 1 to various human diseases (adopted from (1))

1.2 A schematic representation of the aggregation process for 3 Toxic amyloids. Aggregation proceeds by aggregation from the monomeric state to a fibrillar structure through oligomeric intermediates (adapted from (2))

1.3 The geometric isomers of RSV (adapted from (3)) 9

1.4 Shows a typical biological membrane (adapted from (4)) 10

1.5 Shows the model of Aβ-fibril formation on ganglioside clusters 12 in lipid rafts (adopted from (5))

1.6 Shows the two types of oligomers and its ability to interact with 13 the membrane (adopted from (6))

1.7 Schematic representation of interactions of amyloidogenic 15 IAPP (black lines) with SPBs (Supported Planar bilayers), depicting significant reduction in the fluidity of membrane (adapted from (adapted from (7))

1.8 Schematic illustration for the mechanism of TasA interacting 15 with the membrane (adopted from (8))

1.9 Schematic of Aβ interacting with lipids and phase separation 16 in the bilayer shows nonhomogeneity (adopted from (9))

1.10 Illustration of Aβ interacting with gel phase and fluid phase 17 (adapted from (10))

1.11 Illustration of three possible mechanism of Aβ encouraged 19 membrane damage (adopted from (11))

1.12 (A-C) Shows the clustered liposomes distorted by short fibrils. 20 (D) A 3D model of a distorted liposome (blue), surrounding

XII

fibrils visualised by cryo-electron tomography (cET) (adopted from (12))

1.13 Shows the polymorphic phases and corresponding dynamic 21 molecular shapes of a component of phospholipids (adopted from (13-16))

1.14 Shows a typical lamellar phase or phospholipid bilayer (lipids- 22 grey and water-red) (adopted from (17))

1.15 A structure of a liposome (adapted from (18)) 23

1.16 Shows the structure of five different lipid molecules (adapted 24 from (10))

1.17 Schematic representation of fluid (left) and gel (right) phase 25 bilayer (adapted from (10))

1.18 Shows the structure of anionic lipids and cholesterol (adapted 25 from (5, 19))

3.1 Shows the process of nucleation-dependent fibril formation for 36 Toxic vs Functional amyloids (adapted from (5))

3.2 Shows the amino acid sequence and disulphide link for SST- 38 14 (adapted from (20))

3.3 (a) SR-CD spectra for SST-14 co-incubated in vesicles after 3 41 hours and (b) SR-CD spectra for SST-14 co-incubated in vesicles after 72 hours

3.4 SST-14 environment in the bilayer measured by Tryptophan 43 Fluorescence

3.5 (a) Thioflavin T (ThT) assay of SST-14 incubated in NaCl, (b) 46 ThT assay of SST-14 and SST-14 at 5 % (w/w) co-incubated with unsaturated and saturated vesicle with increase in chain length impacting SST fibrillisation, (c) ThT assay of SST-14 and SST-14 co-incubated in DOPC, DOPC vesicle doped with

XIII

cholesterol, DMPC, DMPC doped with cholesterol, DOPC doped with DOPS and DMPC doped with DMPS

3.6 Size distribution of vesicles of various compositions co- 48 incuated with SST-14 (5% w/w) for 1 week as determined by DLS a) DOPC, b) DMPC, c) DOPC doped with cholesterol and d) DOPC doped with DOPS

3.7 Shows the hydrodynamic size of the vesicles with SST-14 over 49 1 week period

3.8 Shows the 1D diffraction pattern of Intensity vs q for bulk 50 lipids at varying SST concentration. Small angle X-ray scattering data were obtained ~7 h after setup of the plates.

3.9 The 1D diffraction pattern of Intensity vs q for bulk lipids at 51 varying SST concentration. Small angle X-ray scattering data were obtained ~7 h after setup of the plates.

3.10 Small angle X-ray scattering (SAXS) of Multilamellar vesicles 53 (MLVs) co-incubated with SST, (a) DOPC vesicle co- incubated with SST and (b) DOPC vesicle doped with cholesterol co-incubated with SST, (c) DMPC vesicles co- incubated with SST

3.11 Atomic Force Microscopy Analysis of 5% (w/w) SST fibrils 56 and SST fibrils co-incubated with DMPC lipid vesicles, (a-b) topographic images of (a) SST fibrils displaying typical amyloid-like unbranched twisted fibrillar structure, (b) SST co- incubated with DMPC vesicles, the formed fibrils display a more heterogeneous structure and evidence of the lipid coating of the fibrils is highlighted by the arrows, (c-d) statistical analysis (approximately 500 fibrils) of fibrillar species formed and (e-f) Persistance length calculations, for (e) SST and (f) SST+DMPC, very little difference, slight increase in stiffness for the SST + DMPC fibril

XIV

3.12 High Resolution AFM images of the change in fibril 57 nanostructure between SST and SST + DMPC fibrils (a-b) topographical AFM images of (a) SST and (b) SST + DMPC fibrils, (c-d) line section through one individual fibril (c) SST and (d) SST + DMPC. (e-f) Direct Fourier Transform analysis of pitch of multiple fibrils (g-h) Auto-Correlation Function analysis of twist angle of multiple fibrils

3.13 An illustration of SST-14 fibril interacting with fluid DMPC 60 bilayer, inspired by AFM image (Figure. 3.9b)

4.1 The amino acid sequences for neurohypophysial hormones (a. 62 AVP and b. OT) (adapted from (21))

4.2 Secondary structures of OT-9, AVP-9 and AVP:OT (1:1) at 1 65 mg mL-1 analysed using Circular Dichroism (CD)

4.3 (a) CD spectra for 1 mg mL-1 (0.01% w/w) OT-9 co-incubated 67 with saturated lipid vesicles after 3 hours, (b) after 72 hours, (c) CD spectra for 1 mg mL-1 OT-9 co-incubated with vesicles doped with various lipid compositions after 3 hours and (d) after 72 hours

4.4 (a) CD spectra for 1 mg mL-1 (0.01% w/w) AVP-9 co- 68 incubated in saturated lipid vesicles after 3 hours, (b) after 72 hours, (c) CD spectra for 1 mg mL-1 AVP-9 co-incubated in vesicles doped with various lipid compositions after 3 hours and (d) after 72 hours

4.5 (a) CD spectra for 1 mg mL-1 (0.01% w/w) AVP:OT (1:1) co- 69 incubated in saturated lipid vesicles after 3 hours, (b) after 72 hours, (c) CD spectra for 1 mg mL-1 AVP:OT (1:1) co- incubated in vesicles doped with various lipid compositions after 3 hours, (b) after 72 hours

4.6 (a) Thioflavin T (ThT) assay for OT-9 in buffer over a 1-week 71 period, (b) ThT assay for OT-9 at 0.5% (w/w) co-incubated

with unsaturated (DOPC) and saturated (DLPC, DMPC, DPPC and DSPC) vesicles with increase in chain length, (c) ThT assay for OT-9 at 0.5% (w/w) co-incubated with unsaturated and saturated vesicle doped with cholesterol and (d) ThT assay for OT-9 at 0.5% (w/w) co-incubated with unsaturated and saturated vesicle doped with charged lipids (DMPS/DOPS)

4.7 (a) ThT assay for vasopressin (AVP-9) in buffer over a 1-week 75 period, (b) ThT assay for AVP-9 at 5% (w/w) co-incubated with unsaturated and saturated vesicle with an increase in chain length, (c) ThT assay for AVP-9 at 5% (w/w) co-incubated with unsaturated and saturated vesicle doped with cholesterol and (d) ThT assay for AVP-9 at 5% (w/w) co-incubated with unsaturated and saturated vesicle doped with charge

4.8 ThT assay for AVP-9 at 5% (w/w) co-incubated with saturated 76 vesicle doped with an increase in the concentration of the charged lipid DMPS (5 – 35 mol%)

4.9 Topographic images of (a) short OT fibrils at 0.5% (w/w) in 77 aqueous solution and (b) image of a single OT fibril

4.10 Topographic image showing no OT fibrils detected with 77 DMPC

4.11 Topographic images of (a) 5 % (w/w) AVP in aqueous solution 78 and (b) AVP incubated with DMPC+DMPS vesicles

5.1 An illustration of the linear structure and amino acid sequence 84 of (a) SP and (b) Deslorelin similar to LHRH, (adapted from (22) (23))

5.2 Secondary structures of SP-11 at 0.3 mg mL-1 (0.03 % w/w) 87 analysed using Synchrotron Circular Dichroism (SR-CD), and Des-9 at 0.3 mg mL-1 analysed using Circular Dichroism (CD)

XVI

5.3 (a) SR-CD spectra for SP-11 at 0.3 mg mL-1 (0.03 % w/w) co- 88 incubated in vesicles after 3 hours, (b) SR-CD spectra for SP- 11 co-incubated in vesicles after 72 hours

5.4 (a) CD spectra for Des-9 at 2 mg mL-1 (0.2% w/w) co- 90 incubated in vesicles after 3 hours, (b) CD spectra for Des-9 co-incubated in vesicles after 72 hours

5.5 ThT assay performed over 1-week period for (a) Substance P 92 (SP-11) in NaCl, (b) 10% w/w or 5 mol% SP-11, co-incubated with unsaturated and saturated vesicle with an increase in chain length, (c) SP (5 mol%) co-incubated with DMPC, DMPC doped with cholesterol and (d) SP (5 mol%) co-incubated with DOPC, DOPC vesicle doped with cholesterol and DOPC doped with DOPS

5.6 ThT assay performed over 1-week period for (a) Deslorelin 95 (Des-9) in NaCl, (b) 10% w/w or 5 mol% Des-9, co-incubated with unsaturated and saturated vesicle with an increase in chain length, (c) Des (5 mol%) co-incubated with DMPC, DMPC doped with cholesterol and (d) Des-9 (5 mol%) co-incubated with DOPC, DOPC vesicle doped with cholesterol and DOPC doped with DOPS

5.7 Atomic force microscopy (AFM) analysis of SP fibrils at 10% 97 (w/w) and SP fibrils co-incubated with DMPC lipid vesicles, (a-b) topographic images of (a) SP fibrils displaying needle- like fibrillar structure and (b) SP co-incubated with DMPC vesicles

5.8 Topographic images of (a) Des fibrils at 10% (w/w) in the 98 absence of DMPC vesicles assemblies can be seen better in the adhesion and (b) deformation channels

5.9 Topographic images of Des fibrils in DMPC displaying a 98 regular looking amyloid (a) with periodic twisting structure (b) the twists can be clearly seen in the deformation channel

XVII

5.10 Illustrations of (a) an electron micrograph image of triptorelin 100 in water (24) and (b) nanotube self-assembly (25)

6.1 Shows the structure of stilbenes (a) RSV (RSV) (26), (b) 104 TMS (TMS) (27) and (c) TS (27)

6.2 Diagram showing RSV location in the bilayer of DPPC/Chol 104 (adapted from (28))

6.3 Shows the Dynamic light scattering (DLS) curves for pure 108 saturated chain vesicles and saturated vesicles doped with 5 mol% RSV (a) DLPC, (b) DMPC, (c) DPPC and (d) DSPC

6.4 DLS curve for pure DMPC+Chol vesicle and vesicle doped 109 with 5 mol% RSV

6.5 DLS curve for (a) Pure DOPC+DOPS vesicle and vesicle 109 doped with 5 mol% RSV and (b) Pure DMPC+DMPS vesicle and vesicle doped with 5 mol% RSV

6.6 (a) CD spectra for SST-14 0.01% w/w (0.1 mg mL-1) co- 111 incubated in saturated vesicles doped with 5 mol% RSV, (b) CD spectra for SST-14 co-incubated in DMPC bilayer doped with cholesterol and 5 mol% RSV and (c) CD spectra for SST- 14 co-incubated in bilayers (DMPC & DOPC) doped with anionic charge (DMPS & DOPS) and 5 mol% RSV

6.7 SST-14 environment in bilayer and bilayer doped with RSV, 112 measured by Tryptophan Fluorescence

6.8 Shows the ThT assay curves for 5% (w/w) SST-14 incubated 114 in buffer (NaCl) vs stilbenes co-dissolved using DMSO

6.9 Shows the ThT assay curves for 5% (w/w) SST-14 co- 115 incubated in saturated chain lipids doped with 5 mol% RSV (a) DLPC, (b) DMPC, (c) DPPC vesicles and (d) DSPC

XVIII

6.10 Shows the ThT assay curves for 5% (w/w) SST-14 co- 116 incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% RSV

6.11 Illustrations of: (a) ThT assay curves for 5% (w/w) SST-14 co- 117 incubated in buffer and DOPC/DOPS vesicles doped with 5 mol% RSV (b) ThT assay curves for 5% (w/w) SST-14 co- incubated in buffer and DMPC/DMPS vesicles doped with 5 mol% RSV

6.12 Shows the Dynamic light scattering (DLS) curves for pure 119 saturated chain vesicles and saturated vesicles doped with 5 mol% stilbenes (a) DLPC, (b) DMPC, (c) DPPC and (d) DSPC

6.13 SST-14 environment in bilayer and bilayer doped with 120 stilbenes, measured by Tryptophan Fluorescence

6.14 Shows the ThT assay curves for 5% (w/w) SST-14 incubated 122 in buffer (NaCl) vs stilbenes co-dissolved using DMSO

6.15 Shows the ThT assay curves for 5% (w/w) SST-14 co- 123 incubated in saturated chain lipids doped with 5 mol% stilbenes (a) DLPC, (b) DMPC, (c) DPPC vesicles and (d) DSPC

6.16 DLS curve for pure DMPC+Chol vesicle and vesicle doped 125 with 5 mol% stilbenes

6.17 Shows the ThT assay curves for 5% (w/w) SST-14 co- 126 incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% stilbenes

6.18 DLS curve for (a) Pure DOPC+DOPS vesicle and vesicle 127 doped with 5 mol% stilbenes and (b) Pure DMPC+DMPS vesicle and vesicle doped with 5 mol% stilbenes

6.19 Illustrations of: (a) ThT assay curves for 5% (w/w) SST-14 co- 128 incubated in buffer and DOPC/DOPS vesicles doped with 5 mol% stilbenes (b) ThT assay curves for 5% (w/w) SST-14 co-

XIX

incubated in buffer and DMPC/DMPS vesicles doped with 5 mol% stilbenes

6.20 AFM analysis of SST-14 fibrils at 10% (w/w) with RSV, 130 topographic images of (a) SST-14 with 5 mol% RSV displaying little or fibrils and (b) SST-14 co-incubated with DMPC vesicles doped with 5 mol% RSV displaying much shorter fibrils

6.21 AFM analysis of SST-14 fibrils at 10% (w/w) with TS, 131 topographic images of (a) SST-14 with 5 mol% TS displaying little or no fibrils and (b) SST-14 with DMPC vesicles doped with 5 mol% TS displaying neat long fibrils

6.22 AFM analysis of SST-14 fibrils at 10% (w/w) with TMS, 132 topographic images of (a) SST-14 with 5 mol% TMS displaying little or no fibrils and (b) SST-14 with DMPC vesicles doped with 5 mol% TMS displaying neat long fibrils

6.23 Diagrammatic illustrations of how RSV inhibits SST 133 fibrillisation by disturbing the lipid monolayer by a dual effect approach (adapted from (29))

S8.1 Shows the DLS curves for liposomes doped with various lipid 160 compositions

S8.2 ThT curve fit for SST (5% w/w) in NaCl 164

S8.3 ThT curve fit for SST (5% w/w) in DOPC 165

S8.4 ThT curve fit for SST (5% w/w) in DLPC 165

S8.5 ThT curve fit for SST (5% w/w) in DMPC 166

S8.6 ThT curve fit for SST (5% w/w) in DPPC 166

S8.7 ThT curve fit for SST (5% w/w) in DSPC 167

S8.8 ThT curve fit for SST (5% w/w) in DOPC+Chol 167

S8.9 ThT curve fit for SST (5% w/w) in DMPC+Chol 168

S8.10 ThT curve fit for SST (5% w/w) in DOPC+DOPS 168

S8.11 ThT curve fit for SST (5% w/w) in DMPC+DMPS 169

S8.12 ThT curve fit for OT (0.1% w/w) in NaCl 169

S8.13 ThT curve fit for OT (0.5% w/w) in NaCl 170

S8.14 ThT curve fit for OT (5% w/w) in NaCl 170

S8.15 ThT curve fit for OT (0.5% w/w) in DOPC 171

S8.16 ThT curve fit for OT (0.5% w/w) in DOPC+Chol 171

S8.17 ThT curve fit for OT (0.5% w/w) in DOPC+DOPS 172

S8.18 ThT curve fit for OT (0.5% w/w) in DLPC 172

S8.19 ThT curve fit for OT (0.5% w/w) in DMPC 173

S8.20 ThT curve fit for OT (0.5% w/w) in DMPC+Chol 173

S8.21 ThT curve fit for OT (0.5% w/w) in DMPC+DMPS 174

S8.22 ThT curve fit for OT (0.5% w/w) in DPPC 174

S8.23 ThT curve fit for OT (0.5% w/w) in DSPC 175

S8.24 ThT curve fit for AVP (5% w/w) in 95DMPC+5DMPS 175

S8.25 ThT curve fit for AVP (5% w/w) in 85DMPC+15DMPS 176

S8.26 ThT curve fit for AVP (5% w/w) in 75DMPC+25DMPS 176

S8.27 ThT curve fit for AVP (5% w/w) in 65DMPC+35DMPS 177

S8.28 ThT curve fit for Des (10% w/w) in DOPC 177

S8.29 ThT curve fit for Des (10% w/w) in DOPC+Chol 178

S8.30 ThT curve fit for Des (10% w/w) in DOPC+DOPS 178

S8.31 ThT curve fit for Des (10% w/w) in DLPC 179

S8.32 ThT curve fit for Des (10% w/w) in DMPC 179

XXI

S8.33 ThT curve fit for Des (10% w/w) in DPPC 180

S8.34 ThT curve fit for Des (10% w/w) in DSPC 180

S8.35 ThT curve fit for Des (10% w/w) in DMPC+Chol 181

S8.36 ThT curve fit for SP (10% w/w) in DLPC 181

S8.37 ThT curve fit for SP (10% w/w) in DMPC 182

S8.38 ThT curve fit for SP (10% w/w) in DPPC 182

S8.39 ThT curve fit for SP (10% w/w) in DSPC 183

S8.40 ThT curve fit for SP (10% w/w) in DOPC+Chol 183

S8.41 ThT curve fit for SP (10% w/w) in DMPC+DMPS 184

S8.42 ThT curve fit for SST (5% w/w) in RSV (RSV) 184

S8.43 ThT curve fit for SST (5% w/w) in TS (TS) 185

S8.44 ThT curve fit for SST (5% w/w) in trimethoxy RSV (TMS) 185

S8.45 ThT curve fit for SST (5% w/w) in DLPC+RSV 186

S8.46 ThT curve fit for SST (5% w/w) in DLPC+TMS 186

S8.47 ThT curve fit for SST (5% w/w) in DLPC+TS 187

S8.48 ThT curve fit for SST (5% w/w) in DMPC+RSV 187

S8.49 ThT curve fit for SST (5% w/w) in DMPC+TS 188

S8.50 ThT curve fit for SST (5% w/w) in DMPC+TMS 188

S8.51 ThT curve fit for SST (5% w/w) in DMPC+Chol+RSV 189

S8.52 ThT curve fit for SST (5% w/w) in DMPC+Chol+TS 189

S8.53 ThT curve fit for SST (5% w/w) in DMPC+Chol+TMS 190

S8.54 ThT curve fit for SST (5% w/w) in DMPC+DMPS+RSV 190

S8.55 ThT curve fit for SST (5% w/w) in DMPC+DMPS+TS 191

S8.56 ThT curve fit for SST (5% w/w) in DMPC+DMPS+TMS 191

XXII

S8.57 ThT curve fit for SST (5% w/w) in DOPC+DOPS+TS 192

S8.58 ThT curve fit for SST (5% w/w) in DOPC+DOPS+TMS 192

S8.59 ThT curve fit for SST (5% w/w) in DPPC+RSV 193

S8.60 ThT curve fit for SST (5% w/w) in DPPC+TS 193

S8.61 ThT curve fit for SST (5% w/w) in DPPC+TMS 194

S8.62 ThT curve fit for SST (5% w/w) in DSPC+RSV 194

S8.63 ThT curve fit for SST (5% w/w) in DSPC+TS 195

S8.64 ThT curve fit for SST (5% w/w) in DSPC+TMS 195

XXIII

Tables Title Page

1.1 Shows the diseases resulting from protein misfolding and 4 subsequent aggregation into amyloid nanofibrils (30, 31)

1.2 Shows a list of functional amyloids in prokaryotic and 5 eukaryotic organisms (2, 31, 32)

1.3 Includes a summary of model membranes involved in 11 fibrillisation of amyloidogenic proteins and peptides (33)

3.1 The secondary structure of selected samples of SST-14 co- 42 incubated with liposomes, analysed using Dichroweb with Contin-LL algorithm method. *Quantitative analysis was performed for DOPC+SST, DOPC+DOPS+SST (3 and 72 hours) and DMPC+SST (72 hours) since NRMSD > 0.250 (in appendix Table S8.3)

3.2 Shows the lag phase, half-time, aggregation rate constant and 47 the final fluorescence intensity of plateau phase for SST-14 at 5% (w/w) in NaCl and vesicles doped with various lipid compositions over 1 week period

3.3 Shows the effect of SST addition (0.01 to 2.5% w/w) on D- 50 spacing of the lamellar phase of DMPC

3.4 Shows the effect of SST addition (0.01 to 2.5% w/w) on D- 51 spacing of the lamellar phase of DSPC

3.5 Lattice parameter or D-spacing for SST co-incubated with 52 vesicle

4.1 Shows the lag phase, half-time, aggregation rate constant and 72 the final fluorescence intensity of plateau phase for OT-9 in the buffer over 1 week

XXIV

4.2 Shows the lag phase, half-time, aggregation rate constant and 73 the final fluorescence intensity of plateau phase for OT-9 at 0.5% (w/w) in buffer and vesicles doped with various lipid compositions over 1 week

4.3 Shows the position, polarity, charge and hydrophobic amino 74 acid residues for OT and AVP

4.4 Shows the lag phase, half-time, aggregation rate constant and 76 the final fluorescence intensity of plateau phase for AVP-9 at 5% (w/w) in DMPC vesicles doped with DMPS at (5-35 mol%)

5.1 The secondary structure of SP-11 in liposomes analysed using 89 Dichroweb with Contin-LL algorithm method

5.2 Shows the lag phase, half-time, aggregation rate constant and 93 the final fluorescence intensity of plateau phase for SP at 10% (w/w) in NaCl and vesicles doped with various lipid compositions over 1 week

5.3 Shows the lag phase, half-time, aggregation rate constant and 96 the final fluorescence intensity of plateau phase for Des at 10% (w/w) in NaCl and vesicles doped with various lipid compositions over 1 week

6.1 Shows the topological polar surface area (tPSA) and partition 105 coefficient (clogP) values for stilbenes

6.2 Shows the lag phase, half-time, aggregation rate constant and 116 the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer, RSV, saturated chain lipids and saturated chain lipids doped with 5 mol% RSV

6.3 Shows the lag phase, half-time, aggregation rate constant and 117 the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% RSV

XXV

6.4 Shows the lag phase, half-time, aggregation rate constant and 118 the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DOPC/DOPS and DMPC/DMPS vesicles doped with 5 mol% RSV

6.5 Shows the lag phase, half-time, aggregation rate constant and 122 the final fluorescence intensity of plateau phase for SST at 5% (w/w) in buffer vs stilbenes

6.6 Shows the lag phase, half-time, aggregation rate constant and 124 the final fluorescence intensity of plateau phase for SST at 5%

(w/w) co-incubated in the buffer, stilbenes, saturated chain lipids and saturated chain lipids doped with 5 mol% stilbenes

6.7 Shows the lag phase, half-time, aggregation rate constant and 126 the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% stilbenes

6.8 Shows the lag phase, half-time, aggregation rate constant and 129 the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DOPC/DOPS and DMPC/DMPS vesicles doped with 5 mol% stilbenes

S8.1 Analysis of hydrophobicity, net charge, and % charged amino 160 acids of SST-14 using PEPTIDE 2.0 (34)

S8.2 Shows the hydrodynamic size of liposomes doped with various 161 compositions

S8.3 Quantitative analysis of the CD spectra for SST incubated in 161 liposomes

S8.4 Shows the effect of concentration on kagg, half-time and lag 162 phase for SST in NaCl

S8.5 Quantitative analysis of the CD spectra for SP incubated in 162 liposomes after 72-hours

XXVI

S8.6 Shows the hydrodynamic size of saturated lipid vesicles with 163 the presence and absence of stilbenes at 5 mol%

S8.7 Shows the hydrodynamic size of DMPC+Chol vesicles with 163 the presence and absence of stilbenes at 5 mol%

S8.8 Shows the hydrodynamic size of anionic charged vesicles with 164 the presence and absence of stilbenes at 5 mol%

XXVII

Abbreviations

Instruments & Techniques:

CD – Circular Dichroism

SRCD – Synchrotron Circular Dichroism

SAXS – Small-angle X-ray scattering

SAXD – Small-angle X-ray diffraction

WAXS – Wide-angle X-ray scattering

AFM – Atomic force microscopy cTEM – Cryo-transmission electron microscopy

ThT – Thioflavin T

DLS – Dynamic light scattering

FTIR – Fourier-transform infrared spectroscopy

SPR – Surface plasmon resonance

XRD – X-ray diffraction

ACF – Autocorrelation function

DFT – Direct fourier transform

MDS – Molecular dynamic simulations

ESR – Electron spin resonance spectroscopy

NMR – Nuclear magnetic resonance spectroscopy

Lipid Derivatives:

Tc – Phase transition temperature

Lα – Liquid crystalline/fluid/liquid disordered bilayer phase

Lo – liquid-ordered phase

Lβ – Gel/solid ordered bilayer phase

XXVIII

MLV – Multi-lamellar vesicles

LUV – Large unilamellar vesicles

SLB – Supported lipid bilayer

PI – Phosphatidylinositol

PIP – Phosphatidylinositol phosphate

PIP2 – Phosphatidylinositol bisphosphate

PE – Phosphatidylethanolamine

PA – Phosphatidic acid

PG – Phosphatidylglycerol

CL – Cardiolipin

SM – sphingomyelin

GM1 – Gangliosides

Unsaturated Lipids: (Total number of carbon atoms : Total number of double bonds)

DOPC (18:1) – 1,2-dioleoyl-sn-glycero-3-phosphocholine

POPC (16:0-18:1) – 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine

Saturated Lipids:

DLPC (12:0) – 1,2-dilauroyl-sn-glycero-3-phosphocholine

DSPC (18:0) – 1,2-distearoyl-sn-glycero-3-phosphocholine

DPPC (16:0) – 1,2-dipalmitoyl-sn-glycero-3-phosphocholine

DMPC (14:0) – 1,2-dimyristoyl-sn-glycero-3-phosphocholine

Anionic lipids

DOPS (18:1) – 1,2-dioleoyl-sn-glycero-3-phospho-L-serine

DMPS (14:0) – 1,2-dimyristoyl-sn-glycero-3-phospho-L-serine

POPS (16:0-18:1) – 1-palmitoyl-2-oleoyl-glycero-3-phospho-L-serine

XXIX

Sterols:

CHOL – cholesterol

Non-lipid derivatives:

Polyphenols:

RSV – Resveratrol (3,5,4’-trihydroxy-trans-stilbene)

TS – Trans-stilbene

TMS – Trimethoxy trans-stilbene or (3,5,4’-trimethoxy resveratrol)

Buffers:

NaCl – Sodium Chloride

DMSO – Dimethyl sulfoxide

PBS – Phosphate buffered saline

List of amino acids:

Pyr (Z) – Pyroglutamic acid

His (H) – Histidine

Trp (W) – Tryptophan

Ser (S) – Serine

Tyr (Y) – Tyrosine

Asp (D) – Aspartic acid

Leu (L) – Leucine

Arg (R) – Arginine

Pro (P) – Proline

Phe (F) – Phenylalanine

Lys (K) – Lysine

Gln (Q) – Glutamine

XXX

Met (M) – Methionine

Cys (C) – Cysteine

Thr (T) – Threonine

Gly (G) – Glycine

Asn (N) – Asparagine

Ala (A) – Alanine

Non-pathological or functional amyloid-forming peptides (FAFP’s):

Cyclic peptides:

SST-14 (aa) – Somatostatin (14 amino acids)

OT-9 (aa) – Oxytocin acetate (9 amino acids)

AVP-9 (aa) – Arginine Vasopressin acetate (9 amino acids)

Linear peptides:

SP-11 (aa) – Substance P acetate (11 amino acids)

Des-10 (aa) – Deslorelin acetate (10 amino acids)

Pathological or toxic amyloids:

AD – Alzheimer’s disease

Aβ – Amyloid-beta

IAPP – Islet amyloid polypeptide

HIAPP – Human Islet amyloid polypeptide

AS or αs – Alpha-synuclein

Prion proteins – PrP

XXXI

Explanatory notes The purpose of these notes is to briefly define the measures taken throughout the preparation of this thesis. They relate to spelling conventions, units of measurements, the expression of analytical results and referencing of literature sources:

1. Australian or British spelling is adopted in the text, for example, fibrillisation (rather than fibrillization), words ending with -ise (rather than -ize) and few other technical terms.

2. Mostly, results are presented in SI units

3. In-text citations are presented in the Vancouver style and are assigned to each reference for example, (35) shown below and applied throughout (see the page. 136)

34. Campolongo P, Ratano P, Ciotti MT, Florenzano F, Nori SL, Marolda R, et al. Systemic administration of substance P recovers beta amyloid-induced cognitive deficits in rat: involvement of Kv potassium channels. PLoS One. 2013;8(11):e78036.

XXXII

CHAPTER 1

XXXIII

1. The role of the lipid bilayer and molecular additives on fibrillisation of functional amyloid-forming peptides

1.1 Introduction

Figure 1.1. Shows amyloid-forming proteins that form β-sheets are linked to various human diseases (adopted from (1))

Failure of a peptide or protein to adopt, or retain, its intrinsic functional conformational state can lead to aggregation into amyloid fibrils, for instance, the twist and internal lateral packing of β-sheets (red lines) (Figure 1.1), which is linked to several devastating disorders, including type II diabetes mellitus, Alzheimer’s Disease, Parkinson’s Disease, Creutzfeldt-Jakob disease and dialysis-related amyloidosis (DRA) (36-38). These neurotic disorders are commonly referred to as protein misfolding diseases (31). Peptides can assemble into amyloid fibrils through a process of nucleated polymerisation (aggregation); with breakage of fibrils providing new ends for elongation (39). This accumulation of fibrils is predicted to have a huge impact on the structure and function of the cell membrane; hence, amyloid proteins have pathological significances (5). The occurrence of amyloid plaques relates to senile plaques (neuritic) disintegration and cell death (40). The role of amyloid protein fibrils in neurodegenerative disease has been the topic of significant research (5, 41). While the details of the molecular mechanism of membrane penetration remain unconfirmed, several models have been proposed. Oligomers of amyloid peptides have been shown to interact with lipid membranes, resulting in a nonspecific ion leakage (42). Peptides can also transform into annular protofibrils that can form ion channel-like structures in the cell membrane (43). Therefore, it is of great significance to recognise both the exact steps behind fibril formation from the monomer state and the precise molecular mechanism of toxicity in these diseases (44).

1.2 Common structural features and mechanism of amyloid formation 1.2.1 Toxic Amyloids Amyloids were first discovered in 1906 when Alois Alzheimer observed proteinaceous deposits in the brain tissue from patients with Alzheimer’s disease (45, 46). The fibrils appear as thin filaments visible by electron microscopy, with a diameter of 10 to 100’s of nm in length (47). Since then numerous proteins and polypeptides have been found to self-assemble into amyloid fibrils under certain conditions. They are associated not only with dozens of devastating diseases but are integral to many biological processes such as hormone storage, signal transduction, and cell surface adhesion (48). Separate from the context of their parent proteins, synthesised peptide segments can self-assemble into amyloid-like fibrils in vitro as well (49, 50). Fibrils formed by diverse proteins and peptides all share a common cross-β structure, composed of interdigitated β-sheets termed “the zipper-like fibril spine” (51). The ability to form amyloid aggregates is a basic property of the polypeptide chain, as the amyloid aggregate structure is based on β-sheets stabilised by hydrogen bonds which contain the polypeptide backbone (31). The aggregation mechanism of amyloid-forming peptides is a nucleation dependent process consisting of a lag phase followed by an exponential growth phase. During this process, protein monomers self-assemble into oligomers, followed by assembly into more intricate structures (38). Initially, during the lag period or phase, there is a transition of soluble protein species into β-sheet nuclei usually involving a large number of metastable oligomers. However, some proteins such as, insulin is biologically active mainly in its monomeric form (52). Certain nuclei then undergo elongation and proliferation until attaining a measurable aggregate concentration, thereby terminating the lag phase as shown in (Figure 1.2) (53). Fibrillisation is therefore slow until a critical oligomeric size capable of nucleating fibril formation is reached. Once a fibril is formed, it cannot revert to a soluble intermediate oligomeric species (54). A rise in the concentration of fibrils has been shown to reduce the lag period and promote fibrillisation (55-57).

Figure 1.2. A schematic representation of the aggregation process for Toxic amyloids. Aggregation proceeds by aggregation from the monomeric state to a fibrillar structure through oligomeric intermediates (adapted from (2))

The characteristic features of amyloid nanofibrils can be observed using Electron Microscopy (EM) or Atomic Force Microscopy (AFM). Fibrils are typically comprised of protofilaments, each approximately 2-5 nm in diameter (58). These protofilaments spiral to form rope-like fibrils that are normally 7-13 nm wide (58-60) or associate in a lateral manner to form long ribbons that are 2-5 nm thick and up to 30 nm wide (61-63). X-ray fibre diffraction data have shown that in each protofilament the protein or peptide molecules are organised in such a way that polypeptide chains form β-strands (60). Specific dyes like Thioflavin-T (ThT), Thioflavin- S and Congo Red (CR), have the ability to adhere to fibrils (50, 64) resulting in green birefringence under cross-polarised microscopy (2, 31, 65-67). It is reported that dyes, such as ThT, adheres to the surface-chain grooves formed from layers of laminated β-sheets (68).

For toxic amyloid-forming peptides, some of which are shown in (Table 1.1) (31), a large body of research suggests that formation of amyloid fibrils and their mechanism of cytotoxicity are membrane linked processes (33). According to Murphy (69), several studies demonstrate that the lipid bilayer actively changes fibril kinetics. For example, the lipid bilayer facilitated nucleation of Amyloid-beta (Aβ) fibrils in one study, with the effect depending on the chemical composition of the lipids (70). It is generally believed that oligomeric aggregates are the most toxic species with a disruptive effect on membrane permeability (71). However, in some instances, the mature fibrils may also induce cell damage by disordering the membrane lipids (72).

Table 1.1. Shows the diseases resulting from protein misfolding and subsequent aggregation into amyloid nanofibrils (30, 31, 73)

Disease Amyloid-forming peptide/protein

Corpora amylacea Kappa-casein

Diabetes mellitus type II Islet amyloid polyprotein (IAPP/Amylin)

AL amyloidosis Immunoglobin light chains or fragments

Familial Mediterranean fever Fragments of serum amyloid A protein

Senile systemic amyloidosis Wild-type transthyretin

Lysozyme amyloidosis Mutants of lysozyme

Fibrinogen amyloidosis Variants of fibrinogen α-chain

Haemodialysis-related amyloidosis Β2-microglobulin

Amylin also called islet amyloid polypeptide Type II diabetes (IAPP)

Medullary carcinoma of the thyroid Calcitonin

Atrial amyloidosis Atrial natriuretic factor

Pituitary prolactinoma Prolactin

Cataract γ-Crystallins

Pulmonary alveolar proteinosis Lung surfactant protein C

Aortic medial amyloidosis Medin

Atherosclerosis Apolipoprotein AI

1.2.2 Functional Amyloids Increasing numbers of proteins with no connection to protein deposition diseases form fibrillar aggregates under some in vitro conditions (36). Many of these are peptide hormones, which are proposed to be stored as amyloids within endocrine secretory granules (Table 1.2). These aggregates have the morphological, structural and dye-binding properties that allow them to be categorised as amyloid fibrils (74, 75). This basic ability has been exploited by living organisms for specific purposes since some of the organisms have been known to convert one or more of their endogenous proteins into amyloid fibrils that possess functional rather than disease associated properties (36, 76). A list of such “functional” amyloid-forming proteins is given in (Table 1.2) with additional functional amyloids listed in the supporting information (Table S1) (77). Experimental evidence for the formation of amyloids by these proteins is provided by Electron Micrograph (EM), CR, ThT, Circular dichroism (CD), Fourier transform infrared (FTIR), Nuclear magnetic resonance (NMR), X-ray diffraction (XRD) and Atomic force microscopy (AFM) (31, 75, 78-80).

Table 1.2. Shows a list of functional amyloids in prokaryotic and eukaryotic organisms (2, 31, 32)

Organisms Non-disease Biological function Refs related amyloid- forming protein

Escherichia coli Curli formation of (67) biofilms

Streptomyces Chaplins modulation of water (81) coelicolor surface tension (development of aerial structures)

Podospora HET-s regulation of (82-93) anserine heterokaryon

formation

Saccharomyces Sup35p regulation of stop cerevisiae codon read-through

Schizophyllum Hydrophobins fungal coat (94) commune formation

Silk worm Chorion protein protecting the oocyte (95) and developing embryo from environmental hazards

Spider Spidroin protein intermediate in the (96) biosynthesis of spider silk

Mammalians Pmel17 assisting with (97) melanin production

Somatostatin-14 regulates growth (98, 99) hormone secretion

Oxytocin-9 stimulation of (99,

uterine contraction 100) Homo sapiens and production of prostaglandins

Vasopressin-9 regulates water (99, content in blood 101)

Deslorelin-10 -a controls (102, synthetic reproduction by 103) analogue of suppressing the luteinising release of hormone (LH) gonadotrophin-

releasing hormone (GnRH)

Substance P-11 responsible for (104) transmission of pain

Amyloid fibres that appear to have a functional role have been observed in bacteria, invertebrates, and humans (105). For instance, Pmel 17 is a functional amyloid that has a vital role in the biosynthesis of the pigment melanin; this process is generally tightly regulated to avoid toxicity (32). However, the unregulated release of functional amyloid seeds could propagate pathological amyloidogenesis (97). It has been shown that injection of functional amyloid fibres, for example, curli (Escherichia coli) and Sup35p (Saccharomyces cerevisiae), into a chronic inflammatory amyloid protein amyloid A, proliferates amyloid formation and development of disease in a reactive amyloidogenesis mouse model (97, 106). This absence of cytotoxicity in functional amyloids is thought to be potentially related to their ability to reverse fibrillisation (32, 107-109). However, the broader principles of regulation of functional amyloidogenesis, and the possible relationship between functional and toxic amyloid formation has not been well characterised to date.

The kinetics and mechanism of amyloid formation can be influenced by interactions with biomolecules and biosurfaces, particularly lipid membranes. These interactions can either promote or inhibit fibrillisation. While the vast majority of the published literature in this area, deals with the effect on fibrillisation of toxic amyloids, some studies have shown that molecules, such as glycosaminoglycans, can impact fibrillisation for functional amyloid- forming peptides (110). The biological aggregation modifiers reviewed in the following sections have been divided into molecular (Section 1.3) and lipid membrane-based (Section 1.4).

1.3 Molecular aggregation modifiers 1.3.1 Impact of Metal Ions Growing evidence supports the binding of metal ions to amyloid species found in the brain of patients suffering from various disease-related amyloid conditions. High concentrations of metal ions such as Cu2+, Zn2+ and Fe3+ have been reported to be bound to Aβ peptides (111, 112). Binding of transition metal ions, including Cu2+ and Zn2+, have been reported to induce Aβ fibrillisation in parts of the neuron which release high concentrations of metal ions such as the synaptic cleft (113, 114). More specifically in the hippocampus, a region of the brain involved in memory and also one of the first regions to be affected by early onset stages of AD (115). Evidence shows deposits of zinc in pancreatic β-cells found in patients with IAPP and iron deposits were reported in the Lewy bodies formed in Parkinson’s disease (116, 117). The nature of the metal ions, Cu(II) was reported to significantly reduce the lag phase and promote the rate of fibril formation for α-synuclein (118). However, Cu(II) was shown to inhibit IAPP fibrillisation, instead of promoting the formation of toxic oligomers via an alternate aggregation pathway (119). For Aβ peptides, Cu (II) is reported to increase the rate of fibril formation by decreasing the lag phase (120, 121). Aβ protofibrils were also detected in the presence of large concentrations of iron, while manganese did not show any binding affinity to Aβ (122, 123). It is reported that Zn2+ can prevent fibrillisation by stabilising the hexamer (124). Thus, transient metal binding to amyloidogenic peptides and fast exchange rates of the metal ions within intra and intermolecular peptide binding sites are currently a growing field in biometal-amyloid interactions (125).

1.3.2 Impact of Heparin/Glycosaminoglycans (GAGs) Polyelectrolytes like GAGs belong to a family of sulfated and intricate polysaccharide molecules that repeat to form heparin (126). Heparin is an anticoagulant known to boost amyloid fibrillisation in vitro (127), including for the toxic Aβ peptides (128-130), Tau protein (131, 132), α-synuclein (133), gelsolin (134, 135), IAPP (136). Heparin also enhances fibrillisation for a range of neuropeptide hormones which form functional amyloids (77). From a kinetic standpoint, heparin significantly reduces the lag phase and increases the rate of the growth phase for many amyloidogenic systems studied. Heparin binds to α-synuclein and not only accelerates the speed of fibrillisation but also increases the fibril yield (133). The mechanism by which heparin acts is primarily through electrostatic interactions, thereby concentrating the oligomers and increasing the degree of order within the aggregates which favours the fibrillisation process (137). It has been reported that a considerable number of

neuropeptide hormones require heparin to induce functional fibrillogenesis in vitro (77). The strong effect of heparin dominates the intermolecular interactions and the morphology of the end structures.

The binding affinity of GAGs to amyloid fibrils occurs principally through electrostatic interactions involving the anionic polyelectrolyte charge and cationic side chain residues of the aggregating protein (138). In addition to biochemical molecules, there is also growing interested in lipid-protein interaction within the membrane that is known to induce amyloid aggregation.

1.3.3 Effects of Resveratrol (RSV) 3,5,4’-trihydroxy-trans-stilbene (RSV), is a polyphenolic substance predominantly found in grapes, red wine, berries and other plant types (139). RSV has been shown to exhibit several biological activities, for instance, anti-inflammatory, antioxidants, anticarcinogenic properties etc. (3, 140). Current evidence suggests that RSV plays a significant part in the prevention and treatment of AD (141-143). RSV exists in two isomers and contains two aromatic rings (Figure 1.3), the trans-RSV, which is most commonly studied than cis-RSV; several studies have shown that trans form is more stable and effective than the inactive cis form (144, 145).

(trans-RSV) (cis-RSV)

Figure 1.3. The geometric isomers of resveratrol (adapted from (3))

The results obtained from research carried out Feng et al. 2009, showed that resveratrol inhibits Aβ42 cytotoxicity and fibril formation, but does not prevent Aβ42 oligomer formation (143). Surface plasmon resonance (SPR) and proton nuclear magnetic resonance (1H NMR) experiments by Fang Ge et al. 2012, showed that RSV could bind to both monomer or fibrillar

Aβ(1-42) (146). In vivo effects of RSV in PC12 cells have shown to protect cells from Aβ25-35 induced toxicity (139). RSV has relatively poor solubility in biological fluids hence lipid carriers are used for drug delivery. In this research, liposomes are chosen due to their biocompatibility and diverse choice of physio-chemical properties (28). The interaction between RSV and functional amyloids have not been explored, until now (Chapter 6).

1.4 Membrane-based aggregation modifiers Amyloid-forming peptides can interact with the lipid bilayer environment of the cell membrane, with a knock-on effect on fibrillisation. Cell membranes consist of a fundamental lipid bilayer structure, a thin membrane consisting of two lipid monolayers stacked back to back. The fluid mosaic model (Figure 1.4) of the lipid membrane, according to Singer and Nicholson, is a two-dimensional fluid in which membrane proteins and other molecules like phospholipids and cholesterol are embedded (147). These sheet-like structures play a crucial role in inhibiting proteins, ions and other molecules from dispersing into surrounding territories. Apart from being the gate of the cells, lipid bilayers are regular binding sites, for several amphipathic molecules, proteins, and peptides and control numerous indispensable biological functions, including cell fusion (148) and ion transport (149).

Figure 1.4. Shows a typical biological membrane (adapted from (4))

Lipid membranes are especially vital to amyloid formation under physiological conditions, as these are abundant in living organisms. Model membrane surfaces of several types have been shown to accelerate amyloid fibrilization and summary included in (Table 1.3). It is believed that the interaction is between the membrane and peptide monomers, or smaller oligomers, rather than mature fibrils. For example, experimental evidence suggests that the monomeric form of Human islet amyloid polypeptide (hIAPP) inserts into phospholipid monolayers, whereas fibrillar hIAPP does not insert (72, 150). The membrane may impact the aggregation mechanism is several ways. For example, local increases in peptide concentration driven by hydrophobic and/or electrostatic interactions of the peptide with the bilayer may increase the rate of fibrillisation. Changes to peptide conformation upon binding to the bilayer may also promote the peptide-peptide contacts necessary for fibrillisation to occur.

Table 1.3. Includes a summary of model membranes involved in fibrillisation of amyloidogenic proteins and peptides (33) Disease Protein or Membrane Summary peptide systems*

Parkinson’s α-synuclein PA/PC, PG/PC, α-helical conformation in bilayer and PS/PC, PG/PE disruption of bilayer due to the formation vesicles of fibrils (151-155)

Alzheimer’s Aβ peptide PC/PG vesicles, α-helical conformation reported in bilayer and Tau PA, PS, PI, PIP, and accelerated the formation of

PIP2, CL, PC, PE, aggregates in ganglioside vesicles and SM, Chol, DG, acidic phospholipids. No fibril formation gangliosides, reported in neutral lipids (156-160) sonicated lipid suspensions

Type II diabetes IAPP PG/PC vesicles, IAPP fibrillisation is observed in a mixed liposomes from bilayer such as PG/PC and reported pancreas lipids disruption to liposome structure (161, and PC, PS/PC 162) vesicles, rat insulinoma tumour cells

Spongiform Prion protein PG, PC, Formation of fibrils reported, however, encephalopathies PC/Chol/SM the disruptive effect is observed mostly in vesicles POPG membrane (163)

Familial Transthyretin PC/PS, PC/PG Anionic bilayers induced rapid formation polyneuropathy, vesicles of fibrils (164) systemic amyloidosis

Thyroid Calcitonin PC, PC/Chol, Membranes consisting of cholesterol and carcinoma PC/PS, ganglioside induced aggregation (165)

PC/ganglioside vesicles

As for molecular aggregation modifiers, the vast majority of the published literature on the effect of lipid membranes on fibrillisation deals with toxic amyloid-forming peptides, including the Aβ peptide, α-synuclein and islet amyloid precursor protein (IAPP). Lipid membranes generally appear to promote, rather than inhibit, fibril growth via interactions with the bilayer. Shen et al. have shown that Aβ misfolds in the presence of surfaces of varying hydrophobicity, and fibrillises at much lower concentrations than those required in bulk. The surface is proposed to permit binding, diffusion and interaction of monomers (166).

Interactions with the membrane can also affect protein conformation. In the case of α- synuclein, interaction with lipids results in an increase in α-helical content and subsequent aggregation. The helical conformation is shown to be more stable in the presence of membranes (167). Similarly, binding of human islet amyloid polypeptide (hIAPP) to membranes containing negatively charged lipids led to the conversion to α-helical form and eventually to β-sheets (168). Current published data show that interactions of hIAPP with membranes results in the formation of the monomeric helical form within the headgroup region, with a subsequent transition to β-aggregates at higher protein concentrations (169).

1.4.1 Ganglioside clusters, Cholesterol and Sphingomyelin The ganglioside (GM1) and cholesterol content, as well as the overall fluidity of the membrane, have been shown to anchor amyloidogenic proteins to membranes, inducing the adoption of β sheet formation and affecting fibrillisation rates (5). Studies have also reported, Aβ amyloidogenesis process is accelerated in the presence of sphingomyelin (SM), ganglioside and cholesterol (Figure 1.5) (170-175).

Figure 1.5. Shows the model of Aβ-fibril formation on ganglioside clusters in lipid rafts (adopted from (5))

The fibrillisation depends on the concentration of cholesterol within the bilayer. Cholesterol has been reported to both induce (176) and constrain (177) the interaction between (Aβ 1-40) and model membranes. This enhanced binding of Aβ and aggressive peptide aggregation with cholesterol has been reported numerous times (5, 178-181). In fact, it has been proposed by Matsuzaki that cholesterol forms sites on the cell membrane that are chosen for Aβ after separating from the amyloid precursor protein (APP) and initiating its fibril formation (182).Cholesterol can also affect acyl chain order, depending on the concentration, which can impact the peptide binding and membrane-mediated fibrillisation process. Evidence from published data also reports that liposomes containing cholesterol accelerate fibril (β-sheet) formation of Aβ (181, 183, 184). A published report has suggested that the toxicity of oligomers can depend on its hydrophobic property (Figure 1.6) (6).

Figure 1.6. Shows the two types of oligomers and its ability to interact with the membrane (adopted from (6))

1.4.2 Charged lipids Anionic lipids seem to be particularly significant in a membrane-assisted conformational change, and electrostatic interactions can play a part in protein-lipid interactions. Previous studies have concluded that hIAPP and α-synuclein require anionic lipids such as phosphatidylserine (PS) for protein binding (152, 161, 167, 168, 185). Studies by Zhao et al. indicated that a range of different proteins formed fibrils in the presence of charged liposomes containing PS (164).

Electrostatic interactions between most cationic amyloid-forming peptides and anionic lipid headgroups is a frequent observation, and this can also facilitate the binding of the peptide to the membrane (179, 186-188). By means of neutron and X-ray scattering techniques, Lee and co-workers have been able to reveal that Aβ40 inserts more readily into anionic DPPG bilayers, compared to neutral DPPC lipid bilayers (5). Anionic liposomes containing PG headgroup have also been shown to promote a β-sheet conformation in Aβ40 in addition to peptide oligomerisation (189); similar observations have been reported with IAPP (185) and α- synuclein (154, 190, 191). Increasing the anionic content of the headgroup in mixed vesicles have shown to accelerate binding of α-synuclein monomers (192) and oligomers (193). It is proposed that interactions between β-amyloid and anionic membranes result in monomeric peptide and that further incubation leads to oligomer formation in the membrane (194, 195). The addition of oligomers to negatively charged lipid vesicles results in fast content release in a bulk permeabilisation assay; in contrast, a much slower loss of content was observed after adding oligomers to vesicles that mimic mitochondrial membranes (196). Even following long incubation, oligomers are unable to induce leakage in artificial plasma membranes with CD experiments confirming minimal changes to the secondary structure (197). Addition of 0.1M NaCl to screen electrostatic interactions demonstrated that the interaction was mainly electrostatic, rather than hydrophobic (198).

Studies show that binding of β-amyloid (1-40) to a supported anionic lipid membrane resulted in a tight binding followed by oligomer formation in the membrane thereby causing a significant decrease in membrane fluidity as shown in Figure 1.7 below (7, 194). Recently, a functional amyloid TasA, which is a major matrix protein found in biofilms of the soil bacterium Bacillus subtilis, demonstrated higher ThT activity with negatively charged lipids compared to neutral membranes. A diagrammatic illustration of its interaction with the membrane is shown in (Figure 1.8) (8).

IAPP

Planar bilayer

Figure 1.7. Schematic representation of interactions of amyloidogenic IAPP (black lines) with SPBs (Supported Planar bilayers), depicting significant reduction in the fluidity of membrane (adapted from (7))

Figure 1.8. Schematic illustration for the mechanism of TasA interacting with the membrane (adopted from (8))

1.4.3 Neutral lipid Phosphatidylcholine (PC) and sphingomyelin (SM) are zwitterionic phospholipids in mammalian cells. Most published data suggest that Aβ (1-40) monomer do no interact with PC (175, 199-202) or SM (158, 175). Although amyloid interactions have been reported with neutral 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) (Figure 1.9) (9) with time in a similar way on the surface of negatively charged 1,2-dioleoyl-sn-glycero-3-phosphoglycerol (DOPG) membrane (203).

Figure 1.9. Schematic of Aβ interacting with lipids and phase separation in the bilayer shows nonhomogeneity (adopted from (9))

1.4.4 Physicochemical properties of the lipid bilayer As summarised in Table 1.3, while fibrillisation in vitro has been reported mainly for negatively charged membranes, enhancement of fibrillisation has also been observed for lipid mixtures composed of phosphatidylcholine (PC) and phosphatidylethanolamine (PE) with cholesterol or gangliosides, and ‘rafts’ consisting of sphingomyelin and cholesterol. Consequently, it is difficult to ascribe the promotion of protein fibrillisation to a specific class of lipids. In addition, physicochemical properties of the lipid bilayer, such as bilayer thickness, fluidity and surface charge, are likely to play a critical role in influencing the mode and extent of the membrane binding of these proteins, in addition to the adoption of an aggregation-prone protein conformation. Lipid-protein interactions may also differ significantly, contingent on the structural characteristic of the proteins or peptides in question and these interactions may either promote or inhibit fibril growth. Any change to physicochemical properties upon peptide incorporation, such as membrane thickness and lipidic phase, may be characterised using a range of techniques including X-ray diffraction and neutron reflectivity experiments (5, 204).

The membrane fluidity, which is related to the level of order in the acyl chains, is one of the most prevalent features of a membrane and it is vital for its biological viability (205). The bilayer thickness is also an important parameter; for a certain lipid head group, the thickness of a liquid-disordered bilayer is determined both by the number of carbons, and the position and number of double bonds in the acyl chains (206, 207). The fluidity depends on both the temperature and the lipid composition within the bilayer, as is readily confirmed in studies of artificial bilayers (208). A model bilayer made from a single type of phospholipid varies from a fluid state to a two-dimensional gel or crystalline state at the melting point, Tm (209). This phase transition temperature is typically lower for shorter chains, or chains containing one or more unsaturated double bonds (210). Shorter chains are associated with weaker hydrophobic

and Van der Waals interactions between the hydrocarbon chains, while the presence of unsaturation creates kinks in the hydrocarbon chains, reducing packing efficiency and promoting the retention of membrane fluidity at lower temperatures (211). Yeasts, Bacteria and other organisms can maintain the fluidity of their membrane by adjusting the fatty acid composition of their membrane lipids with that of their environment (212).

Enhancement of Aβ fibrillation has been demonstrated in fluid phase membranes due to increased mobility, whereas mobility is restricted in a gel phase with concomitant retardation in fibrillation (Figure 1.10) (10). Whereas α-synuclein (AS) bound to membranes in gel phase (DPPC) and 1,2-dipalmitoyl-sn-glycero-3-phosphoglycerol (DPPG), indicated deeper insertion of the protein into the bilayer due to the electrostatic interaction between positively charged N-terminal of α-synuclein and negatively charged surface of the lipid (213). The deepest immersion of C-terminus was observed in the presence of rigid neutral DPPC probably due to the absence of electrostatic repulsion (214). To allow measurement of a higher order of Bragg peaks, the membranes were studied in their gel states using high-resolution X-ray diffraction (215).

Figure 1.10. Illustration of Aβ interacting with gel phase and fluid phase (adapted from (10))

The physicochemical properties of the bilayer can be impacted by the lipid composition. For example, the eukaryotic plasma membrane contains significant levels of cholesterol, up to one molecule for every phospholipid molecule (211). The cholesterol molecule is reported to orient with their hydroxyl groups close to the polar head groups of the phospholipid molecules (216). In this position, their rigid, platelike steroid rings interact with and partially restrain the hydrocarbon chains nearby to the polar head groups, which can decrease membrane fluidity, increase local bilayer rigidity, and decrease the permeability of the bilayer to small water- soluble molecules (217).

In addition, anionic phospholipids in cellular membranes can regulate electrostatic potential in the membrane, and drive interaction between peripheral proteins (218). It has been reported that the ordering of PC chains and stiffness of the bilayer increases with an increase in negative PS content (219). Several amyloid studies have been investigated using negatively charged lipid molecules since it is known to induce severe aggregation or fibril formation via electrostatic interactions (220).

1.5 Effect of amyloid-forming peptides on the structure of the lipid bilayer It is generally accepted that the interaction between toxic amyloid aggregates and membrane can induce disruption of the barrier function of the membrane (164). However, the interaction between non-toxic or functional amyloid-forming peptides and lipid membranes are in fact not well explored (5).

It has been reported in many published papers that peptides including antimicrobial peptides (gramicidin) and α-helical peptides (alamethicin) can disturb the bilayer architecture by inducing a transition to a non-lamellar phase, e.g. the hexagonal (HII) phase. Such a phase transition results from fundamental changes to bilayer curvature induced by binding of the peptide (221). It has been suggested that Aβ can also drive changes in membrane curvature which cause a lamellar to a non-lamellar phase transition and that the specific transition can differ depending upon the relative concentration of lipids in the bilayer (221, 222). Studies on liposomes exposed to IAPP shows that the membrane-peptide interaction causes pores in the membrane, as the oligomeric state of the peptide expands and it is known to be stable for days (223). This result agrees with the hypothesis that antimicrobial peptides and amyloids share a mutual pore-forming ability.

Experimental findings suggest that there can be more than one mechanism involved for amyloid peptides interacting with the membrane, thereby leading to events concomitant with amyloid diseases.

Figure 1.11. Illustration of three possible mechanism of Aβ encouraged membrane damage (adopted from (11))

There are three proposed models for toxic peptide interaction with membranes (Figure 1.11). 1. The carpet model, where carpeting of the peptide on one side of the membrane surface, causing asymmetric pressure between two sides and leakage of small molecules. 2. Formation of stable pores and ion channels (by forming Ca2+ channels in lipid bilayers upon incorporating Aβ40). 3. Detergent-like effects, where membrane association of the amyloid-forming peptides in the form of micelle-like structures (183).

Studies carried out by Atomic force microscopy (AFM) on supported bilayer lipid membranes subjected to amyloidogenic peptide (HypF), either in its monomeric or in the oligomeric aggregation form, have revealed that the amyloid-like peptide can disrupt the bilayer causing membrane thinning (224). Alternatively, amyloid fibrils, proliferating at membranes, may remove lipids by interacting with an outer leaflet (225). Using experimental surface-sensitive techniques, Reynolds et al. and co-workers were able to conclude that α-synuclein monomers massively disrupted the membrane by removing lipids followed by membrane thinning and subsequently, the membrane integrity is lost (37). In summary, published data shows that both oligomers and fibrils can disturb the order of the membrane and can give rise to cell disruption.

Sparr et al. indicated that hIAPP interacts with liposomes during amyloid formation, inducing uptake of lipids and resulting in the loss of the barrier properties of the membrane (162). hIAPP fibrils were shown to grow in an ordered manner on the surface of distorted lipid vesicles. Relating the similarity of the kinetic profiles of dye leakage and thioflavin T (ThT)

fluorescence, the authors suggested that fibril growth is associated with membrane damage through a process that also implicates alterations in membrane curvature and lipid packing (72). A similar behaviour was described by Milanesi et al. in an experiment performed exposing liposomes to β2-microglobulin fibrils formed in vitro. The result from this experiment shows that fibril tips extract lipids from the membrane at the point of contact, causing distortions in the liposome shape (Figure 1.12) (12).

Figure 1.12. (A-C) Shows the clustered liposomes distorted by short fibrils. (D) A 3D model of a distorted liposome (blue), surrounding fibrils visualised by cryo-electron tomography (cET) (adopted from (12))

1.6 Model Lipid Bilayers Determining the effect of the cell membrane on the amyloid formation and determining the converse effect of amyloid formation on the structure of the cell membrane, can be difficult due to the inherent complexity of cell membranes, and difficulties in working with biological cell-lines. Therefore, model membranes, where membrane parameters may be more easily controlled, are typically used to determine the effect of specific membrane properties on amyloid fibrillogenesis. Model systems, which can be based on lipid monolayers or bilayers, include the bulk lipid lamellar phase, liposomes (or multi-lamellar vesicles) and supported lipid bilayers. All these model membranes have shown to impact fibrillogenesis of toxic amyloid- forming peptides, although their effect on functional amyloid-forming peptides remains essentially uncharacterised.

A monolayer is one of the simplest ways to mimic the outer leaflet of a biological membrane. The hydrophobic tails of the lipid chains are in contact with air (interface), and the polar lipid headgroups are directed towards the aqueous phase (5). Several studies have investigated fibrillisation in the presence of lipid monolayers (5, 150, 224, 226). X-ray reflectivity study

has shown fibril nucleation in IAPP only occurs in the presence of DOPC/DOPG monolayers, but not with pure neutral DOPC monolayers (29).

Supported lipid bilayers (SLBs) are made via Langmuir transfer or fusion of vesicle onto a suitable surface (227, 228). Literature has shown progressive accumulation of oligomers and fibrils of Aβ1-42 using a SLB (194, 229). Due to the positive charge of the peptide and fluid phase of the lipid bilayer is known to induce aggregation (203).

1.6.1 Bulk Lamellar Phase The lipids of the cell membrane can self-assemble into numerous structures (Figure 1.13) of which the bulk lamellar phase is most applicable to biomembranes. The lamellar phase consists of a 1-D stack of lipid bilayers, separated by water layers (Figure 1.14). However, some cell membrane lipids do not adopt the lamellar phase but instead, form more curved non-lamellar phases. Such phases, which include the hexagonal and bicontinuous cubic phases, might relate to temporary events in biomembrane, such as fusion, fission and pore formation (230, 231). However, in this thesis, we focus on the lamellar phase and liposome systems.

Types of Lipids Lipid Phases Lipid Shapes Diagram

Figure 1.13. Shows the polymorphic phases and corresponding dynamic molecular shapes of a component of phospholipids (adopted from (13-16))

Figure 1.14: Shows a typical lamellar phase or phospholipid bilayer (lipids-grey and water-red) (adopted from (17))

1.6.2 Liposomes Liposomes are vesicles with an aqueous core entrapped by one or more than one lipid bilayer (Figure 1.15). Liposome properties vary significantly with lipid composition, size, surface charge and method of preparation (sonication vs extrusion) (232). Moreover, the choice of the bilayer components controls the ‘rigidity’ or ‘fluidity’ and the charge of the bilayer. For example, unsaturated phosphatidylcholine molecules from natural sources (soybean and egg) give much more permeable and disordered bilayers, whereas saturated phospholipids with long acyl chains (DPPC or DSPC) form a rather stiff and impermeable bilayer structure (233-235). Mostly, liposomes are spherical with particle size ranging from 30 nm to several micrometres and can be classified as: (1) multilamellar vesicles (MLV) and (2) unilamellar vesicles. Unilamellar vesicles can be further divided into two categories: (1) large unilamellar vesicles (LUV) and small unilamellar vesicles (SUV) (236). In unilamellar liposomes, the vesicle consists of a single phospholipid bilayer sphere enclosing the aqueous solution. In multilamellar liposomes, vesicles are arranged in an onion ring-like fashion (232). These vesicles can transport hydrophilic, hydrophobic and lipophilic compounds to their designated target sites without causing any decomposition of the entrapped molecules (237). Such characteristics make them a perfect delivery vehicle in a broad range of pharmaceutical applications (238) and a model membrane in several amyloid research studies (5, 239, 240).

Figure 1.15. A structure of a liposome (adapted from (18))

1.7 Effect of lipid bilayer composition on the fibrillisation of functional amyloid- forming peptides In this thesis, I investigate the effect of model lipid membranes, mainly focused on liposomes, on the fibrillisation of functional amyloid-forming peptides. In contrast to the extremely large body of literature investigating the effect of model lipid membranes on amyloid aggregation in disease-related peptides (5, 241, 242), almost no studies to date have investigated the effect of the lipid bilayer on the fibrillisation of functional amyloid-forming peptides.

Five different functional amyloid-forming peptides (Somatostatin-14, Oxytocin-9, Vasopressin-9, Deslorelin-9 and Substance P-11) has been studied. Specifically, the effect of lipid chain length and unsaturation, charged lipids, cholesterol and addition of polyphenol additives such as resveratrol, trans-methoxy stilbene and trans-stilbene which are known to inhibit or influence amyloid aggregation for toxic amyloids have been investigated.

In this thesis, model membranes used to determine the effect of the lipid bilayer on the kinetics and mechanism of fibrillisation were typically multi-lamellar vesicles or liposomes. The bulk lamellar phase was also used in SAXS studies to determine any reciprocal effects on the structure of the lipid bilayer.

Five different lipid molecules, which either form a fluid or gel phase at room temperature, were used to form the lipid bilayer. These were: 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2-dipalmitoyl-sn-glycero- 3-phosphocholine (DPPC), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2- dioleoyl-sn-glycero-3-phosphocholine (DOPC) as shown in (Figure 1.16). These five lipids have the same phosphatidylcholine headgroup but differ in the architecture of the hydrophobic

o alkyl tails. The phase transition temperature (Tc) of the lipids are: DLPC (-57 C), DOPC (-47 oC), DMPC (23 oC), DPPC (41 oC) and DSPC (55 oC) indicating that DLPC, DMPC and DOPC are fluid at room temperature (although we note that DMPC is close to the del-fluid transition) (10). This is in agreement with Total Internal Reflection Fluorescence Microscopy (TIRFM) studies performed on the lipids, show that lipid molecules within DLPC, DOPC and DMPC spreads rapidly (243), and with molecular dynamics (MD) studies (209, 244). DMPC is best- characterised model lipid due to its phase transition behaviour between fluid and gel, contributing higher structural resolution (70). A highly fluid membrane would be made up of well-spaced lipids with a greater degree of freedom in acyl tail motion, while a gel membrane would be more rigid (such as DPPC and DSPC) (70) (Figure 1.17).

(DOPC C18:1)

(DLPC C12:0)

(DMPC C14:0)

(DPPC C16:0)

(DSPC C18:0)

Figure 1.16. Shows the structure of five different lipid molecules (adapted from (10))

Figure 1.17. Schematic representation of fluid (left) and gel (right) phase bilayer (adapted from (10))

To investigate the interaction between membrane surface charge and the fibrillisation of functional amyloids, we investigated the model membranes doped with anionic lipids including 1,2-dioleoyl-sn-glycero-3-phosphoserine (DOPS) and 1,2-dimyristoyl-sn-glycero-3- phosphoserine (DMPS), since ‘PS’ is an excellent substitute for naturally occurring brain PS. The effect of membrane fluidity was investigated using cholesterol (Chol) (Figure 1.18) as a membrane additive. Evidence suggests that cholesterol in the membrane increases stiffness (5) and also induce liquid-ordered phase domain (Lo), which creates the likelihood of favourable interface of ligands with distinct phase (245, 246).

(DOPS C18:1)

(DMPS C14:0)

(Cholesterol)

Figure 1.18. Shows the structure of anionic lipids and cholesterol (adapted from (5, 19))

The functional amyloid-forming peptides studies include the peptide hormone Somatostatin (SST-14), which is involved in regulating the central nervous system (CNS) and endocrine system (ES), has a cyclic structure resulting from a disulphide bridge between Cysteine (Cys) (Cys3-Cys14) (247). It has been reported to form amyloid fibrils in vitro (77, 248, 249). Recently, Dharmadana et al. have demonstrated that aggregation enhancers, such as heparin, commonly used to increase the aggregation rate result for amyloid-forming peptides result in clear morphological differences in the resulting fibrils for SST (98). Two additional cyclic functional amyloid-forming peptides are investigated in Chapter 3, Oxytocin (OT) and Arginine Vasopressin (AVP). Both peptides are neurohypophysial hormones, which are involved in sexual behaviours and social recognition, and are implicated in neuropsychiatric disorders including intellectual disability (ID), autism spectrum disorders (ASDs) and bipolar disorder (BD) (250, 251). Studies have concluded that neurohypophysial hormones contain sequences which are likely to form pathological aggregates due to mutations. Such aggregation of peptide hormones into secretory granules corresponds to the formation of functional amyloids (77). Immunohistochemical studies conducted on mouse pituitary tissue has confirmed that oxytocin and vasopressin, are stored as amyloid-like aggregates in their secretory granules, located in the posterior lobe of the pituitary (99, 100).

In order to compare the effects of peptide structure (cyclic vs linear) in inducing amyloid aggregation, we additional studied the following linear peptides in Chapter 5, Substance P (SP) (252) and Deslorelin (DES) (253). SP belongs to a neuropeptide group called tachykinins, and it is made up of 11 amino acids (254, 255). It has a cationic charge of +3 at neutral pH (256), and Its activity spans across the central and peripheral nervous system (257), involved in different functions such as pain transmission (254), regulation of blood flow (258) and gastrointestinal motility (259). Evidence, reports reduced levels of SP, is capable of antiapoptotic and neuroprotective properties, have been found in the brain and spinal fluid of Alzheimer’s disease (AD) patients (35). Literature evidence suggests that SP plays an important role in preventing brain ageing-related memory decline (255, 260-262). Recently, SP is also shown to form nanotubes (Durga et al. 2018) through Cryo-Transmission Electron Microscope (Cryo-TEM), and Fourier-Transform Infrared spectroscopy (FTIR) and this study would illustrate structural characteristics of peptide self-assembly process. The secondary structure of SP has been reported from unfolded to aggregates or β-sheets (254) and X-ray, and neutron diffraction studies have also revealed β-sheet for aggregated SP (263). Deslorelin (Des) is nonapeptide synthetic analogue of gonadotropin-releasing hormone (GnRH). Des regulates

the follicle stimulating hormone (FSH) and luteinising hormone (LH) (264, 265). However, no evidence of β-sheets or aggregates reported to date.

Finally, in Chapter 6, the effects of RSV doped in liposomes of various lipid compositions on the aggregation profile of SST-14 was investigated. RSV is a natural product found in red wine, which has been shown to have an inhibitory effect on the aggregation of toxic amyloid species reported in the literature (266, 267). Also, this effect was further compared with two additional stilbenes analogues of RSV, TMS and TS however, no literature is available reporting the effects of these analogues of stilbenes with amyloids, until now.

CHAPTER 2

2. Materials and Methods This chapter provides an overview of the experimental protocols used for the research presented in this thesis, and all experiments were performed at room temperature 25oC. All experiments are limited on a certain amount of physical restrictions be it instrumentation failure, human error, time and current state of knowledge.

2.1 Materials 2.1.1 Lipids 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn-glycero-3-phospho-L- serine (DOPS-sodium salt), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2- distearoyl-sn-glycero-3-phosphocholine (DSPC), 1,2-dipalmitoyl-sn-glycero-3- phosphocholine (DPPC), 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC) and 1,2- dimyristoyl-sn-glycero-3-phospho-L-serine (DMPS-sodium salt) of >99% purity were purchased from Avanti polar lipids, INC, United States of America (USA). Cholesterol ≥99% purity was purchased from Sigma-Aldrich (Australia).

2.1.2 Non-lipid materials Thioflavin T (ThT), Sodium Chloride (NaCl) and Phosphate buffered saline (PBS), and Dimethylsulfoxide (DMSO) were purchased from Sigma-Aldrich (Australia). Resveratrol (RSV) 3,4,5-Trihydroxy-trans-stilbene of >99%, HPLC grade was purchased from AK Scientific, INC (USA), 3,4,5-Trimethoxy-trans-stilbene (TMS) of >99% purity was purchased from SAPPHIRE Bioscience Pty. Ltd. (Australia), and trans-stilbene (TS) was synthesised and supplied by A/Prof. Helmut Hugel, School of Science, RMIT. The partition coefficient (LogP) and topological surface area (tPSA) for stilbenes were calculated using Molinspirataion Cheminformatics (MC) software (268) (included in Table 6.1, Chapter 6).

2.1.3 Peptides Somatostatin (SST, AGCKNFFWKTFTSC, Cys3-Cys14 and molecular weight (MW) 1637.9 1 6 Da) and Vasopressin acetate (AVP, CYFQNCPRG-NH2, Cys -Cys and MW 1084.2 Da), purity of 98.0%, were purchased from polypeptide group the USA. Substance P acetate (SP,

RPKPQQFFGLM-NH2 and MW 1347.6 Da), purity of 97.39%, was purchased from CHEM PEP INC USA. Oxytocin acetate (OT, CYIQNCPLG-NH2, Cys1-Cys6 and MW 1007.2 Da), purity of 98.15% and Deslorelin acetate (Des, XHWSYdWLRP-NHEt, MW 1282.5 Da), purity of 99.14%, were purchased from MIMOTOPES, Melbourne, Australia. The % hydrophobicity,

% charged amino acid residue, and a net charge for all functional amyloid-forming peptides was analysed using peptide 2.0 (34), (table included in appendix S8.1).

2.2 Methods 2.2.1 Liposome Preparation Multilamellar vesicles (MLV) with compositions (100 mol% DOPC, 100 mol% DLPC, 100 mol% DMPC, 100 mol% DPPC, 100 mol% DSPC, 70 mol% DOPC/30 mol % Chol, 70 mol% DMPC/30 mol% Chol, 65 mol% DOPC/35 mol% DOPS, 65 mol% DMPC/35 mol% DMPS were combined in appropriate ratios and prepared by dissolving lipids in chloroform. The solution was gently dried under a stream of N2, and the film was then thoroughly dried under high vacuum overnight to remove any traces of residual solvent. The obtained dried film rehydrated with buffer (150 mM NaCl), to yield a final concentration of 20 mM and dispersed by sonication for 30 seconds using a 130 W (Vibra-Cell, VCX 130, Sonics, Newton, CT, USA) with 35% amplitude, pulse for 5 secs on/off. The lipid suspension was extruded 10 times through a 0.1 µm or 100 nm polycarbonate Nucleopore membrane filter (Whatman) fixed on a mini extruder (Avanti Polar Lipids, Inc.) to obtain MLV with an average diameter of 100 nm. Polydispersity Index (PDI) and particle size measured using Dynamic Light Scattering (DLS) technique, with instrument Zetasizer Nano ZS in PC2 lab at the Micro Nano Research Facility (MNRF) at RMIT (City campus). The hydrodynamic size data was analysed using zeta sizer software, and file exported .csv and plotted intensity vs size (nm) (include in appendix Figure S8.1 and Table S8.2). DLS measurements are very sensitive to temperature and solvent viscosity. Therefore, the temperature must be kept constant and solvent viscosity must be known for a reliable DLS experiment (270).

2.2.2 Thioflavin T (ThT) Binding Assay Peptides were prepared in 25 µM ThT and liposomes doped with various lipid compositions (specified in 2.2.1). Therefore the peptide aggregation was measured in a series of concentration as described in each Chapter. The initial fluorescence (t = 0) was immediately recorded after sample preparation and at different time points throughout 168 hours or 1-week. The fluorescence intensity made in 6 replicates using PerkinElmer plate reader (excitation wavelength at 440 nm, emission wavelength at 482 nm respectively). The fluorescence intensity recorded from 200 µL samples contained into 96 microwell plates (Greiner Bio-one). The data is background subtracted and expressed as mean ± Standard Deviation (SD). To calculate aggregation rate constant kagg, the relative ThT fluorescence intensity as a function of

time was fitted to the following exponential equations (1) and (2). Equation 1 (one-phase association) was used when no lag phase was recorded, where y0 is the relative fluorescence intensity at t = 0, Plateau is the fluorescence intensity value at infinite t, k is the aggregation rate constant. Equation 2 (plateau followed by a one-phase association) was used when the lag phase was recorded, where t0 is the time at which fluorescence intensity increases after a lag phase (271).

−퐾푡 (1) 퐹 (푡) = 푦0 + (Plateau − 푦0) × (1 − 푒 ) −퐾(푡−푡0) (2) 퐹 (푡) = (Plateau − 푦0) × (1 − 푒 )

The half-time or (t1/2) for each sample obtained from the midpoint of ThT fluorescence a region that represents the pre and post-aggregation stage (272). Limitations, ThT assay does not recognise amyloid fibrils per se, but binds to the molecular groove found typically on the amyloid fibril surface and becomes activated. ThT is also unable to detect early aggregates i.e. oligomers and protofibrils (273).

2.2.3 Tryptophan Fluorescence Fluorescence emission spectra were measured using FluoroMax (Horiba) spectrofluorimeter at an excitation wavelength of 290 nm and emission scan ranging 295 to 500 nm. The sample concentration described in Chapter 3 and 6 were excited at 290 nm and emission scan ranging 295 to 500 nm. The excitation and emission slit width were both set to 5 nm. The background fluorescence of buffer or liposomes was subtracted out from the sample reading. The data was exported as a .csv file and plotted λmax versus sample (lipid + peptide). Fluorescence emission maxima (λmax) were calculated by averaging the mean of three measurements that exhibited the highest signal. The shift in the wavelength indicates the hydrophobicity and hydrophilicity of the peptide, blue shift denotes lower wavelength, or hydrophobic and red shift depicts higher wavelength or hydrophilic.

2.2.4 Atomic Force Microscopy (AFM) All AFM imaging was performed in soft tapping mode in PBS to ensure that the fidelity of the lipid bilayers investigated was not affected by dehydration. The samples (as described in each Chapter) were pipetted onto a disk of freshly cleaved mica and left for 10 min, at which time the disk was gently washed in PBS before mounting on the AFM in a fluid cell (Bruker). The concentration of fibrils and lipids described in the subsequent Chapters. Imaging was performed on a Multimode 8 AFM using a Nanoscope V controller (Bruker) using an ultra- sharp silicon nitride cantilever optimised for fluid imaging (SCANASSYST FLUID +, Bruker).

The tips possessed a nominal end radius of 2 nm, and a resonant frequency of 150 kHz (in air). All images were flattened using the first order flattening algorithm in the Nanoscope analysis software and no further image manipulation occurred. Statistical analysis of the AFM images was performed in the software FiberApp (274) from data sets of around 500 individual fibrils. In some cases, fibril numbers were too low (<500) to perform statistical analysis. Note that all the samples were prepared at RMIT and the imaging was performed at Swinburne University. Limitations of AFM, relatively slow scan time, can lead to thermal drift on the sample and time-resolved kinetic studies of fibril formation is not feasible using this technique and the resolution can be impacted if the probe is not sharp enough (275).

2.2.5 Circular Dichroism (CD) CD spectroscopy was performed on Jasco J-1500, spectrophotometer. Measurements were conducted in the 280-190 nm region and measured every 1 nm at a scan speed of 50 nm/min using 2s equilibration time per step. Each spectrum is the average of 6 scans and was smoothed with a 7-point Savitsky-Golay filter to increase the signal-to-noise ratio. The mesophase was prepared by mixing peptides with liposomes at a ratio of 1:33, the final concentration of 6 mM vesicles and appropriate concentration was selected for functional amyloid-forming peptides (as described in Chapter 4, 5 and 6), to obtain a good signal to noise ratio with a high tension (HT) ≥ 800. The peptide only spectra were collected in 1 mm path quartz cuvette according to Beer’s Law equation: 퐴푏푠표푟푏푎푛푐푒 = 푒 푥 퐿 푥 푐; (e) molar extinction; (L) path length of cell holder and (c) concentration of solution. While the peptides co-incubated in lipids were recorded in a 0.1 mm path length sandwich cuvette (due to high scattering). Each sample was recorded after 3-hour and 72-hour incubation in buffer or lipid mesophase, in which the spectrum of a blank sample of the related lipid mesophase without peptide was deducted from that with incorporated peptide. A quantitative approach was taken to predict the secondary conformation of peptides due to high normalised root mean square errors (NMRSD) >0.250. Limitations, secondary structure cannot be accurately determined for the majority of proteins whose 3D structures have been solved and restriction to detect protein structures in far UV range (i.e. below 180 nm) (276). Difficult to detect peptide conformational changes in smaller aliquots, especially in amyloid fibril studies.

2.2.6 Synchrotron Circular Dichroism (SR-CD) (ASTRID II facility in Aarhus Denmark) sample measurements and characterisation The measurements were conducted in the 170-280 nm region using the CD beamline of the synchrotron radiation source ASTRID2 at ISA, a centre for storage ring facilities at University of Aarhus (Denmark). Suprasil Quartz cells (Hellma GmbH & Co., Germany) of 0.1 mm path length, requiring a ~30 µL sample volume, were used in a temperature-regulated sample holder. The measurements were obtained at 1 nm increments in wavelength, 2.15 s averaging time, and 6 scans for each baseline and sample data. Data were smoothed with a 7-point Savitzky- Golay filter (277). Liposomes in the absence of peptide were used as the baseline. The spectra are presented as molar ellipticity (M-1 cm-1), calculated with the mean residue weight (MRW) 푀표푙푒푐푢푙푎푟 푀푎푠푠 표푓 푝푒푝푡푖푑푒 of the peptide, using the equation 푀푅푊 = ; N = number of amino 푁−1 acids” (278), the path lengths of CD cells measured (102.3 and 97.7 µm). The data analysis performed using the online the DICHROWEB server (279) using CONTIN-LL algorithm. DICHROWEB analysis for samples with high NMRSD >0.250, a similar approach as described in section 2.2.5 was taken to predict the secondary conformation of peptides, included in appendix Table S8.3 and S8.5). Limitations, samples such as proteins can be damaged by radiation over longer exposure.

CHAPTER 3

3. Effect of lipid composition on fibrillisation of the functional amyloid-forming peptide Somatostatin-14 (SST-14)

3.1 Introduction The ability of amyloid-forming peptides to self-assemble into larger aggregates, characterised by a β-sheet rich structure, is a trademark of several neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease and type II diabetes mellitus (280). Recently, neuropeptide hormones, including somatostatin, oxytocin, and vasopressin have been shown to self-assemble into fibrils with similar morphological features to those found in toxic amyloids (99). These so-called “functional amyloids” are stored as amyloids within endocrine secretory granules and possess physiologically normal functions in living systems (5).

The mechanism of amyloid fibril growth has been proposed for human amyloidogenic peptides, including islet amyloid polypeptide (IAPP) and Amyloid-β peptide (Aβ), and as described in Chapter 1, normally occurs through a nucleation-dependent polymerisation process made up of two phases: a nucleation (lag phase) and an extension (growth phase) (Figure 3.1) (281). During the lag phase, peptide monomers undergo a conformational shift to a β-sheet conformation, with subsequent self-assembly into oligomers and higher-order protofibrils (5, 282). This conformational change is unfavourable thermodynamically and a rate-limiting step of the fibrillisation process. Once the nucleus has formed, the further addition of monomers to the nucleus results in the extension and growth of amyloid fibrils (281). The mechanism of amyloid formation is thought to be similar for functional amyloids, although (unlike for toxic amyloids) amyloid formation is a reversible process (5, 98).

Figure 3.1. Shows the process of nucleation-dependent fibril formation for Toxic vs Functional amyloids (adapted from (5))

Numerous studies have shown that membrane surfaces can accelerate fibril growth for toxic amyloids. Several mechanisms have been proposed for this effect (5, 158, 283). Peptide- binding to the membrane can induce a conformational change to the peptide, promoting fibril growth (245). A conformational change to a α-helical structure has been observed for Aβ, IAPP and α-synuclein following binding to the membrane (152, 284-287). Local increases in peptide concentration (referred to as “molecular crowding”) can then facilitate aggregation (5). It has also been suggested that the dielectric constant of the solvent is locally reduced in the microenvironment of the membrane and that this can promote the formation of intermolecular hydrogen-bonding or hydrophobic interactions between the peptides (283). The fibrillisation process can, in turn, significantly impact the structural integrity of the lipid membrane, via membrane thinning and pore formation, an effect which is strongly correlated with toxicity for toxic amyloids (288).

Research shows that the rate of fibril formation in toxic amyloidogenic peptides is influenced by a range of membrane parameters, such as lipid composition, surface charge and intrinsic membrane curvature (5). Anionic lipids have been shown to accelerate fibril growth for a range of cationic amyloid-forming peptides including Aβ (215, 288), IAPP (289) and α-synuclein (290), presumably due to increased electrostatic interactions (291). Neutron and X-ray scattering studies by Lee and co-workers demonstrated that Aβ40 inserts more readily into anionic DPPG bilayers compared to neutral DPPC lipid bilayers (5). Liposomes containing the anionic PG lipid were also found to promote a β-sheet conformation for Aβ40 (189), IAPP (185) and α-synuclein (154, 190, 191). α-synuclein bound to membranes in gel phase (DPPC/

DPPG), indicated a deeper penetration of the peptide within the bilayer as a result of peptide- lipid bilayer interaction (37, 153, 197, 290). An increase in surface curvature of the membrane has also been linked with increased binding and aggregation rates (292).

Cholesterol along with sphingomyelin and ganglioside lipids, which form ordered lipid-raft microdomains have been shown to enhance peptide binding and subsequent fibrillisation for Aβ, IAPP and prion proteins (170-175, 293). In contrast, α-synuclein however, preferentially binds to disordered bilayers. Binding to such domains has been shown to result in a conformational change for α-synuclein, from random coil to α-helix (at low peptide-lipid ratio) (294-296) and to β-sheets (at high peptide-lipid ratio) (297).

Recently, a study on TasA, a model functional amyloid protein from the soil bacterium Bacillus subtilus, suggested, for the first time, that lipid membranes may play a similar role in the fibrillogenesis of functional amyloids (8). The interaction of TasA with a bacterial model membrane composed of a ternary mixture of PE, PG and cardiolipin resulted in disordered aggregates, displaying clear morphological differences to the fibrils formed in contact with a PC, Cholesterol and Sphingomyelin mixture designed to mimic a eukaryotic cell membrane. This initial study (and the body of literature on toxic amyloids) tentatively suggests that the ability of amyloidogenic proteins and peptides to interact with eukaryotic membranes is a ubiquitous feature which tends to promote amyloid assembly (8). However, in contrast to the vast body of literature on toxic amyloid-forming peptides, this study remains, to the best of our knowledge, the only investigation into the effect of eukaryotic lipid membranes on fibrillisation of functional amyloids.

In Chapter 3, the effect of the lipid membrane composition on the formation of fibrils has been investigated for the functional amyloid-forming peptide SST-14 using multi-lamellar vesicles (MLVs). The peptide hormone SST-14, which is involved in regulating the central nervous system (CNS) and endocrine system (ES), has a cyclic structure resulting from a disulphide bridge (Cys3-Cys14) (Figure 3.2) (247).

Figure 3.2. Shows the amino acid sequence and disulphide link for SST-14 (adapted from (20))

It has been reported to form amyloid fibrils in vitro (77, 248, 249). Recently, Dharmadana et al. have demonstrated that aggregation enhancers, such as heparin, commonly used to increase the aggregation rate for amyloid-forming peptides result in clear morphological differences in the resulting fibrils for SST (98). The fibrillisation of SST was characterised in the presence of liposomes of various lipid compositions. Saturated lipids with chain lengths from C12 to C18 were investigated, as well as the singly unsaturated C18 lipid DOPC. The effect of cholesterol and the anionic lipids DMPS and DOPS were also investigated. Results are rationalised in terms of the effect of membrane parameters including membrane surface charge and fluidity, hydrocarbon chain packing, and lipid solubility on the fibrillisation of functional amyloid- forming peptides.

3.2 Materials and Methods 3.2.1 Liposome preparation and analysis The liposome preparation and analysis are described in Chapter 2, section 2.2.1.

3.2.2 Small Angle X-ray Diffraction (SAXD) on bulk lamellar sample, preparation and data analysis Self-assembled lipid mesophase (gel phase) of 100 mol% DMPC and 100 mol% DSPC prepared from a chloroform solution. The solution was dried under fume hood overnight to evaporate and remove any residual solvent and lyophilised for 24 hours. Approximately, 80 wt% of Milli-Q water was added to dried films followed by 0.01% w/w to 2.5% w/w SST at a 푚표푙 푆푆푇 ratio of (1:97). Mol% is defined as 100 푋 . The peptide and self-assembled 푚표푙 푆푆푇+푚표푙 푙푖푝푖푑 lamellar phase (gels) allowed to equilibrate overnight before SAXD experiments. SAXS experiments were run at the SAXS/WAXS beamline of the Australian Synchrotron. A beam

current of 10.0 keV was used with dimensions 200 µm x 300 µm and a typical flux of 1 x 1013 photons s-1. The sample to detector distance was 1.5 m, with an exposure time of 1 s. Two- dimensional diffraction patterns were recorded on a Pilatus 1 M detector, which offers minimum noise and collection of data over a large area. The samples were deposited in gel plates which consist of 40 wells and kapton tape was used to retain moisture in the samples. The 2D diffraction images are converted to 1D files of intensity vs q using the scatterbrain program. This was used in combination with IDL-based AXcess software package, developed at Imperial College London, which permits individual analysis of X-ray images. Calibration of the program is carried out with a known standard (silver behenate, d = 58.38 A). Followed by assigning the sample peaks, the program then calculates the d-spacing or lattice parameter for a specified phase. The D-spacing of more than one or co-existing phases are calculated 2휋 separately using the equation: 푞 = . 퐷

3.2.3 Small Angle X-ray Scattering (SAXS) Liposomes used in this experiment and preparation is described in Chapter 2, section 2.2.1 with the same ratios but only selected vesicles were used (DOPC, DOPC/Chol and DMPC). The beamline and its configurations are similar to 3.2.2, with changes to the sample and plate setup. The peptide (SST-14) co-incubated at 5% w/w or 2 mol% in buffer (150 mM NaCl) and liposomes at a similar ratio to bulk samples (3.2.2) and the final concentration of vesicles were 6 mM. 100 µl of sample was deposited in 96 well polymerase chain reaction (PCR) plates. The plate was sealed with a silicon rubber lid to prevent evaporation and pre-cut to allow smooth penetration of the sample loading needle. The plates were positioned in the plate holder and standard safety, the procedure is administered before X-rays were directed on to the samples. An auto-loader set-up was used and the plate holder custom built can accommodate 2 plates of dimensions (85 x 127 mm, 96 or 384 well) and is fixed in front of the beam, coordinates mapped for samples and an automated scan of each well is initiated with an approximate run time ~1 hour per plate. The scatterbrain software developed by SAXS/WAXS beamline scientists (Dr. Stephen Mudie and Dr. Nigel Kirby) was used to perform averaging, buffer subtraction and preliminary analysis on samples. The data was exported as a .csv file and plotted as intensity vs q, and the D-spacing equation was used in 3.2.2 to calculate D-spacing (D) or lattice parameter. Limitations are radiation damage to samples e.g. proteins and spatial averaging can occur due to random orientation of dissolved samples, which leads to loss of information (298).

3.2.4 Preparation of Functional amyloid-forming peptides (FAFP’s) 150 mM NaCl was added to a weighed fraction of SST-14. The solution was vortexed for a few seconds for smaller aliquots or a minute for bigger aliquots to produce solutions having a final concentration ranging from 0.01% w/w to 5% w/w.

3.2.5 Thioflavin T (ThT) Binding Assay The concentration of SST-14 in the buffer is described in 3.2.4 and 5% w/w or 2 mol% is used in liposomes (preparation and analysis described in Chapter 2, section 2.2.1). ThT assay protocol is described in Chapter 2, section 2.2.2.

3.2.6 Time-resolved kinetic studies using Dynamic Light Scattering (DLS) 5% w/w SST was co-incubated in liposomes at an appropriate ratio (1:97) with selected vesicles (DOPC, DMPC, DOPC/Chol and DOPC/DOPS). The liposome preparation and DLS protocol and analysis are described in Chapter 2, section 2.2.1.

3.2.7 Atomic Force Microscopy (AFM) 5 % (w/w) solution of SST fibrils or fibrils + liposomes (SST + DMPC) at ratio (1:97). Imaging and analysis are carried out as described in Chapter 2, section 2.2.4.

3.2.8 Tryptophan Fluorescence The peptide (SST-14) prepared at 0.01% w/w in buffer (NaCl) and co-incubated in liposomes doped with various lipid compositions at a ratio (1:97) (described in Chapter 2, section 2.2.1). The experimental protocol and analysis are described in Chapter 2 section 2.2.3.

3.2.9 Synchrotron Circular Dichroism (SR-CD) (ASTRID II facility in Aarhus Denmark) sample measurements and characterisation A weighed fraction of 0.03% w/w or 0.3 mg mL-1 SST was dissolved in buffer (MilliQ water) and co-incubated in selected liposomes (DOPC, DMPC, DOPC/Chol and DOPC/DOPS (lipid at an appropriate ratio (1:97) as described in Chapter 2, section 2.2.1 with a minor change to this protocol for liposome preparation (Milli-Q water instead of 150 mM NaCl to rehydrate vesicles). The experimental parameter and SRCD spectra analysis are described in Chapter 2, section 2.2.6.

3.3 Results 3.3.1 Interaction with the lipid bilayer changes the secondary structure of SST- 14 and the kinetics of fibril growth The secondary structure of SST-14 was studied using synchrotron circular dichroism (CD) in the presence and absence of liposomes. Note that a typical peptide concentration for CD experiments was used, 0.3 mg mL-1, or (0.03 % w/w) which is below the minimum SST-14 concentration at which amyloid fibril formation is observed (59). The CD spectrum of SST-14 in water contains a large minimum at 201 nm, consistent with a secondary structure which is a mostly random coil (Figure 3.3a). A shoulder at around 195 nm indicates the presence of β- sheet structures, presumably due to intra-peptide hydrogen bonds constrained by the cyclic molecular structure and potentially to inter-peptide networks within small soluble oligomers. This is confirmed by analysis of the spectrum using Dichroweb, which yields 54% random coil and 42% β-sheets and turns (Table 1), in agreement with previous ATR-FTIR and CD studies on SST-14 in water (299).

a) b)

Figure 3.3. (a) SR-CD spectra for SST-14 co-incubated in vesicles after 3 hours and (b) SR- CD spectra for SST-14 co-incubated in vesicles after 72 hours

Table 3.1. Secondary structure of selected samples of SST-14 co-incubated with liposomes, analysed using Dichroweb with Contin-LL algorithm method. *Quantitative analysis was performed for DOPC+SST, DOPC+DOPS+SST (3 and 72 hours) and DMPC+SST (72 hours) due to NRMSD > 0.250 (in appendix Table S8.3)

Helix Sheet Turns Unordered Samples NRMSD Ref Set % % % % After 3 hours Pure SST 4 27 15 54 0.209 7 DMPC+SST 15 39 26 6 0.191 7 70DOPC:30CHOL+SST 8 53 31 3 0.161 7 After 72 hours Pure SST 5 26 14 55 0.219 7 70DOPC:30CHOL+SST 1 45 27 26 0.120 7

Following three hours incubation with liposomes of various compositions, a shift in the CD spectrum was observed for all samples relative to pure SST (Figure 3.3a). The minimum at 201 nm becomes less intense and moves to higher wavelength, while the shoulder feature at 195 nm becomes more pronounced, consistent with an increase in β-sheets and a concomitant reduction in the proportion of the structure that is the random coil. The effect is dependent on the lipid composition and is most pronounced for the lipid bilayer doped with the anionic lipid DOPS, presumably due to an electrostatic interaction with the positively charged peptide. Dichroweb analysis, which was only possible for selected samples due to large errors in data- fitting (NRMSD > 0.250), confirmed a significant reduction in the percentage of random coil structures and increased in β-sheets and turns (Table 3.1). This effect appears to become more pronounced with time, particularly for the anionic lipid bilayer, which displays a characteristic β-sheet rich spectrum of a strong negative peak at around 210 nm and a maximum of comparable intensity at 195 nm after 72 hours, with no residual trace of random coil structures visually observed in the spectrum (Figure 3.3b).

Intrinsic tryptophan fluorescence measurements, (Figure 3.4), provide information on the environment around the single tryptophan (Trp) residue in SST-14. In buffer, λmax has a value of approximately 357nm. Following incubation with liposomes of various lipid compositions, a characteristic blue shift of λmax to lower wavelength is typically observed. This is

characteristic of the presence of a more hydrophobic environment around the tryptophan residue, consistent with some penetration of the SST into the liposomal lipid bilayer. The magnitude of this blue shift is larger for liposomes formed from the unsaturated lipid DOPC, and smaller for liposomes formed from saturated chain lipids, indicating greater penetration of the peptide into the unsaturated bilayer. Addition of the anionic PS lipid resulted in an increased blue shift relative to the analogous uncharged bilayer for unsaturated and saturated chain liposomes, suggesting that the electrostatic interaction has also promoted deeper penetration into the bilayer. The effect of cholesterol was not consistent; an increased blue shift was observed for DOPC, but a reduced blue shift for DMPC.

Figure 3.4. SST-14 environment in buffer (control) and in bilayer measured by Tryptophan Fluorescence

The fluorescent dye Thioflavin T (ThT) specifically binds to β-sheets or fibrils, which results in an increase in its fluorescence intensity. ThT fluorescence binding assays were used to monitor the kinetics of SST-14 amyloid fibrils formation in the presence of liposomes. The increase in ThT relative fluorescence with time is plotted in (Figure 3.5a) for SST-14 at various concentrations in the presence of NaCl. In the absence of the lipid bilayer, ThT binding was detected for SST concentrations of > 2.5% (w/w). An initial lag phase of 12 hours, characterised by a weak increase in ThT fluorescence, was followed by a strong growth phase.

This growth phase can be used to calculate the rate constant for aggregation (kagg) by fitting a mono-exponential decay (characteristic of first-order kinetics) to the data, as described in (98).

The kagg values for SST-14 in NaCl was obtained for concentrations 2.5 and 5% w/w, i.e. 0.0037 s-1 to 0.019 s-1 (in appendix Table S8.4).

The increase in ThT relative fluorescence for SST-14 (2 mol%) incubated with liposomes of various lipid compositions, including saturated chain lipids of various lengths (DLPC (C12); DMPC (C14); DPPC (C16) and DSPC (C18)) and the unsaturated lipid DOPC (C18:1) is shown in (Figure 3.5b). The values of lag phase, kagg, t1/2, and final fluorescence intensity of the plateau phase were extracted from the ThT curves are provided in Table 3.2. Incubation with saturated chain lipids increased the value of kagg, relative to that observed in the buffer, in all cases, with no strong correlation between kagg and the alkyl chain length (Table 3.2). These results show that the presence of saturated lipids accelerates the SST-14 aggregation process. Saturated chain lipids also reduced the lag period relative to that seen in NaCl and resulted in slightly higher fluorescence plateau values, supporting that fibril nucleation occurs earlier and more aggregates are formed in the presence of these liposomes. However, the half-time ( for

SST fibrils in DLPC liposomes (t1/2 = 6.3 hours) was much shorter compared to other saturated chain lipids, (Table 3.2), consistent with a previous study which showed accelerated Aβ aggregation with short-chain lipid (300). The most significant effect is observed with the saturated chain lipid DMPC which has a significantly highly fluorescence plateau value (Table 3.2) than any other saturated chain lipid, suggesting the greatest number of fibrils are formed with this combination.

In contrast, the unsaturated chain lipid DOPC appeared to slow down fibril aggregation, with both an increase in the lag period, a reduction in the t1/2 value, and a reduction in the value of kagg (Table 3.2). The fluorescence plateau value was also much lower in the presence of DOPC, suggesting many fewer aggregates are formed. The reduction in aggregation may be linked to the deeper penetration in the DOPC bilayer suggested by the tryptophan fluorescence results presented above and is discussed further in the discussion.

The effect of the addition of cholesterol to the DOPC and DMPC lipid bilayers on fibrillisation was also investigated (Figure 3.5c). In a saturated chain bilayer (DMPC), cholesterol appeared to significantly increase the fluorescence intensity of the final plateau, with a slight decrease -1 of the kagg value (0.04 s vs 0.03 with cholesterol). The t1/2 value (23h) was slightly longer than that for pure DMPC (13h) (Table 3.2). This result supports that more peptide fibrils are formed in the presence of cholesterol in the saturated bilayer. In the unsaturated DOPC lipid bilayer, while addition of cholesterol did not significantly increase the plateau value, it did reduce the lag phase from 48 hours to 24 hours and slightly increased the kagg value from 0.024 to 0.029, suggesting that the kinetics of the fibril formation process is affected, but not the number of fibrils.

The greatest impact on fibril growth was observed for the charged lipids DOPS and DMPS, in agreement with numerous studies on the strong electrostatic interaction between cationic peptides and anionic lipid membranes (5, 37, 214, 301). The effect of charge was also observed in a recent study on the functional amyloid TasA, which interacted more strongly with a charged lipid membrane composed of PE/PG/CL (302). We note that the lag period was reduced, and kagg and final fluorescence plateau values increased significantly when the bilayer was doped with 35 mol% anionic lipids, which is also in agreement with the shortest half-time values (Table 3.2). While the rate of growth was similar for both anionic liposomes, the volume of fibrils formed was significantly greater for the anionic DMPC/DMPS liposomes.

Figure 3.5. (a) Thioflavin T (ThT) assay of SST-14 incubated in NaCl, (b) ThT assay of SST- 14 and SST-14 at 5 % (w/w) co-incubated with unsaturated and saturated vesicle with increase in chain length impacting SST fibrillisation, (c) ThT assay of SST-14 and SST-14 co-incubated in DOPC, DOPC vesicle doped with cholesterol, DMPC, DMPC doped with cholesterol, DOPC doped with DOPS and DMPC doped with DMPS

Table 3.2. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST-14 at 5% (w/w) in NaCl and vesicles doped with various lipid compositions over 1 week period

-1 1or2 Conditions Lag phase t50 or (t1/2) kagg (s ) Final (hours) (hours)3 fluorescence intensity of plateau phase (x106)

Buffer

NaCl 12 51 0.01862 2.7

Unsaturated

DOPC 48 85 0.02382 1.4

Effect of chain length or bilayer thickness

DLPC 0 6.3 0.05891 3.8

DMPC 0 13 0.03801 6.2

DPPC 24 41 0.06312 4.9

DSPC 12 28 0.05112 4.8

Effect of Cholesterol

70DOPC+30CHOL 24 36 0.02872 3.3

70DMPC+30CHOL 0 23 0.03021 11.0

Effect of Charge

65DOPC+35DOPS 0 2.6 0.3251 5.6

65DMPC+35DMPS 0 0.3 0.2191 12.3

One-phase association1, Plateau followed by one-phase association2 and half-time3

3.3.2 The peptide and bilayer interaction reveals no significant changes to bilayer size and shape The average size distribution is almost the same which indicates that the vesicles do not change significantly with the presence of the peptide (Figure 3.6). The broadness of the peaks is observed to be unchanged which further suggests that the interaction between peptide and lipid had not occurred. Therefore, following (Figure 3.7) shows the hydrodynamic size change with respect to time (hours), and the results clearly suggest no change in the overall hydrodynamic size.

Figure 3.6. Size distribution of vesicles of various compositions co-incuated with SST-14 (5% w/w) for 1 week as determined by DLS a) DOPC, b) DMPC, c) DOPC doped with cholesterol and d) DOPC doped with DOPS

Figure 3.7. Shows the hydrodynamic size of the vesicles with SST-14 over 1 week period

3.3.3 The interaction between the peptide and the lipid bilayer do not significantly alter the bilayer structure The effect of SST on the bulk lamellar structure was investigated for DMPC and DSPC phases using Synchrotron Small-Angle X-ray diffraction (SAXD). The well-defined lamellar peaks are characteristic of a bilayer structure. Upon incubating SST with DMPC (Figure 3.8) or DSPC (Figure 3.9), no changes to the lamellar peaks and D-spacing (Table 3.3 and Table 3.4), which indicates that the sharp lamellar peaks were retained at all concentrations, which suggests that the bilayer structure is unpertubed.

Figure 3.8. Shows the 1D diffraction pattern of Intensity vs q for bulk lipids at varying SST concentration. Small angle X-ray scattering data were obtained ~7 h after setup of the plates.

Table 3.3. Shows the effect of SST addition (0.01 to 2.5% w/w) on D-spacing of the lamellar phase of DMPC

Lipids mol% Peptide % w/w Phase D-spacing (A)

100 DMPC No peptide Lα1 62

100 DMPC 0.01 Lα1 62

100 DMPC 0.05 Lα1 62

100 DMPC 0.1 Lα1 62

100 DMPC 0.5 Lα1 62

100 DMPC 1 Lα1 61

100 DMPC 2.5 Lα1 62

Figure 3.9. The 1D diffraction pattern of Intensity vs q for bulk lipids at varying SST concentration. Small angle X-ray scattering data were obtained ~7 h after setup of the plates.

Table 3.4. Shows the effect of SST addition (0.01 to 2.5% w/w) on D-spacing of the lamellar phase of DSPC

Lipids mol% Peptide % w/w Phase D-spacing (A)

100 DSPC No peptide Lα1 67

100 DSPC 0.01 Lα1 67

100 DSPC 0.05 Lα1 67

100 DSPC 0.1 Lα1 68

100 DSPC 0.5 Lα1 69

100 DSPC 1 Lα1 68

100 DSPC 2.5 Lα1 67

The effect of SST on the lipid bilayer structure of multi-lamellar vesicles (MLVs) was investigated. Liposomes formed by DMPC, DOPC and DOPC+Chol mixtures displayed x-ray scattering patterns consistent with the formation of multi-lamellar liposomes, (Figure 3.10). Two clear correlation peaks are seen in each case corresponding to lattice parameter or D- spacings of 64 Å, 62 Å and 65 Å (Table 3.5) respectively, in good agreement with the lamellar

phase formed by these systems (303-305). Upon addition of peptide (peptide : lipid ratio = 1:97), a decrease in the scattered intensity was observed. However, two clear correlation peaks are still observed for DOPC and DOPC+Chol suggesting that the bilayer structure of the liposomes has not been impacted significantly, despite CD and tryptophan fluorescence studies showing significant interaction between the peptide and these bilayers. For DMPC liposomes the second correlation peak has been replaced by a single broad peak in the scattering pattern, (Figure 3.8c), indicating that the multi-lamellar structure has been compromised, suggesting a phase transition from lamellar to micelles.

Table 3.5. Lattice parameter or D-spacing for SST co-incubated with vesicle

Liposomes SST (Mol%) Peptide : lipid D-spacing

DOPC 0 0 62

70DPC+30CHOL 0 0 65

DMPC 0 0 64

DOPC+SST 1 1:97 60

70DOPC+30CHOL+SST 1 1:97 63

DMPC+SST 1 1:97 0

Figure 3.10. Small angle X-ray scattering (SAXS) of Multilamellar vesicles (MLVs) co- incubated with SST, (a) DOPC vesicle co-incubated with SST and (b) DOPC vesicle doped with cholesterol co-incubated with SST, (c) DMPC vesicles co-incubated with SST

3.3.4 Clear morphological differences are observed between fibrils formed in the buffer and those incubated with liposomes Results so far suggest that the addition of liposomes formed from DMPC lipids (compared to other zwitterionic lipids studied) result in faster fibrillisation kinetics for SST-14 and a significant increase in total volume of fibrils formed. SAXS results suggest that the bilayer structure of DMPC liposomes is impacted by this interaction. We, therefore, used DMPC liposomes to perform a comparative analysis of the fibrils formed in the presence of liposomes, with those formed in the buffer using Atomic Force Microscopy. Analysis was carried out on SST fibrils and SST fibrils co-incubated with DMPC lipid vesicles (1:97), (Figure 3.11). Clear differences in SST fibril morphology were observed between samples formed in NaCl and in the presence of liposomes. SST fibrils form typical amyloid-like unbranched twisted fibrillar structures, (Figure 3.11a). In contrast, the fibrils formed by SST co-incubated with DMPC vesicles, (Figure 3.11b), displayed a more heterogeneous structure, likely due to an uneven coating of lipids decorating the fibrils. Evidence of the lipid coating of the fibrils is highlighted by the arrows in (Figure 3.11b). The distribution in height (Figure 3.11c) and length (Figure 3.11d) of individual SST and SST-DMPC fibrils were determined by a statistical analysis of more than 500 fibrils using the open source software FiberApp (306). The contour length distributions show that the bare fibrils display a relatively broad single population of contour lengths centred around 1000 nm, with a small additional population of very long fibrils (> 8000 nm). The distribution of contour lengths for fibrils coated with DMPC, on the other hand, is very heterogeneous with seemingly multiple sub-populations co-existing. From the histograms in (Figure 3.11c) it is obvious that the fibrils coated with DMPC have a tendency to be longer with large populations with contour lengths > 5000 nm, this tendency for longer fibrils to be formed in the presence of DMPC is consistent with the increase in kinetics and overall fluorescence levels in the ThT analysis (Figure 3.5b). The height distribution statistics (Figure 3.11d) indicate that fibrils formed in the absence of lipids have a relatively small distribution of heights with a single population centred around 2- 3 nm. Fibrils formed in the presence of DMPC liposomes displayed a more heterogeneous distribution of heights with what appears to be two populations centred around 3.5 and 6.5 nm. Presumably, these two populations relate to partially lipid coated fibrils (with a mean diameter only 0.5 nm larger than the bare fibrils) and a second population of lipid coated fibrils (with an average fibril diameter around 3.5-4.5 nm larger than the uncoated fibrils). This 3 nm increase

in fibril height is too small to correspond to a full bilayer surrounding the fibril (typically a full bilayer is approximately 5 nm in width, multiplied by two to account for the coating of the top and the bottom of the fibril), suggesting that upon co-aggregation of the SST and the DMPC lipids the bilayer structure is disrupted and something approximating the height of a lipid monolayer (~2.5 nm) is surrounding the fibrils on all sides (in agreement with the suggested bilayer disruption from the SAXS). Persistance length calculations suggest that SST-DMPC fibrils display a slight increase in stiffness (Figure 3.11e, f). The high-resolution AFM images of the fibrils formed in NaCl display a commonly observed mature amyloid twisted fibril structure (Figure 3.12a) (307). However, in the presence of the DMPC lipids, this twisted morphology seems to be obscured (Figure 3.12b). The representative height profiles of bare SST and SST + DMPC fibrils confirm this is showing a very regular periodicity for the bare fibrils (Figure 3.9c), and no obvious periodicity for the DMPC coated fibrils. Through Direct Fourier Transform (DFT) (Figure 3.12e, f) and autocorrelation function (ACF) (Figure 3.12g, h) analysis of the populations of fibrils we can very accurately calculate the peak to peak distance of these periodic structures. For the bare fibrils, DFT analysis provides a pitch of 64.4 nm while ACF shows 63 nm for the bare SST fibrils, and as expected no periodicity is detected when the DMPC liposomes are added.

Figure 3.11. Atomic Force Microscopy Analysis of 5% (w/w) SST fibrils and SST fibrils co- incubated with DMPC lipid vesicles, (a-b) topographic images of (a) SST fibrils displaying typical amyloid-like unbranched twisted fibrillar structure, (b) SST co-incubated with DMPC vesicles, the formed fibrils display a more heterogeneous structure and evidence of the lipid coating of the fibrils is highlighted by the arrows, (c-d) statistical analysis (approximately 500

fibrils) of fibrillar species formed and (e-f) Persistance length calculations, for (e) SST and (f) SST+DMPC, very little difference, slight increase in stiffness for the SST + DMPC fibrils

Figure 3.12. High Resolution AFM images of the change in fibril nanostructure between SST and SST + DMPC fibrils (a-b) topographical AFM images of (a) SST and (b) SST + DMPC fibrils, (c-d) line section through one individual fibril (c) SST and (d) SST + DMPC. (e-f) Direct Fourier Transform analysis of pitch of multiple fibrils (g-h) Auto-Correlation Function analysis of twist angle of multiple fibrils

3.4 Discussions Co-incubation of the SST-14 peptide with lipid membranes results in clear differences in the rate of fibrillisation, and the volume and morphology of the formed fibrils. The effect is dependent on the lipid bilayer composition. As expected, the addition of an anionic lipid into the bilayer strongly promoted fibril growth, with both a fast rate of fibrillisation and a large volume of formed fibrils. This is consistent with the known effect of anionic lipids on toxic amyloid-forming peptides and results from increased electrostatic interaction between the cationic peptide and the anionic membrane (5, 153, 160, 284, 308). More unexpectedly the effect of chain unsaturation in the lipid also had a significant effect. In all cases, lipids having saturated hydrocarbon chains both increased the aggregation rate and resulted in a greater number of fibrils formed. We note that the fastest aggregation rates (shortest t1/2 times) were observed for the short-chain lipids, DLPC (C12:0) and DMPC (C14:0), although all four saturated chain lipids display similar kagg values during the fibril growth phase. Interestingly, a fast aggregation rate does not always correlate with the greatest volume of fibrils formed; while the fluorescence plateau value is correspondingly high for DMPC, it is low for DLPC. These data are in agreement with a recent study on α-synuclein, which also showed the fastest aggregation rates for the charged analogues DLPS and DMPS (309). The authors rationalised this effect in terms of the relatively high solubility of these short-hydrocarbon chains (in general the free energy of solubilisation of lipids in water is proportional to chain length). They suggested that the mechanism of membrane-mediated fibrillisation may, therefore, involve not just binding of the peptide to the membrane, but a reciprocal transfer of lipids to the local environment of the peptide. They noted the physiological relevance of this effect due to the generation of short-chain lipids from the peroxidation of poly-unsaturated lipids, which is known to be damaging to cells. It is interesting to note an identical effect for the uncharged lipids investigated here, suggesting that the effect of lipid solubility may be comparable to the known effect of charge. Both CD and intrinsic tryptophan fluorescence studies confirm the interaction between the peptide and the lipid bilayer, although the strength of this interaction is not fully correlated with the effect on fibrillisation. The change in secondary structure of the peptide in the presence of the liposomes indicates interaction is associated with changes to the secondary structure of the peptide consistent with a reduction in random coil structure and an increase in β-sheets and is most pronounced for the charged lipid bilayer DOPC:35%DOPS. Tryptophan fluorescence studies indicate that the peptide resides in a more hydrophobic environment in the presence of

liposomes, confirming that the peptide inserts, at least partially, into the bilayer. The blue shift is most pronounced for the unsaturated bilayer DOPC, and more minor for the saturated chain bilayers suggesting that SST inserts more deeply into the unsaturated bilayer. The deeper penetration into the DOPC bilayer is consistent with recent studies suggesting that the increased lateral bilayer pressure associated with the saturated hydrocarbon chains can hinder peptide insertion (310) and may inhibit aggregation by blocking potential aggregation sites for the peptide. We also note clear differences in peptide-membrane interactions between gel (DPPC, DSPC) and fluid lamellar bilayers (DLPC, DMPC, DOPC). Tryptophan fluorescence experiments suggest that the peptide resides in a more hydrophobic environment for the fluid membranes, indicative of deeper penetration into the bilayer. The rate of diffusion of the amyloid-beta peptide over the 2-D membrane surface has been shown to be slower in the gel phases (243), consistent with a recent study which showed that diffusion rates for membrane-associated peptides are strongly linked to the underlying lipid self-diffusion rates (311) For the amyloid- beta peptide, a reduction in diffusion was linked with a retardation in fibril growth. Here, a similar effect was noted for the saturated hydrocarbon chains, aggregation of SST in the gel phase was slower than that in the fluid lamellar phase. However, the effect is more minor than that of the lipid solubility as evidenced by the large difference in aggregation rates for saturated DLPC and unsaturated DOPC. A similar effect was noted by (309) for α-synuclein; while the binding of α-synuclein was greatest to fluid membranes, this was not correlated with aggregation rate (309). To summarise, a combination of CD, intrinsic tryptophan fluorescence, ThT assays, SAXS and AFM results suggest that membrane surface charge is the parameter which enhances fibril formation most, in line with published literature on toxic amyloid-forming peptides. Saturated chain lipids appear to enhance fibrillisation while unsaturated lipids inhibit fibrillisation, and this effect is correlated with deeper penetration of the peptide into the unsaturated bilayer due to reduced lateral pressure. Aggregation of peptides may be enhanced on the membrane surface (as opposed to deeper in the bilayer) due to the probability of contact with other surface- associated peptides resulting in amyloid formation. A more minor effect is seen with lipid solubility; shorter chain lipids appear to have the fastest aggregation rates, potentially linked to the reciprocal transfer of lipids out of the bilayer. The short chain, saturated lipid DMPC is a particularly interesting case study, with an extremely high volume of fibrils formed with highly heterogeneous contour lengths.

Figure 3.13. An illustration of SST-14 fibril interacting with fluid DMPC bilayer, inspired by AFM image (Figure 3.9b) Statistical analysis of the AFM data suggests that the DMPC lipid bilayer is disrupted and a lipid monolayer sheath is formed around the growing SST fibrils obscuring their natural twisted morphology. Presumably, this bilayer disruption is made possible by the highly soluble nature of the short chain DMPC lipids. The reason this effect is not also observed by SAXS for the equally short chain length DLPC is unclear and warrants further study. Here it is worth highlighting the power of combining high-throughput ensemble techniques like SAXS and macromolecular imaging techniques such as AFM. By taking this approach, we were able to screen the entire library of lipids and identify unusual behaviour in the DMPC lipids. An in- depth statistical analysis of DMPC and SST co-aggregation then allowed us to detect fibrils with two sub-populations of diameters with a difference corresponding to one lipid monolayer thickness. The fibrils appeared to have a distinct “pearl necklace” formation seen in AFM and its interaction with DMPC bilayer is shown in a schematic representation (not to scale, Figure 3.13). If we did not take such a statistical approach to high-resolution macromolecular imaging in order to determine the diameter of our fibrils, and instead relied on ensemble techniques for colloidal analysis (i.e. Dynamic Light Scattering) these sub-populations with subtle differences in fibril diameter would have been indistinguishable.

CHAPTER 4

4. The interaction between the cyclic neurohypophysial hormone-like peptides Vasopressin (AVP-9) and Oxytocin (OT-9) and model membranes

4.1 Introduction The neurohypophysial hormones AVP-9 and OT-9 (312) are part of a family of structurally and functionally related peptide hormones with a conserved intramolecular ring, comprised of 6 amino acid (aa) residues and a C-terminal extension of 3 aa residues (250). They are formed as large precursor proteins that are consequently processed to give rise to mature nonapeptides. These peptides differ in the amino acids at positions 3 and 8 (Figure 4.1), contributing to significant differences in their activity (313). Neurohypophysial hormones are involved in sexual behaviours and social recognition and contribute to neuropsychiatric disorders including intellectual disability (ID), autism spectrum disorders (ASDs) and bipolar disorder (BD) (250, 251). The primary role of AVP is in the regulation of blood pressure and urine concentration (313-315). It is also responsible for the stability of body temperature (316) and the stimulation of adrenocorticotropin secretion (317). OT is involved in labor and lactation and plays an indirect, facilitatory role in the release of luteinising hormone (LH) (313).

AVP-9 OT-9

Figure 4.1. The amino acid sequences for neurohypophysial hormones (a. AVP and b. OT) (adapted from (21))

Studies have concluded that neurohypophysial hormones contain sequences which are likely to form aggregates due to mutations (250) resulting in the formation of functional amyloids in secretory granules(77). Immunohistochemical studies conducted on mouse pituitary tissue have confirmed the existence of amyloid-like fibrils of OT-9 and AVP-9 in secretory granules located in the posterior lobe of the pituitary (99, 100). AVP receptors are typically found located in the same brain regions as Aβ fibrils and in vivo studies conducted by Jing and co-

workers, have reported that AVP can prevent Aβ induced impairment (101). More recently, Atosiban, a nonapeptide drug which is an OT-9 antagonist and shares most of its features, has been reported to self-assemble into 7-nm wide amyloid fibrils above a critical concentration of 2-10% (w/w) (100). Atosiban is also an inhibitor of OT-9 and AVP-9 receptors and is directed to delay premature labor (318).

In Chapter 3, we demonstrated that interaction with lipid membranes significantly impacted the rate of fibrillisaton and fibril morphology for the functional amyloid-forming peptide SST. The lipid composition was shown to modulate this effect, with a surface charge, and lipid chain saturation both playing an important role. Herein, the effect of lipid composition on the formation of fibrils has been investigated for AVP-9 and OT-9, which exhibit a high degree of structural resemblance (319). These peptides are “cyclic” with Cys residue at positions 1 and 6, linked by a disulphide bond and tripeptide tail (21, 319). They are similar to SST-14 in their cyclic structure but differ in their disulphide link (Cys1-Cys6 for OT and AVP; Cys3-Cys14 for SST) and the fact that they consist of a smaller number of amino acids overall. The peptide cyclicity may play an important role in the interaction of these peptides with the lipid membrane and their subsequent aggregation. Saturated lipids with chain length from C12 to C18 were investigated, as well as the singly unsaturated C18 lipid DOPC. The effect of cholesterol and the anionic lipids DMPS and DOPS were also investigated. Results were reasoned in terms of the structure of the peptide, as well as the effect of membrane parameters including membrane surface charge and fluidity, hydrocarbon chain packing, and lipid solubility on the fibrillisation of functional amyloid-forming peptides.

While the effect of lipid membranes on fibrillisation has not been characterised for these peptides, studies carried out by Sikorska et al. and co-workers have confirmed that neurohypophysial hormone-like peptides, do interact with lipid membranes. Electrostatic and hydrophobic interactions between AVP/OT and anionic liposomes composed of a DPPC:DPPG mixture were found to determine the overall conformation of the peptide (313). Both AVP and OT showed no change in conformation in water however, in a liposome (DPPC:DPPG) an increase in β-sheets and reduction in disordered, random coil structure was reported (313).

4.2 Materials and Methods 4.2.1 Liposome preparation The preparation and analysis are described in Chapter 2, section 2.2.1.

4.2.2 Preparation of Functional amyloid-forming peptides (FAFP’s) 150 mM NaCl was added to a weighed fraction of OT and AVP. The solution was vortexed for a few seconds for smaller aliquots or a minute for bigger aliquots to produce solutions having a final concentration ranging from 0.01% w/w to 5% w/w.

4.2.3 Thioflavin T (ThT) Binding Assay The concentration of AVP and OT in the buffer is described in 4.2.2, and in liposomes, 0.5% w/w OT and 5% w/w AVP was used (preparation and analysis described in Chapter 2, section 2.2.1). ThT assay protocol is described in Chapter 2, section 2.2.2.

4.2.4 Atomic Force Microscopy (AFM) 0.5 % (w/w) solution of OT fibrils or fibrils + liposomes (OT + DMPC) and 5 % (w/w) solution of AVP fibrils or fibrils + liposomes (AVP + DMPC+DMPS) at ratio (1:97). Imaging protocol is described in Chapter 2, section 2.2.4 and statistical analysis not possible with a low number of fibril (<500).

4.2.5 Circular Dichroism (CD) A weighed fraction of 0.1% w/w or 1 mg mL-1 AVP or OT was dissolved in buffer (MilliQ water). An equimolar (1:1) mixture of both peptides at similar concentration was also studied. The peptides were co-incubated in liposomes (DOPC, DOPC/Chol, DOPC/DOPS, DLPC, DMPC, DPPC and DSPC) at an appropriate ratio (1:97) with a slight change to the re-hydration buffer and are described in Chapter 2, section 2.2.1. The protocol for CD measurements and spectra analysis are described in Chapter 2, section 2.2.5.

4.3 Results 4.3.1 Effect of the lipid bilayer on the secondary structure of cyclic neurohypophysial hormones The secondary structure of the cyclic nonapeptides OT-9 and AVP-9 at 1 mg mL-1 (0.1% w/w) was determined using a benchtop CD in the presence and absence of liposomes. An equimolar (1:1) mixture of both peptides was also studied, as they are known to both originate from the same secretory granules. Spectra for all three samples in the absence of liposomes are presented in Figure 4.2. The spectra for both AVP and OT are characterised by a large negative band at 195nm and 198nm, respectively, consistent with a disordered, random coil structure in agreement with a recent study on these peptides in water (313).

The CD spectrum of an equimolar ratio of AVP and OT (1:1) is also dominated by a negative band around 195 nm, consistent with a disordered structure, indicating that the conformation of the peptides does not change significantly on mixing. The estimation of secondary structure for these peptides is complicated due to limitation of the UV range not accessible in most bench-top CD compared to synchrotron CD (chapter 3). Therefore, dichroweb analysis had resulted in large errors in data-fitting (NRMSD > 0.250).

Figure 4.2. Secondary structures of OT-9, AVP-9 and AVP:OT (1:1) at 1 mg mL-1 analysed using Circular Dichroism (CD)

The CD spectra of OT following incubation with liposomes based on saturated chain lipids are presented in Figure 4.3 after 3 hours incubation (Figure 4.3a) and 72 hours incubation (Figure 4.3b). No significant changes to the CD spectrum observed for OT-9 in water were observed following incubation of up to 72 hours with liposomes indicating either that the peptide has limited interaction with the lipid bilayer, or that this interaction does not result in a conformational change of the peptide.

The CD spectra of OT following incubation with DOPC-based liposomes of various compositions are presented in Figure 4.3 after 3 hours incubation (Figure 4.3c) and 72 hours incubation (Figure 4.3d). Changes to the CD-spectra were observed for the majority of samples consistent with changes to the secondary structure following incubation with the lipid membrane. Following 3 hours incubation with DOPC, and mixtures of DOPC with Chol or DOPS, the minimum at 198 nm becomes less intense and moves to a higher wavelength (~200 - 204nm), consistent with an increase in β-sheets and an associated reduction in the proportion of the structure that is the random coil. The effect is dependent on the lipid composition and is more pronounced for the lipid bilayer doped with the anionic lipid DOPS, presumably due to an electrostatic interaction with the positively charged OT peptide. For charged liposomes, these effects appear to become more pronounced with time (Figure 4.3d). A similar effect was noted for the addition of OT to charged DPPC-DPPG liposomes and was also correlated with an increase in β-sheets at the expense of random coil structure (313).

a) b)

c) d)

Figure 4.3. (a) CD spectra for 1 mg mL-1 (0.01% w/w) OT-9 co-incubated with saturated lipid vesicles after 3 hours, (b) after 72 hours, (c) CD spectra for 1 mg mL-1 OT-9 co-incubated with vesicles doped with various lipid compositions after 3 hours and (d) after 72 hours

Similar to results obtained with OT-9, co-incubation of AVP-9 with saturated chain lipids did not result in any significant changes to the CD spectra even at 72-hour period, again suggesting no conformational changes to the peptide after interaction with the saturated chain membrane (Figure 4.4a-b).

Upon three hours incubation of AVP-9 with DOPC-based liposomes of various compositions, a slight shift in the CD-spectrum was observed for all samples, (Figure 4.4c). The minimum at 198 nm becomes more intense and moves towards higher wavelength (~200nm). A second minimum at 205 -210 nm was observed, particularly for DOPC-Chol and DOPC-DOPS samples, which is consistent with an increase in β-sheets and the associated reduction in the proportion of structure that is a random coil as previously observed for AVP upon interaction with charged liposomes (313). This observation is in agreement with the known effect of cholesterol and charge on some toxic amyloid-forming peptides; both have been shown to

induce conformational changes, e.g. in α-synuclein (284, 320). Although the effect becomes slightly more noticeable with time, particularly for cholesterol doped liposomes, (Figure 4.4d), in general, the conformation of AVP is not significantly affected by interaction with uncharged liposomes. A similar effect was noted by Sikorska et al. (313), who found that the conformation of AVP did not change following interaction with liposomes. An NMR study by this group suggested that the positively charged Arg residue at position 8 of the AVP peptide promotes the same extended conformation of the C-terminus regardless of whether the peptide is in solution or in SDS micelles (321). A similar effect has been reported upon transfer of AVP to hexafluoroacetone (structure-stabilising co-solvent) (322, 323).

a) b)

c) d)

Figure 4.4. (a) CD spectra for 1 mg mL-1 (0.01% w/w) AVP-9 co-incubated in saturated lipid vesicles after 3 hours, (b) after 72 hours, (c) CD spectra for 1 mg mL-1 AVP-9 co-incubated in vesicles doped with various lipid compositions after 3 hours and (d) after 72 hours

Since AVP and OT are found in the same secretory granules, the peptide conformation for the equimolar mixture was investigated in lipid vesicles (99). Upon incubation, we note no change in conformation for (AVP:OT) in saturated lipid vesicles, (Figure 4.5a or b), or unsaturated DOPC liposomes of various compositions, (Figure 4.5c or d) up to 72 hrs incubation.

a) b)

c) d)

Figure 4.5. (a) CD spectra for 1 mg mL-1 (0.01% w/w) AVP:OT (1:1) co-incubated in saturated lipid vesicles after 3 hours, (b) after 72 hours, (c) CD spectra for 1 mg mL-1 AVP:OT (1:1) co- incubated in vesicles doped with various lipid compositions after 3 hours, (b) after 72 hours

4.3.2 Effect of the lipid bilayer on the kinetics of fibril growth of OT-9 The kinetics of aggregation was characterised using a ThT fluorescence assay for OT and AVP in the presence and absence of liposomes of various compositions. The increase in ThT relative fluorescence with time is plotted in Figure 4.6 for OT-9 dissolved in 150 mM NaCl at various concentrations. No lag phase was observed however, a strong growth phase was observed from t=0. This growth phase can be used to calculate the rate constant for aggregation (kagg) by fitting a mono-exponential decay to the data (98).

In the absence of the lipid bilayer, ThT binding was detected for OT-9 at all concentrations studied (0.1 – 5% w/w) (Figure 4.6a), suggesting that OT-9 readily aggregates into -sheet rich structures. This is interesting when compared to SST-14 (Chapter 3), where aggregates were only observed above a critical concentration of 2.5% (w/w). A decrease in half-time and increase in kagg with increasing concentration was observed (Table 4.1), although a lag phase was not observed at any concentration.

The evolution of ThT relative fluorescence with time for OT-9 0.5% (w/w) incubated with liposomes of various lipid compositions, including saturated chain lipids of various lengths (DLPC (C12); DMPC (C14); DPPC (C16) and DSPC (C18)) and the unsaturated lipid DOPC (C18:1) is presented in Figure 4.6b. Co-incubation of OT-9 with liposomes formed from saturated chain lipids resulted in a significant increase in the intensity of the fluorescence plateau phase, relative to aggregation in buffer (Figure 4.6b), indicating an increase in the amount of beta-sheet rich assemblies formed. The final fluorescence intensity of plateau was significantly higher in the presence of longer chain lipids (DPPC/ DSPC) compared to shorter chain lipids (DLPC/DMPC) suggesting that gel-phase (DPPC/DSPC) promotes aggregation of this peptide more than fluid-phase (DLPC/DMPC) bilayers. This observation contrasts with that seen for SST-14, where aggregation was increased in the presence of shorter-chain lipids.

In contrast, the presence of liposomes formed from the unsaturated DOPC did not have a significant effect on aggregation. The final fluorescence intensity was similar to that observed in the buffer, although the kagg value for the growth phase had increased (Table 4.2) and was similar to that seen for some saturated chain liposomes (Figure 4.6c).

The effect of cholesterol (doped at 30mol% within DOPC and DMPC bilayers) was also investigated. Addition of cholesterol did not significantly affect the kagg values (Table 4.2), although a slight increase in fluorescence intensity was observed in the presence of cholesterol in both cases (Figure 4.6c). Addition of cholesterol was observed to have a similar effect on fibrillisation of SST-14.

As for SST-14, the addition of lipids having a charged PS headgroup to the bilayer had the greatest effect on aggregation. The intensity of the fluorescence plateau was highest in the presence of charged liposomes, indicating that the vesicles containing lipids with a charged headgroup promoted the formation of −sheet containing assemblies to a greater extent than another lipid compositions tested. Interestingly, the fluorescence intensity (Figure 4.6d) are similar for charged liposomes formed from either DOPC or DMPC suggesting that the effect

of charge is dominant, and counteracts the observed effect of unsaturated liposomes to either slow down or have no effect on fibril growth. This observation is consistent with our results for SST-14 in Chapter 3, as well as the known effect of charge on the fibrillisation of toxic amyloids (5) and the recently studied functional amyloid TasA (8).

Figure 4.6. (a) Thioflavin T (ThT) assay for OT-9 in buffer over a 1-week period, (b) ThT assay for OT-9 at 0.5% (w/w) co-incubated with unsaturated (DOPC) and saturated (DLPC, DMPC, DPPC and DSPC) vesicles with increase in chain length, (c) ThT assay for OT-9 at 0.5% (w/w) co-incubated with unsaturated and saturated vesicle doped with cholesterol and (d) ThT assay for OT-9 at 0.5% (w/w) co-incubated with unsaturated and saturated vesicle doped with charged lipids (DMPS/DOPS)

Table 4.1. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for OT-9 in the buffer over 1 week

2 -1 1 OT % Lag Phase t50 or (t1/2) kagg (s ) Final fluorescence (w/w) (hours) (hours) intensity of plateau phase (x106) 0.1 0 5.4 0.11 6.9 0.5 0 5.0 0.14 12.7 5 0 0.5 0.69 20.1 One-phase association1 and half-time2

Table 4.2. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for OT-9 at 0.5% (w/w) in buffer and vesicles doped with various lipid compositions over 1 week

2 -1 1 Conditions Lag phase t50 or (t1/2) kagg (s ) Final fluorescence (hours) (hours) intensity of plateau phase (x106) NaCl 0 0.5 0.1401 12.7 Unsaturated DOPC 0 0.9 0.2831 14.3

Effect of chain length or bilayer thickness DLPC 0 0.5 0.3471 21.3

DMPC 0 0.5 0.3741 21.4

DPPC 0 0.5 0.3321 27.0

DSPC 0 1.0 0.3761 26.3

Effect of Cholesterol 70DOPC+30CHOL 0 0.2 0.2491 16.1

70DMPC+30CHOL 0 0.5 0.2181 24.6 Effect of Charge 65DOPC+35DOPS 0 0.5 0.2581 24.6

65DMPC+35DMPS 0 0.5 0.3591 23.7

One-phase association1 and half-time2 (324)

4.3.3 Effect of the lipid bilayer on the kinetics of fibril growth for AVP A similar study was carried out on AVP-9, which differs from OT-9 only in the amino acids at positions 3 and 8 (Figure 4.1).

Table 4.3. Shows the position, polarity, charge and hydrophobic amino acid residues for OT and AVP

OT-9 AVP-9 Position 3 8 3 8 Amino acid residue Phe Arg Ile Leu Polarity Non-polar Polar Non-polar Non-polar Charge Neutral +ve Neutral Neutral Hydrophobic side- Yes No Yes Yes chain

For AVP-9 in solution, overall low fluorescence intensity was observed over 1 week, (Figure 4.7a), even at the highest concentration of 5% (w/w) suggesting that no appreciable aggregation occurred over this time period.

Following incubation of AVP-9 with liposomes formed from saturated chain lipids, no significant changes in fluorescence intensity were observed (Figure 4.7b). A small increase in fluorescence intensity was observed for the longest chain lipids (DPPC and DSPC) suggesting that a small number of −sheet containing assemblies may have formed for these samples. A similar effect was noted for liposomes formed from the unsaturated lipid DOPC, and mixtures of DOPC and Chol (Figure 4.7c).

However, upon addition of 35 mol% of the charged lipid DMPS to DMPC liposomes, an appreciable increase in fluorescence was noted over approximately 1 week indicating that the presence of the charged liposome has accelerated aggregation for this peptide (Figure 4.7d). Interestingly, a similar effect was not observed for charged liposomes formed from a mixture of DOPC and DOPS. We note that the intensity of the fluorescence plateau and kagg values (Table 4.3) are significantly lower compared to those for OT-9 (Table 4.2).

Figure 4.7. (a) ThT assay for AVP-9 in buffer over a 1-week period, (b) ThT assay for AVP-9 at 5% (w/w) co-incubated with unsaturated and saturated vesicle with an increase in chain length, (c) ThT assay for AVP-9 at 5% (w/w) co-incubated with unsaturated and saturated vesicle doped with cholesterol and (d) ThT assay for AVP-9 at 5% (w/w) co-incubated with unsaturated and saturated vesicle doped with charge

We further investigated the effect of bilayer charge by varying the percentage of the charged lipid DMPS from 5 to 35 mol% in the DMPC bilayer. A characteristic growth curve consistent with amyloid fibril formation was observed at all levels of charge (Figure 4.8). A consistent increase in final fluorescence intensity of the plateau phase with increasing % of DMPS was observed, suggesting that more fibrils are formed as the bilayer charge increases. In addition, -1 -1 kagg values increase slightly 0.020 s (5 mol% DOPS) to 0.032 s (35 mol% DOPS) (Table 4.4).

Figure 4.8. ThT assay for AVP-9 at 5% (w/w) co-incubated with saturated vesicle doped with an increase in the concentration of the charged lipid DMPS (5 – 35 mol%)

Table 4.4. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for AVP-9 at 5% (w/w) in DMPC vesicles doped with DMPS at (5- 35 mol%)

2 -1 1 Conditions Lag Phase t50 or (t1/2) kagg (s ) Final fluorescence (hours) (hours) intensity of plateau phase (x106) DMPC 0 0 0 0.007 95DMPC+5DMPS 0 36.5 0.0200 0.047 85DMPC+15DMPS 0 26.8 0.0222 0.066 75DMPC+25DMPS 0 14.8 0.0281 0.12 65DMPC+35DMPS 0 3.76 0.0325 0.16 One-phase association1 and half-time2

4.3.4 AFM a surface technique used to study aggregate deposited from solution both in the absence and presence of liposomes AFM analysis was carried out on OT fibrils formed in the presence or absence of DMPC lipid vesicles. For OT fibrils formed in the buffer, a small number of short fibrils were observed via AFM (Figure 4.9). Such fibrils contrast with those formed by SST in the buffer, which is much

longer and more numerous and display the typical amyloid-like unbranched twisted fibrillar structures.

Upon co-incubating OT with DMPC vesicles, large amorphous aggregates are detected (Figure 4.10), which we believe are due to co-assemblies of lipids and a non-fibrillar beta sheet containing OT assemblies that form planer films upon adsorbing to the positively charged mica substrate. No fibrils could be detected at this concentration (0.5% w/w).

Figure 4.9. Topographic images of (a) short OT fibrils at 0.5% (w/w) in aqueous solution and (b) image of a single OT fibril

Figure 4.10. Topographic image showing no OT fibrils detected with DMPC

No fibrils were detected in the AFM analysis on AVP in buffer (Figure 4.11a) in agreement with the ThT assay results. As for the OT samples, large planar films were also detected for AVP co-incubated with charged DMPC+DMPS vesicles (Figure 4.11b). Like in the case of OT we suggest these films are due to co-assemblies of non-fibrillar aggregates of AVP and anionic lipid vesicles which have been converted to planar films due to strong electrostatic interactions with the cationic mica substrate.

Figure 4.11. Topographic images of (a) 5 % (w/w) AVP in aqueous solution and (b) AVP incubated with DMPC+DMPS vesicles

4.4 Discussion In this chapter the effect of lipid membranes on the aggregation of two cyclic peptides, OT-9 and AVP-9 was investigated. Despite differing only in the amino acids at position 3 and 8 (Table 4.3), aggregation of these peptides was markedly different. ThT assay results suggest that OT aggregates readily at all concentrations studied, with fast aggregation rates and no lag period observed. In contrast, no appreciable increase in fluorescence intensity was observed for AVP, except in the presence of highly anionic liposomes.

The significant increase in propensity for aggregation for OT relative to AVP may reflect the phenylalanine (Phe) amino acid at position 3, compared to Isoleucine (Ile) for AVP. Previous studies have shown that the substitution of a single amino acid can have a significant effect on

the aggregation propensity of a peptide. Chiti et al. have demonstrated that aggregation rates can span a range of up to 1000 following an amino acid mutation, with aggregation rates up to 30 times faster and 15 times slower than the wild-type peptide observed (325). The effects of a specific mutation were shown to correlate strongly with the hydrophobicity and charge of the peptide, as well as the resultant secondary structure (325). Here, the substitution of the Phe residue can promote π – π stacking interactions, and leads to increased hydrophobicity of the peptide, driving aggregation. Furthermore, Phe is known to be critical amyloid forming amino- acid, and the short Phe-Phe sequence is known to be the minimal amyloid-forming peptide sequence within Aβ (326).

AFM data indicate that neither of these peptides forms the typical amyloid-like twisted fibrillar structures, with only a few short fibril-like structures observed for OT, and no aggregates observed for AVP. This is consistent with a study by Beuret and co-workers, who were unable to stain pro-AVP-9 aggregates in the secretory granules of the endoplasmic reticulum (ER) (327). It is possible that the AFM washing steps used, typical for the formation of large fibrils, may have washed off the smaller AVP fibrillar aggregates. Further AFM studies are therefore required to confirm the morphology of the beta-sheet rich aggregates observed in the ThT assays. In contrast, numerous fibrils were observed via AFM for SST, and these had the characteristic amyloid-like unbranched twisted fibrillar structures. Both the fluorescence plateau value and the rate constant for aggregation were significantly higher for OT than for SST at the same concentration (5% w/w). In addition, no lag period was observed for OT is probably indicative of a lack of nucleation phase (transition from monomers to oligomers), while a lag period of approximately 12 hours was observed for SST. Typically, amyloid kinetic growth curves possess a lag phase whereby the concentration of non-fibrillar, but insoluble, oligomers are increasing to a sufficient level where it can seed the exponential growth of fibrillar structures (328). The lack of observable lag phase, high ThT plateau, and absence of fibrils observed in the AFM images suggests a rapid formation of many beta-sheet containing aggregates of OT that are not capable of seeding the exponential growth of fibrillar structures. This is presumably due to increased solubility of these OT aggregates. Such aggregates have previously been hypothesised as an off-pathway product of the amyloid fibril mechanism (329, 330).

The lack of OT fibrils seen here by AFM, despite the relatively high intensity of the ThT assay fluorescence plateau, suggests that the dominant OT aggregates formed, are non-fibrillar beta- sheet containing assemblies (ThT is sensitive to all β-sheet containing assemblies, not just

fibrillar structures). Despite the absence of fibrils, the effect of the lipid bilayer on the rate and extent of aggregation for OT was qualitatively similar to that observed for SST in the previous chapter. Bilayers formed from saturated chain lipids appeared to promote aggregation for OT, with an increase in the rate constant for aggregation and the intensity of the ThT fluorescence plateau for saturated chains from C12 to C18. This is qualitatively similar to the behaviour of SST in the presence of liposomes. However, for SST, the short-chain lipids DLPC and DMPC had the greatest effect on fibril growth, while for OT, the longer-chain lipids DPPC and DSPC promote aggregation more. While saturated chain lipids did not significantly enhance aggregation of AVP, we do, however, note that the longer-chain lipids also appeared to have the most promoting effect on aggregation for AVP. A previous study on Aβ (1-42) has also suggested that this peptide interacts more with gel-like bilayers (including DPPC), thereby enhancing aggregation (331). This may suggest that the mechanism of aggregation may differ depending on whether fibrils or amorphous aggregates are formed. The increase in aggregation for AVP and OT appears to be driven by a general increase in the interaction of the peptide and the bilayer. In contrast, the typical amyloid-fibril formation mechanism observed for SST may involve a reciprocal transfer of lipids from the membrane to the growing amyloid (37).

For OT and AVP, bilayers formed from the unsaturated chain lipid DOPC did not promote aggregation. A similar effect was observed for SST and appeared to be strongly correlated with the depth of penetration of the peptide into the bilayer, as determined via intrinsic tryptophan fluorescence from the Trp residue in SST. While a similar analysis was not possible for OT or AVP due to the absence of the Trp amino acid in both peptides, it is likely that a similar effect occurs for these peptides. In general, insertion of a peptide having some hydrophobic character into a lipid bilayer can be modified by the lateral pressure profile in the bilayer. Highly branched, or saturated hydrocarbon chains, are associated with higher lateral pressure in the bilayer which can prevent insertion of the hydrophobic segment of the peptide into the bilayer (332). The reduced lateral pressure associated with the DOPC bilayer can result in deeper penetration of the peptide, and this was confirmed for SST. This indicated that aggregation is promoted by a more peripheral association of the peptide with the lipid membrane. This association does not seem to promote a significant conformational change in either OT or AVP. This is in contrast to a previous study on interactions between OT and AVP, and their receptors, which suggested that the peptide only adopts the bioactive conformation required for receptor docking following association with the membrane (313). There is no evidence. Therefore, that membrane-associated conformational change drives fibril growth. Rather, the data seem to

point towards the “molecular crowding” mechanism where a high local concentration of peptides at the membrane surface may be driving the increase in the rate of aggregation.

For all three cyclic peptides studied to date, the addition of anionic lipids to the bilayer drives the greatest increase in rate consistent with a strong electrostatic interaction between the cationic peptide and the anionic bilayer. A similar observation is seen with AVP and the charged bilayer DPPC:DPPG (309, 313). The secondary structure of OT-9 also modified in the presence of DOPC/DOPS liposomes, with a transition from random coil towards β-sheets observed. Some studies suggest that this reflects the tyrosine (Tyr) side chain being more exposed to the outside environment (313, 333).

In summary, significant differences were noted in the aggregation of all three cyclic peptides investigated in Chapters 3 and 4, despite two of these peptides having a similar amino acid sequence, structure and location within the secretory granules. Previous studies have shown that transplanting a small sequence stretch of six amino acids into a non-amyloid forming protein, can result in a protein which is capable of forming amyloid fibrils (334). Herein, a variation of a single amino acid has been shown to have a marked influence on the ability of a peptide to form amyloid fibrils or other assemblies. This is consistent with single homomutants of the amyloid-forming protein α-synuclein being responsible for significant changes in aggregation, toxicity and mortality in familial cases of Parkinson’s disease (335, 336). This change does not seem to be influenced by interactions with the lipid membrane, which are similar for both peptides. We suggest that the role of the membrane may be to speed up the rate of aggregation by bringing the peptides into contact on the lipid membrane surface. However, intrinsic peptide-peptide interactions still appear to fundamentally drive aggregation or otherwise.

CHAPTER 5

5. Effect of lipid composition on fibrillisation in linear functional amyloid- forming peptides, Substance P-11 (SP-11) and Deslorelin-10 (Des-10)

5.1 Introduction In the previous two Chapters, I have shown that interaction with the lipid bilayer drives fundamental changes in the kinetics of amyloid formation and the aggregate morphology for three functional amyloid-forming peptides, somatostatin, oxytocin and vasopressin. All of these peptides have a cyclic structure, which has been shown to stabilize certain peptides (337, 338). In contrast, linear peptides have been reported in some cases to adopt a more flexible conformation (339). While the driving force for self-assembly of linear and cyclic peptides are mainly hydrophobic and π-π stacking interactions between tryptophan residues (339, 340). Fundamental differences in linear and cyclic peptides may affect their aggregation and fibrillisation. For example, a TEM study by Choi et al. and co-workers demonstrated that linear peptides were inclined to aggregate into irregular objects, while cyclic peptides preferred to form stable regular spheres (339).

In this chapter, I have investigated the effect of the lipid bilayer on the fibrillisation of two linear peptides, SP-11 (SP), and Des-10 (Des), Figure 5.1. The 11 amino-acid peptide, SP, is a linear peptide (341) that belongs to a neuropeptide group called tachykinins, (254, 255) (Figure 5.1). It has a cationic charge of +3 at neutral pH (256). Its activity spans across the central and peripheral nervous system (257); it is involved in different functions including pain transmission (254), regulation of blood flow (258) and gastrointestinal motility (259). SP may have neuroprotective properties and play a role in inhibiting amyloid formation for the Aβ peptide in AD, as well as preventing brain ageing-related memory decline (255, 260-262). Reduced levels of SP have been observed in the brain and spinal fluid of Alzheimer’s disease (AD) patients (35). Administration of SP was found to reduce the potassium channels overexpressed by Aβ in a rat model of AD (35). In addition, Kowall and co-workers reported that SP, co-administered with Aβ peptide into the cerebral cortex of rat, prevented amyloid- induced neuronal loss (342).

a) SP-11

b) Des-10

Figure 5.1. An illustration of the linear structure and amino acid sequence of (a) SP-11 and (b) Des-10 (adapted from (22) (23))

SP is probably aggregated in vivo within storage granules (343). Aggregation of SP has been reported by Choo et al. at concentrations > 10 mg ml-1 (344). NMR studies also demonstrated aggregation of SP at both basic and acidic pH (104). In contrast, Chassaing and co-workers found that aggregation only occurred at neutral pH in the presence of NaCl (345). It was reported to aggregate into larger fibrils which were detected by electron microscopy (346). Recently, a structural study of the SP assembly process (using a combination of SAXS and FTIR) by Dharmadana and co-workers (Dharmadana et al., 2019, manuscript in preparation) has demonstrated that SP can self-assemble into nanotubes.

In solution, SP has been reported as being reasonably disordered, (254, 347-349). It has been suggested that the presence of two proline (pro) residues at positions 2 and 4 does not allow long-range hydrogen bonds, thereby promoting a random-coil conformation for this peptide in solution (349). However, a range of secondary structures has been reported for SP, from unfolded to aggregates or β-sheets (254). X-ray and neutron diffraction studies have also revealed β-sheet structures for aggregated SP (263). CD studies have shown that SP can adopt a partial α-helix in the presence of amphipathic mediators with negatively charged head groups; no helix has been reported in the presence of positively charged head groups (254, 256, 301). The interaction of cationic peptides with membranes or micelles is greatly dependent upon the charge distribution along the peptide chain and nature of the headgroup (350). In fact, this has been claimed to be an essential prerequisite for the biological activity of SP (351, 352).

Des-10 (Des) is a decapeptide synthetic analogue of Luteinising hormone-releasing hormone (LHRH), which is also a linear peptide (253). Des-10 consists of modified amino acid residues at positions 6 (replacing Gly with D-Trp) and at position 10 swapping Gly with ethylamide (Figure 5.1). Synthetic Des-10 is 144 times more potent that native LHRH (353), whose short half-life (~hours) prevents its use in targeting cancer cells (354, 355). Des-10 is stated to induce

ovulation by targeting the pituitary-gonadal axis, which regulates the follicle stimulating hormone (FSH) and luteinising hormone (LH) (264). (264, 265). Des also suppresses testosterone which leads to a reduction in sperm count and eventually deteriorating the testes (102, 356).

The secondary structure of Des is reported to be disordered (265), with no evidence of β-sheets or aggregates reported to date. While no aggregation has been investigated for Des, Triptorelin, which is another LHRH analogue with a highly similar sequence to Des, has been reported to form small nanotubes at low pH or large nanotubes at high pH (24).

The lack of secondary structure in several peptide hormones, including SP-11and Des-10, has led to the suggestion that the peptide must undergo some conformational change to allow it to be recognised by the receptor; i.e. the conformation of the peptide in solution may not be the active conformation in vivo. It has been previously observed that interaction with a lipid bilayer may induce a more biologically relevant conformation for SP (357). Some studies have indicated that SP aggregates in contact with a lipid bilayer, resulting in a loss of penetrability of the lipid vesicles (256, 358).

In this chapter, I have therefore investigated both the secondary structure and the aggregation process for SP and Des, in the presence of lipid membranes (liposomes of various lipid compositions). Saturated lipids with chain length C12 to C18, as well as singly unsaturated C18 DOPC were investigated. The effect of cholesterol and the anionic lipid DMPS and DOPS were also studied. Based on results from Chapters 3 and 4, the following membrane parameters were investigated: membrane surface charge, bilayer thickness and fluidity, the degree of unsaturation in the bilayer and addition of cholesterol. To the best of our knowledge, information is limited on linear peptides aggregating or forming amyloids upon interacting with eukaryotic membrane or model systems.

5.2 Materials and Methods 5.2.1 Liposome preparation Liposomes were prepared as described in Chapter 2, section 2.2.1.

5.2.2 Preparation of Functional amyloid-forming peptides (FAFP’s) 150 mM NaCl was added to a weighed fraction of SP and Des. The solution was vortexed for a few seconds for smaller aliquots or a minute for bigger aliquots to produce solutions having a final concentration ranging from 0.01% w/w to 10% w/w.

5.2.3 Thioflavin T (ThT) Binding Assay ThT assay was carried out as described in Chapter 2, section 2.2.3.

5.2.4 Atomic Force Microscopy (AFM) 10 % (w/w) solution of SP fibrils or fibrils + lipid (SP + DMPC) and 10 % (w/w) solution of Des fibrils or fibrils + lipid (Des + DMPC) at ratio (1:97). Imaging was carried out as described in Chapter 2, section 2.2.4 and statistical analysis not possible with a low number of fibril (<500).

5.2.5 Circular Dichroism (CD) A weighed fraction of 0.03% w/w or 0.3 mg mL-1 Des-10 was dissolved in MilliQ water (buffer), and liposomes at ratio (1:97) were prepared as described in Chapter 2, section 2.2.1 with a slight change to the re-hydration buffer, Milli-Q water was used instead of 150 mM NaCl. The measurements and analysis are described in Chapter 2, section 2.2.5.

5.2.6 Synchrotron Circular Dichroism (SR-CD) (ASTRID II facility in Aarhus Denmark) sample measurements and characterisation SRCD sample preparation for SP-11 is followed as described in 5.2.5. The measurements and analysis are described in Chapter 2, section 2.2.6.

5.3 Results 5.3.1 Interaction with the lipid bilayer changes the secondary structure of SP-11 and Des-10 Substance P

The secondary structure of SP-11 at 0.3 mg mL-1 (0.03% w/w) was studied using SRCD in the presence and absence of liposomes. The CD spectrum for SP-11 in water (Figure 5.2) contains a double minimum at 200 nm and 193 nm, consistent with a secondary structure which is a mostly random coil (unfolded), with some beta-sheet structures. This was confirmed by dichroweb analysis on this sample which yields 72% random coil and 26% β-sheets and turns (Table 5.1), in agreement with previous FTIR and CD studies on SP-11 (254, 359) (360) (290, 361). (Note that SP concentrations used for FTIR studies are one or two orders of magnitude more than those used for CD experiments).

Figure 5.2. Secondary structures of SP-11 at 0.3 mg mL-1 (0.03 % w/w) analysed using Synchrotron Circular Dichroism (SR-CD), and Des-10 at 0.3 mg mL-1 analysed using Circular Dichroism (CD)

The CD spectra of SP-11 following three hours incubation with liposomes of various compositions is shown in Figure 5.3a. Dichroweb analysis, which was only possible for selected samples due to large errors in data-fitting (NRMSD > 0.250) is provided in Table 5.1. For all uncharged liposomes, a decrease in the intensity of the minimum at 200 nm was observed, with a concomitant increase in the small maximum at 193 nm, consistent with an increase in β-sheets structures. Dichroweb analysis (Table 5.1) confirmed an increase in the percentage of β-sheets and turns, to approximately 50-60%, and a corresponding decrease in the percentage of unordered structure from approximately 70% to approximately 30-40%. No major differences were observed between saturated and unsaturated lipids or following the addition of cholesterol. The change in secondary structure is most pronounced for the charged liposome (DOPC doped with 35 mol% DOPS) (Figure 5.4a). Dichroweb analysis suggests that SP forms a partial alpha-helix following interaction with the anionic bilayer. This is in agreement with previous CD studies demonstrating that SP forms a partial helix in the presence of negatively charged amphiphiles including SDS micelles (254), and lipids having anionic PG or PS headgroups (254, 346), No significant increase in α-helical structure has been reported in the presence of positively charged headgroups (343). Similarly, Shang et al. and co-workers have also reported this observation of conformational change to α-helix using CD (362).

After 72 hours, it is indicative that a slight increase in the positive maximum at approximately 195 nm is observed for SP incubated with DOPC+DOPS (Figure 5.3b). This can also be confirmed from dichroweb analysis, which shows an increase in % α-helical structure (Table 5.1), suggesting that with time, the α-helix structure is favoured. This observation has previously been reported for α-synuclein with SDS (363).

a) b)

Figure 5.3. (a) SR-CD spectra for SP-11 at 0.3 mg mL-1 (0.03 % w/w) co-incubated in vesicles after 3 hours, (b) SR-CD spectra for SP-11 co-incubated in vesicles after 72 hours

Table 5.1. Secondary structure of SP-11 in liposomes analysed using Dichroweb with Contin- LL algorithm method

Helix Sheet Turns Unordered Samples NRMSD Ref Set % % % % After 3 hours Pure SP 3 14 12 72 0.092 7 DMPC+SP 5 28 25 43 0.202 1 DOPC+SP 5 37 25 33 0.259 1 65DOPC:35DOPS+SP 15 16 30 40 0.222 1 70DOPC:30CHOL+SP 6 38 23 33 0.213 1 After 72 hours Pure SP 4 13 14 70 0.098 7 DMPC+SP 5 26 25 44 0.185 1 65DOPC:35DOPS+SP 17 14 29 40 0.231 1 70DOPC:30CHOL+SP 8 36 22 35 0.188 1

Deslorelin

The secondary structure of Des-10 was studied using the bench-top CD in the presence and absence of liposomes, Figure 5.4. Due to large errors in data-fitting (NRMSD > 0.250), dichroweb analysis was not possible for these samples.

The CD spectrum for the peptide in its native conformation (Figure 5.2), has a well-defined minimum at 218 nm, and a double minimum at ~203 nm and ~198 nm, which could suggest a turn or bent-like conformation, along with some random coil. By analogy with Triptorelin, which shares the same sequence as Des but with the addition of a glycine amino acid, a globular conformation was observed at higher pH values (pH > 7.5), while a more extended β-sheets based conformation was observed at pH < 6.5, with a turn. (24) A similar observation has been reported with LHRH (278, 364). Recently, Kapoor et al. and co-workers have reported no major broadening of band or shift in the peak for Des using IR, which indicates no significant structural perturbation (265).

Following three hours incubation of Des-10 with liposomes with various compositions, changes to the spectra may be observed, (Figure 5.4a). For DOPC liposomes and DOPC/Chol liposomes, the double minimum at ~203 nm and ~198 nm is replaced by a single minimum and moves to higher wavelengths, which may suggest an increase in β-sheet structures. Changes to the DMPC spectrum are more minor, suggesting that this interaction has had less effect on the Des conformation. In contrast, the addition of charged liposomes had the biggest effect, presumably due to electrostatic interactions with the cationic peptide, although the spectrum is complex and difficult to deconvolute.

a) b)

Figure 5.4. (a) CD spectra for Des-10 at 2 mg mL-1 (0.2% w/w) co-incubated in vesicles after 3 hours, (b) CD spectra for Des-10 co-incubated in vesicles after 72 hours

5.3.2 Effect of the lipid bilayer on the kinetics of aggregation for SP-11 and Des-10 The fluorescent dye Thioflavin T (ThT) specifically binds to the β-sheets structures present in amyloid fibrils, resulting in an increase in its fluorescence intensity. ThT fluorescence binding assays were used to monitor the kinetics of aggregation for SP and Des in solution, and in the presence of liposomes. The increase in ThT relative fluorescence with time is plotted in Figure 5.5a for SP at various concentrations in NaCl solution. No increase in fluorescence was detected at any concentration studied, indicating that aggregation into β-sheet rich structures was not observed.

The increase in ThT relative fluorescence for SP (5 mol%) incubated with liposomes of various lipid compositions, including saturated chain lipids of various lengths (DLPC (C12); DMPC (C14); DPPC (C16) and DSPC (C18)) and the unsaturated lipid DOPC (C18:1) is shown in Figure 5.5b. An increase in ThT fluorescence, characteristic of the growth of aggregates rich in β-sheets was observed for some samples. An initial lag phase (absent for some samples), characterised by a weak increase in ThT fluorescence, was followed by a strong growth phase.

This growth phase can be used to calculate the rate constant for aggregation (kagg) by fitting a mono-exponential function (characteristic of first-order kinetics) to the data, as described in

(98). Values of kagg, t1/2, lag phase and final fluorescence intensity of plateau phase are provided in Table 5.2. Incubation with saturated chain lipids resulted in the formation of aggregates, while the unsaturated lipid DOPC did not cause an increase in fluorescence. The intensity of the fluorescence plateau was highest for the longer-chain lipids DPPC and DSPC, and lower for the shorter-chain lipids, DLPC and DMPC, suggesting that the greatest number of fibrils are formed with DPPC and DSPC liposomes. There appeared to be no correlation between chain length and the value of kagg or t1/2, similar to the trend observed for SST in chapter 3. These results show that the presence of saturated lipids initiates the SP aggregation process.

Figure 5.5. ThT assay performed over 1-week period for (a) SP-11 in NaCl, (b) 10% w/w or 5 mol% SP-11, co-incubated with unsaturated and saturated vesicle with an increase in chain length, (c) SP (5 mol%) co-incubated with DMPC, DMPC doped with cholesterol and (d) SP (5 mol%) co-incubated with DOPC, DOPC vesicle doped with cholesterol and DOPC doped with DOPS

Table 5.2. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SP at 10% (w/w) in NaCl and vesicles doped with various lipid compositions over 1 week

-1 1or2 Conditions Lag phase t50 or (t1/2) kagg (s ) Final fluorescence (hours) (hours)3 intensity of plateau phase (x106) Buffer NaCl No fitting Unsaturated DOPC No fitting Effect of chain length or bilayer thickness DLPC 0 19 0.0371 0.22 DMPC 12 66 0.0182 0.15 DPPC 0 35 0.021 0.58 DSPC 0 26 0.031 0.55 Effect of Cholesterol 70DOPC+30CHOL 0 31 0.0221 0.37 70DMPC+30CHOL No fitting Effect of Charge 65DOPC+35DOPS No fitting 65DMPC+35DMPS 0 13 0.1571 0.61 One-phase association1, Plateau followed by one-phase association2 and half-time3

Addition of cholesterol to the DMPC and DOPC liposomes did not have a consistent effect on fibril growth. A decrease in fluorescence plateau was seen for DMPC/Chol liposomes, relative to pure DMPC (Figure 5.5c), while a slight increase was observed for DOPC/Chol liposomes, relative to pure DOPC.

Addition of the anionic lipid DMPS to the DMPC bilayer resulted in the highest fluorescence plateau value observed suggesting the greatest number of fibrils were formed with this -1 combination (Figure 5.5d). This is also in agreement with the highest kagg (0.157 s ) (Table 5.2). This indicates that electrostatic interactions play a role in the interaction between the highly charged SP (+3) and anionic bilayer. However, the characteristic ThT growth curve was

not observed at early time-points, with a large decrease in fluorescence intensity observed. This seems to be an artefact of the ThT assay for SP, and was observed for samples (e.g., SP in solution) even when the sample did not form appreciable aggregates.

Similarly, a ThT assay was used to detect the formation of β-sheets rich aggregates for Des-10 in the presence and absence of liposomes, Figure 5.6. The fluorescence intensity remained low throughout 1 week for Des-10 in NaCl, suggesting that no aggregation occurred (Figure 5.6a).

The ThT assay for Des-10 incubated with liposomes of various lipid compositions, including saturated chain lipids of various lengths (DLPC (C12); DMPC (C14); DPPC (C16) and DSPC (C18) and unsaturated lipid DOPC is shown in (Figure 5.6b). As for SP, an appreciable increase in fluorescence was observed for saturated chain lipids, while little to no fluorescence was observed in the presence of the unsaturated lipid DOPC. The highest fluorescence of the final plateau phase was obtained for Des-10 with the short-chain lipids DLPC and DMPC (Figure

5.6b). These lipids also had slightly lower (t1/2) or half-time values (20.5 and 23.8 hours), although there was no strong correlation between kagg values and chain length (Table 5.3).

The effect of cholesterol on the DOPC and DMPC lipid bilayers was also investigated. In Des- 10, it can be seen that the final fluorescence intensity of the plateau region is reduced compared to pure DMPC (Figure 5.6c). This can be confirmed from the increase in half-time (23.8 to 30.1 hours) (Table 5.3). This is suggesting that cholesterol has reduced the fibrillisation of Des- 10 compared to pure DMPC. In contrast, the addition of cholesterol to DOPC had little effect on the final fluorescence plateau value.

In contrast to all other peptides studied, the addition of anionic DMPS into DMPC bilayers did not increase fibrillisation kinetics (Figure 5.6d). Rather, this lipid appeared to have an inhibiting effect. In contrast, an appreciable increase in the final fluorescence intensity of the plateau region was observed for unsaturated DOPC doped with DOPS. This also resulted in -1 the highest kagg value (0.089 s ), compared to other bilayers.

Figure 5.6. ThT assay performed over 1-week period for (a) Des-10 in NaCl, (b) 10% w/w or 5 mol% Des-10, co-incubated with unsaturated and saturated vesicle with an increase in chain length, (c) Des (5 mol%) co-incubated with DMPC, DMPC doped with cholesterol and (d) Des-10 (5 mol%) co-incubated with DOPC, DOPC vesicle doped with cholesterol and DOPC doped with DOPS

Table 5.3. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for Des at 10% (w/w) in NaCl and vesicles doped with various lipid compositions over a 1-week period

-1 1or2 Conditions Lag phase t50 or (t1/2) kagg (s ) Final fluorescence (hours) (hours)3 intensity of plateau phase (x106) Buffer NaCl No fitting Unsaturated DOPC 12 95 0.0032 0.083 Effect of chain length or bilayer thickness DLPC 0 21 0.0331 0.40 DMPC 0 24 0.0231 0.34 DPPC 0 28 0.0251 0.12 DSPC 0 28 0.0201 0.14 Effect of Cholesterol 70DOPC+30CHOL 0 46 0.0151 0.084 70DMPC+30CHOL 0 30 0.0291 0.16 Effect of Charge 65DOPC+35DOPS 0 78 0.0891 0.15 65DMPC+35DMPS Fitting not possible One-phase association1, Plateau followed by one-phase association2 and half-time3

5.3.3 Clear morphological differences are observed between aggregates formed in the buffer and those incubated with liposomes Atomic Force Microscopy (AFM) analysis was carried out on SP-11 and Des-10 using deposited layers, (Figure 5.7b and 5.9a-b). In solution, needle-like fibrillar structures were observed for SP, Figure 5.7a. TEM analysis carried out on similar structures (Durga Dharmadna, unpublished PhD thesis, RMIT University) were determined to have a fibrillar morphology, with the lateral alignment of fibrils into bundles. SAXS analysis (Durga Dharmadna, unpublished PhD thesis, RMIT University) demonstrated that these structures have a characteristic nanotube symmetry. In contrast, no aggregate structures were detected for SP-11 co-incubated with DMPC vesicles, (Figure 5.7b) consistent with ThT assay, which displayed relatively weak fluorescence intensity in the presence of DMPC liposomes. Due to the small number of fibrils (<500), fibre app analysis on SP-11 could not be performed.

Figure 5.7. Atomic force microscopy (AFM) analysis of SP fibrils at 10% (w/w) and SP fibrils co-incubated with DMPC lipid vesicles, (a-b) topographic images of (a) SP fibrils displaying needle-like fibrillar structure and (b) SP co-incubated with DMPC vesicles

AFM images of Des-10 in solution show numerous multilaminar crystalline structures in the absence of DMPC vesicles, (Figure 5.8a). The lack of appreciable fluorescence intensity in the ThT assay suggests that these structures have a low binding affinity to ThT dye molecules. These multi-laminar structures can be seen clearly with adhesion and deformation channels. These nanotubes assemblies are also reported with triptorelin (24) as shown in Figure 5.10a. In contrast, in the presence of DMPC vesicles, typical amyloid-like structures with twists can

be seen (Figure 5.9a). These twists can be clearly seen in the deformation channel (Figure 5.9b). As for SP, due to the small number of fibrils (<500), fibre app analysis was not possible.

Figure 5.8. Topographic images of (a) Des fibrils at 10% (w/w) in the absence of DMPC vesicles assemblies can be seen better in the adhesion or (b) deformation channels

Figure 5.9. Topographic images of Des fibrils in DMPC displaying a regular looking amyloid (a) with periodic twisting structure (b) the twists can be clearly seen in the deformation channel

5.4 Discussion In contrast, to the earlier cyclic peptides studied, the formation of amyloid fibril structures was not detected for either SP or Des-10 in solution. ThT assay results did not detect any appreciable increase in fluorescence consistent with the formation of β-sheets rich aggregates, and only a few needle-like fibril structures were detected for SP in solution. In contrast, recent work on SP in NaCl solution has found that this peptide aggregates into nanotubes, with the internal nanotube symmetry, confirmed via SAXS measurements (Durga Dharmadana, Unpublished PhD Thesis, RMIT University).

Addition of liposomes did affect the kinetics of aggregation, with a clear increase in fluorescence for some samples. However, these aggregates were not observed in AFM images. Addition of saturated liposomes tended to increase the aggregation rate, as has been consistently seen for all peptides studied. CD studies suggest that the peptide conformation changes less following interaction with the saturated liposome DMPC than with liposomes based on the unsaturated DOPC lipid. This may suggest that the aggregation mechanism again depends on a more peripheral attachment of the peptide to the membrane surface, with deeper penetration and/or conformational change tending to reduce the tendency to aggregate. This may also explain why the final fluorescence plateau is highest for the longer-chain lipids DPPC and DSPC. Both of these lipids adopt a gel-like bilayer with close packing of the hydrocarbon chains which may prevent the peptide from inserting too deeply into these bilayers.

The lack of interaction between SP and liposomes based on neutral lipids may reflect the high net positive charge of SP (+3), is inducing charge repulsion which can lead to relatively weak hydrogen bonds (349). Keire and Fletcher (1996), also found no change in conformation of SP following interaction with anionic or zwitterionic micelles (365). It has been suggested that SP adopts an extended conformation at the surface of the bilayer, forming hydrogen bonds with its neighbours (254) and such observation is also reported with linear AMPs (366).

AFM results show that Des-10 aggregates into multi-laminar structures in solution, although ThT binding to these aggregates was not observed. Previous studies on the analogue of Des- 10, triptorelin is also reported to self-assemble into nanotubes (24). For triptorelin, at pH values <8, bundles of these nanotubes were observed to laterally associate into fibres (24). The formation of the nanotubes was shown to be highly pH dependent and related to the protonation state of the Histidine residue (pKa = 6.1) of Triptorelin (24). At high ph values (>8) Triptorelin was found to adopt a globular conformation partially stabilised by an H-bond between the

histidine and serine residues, as well as aromatic packing between the histidine residues. At lower pH (<6.5) the histidine residue becomes protonated disrupting these interactions, and promoting the formation of β-sheets, within nanotube bundles. For Des-10, which also contains His and Ser residues in close proximity, similar differences between aggregates may be observed depending on pH. While this was not explicitly investigated during the thesis, at pH values close to neutral, the His residue should be protonated and promote the formation of nanotubes. The aggregates detected for Des-10 via AFM are potentially consistent with bundles of nanotubes although there is no definitive evidence for this.

Interestingly, the addition of liposomes results in a completely different aggregate morphology, as determined via AFM. The twisted aggregates observed are consistent with twisted or helical ribbons. Such structures have been proposed as structural intermediates during the formation of nanotubes, Figure 5.10 (25). It is possible that the addition of lipids could be preventing lateral association of the protofilaments into nanotubes, resulting in the system remaining in the twisted or helical ribbon intermediate stage. This theory is consistent with SST results presented in Chapter 3, where lipids were observed to decorate the amyloid fibrils in a pearl necklace arrangement.

Figure 5.10. Illustrations of (a) an electron micrograph image of triptorelin in water (24) and (b) nanotube self-assembly (25)

100

As for SP, the addition of liposomes had a significant effect on aggregation kinetics, as determined by ThT binding assays. In common with SP, the addition of saturated chain lipids tended to increase the final fluorescence plateau suggesting a greater volume of aggregates were formed. While this effect was consistently seen for 9 duplicate samples it is an anomalous result and at present we do not have an explanation. Although, saturated charged surface (DMPC/DMPS) did not show any appreciable rise in fluorescence intensity, suggests that the interaction is dependent on the headgroup. However, for Des-10, the shorter chain lipids (DLPC/DMPC), which adopt a fluid-lamellar phase at room temperature, appear to promote more aggregation than the gel-phase longer lipids DPPC and DSPC. This is the opposite effect to that observed for SP and suggests that the aggregation mechanism may be distinct for each peptide.

Overall results suggest that while Des-10 and SP do not form typical amyloid-like fibrils, they may form structures consistent with nanotubes observed to self-assemble in analogies of these peptides. Although a typical amyloid formation is not observed, interaction with lipid membranes appears to promote or inhibit aggregation in a similar manner to that observed for amyloid formation, suggesting that these interactions may be independent of the exact mechanistic pathway.

101

CHAPTER 6

102

6. Effects of Stilbenes on functional amyloid-forming peptide Somatostatin-14 (SST-14)

6.1 Introduction Resveratrol (RSV), which is a stilbene and a natural polyphenol (Figure 6.1a) that is found in red wine and several plant species (28, 367), has received significant interest due to its immense pharmacological activities including antineuroinflammatory against neurodegenerative and cardiovascular diseases (139, 249, 368-370).

However, RSV is poorly soluble in water, demonstrating passive diffusion via plasma membranes (371). To overcome this solubility issue, lipid-based carriers like liposomes serve as an excellent vehicle for delivering hydrophobic molecules (28, 372). Lipid membranes and polyphenols are suggested to interact mainly through hydrogen bonding and hydrophobic or electrostatic interactions (373, 374), due to the amphiphilic nature of phospholipids, although the hydrophobic interaction is strongest (375-377). A study conducted by Fei et al. 2018, showed that RSV increases the surface area per lipid, subsequently decreasing its membrane thickness, which is the opposite effect observed with cholesterol (26). Neves et al. 2016, has reported broader Bragg peaks and lower D-spacing or lattice parameter for DMPC (1,2- dimyristoyl-sn-glycero-3-phosphocholine) in the presence of 10 mol% RSV using small-angle x-ray scattering (SAXS) and wide-angle x-ray scattering (WAXS) (375). RSV has a high affinity for the lipid part of the membrane which is determined by its corresponding partition coefficient (LogP) (defined as, the logarithmic ratio of the concentration of solute between two solvents in a biphasic system). Partition coefficient values for RSV in a range of lipid bilayers have been reported as 2.63 in DMPC (378), 3.07 in DPPC (1,2-dipalmitoyl-sn-glycero-3- phosphocholine) and 3.11 in DSPC (1,2-distearoyl-sn-glycerophosphocholine) vesicles (26, 379).

Moreover, lipid biomembranes show mainly two phases, depending on the composition, water content and temperature (380). The fluid lamellar phase (Lα) is known to be the most common lipid arrangement in cells. However, a more ordered or gel phase (Lβ) can also coexist with a fluid phase (381). It is widely accepted that the hydroxyl groups (OH) of RSV is responsible for interacting with head groups of phospholipids (26, 382, 383). Selvaraj et al. 2013, showed that high doses of RSV localise within the outer leaflet of the bilayer (384), although their

103

precise location and orientation remains a mystery. Molecular docking studies have suggested that the ideal orientation of RSV is nearly planar (surface-level) (Figure 6.2) in the bilayer (385), but previous studies have shown that it can also be tilted in the bilayer (386) and others have claimed that it is located at the phospholipid-water interface (383, 387). Numerous literature studies have examined the location of RSV within the model membrane using variety of techniques like differential scanning calorimetry (DSC), nuclear magnetic resonance spectroscopy (NMR), X-ray, neutron diffraction, electron spin resonance spectroscopy (ESR) and molecular dynamic simulations (MDS) (382, 383, 385, 386).

Figure 6.1. Shows the structure of stilbenes (a) RSV (26), (b) TMS (27) and (c) TS (27)

Addition of RSV can impact the physicochemical properties of the bilayer. For example, on DMPC and model membranes consisting of a mixture of phosphatidylcholine (PC), sphingomyelin (SM) and Chol (375, 388, 389). Fei et al. 2018, have suggested that increasing the concentration of RSV (0 to 33 mol%) reduces the chain melting temperature (Tm) of DPPC and DSPC (26). Sarpietro et al. 2007, have also shown similar observation, upon interacting

RSV with DMPC (387). Moreover, a consistent decrease in Tm is also reported with POPC/POPS (1-palmitoyl-2-oleoyl-sn-glycerophosphocholine/phosphoserine) monolayer membranes (390).

RSV

Chol

104

Figure 6.2. Diagram showing RSV location in the bilayer of DPPC/Chol (adapted from (28))

RSV has been shown to inhibit amyloid formation, for Islet Amyloid Polypeptide (IAPP) (391) and Amyloid-beta (Aβ) aggregation (3, 146, 392, 393). Literature information is scarce to confirm whether RSV can protect cells via direct inhibition of the interaction between IAPP toxic species (266, 391, 394, 395) and little is known about the mechanism with which RSV interacts with amyloids. Though some light is shed with present nuclear magnetic resonance (NMR) studies, which has reported that Lysine (Lys) at peptide position 1 and Histidine (His) at peptide position 18 of a nonphysiological analogue of IAPP is involved in the binding of RSV (396). The resonance recorded from NMR spectra revealed the largest change for His-18 occurred during titration with RSV. Therefore, it was concluded that His-18 was critical for RSV-IAPP interaction. However, Hsien et al. 2015 have reported that phenylalanine (Phe) at position 15 also has an important role in the interaction between RSV and IAPP (391). The aromatic rings of polyphenols have been reported to be involved in the binding and interaction with cationic charged sites in IAPP (391).

In this chapter, we compare the effect of RSV, with that of two other stilbene derivatives: trans- 3,5,4’-trimethoxy stilbene (TMS) (Figure 6.1b) and trans-stilbene (TS) (Figure 6.1c), which differ mainly in their functional groups. The topological polar surface areas and octanol-water partition coefficients (obtained using the Molinspiration Cheminformatics (MC) software (397)) for all three polyphenols are listed in Table 6.1.

Table 6.1. Shows the topological polar surface area (tPSA) and partition coefficient (clogP) values for stilbenes

Stilbenes tPSA (A2) clogP RSV 60.68 2.99 TMS 27.70 4.59 TS 0.00 4.50

TMS has demonstrated to be more potent than RSV, concerning its anticancer and anti- inflammatory properties in many cell systems (398-401). However, its effect on amyloid aggregation has not been investigated.

The effect of these stilbenes, both in solution and within a lipid bilayer, on the fibrillisation of SST-14 has been investigated in this chapter. SST-14 is a cyclic functional amyloid-forming

105

peptide; its amyloid formation and secondary structure conformation in vesicles have been explored in depth in Chapter 3. The use of RSV and two additional stilbene-based analogues of RSV allows for the comparison of the effects of different functional groups, and membrane interactions, on the effect of stilbenes on amyloid formation. To our knowledge, this is the first study to look at the impact of a series of stilbenes on functional amyloid aggregation. The conformation of SST in the presence of the stilbenes both in solution and bilayer-based was investigated using circular dichroism (CD). Intrinsic tryptophan fluorescence measurements provided information on the local environment of the SST. Thioflavin-T (ThT) assays were used to assess the impact of the stilbenes on fibril growth, while AFM images provided complementary images on the number of the morphology of formed fibrils deposited on a substrate. In this work, saturated chain lipids (DLPC-C12, DMPC-C14, DPPC-C16 and DSPC- 18) and charge (DMPS/DOPS) were mainly used, since we have demonstrated in Chapter 3, that saturated chain lipids and the anionic charge are responsible for inducing the highest rate of fibril formation in SST-14.

6.2 Materials and Methods 6.2.1 Liposome preparation The preparation and analysis are described in Chapter 2, section 2.2.1.

6.2.2 Preparation of stilbenes and liposome doped with stilbenes (RSV, TMS and TS) The stilbenes were doped in liposomes (DLPC, DMPC, DPPC, DSPC, DMPC/Chol, DOPC/DOPS) at a ratio of 18:1, co-dissolved in chloroform/ethanol mixture (2:1, v/v) in glass Moles of RSV vials. Mol% is defined as, 100 푋 . The preparation of vesicles doped moles of RSV + moles of lipid with stilbenes and analysis of size using DLS are described in Chapter 2, section 2.2.1.

6.2.3 Preparation of Functional amyloid-forming peptides (FAFP’s) 150 mM NaCl was added to a weighed fraction of SST-14at 5% w/w.

6.2.4 Thioflavin T (ThT) Binding Assay A weighed fraction of stilbenes (RSV, TS and TMS) was prepared in DMSO and added to SST-14 at 5% (w/w) at a ratio of 1:33 and stilbenes doped with liposomes is described in section 6.2.2. The experimental procedure and analysis are described in Chapter 2, section 2.2.2.

106

6.2.5 Tryptophan Fluorescence The sample preparation is described in Chapter 3, section 3.2.7 with addition and preparation of RSV and TMS described in 6.2.4. The ratio (SST-14 + lipid + stilbenes) is at (33:1:18), and the protocol is similar to Chapter 3, section 3.2.7 and experimental protocol and analysis are described in Chapter 2, section 2.2.3.

6.2.6 Circular Dichroism (CD) A weighed fraction of 0.01% w/w or 0.1 mg mL-1 SST was dissolved in buffer (Milli-Q water) and co-incubated in liposomes doped with RSV (33:1:18) (procedure described in 6.2.2 with a minor change, Milli-Q water used instead of 150 mM NaCl to rehydrate vesicles). The experimental protocol and analysis of CD spectra are described in Chapter 2, section 2.2.5.

6.2.7 Atomic Force Microscopy (AFM) 5 % (w/w) solution of SST fibrils were used and sample preparation is described in section 6.2.4 and SST + lipid + stilbenes (33:1:18) and experimental protocol described in Chapter 2, section 2.2.4 and no statistical analysis performed due to low fibril number (<500).

6.3 Results 6.3.1 Addition of RSV to liposomes DLS data confirm that addition of RSV to liposomes of various lipid compositions has only a minor impact on the particle size and distribution, Figure 6.3-6.5 (Table included in appendix S8.6 to S8.8).

107

Figure 6.3. Shows the Dynamic light scattering (DLS) curves for pure saturated chain vesicles and saturated vesicles doped with 5 mol% RSV (a) DLPC, (b) DMPC, (c) DPPC and (d) DSPC

108

Figure 6.4. DLS curve for pure DMPC+Chol vesicle and vesicle doped with 5 mol% RSV

Figure 6.5. DLS curve for (a) Pure DOPC+DOPS vesicle and vesicle doped with 5 mol% RSV and (b) Pure DMPC+DMPS vesicle and vesicle doped with 5 mol% RSV

6.3.2 SST-14 interaction with the lipid bilayer doped with RSV CD measurements were performed to investigate conformational changes of SST-14 with or without RSV in the bilayer. Note that a typical peptide concentration 0.1 mg mL-1 was used in these experiments. As amyloid formation only occurs at concentration above 50 mg mL-1 for

109

SST (Chapter 3), SST-14 in water, is mainly in a random coil conformation with some β-sheets and turns (Figure 6.6a), which is in agreement with the literature (98, 402).

Following the addition of RSV in solution (i.e. in the absence of lipids), no significant shift in the spectrum relative to pure SST (in water) is observed, suggesting that no major conformational change had occurred. However, a reduction in the small shoulder at ~196 nm could indicate a minor reduction in β-sheet structure. Addition of saturated chain liposomes doped with RSV, (DLPC-C12, DMPC-C14, DPPC-C16 and DSPC-C18), and the presence of cholesterol also produced only minor changes to the spectra, (Figure 6.6a and b). The minimum at 203nm is observed to move to slightly higher wavelength, consistent with a small increase in β-sheet structure. This is in contrast to the behaviour of pure SST, where a more significant increase in β-sheets and reduction in random coil structures were observed following incubation with saturated chain liposomes. It suggests that the presence of RSV embedded within the bilayer may inhibit interaction with SST.

Incubation with a charged bilayer (both in the presence and absence of RSV) was shown to impact the CD spectrum to the highest degree (Figure 6.6c). The minimum at 203nm shifts to higher wavelengths (β-sheets), presumably due to electrostatic interaction with the positively charged peptide.

110

Figure 6.6. (a) CD spectra for SST-14 0.01% w/w (0.1 mg mL-1) co-incubated in saturated vesicles doped with 5 mol% RSV, (b) CD spectra for SST-14 co-incubated in DMPC bilayer doped with cholesterol and 5 mol% RSV and (c) CD spectra for SST-14 co-incubated in bilayers (DMPC & DOPC) doped with anionic charge (DMPS & DOPS) and 5 mol% RSV

111

Intrinsic fluorescence from the Trp residue in the SST peptide may be used to provide additional information on the local environment around the peptide. In Chapter 3, I showed that the depth of penetration of the SST into the lipid bilayer depended on the lipid composition.

Here, the hydrophobicity of the local environment of the tryptophan residue present in SST-14 was studied in buffer and RSV using intrinsic tryptophan fluorescence assay. A significant red shift, i.e. to the higher wavelength (Figure 6.7) was observed for SST co-incubated with RSV compared to SST in solution. This suggests that interaction with the RSV has caused the Trp residue to be more hydrophilic. This could suggest that, a change in peptide conformation however, not observed via CD.

Figure 6.7. SST-14 environment in bilayer and bilayer doped with RSV, measured by Tryptophan Fluorescence

112

A corresponding red shift or higher λ max is observed upon addition of SST to bilayers doped with RSV (5 mol%) (Figure 6.7). This suggests that the presence of the RSV prevents the SST from penetrating into the lipid bilayer, in contrast to the CD results which suggested that the interaction was less pronounced. As for SST in liposomes (no RSV), the effect is dependent on the lipid composition of the bilayer. However, in the presence of RSV, the deepest penetration seems to be into the DPPC and DSPC bilayers. This is again in contrast to bilayers without RSV, where charged bilayers tended to result in the deepest penetration of the peptide.

6.3.3 Effect of RSV on fibrillisation of SST-14 ThT fluorescence binding assays were used to monitor the kinetics of SST-14amyloid fibril formation in the presence of RSV, both in solution and doped within a lipid bilayer.

A significant reduction in the final fluorescence intensity of the plateau phase was observed following addition of RSV in solution (Figure 6.8) suggesting fewer fibrils are formed. This is in agreement with literature reports on the effect of RSV on fibrillisation of toxic amyloid species like Aβ and IAPP (146, 391). However, a slight decrease in the lag phase and increase in kagg was observed (Table 6.2) relative to in buffer, suggesting that the rate of fibril growth actually increased in the presence of the RSV.

The effect of RSV doped within liposomes formed from saturated hydrocarbon chains was also investigated Figure 6.9. In the absence of RSV, these liposomes were shown to have a strong promoting effect on fibrillisation of SST in Chapter 3. For liposomes doped with RSV, clear differences in behaviour were observed between the short-chain lipids (DLPC, DMPC) which from fluid bilayers and the longer-chain lipids (DPPC, DSPC) which form gel-phase bilayers. For all saturated chain lengths, the presence of RSV in the bilayer reduced the intensity of the final fluorescence plateau relative to the bilayer without RSV, indicating that the RSV is acting against the strong promoting effect of the bilayer, and acting to inhibit fibril growth as seen in solution. However, for the longer chain lipids (DPPC and DSPC) RSV-doped bilayers also increased the lag time, suggesting that they are inhibiting nucleation. In contrast, no reduction in lag time is seen for the shorter-chain lipids, relative to the bilayer without RSV. This backs up the theory presented in Chapter 3 that a different aggregation mechanism may be at play for the longer-chain lipids.

As observed throughout the thesis, the effect of cholesterol doped within the bilayer was not significant, Figure 6.10, with similar behaviour to that seen for the pure saturated chain lipids. For charged liposomes, a similar effect was also observed, Figure 6.11. However, while the

113

overall fluorescence plateau was reduced in the presence of RSV, the rate of aggregation remains very fast, indicating that the effect of the charged bilayer is dominant in this case. This effect may also be due to acid-base interaction between RSV and the negatively charged bilayer.

Figure 6.8. Shows the ThT assay curves for 5% (w/w) SST-14 incubated in buffer (NaCl) vs stilbenes co-dissolved using DMSO

114

Figure 6.9. Shows the ThT assay curves for 5% (w/w) SST-14 co-incubated in saturated chain lipids doped with 5 mol% RSV (a) DLPC, (b) DMPC, (c) DPPC vesicles and (d) DSPC

115

Table 6.2. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer, RSV, saturated chain lipids and saturated chain lipids doped with 5 mol% RSV

-1 1or2 Conditions Lag t50 or (t1/2) kagg (s ) Final fluorescence phase (hours)3 intensity of plateau (hours) phase (x106) Buffer SST 12 51 0.01862 2.7 SST+RSV 0 12 0.04021 1.3 Effect of chain length or bilayer thickness SST+DLPC 0 6.3 0.05891 3.8 SST+DLPC+RSV 5 mol% 0 1.9 0.2231 2.2 SST+DMPC 0 13 0.03801 6.2 SST+DMPC+RSV 5 mol% 0 2.4 0.1351 2.2 SST+DPPC 24 41 0.06312 4.9 SST+DPPC+RSV 5 mol% 24 15 0.0212 1.9 SST+DSPC 12 28 0.05112 4.8 SST+DSPC+RSV 5 mol% 24 36 0.04252 2.0 One-phase association1, Plateau followed by one-phase association2 and half-time3

Figure 6.10. Shows the ThT assay curves for 5% (w/w) SST-14 co-incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% RSV

116

Table 6.3. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% RSV

-1 1or2 Conditions Lag phase (t1/2) or t50 kagg (s ) Final (hours) (hours)3 fluorescence intensity of plateau phase (x106) SST 12 51 0.01862 2.7 SST+RSV 0 12 0.04021 1.3 SST+DMPC+CHOL 0 23 0.03021 11.0 SST+DMPC+CHOL+RSV 5 mol% 0 14 0.02011 5.2 One-phase association1, Plateau followed by one-phase association2 and half-time3

Figure 6.11. Illustrations of: (a) ThT assay curves for 5% (w/w) SST-14 co-incubated in buffer and DOPC/DOPS vesicles doped with 5 mol% RSV (b) ThT assay curves for 5% (w/w) SST- 14 co-incubated in buffer and DMPC/DMPS vesicles doped with 5 mol% RSV

117

Table 6.4. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in buffer and DOPC/DOPS and DMPC/DMPS vesicles doped with 5 mol% RSV

-1 1or2 Conditions Lag (t1/2) or t50 kagg (s ) Final phase (hours)3 fluorescence (hours) intensity of plateau phase (x106) SST 12 51 0.01862 2.7 SST+RSV 0 12 0.04021 1.3 SST+DOPC+DOPS 0 2.6 0.3251 5.6 SST+DOPC+DOPS+RSV 5 mol% 0 0.5 Fitting not 9.8 possible SST+DMPC+DMPS 0 0.3 0.2191 12.3 SST+DMPC+DMPS+RSV 5 mol% 0 1.0 0.4771 8.7 One-phase association1, Plateau followed by one-phase association2 and half-time3

RSV has been shown to inhibit aggregation, both in solution, and doped within a lipid bilayer, where it appears to act against the promoting effect of the lipid bilayer.

A similar study was then carried out on two additional stilbene analogues of RSV, TS (TS) and TMS (TMS). Both of these molecules have a much lower topological polar surface area than RSV, and a higher octanol-water partition coefficient, suggesting that they are more likely to reside within the lipid bilayer environment of the liposomes.

6.3.4 SST-14 interaction with the lipid bilayer doped with tri-methoxy RSV and trans-stilbene DLS data show some changes to the liposome hydrodynamic particle size and distribution upon the addition of stilbenes, TS and TMS which has impacted the size of the vesicles slightly compared to RSV (Figure 6.12) (in appendix Table S8.6). However, as these changes are not consistently to higher or lower average liposome size, it is likely these represent some experimental variation in liposomes size which is unlikely to affect the results.

118

Figure 6.12. Shows the Dynamic light scattering (DLS) curves for pure saturated chain vesicles and saturated vesicles doped with 5 mol% stilbenes (a) DLPC, (b) DMPC, (c) DPPC and (d) DSPC

119

Figure 6.13. SST-14 environment in bilayer and bilayer doped with stilbenes, measured by Tryptophan Fluorescence

The hydrophobicity of the local environment of the tryptophan residue present in SST-14 was studied in TMS using intrinsic tryptophan fluorescence assay (Figure 6.13). (Note that data for TS are not available). As for RSV, a significant red shift, i.e. to the higher wavelength (Figure 6.13) was observed for SST co-incubated with TMS in solution, suggesting a conformational change has occurred upon interaction with TMS. Again, the addition of the liposomes caused a corresponding blue shift or decrease in λ max. This blue shift is greater in the presence of TMS, which may be related to the higher hydrophobicity of this molecule compared to RSV. Variations between the different lipid compositions were reduced compared to that for pure bilayers or RSV-doped bilayers. However, the data again suggest that the SST peptide has interacted with the lipid bilayer, with some penetration of the peptide into the bilayer.

120

6.3.5 Effect of the stilbenes TMS and TS on fibrillisation of SST-14 ThT fluorescence binding assays were used to monitor the kinetics of SST-14 amyloid fibril formation in the presence of TMS and TS, both in solution and doped within a lipid bilayer.

A significant reduction in the final fluorescence intensity of the plateau phase was observed following addition of either TMS or TS in solution (Figure 6.14) suggesting fewer fibrils are formed, in agreement with results for RSV and previous data on the effect of RSV on the fibrillisation of toxic amyloid-forming peptides (146, 391). However, in contrast to RSV, both TMS and TS increased the lag phase to 24 and 54 hours, respectively. TS also significantly -1 -1 reduced the value of kagg (0.003 s ) (Table 6.5) relative to in buffer (0.019 s ), suggesting that it not only reduced the final volume of fibrils formed, but also the rate of fibril growth.

The effect of TMS and TS doped within liposomes formed from saturated hydrocarbon chains was also investigated Figure 6.15. In the absence of the stilbenes, these liposomes are known to have a strong promoting effect on fibrillisation of SST. For all saturated chain liposomes doped with TS, the intensity of the final fluorescence plateau was reduced (or remained relatively constant) relative to the bilayer without TS, similar to what was seen for RSV. However, the addition of TMS within these bilayers resulted in an increase to the fluorescence plateau, the opposite effect to that of TMS in solution. This effect was very marked, with the plateau increasing by approximately an order of magnitude relative to SST in solution.

Addition of TMS to the bilayer also reduced the lag time, suggesting that it both increases the nucleation rate and the final volume of fibrils formed. In contrast, the effect of bilayer- encapsulated TS again depended on the lipid chain length. For short-chain lipids, the addition of TS reduced the lag rate whereas for longer chain lipids it increased the lag rate.

Aggregation rate constants (kagg) did not vary considerably between the liposomes with and without stilbenes suggesting that while the peptide-bilayer interactions may impact nucleation rates and volume of fibrils formed, they have a lesser effect on the actual rate of growth of the fibrils.

121

Figure 6.14. Shows the ThT assay curves for 5% (w/w) SST-14 incubated in buffer (NaCl) vs stilbenes co-dissolved using DMSO

Table 6.5. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) in buffer vs stilbenes

-1 1or2 Conditions Lag (t1/2) or t50 kagg (s ) Final fluorescence phase (hours)3 intensity of plateau (hours) phase (x106) SST 12 51 0.01861 2.7 SST+RSV 5 mol% 0 12 0.04021 1.3 SST+TS 5 mol% 54 104 0.003412 0.3 SST+TMS 5 mol% 24 46 0.02782 1.7

122

Figure 6.15. Shows the ThT assay curves for 5% (w/w) SST-14 co-incubated in saturated chain lipids doped with 5 mol% stilbenes (a) DLPC, (b) DMPC, (c) DPPC vesicles and (d) DSPC

123

Table 6.6. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer, stilbenes, saturated chain lipids and saturated chain lipids doped with 5 mol% stilbenes

-1 1or2 Conditions Lag t50 or (t1/2) kagg (s ) Final fluorescence phase (hours)3 intensity of plateau (hours) phase (x106) Buffer SST 12 51 0.01862 2.7 SST+RSV 0 12 0.04011 1.3 SST+TS 54 104 0.003412 0.3 SST+TMS 24 46 0.02782 1.7 Effect of chain length or bilayer thickness SST+DLPC 0 6.3 0.05891 3.8 SST+DLPC+RSV 5 mol% 0 1.9 0.2231 2.2 SST+DLPC+TS 5 mol% 0 7.6 0.06601 3.1 SST+DLPC+TMS 5 mol% 6 23 0.03192 21.2 SST+DMPC 0 13 0.03801 6.2 SST+DMPC+RSV 5 mol% 0 37 0.1351 2.2 SST+DMPC+TS 5 mol% 0 26 0.01581 8.3 SST+DMPC+TMS 5 mol% 0 33 0.02081 19.4 SST+DPPC 24 41 0.06312 4.9 SST+DPPC+RSV 5 mol% 24 56 0.0212 1.9 SST+DPPC+TS 5 mol% 12 36 0.02762 3.1 SST+DPPC+TMS 5 mol% 0 21 0.04871 9.1 SST+DSPC 12 28 0.05112 4.8 SST+DSPC+RSV 5 mol% 24 37 0.04252 2.0 SST+DSPC+TS 5 mol% 12 32 0.02802 3.9 SST+DSPC+TMS 5 mol% 0 21 0.03051 7.3 One-phase association1, Plateau followed by one-phase association2 and half-time3

124

The DLS data is showing a hydrodynamic particle size and distribution which suggests that stilbenes have intercalated well within the bilayer (Figure 6.16) (in appendix Table S8.7).

Figure 6.16. DLS curve for pure DMPC+Chol vesicle and vesicle doped with 5 mol% stilbenes

Addition of cholesterol to the liposomes did not significantly affect the results. As for samples without cholesterol, the greatest fibril growth is observed with TMS (Figure 6.17) doped in DMPC/Chol, which shows the highest final fluorescence intensity before reaching a plateau. -1 This is also consistent with the highest kagg value (0.06 s ) and the lowest t1/2 (11 h) (Table 6.7).

125

Figure 6.17. Shows the ThT assay curves for 5% (w/w) SST-14 co-incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% stilbenes

Table 6.7. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DMPC/Chol vesicles doped with 5 mol% stilbenes

-1 1or2 Conditions Lag (t1/2) or t50 kagg (s ) Final fluorescence phase (hours)3 intensity of plateau (hours) phase (x106) SST 12 51 0.01862 2.7 SST+RSV 0 12 0.04021 1.3 SST+TS 54 104 0.003412 0.3 SST+TMS 24 46 0.02782 1.7 SST+DMPC+CHOL 0 23 0.03021 11.0 SST+DMPC+CHOL+RSV 5 mol% 0 14 0.02011 5.2 SST+DMPC+CHOL+TS 5 mol% 12 30 0.03882 9.6 SST+DMPC+CHOL+TMS 5 mol% 0 11 0.06061 15.2 One-phase association1, Plateau followed by one-phase association2 and half-time3

126

6.3.6 SST-14 interaction with anionic charged lipid bilayer doped with trimethoxy RSV and trans-stilbene The DLS data is suggesting only minor changes to the particle size distribution on the addition of TMS or TS (Figure 6.18) (in appendix Table S8.8).

Figure 6.18. DLS curve for (a) Pure DOPC+DOPS vesicle and vesicle doped with 5 mol% stilbenes and (b) Pure DMPC+DMPS vesicle and vesicle doped with 5 mol% stilbenes

Even for the anionic bilayers, which strongly promote fibril formation, a significant increase in the final fluorescence intensity of the plateau phase (Figure 6.19) is observed upon incorporating TMS. This is suggesting that hydrophobic and electrostatic interactions are strongly promoting aggregation in SST-14.

127

Figure 6.19. Illustrations of: (a) ThT assay curves for 5% (w/w) SST-14 co-incubated in buffer and DOPC/DOPS vesicles doped with 5 mol% stilbenes (b) ThT assay curves for 5% (w/w) SST-14 co-incubated in buffer and DMPC/DMPS vesicles doped with 5 mol% stilbenes

128

Table 6.8. Shows the lag phase, half-time, aggregation rate constant and the final fluorescence intensity of plateau phase for SST at 5% (w/w) co-incubated in the buffer and DOPC/DOPS and DMPC/DMPS vesicles doped with 5 mol% stilbenes

-1 1or2 Conditions Lag phase (t1/2) or t50 kagg (s ) Final fluorescence (hours) (hours)3 intensity of plateau phase (x106) SST 12 51 0.01862 2.7 SST+RSV 0 12 0.04021 1.3 SST+TS 54 104 0.00342 0.3 SST+TMS 24 46 0.02782 1.7 SST+DOPC+DOPS 0 2.6 0.3251 5.6 SST+DOPC+DOPS+RSV 5 0 0.5 Fitting not 9.8 mol% possible SST+DOPC+DOPS+TS 5 0 17 0.03161 6.9 mol% SST+DOPC+DOPS+TMS 5 0 2.4 0.2181 9.9 mol% SST+DMPC+DMPS 0 0.3 0.2191 12.3 SST+DMPC+DMPS+RSV 5 0 1.0 0.4771 8.7 mol% SST+DMPC+DMPS+TS 5 0 1.5 0.4211 10.8 mol% SST+DMPC+DMPS+TMS 5 0 2.2 0.3331 21.1 mol% One-phase association1, Plateau followed by one-phase association2 and half-time3

Since it has been reported that RSV has no significant absorption within the range of ThT fluorescence (391), nevertheless this does not imply that it will not interfere with ThT assays. Thus, complementary AFM study was conducted to confirm that results are not influenced by interference between ThT and stilbenes.

129

6.3.7 Clear morphological differences are observed between fibrils formed in the buffer and RSV doped with lipid vesicles In Chapter 3, we concluded that SST aggregates into typical amyloid-like unbranched twisted fibrillar structures in solution. In contrast, the fibrils formed by SST co-incubated with DMPC vesicles were much longer and had slightly greater stiffness, suggesting that SST fibrils form heterogeneous structure, likely due to an uneven coating of lipids decorating the fibrils.

AFM images were obtained of SST samples incubated with the three stilbene derivatives in solution and encapsulated within DMPC bilayers.

Following the incubation of SST-14 fibrils in RSV (absence of lipid) no fibrils were detected, (Figure 6.20a) in agreement with ThT assay results which showed a decrease in the fluorescence intensity. Both techniques suggest that RSV had indeed inhibited SST fibril formation (Figure 6.8). However, in the presence of a bilayer doped with RSV, fibrils were observed tightly associated into clusters of lipid aggregates (Figure 6.20b). The few individual fibrils that can be detected are much shorter fibrils, indicating that RSV had impacted fibril growth also when encapsulated within the bilayer.

Figure 6.20. AFM analysis of SST-14 fibrils at 10% (w/w) with RSV, topographic images of (a) SST-14 with 5 mol% RSV displaying little or fibrils and (b) SST-14 co-incubated with DMPC vesicles doped with 5 mol% RSV displaying much shorter fibrils

130

AFM analysis of SST-14 fibrils in TS (absence of lipids) again showed no evidence of fibrils, but tiny aggregates clumped together (Figure 6.21a) which is also in agreement with the lowest fluorescence intensity observed with ThT assay (Figure 6.14). Whereas, long SST fibrils showing twists (Figure 6.21b) were observed in DMPC vesicles doped with TS. This is consistent with the ThT assay which indicated that fibrils were still formed in the presence of the TS-doped liposomes (Figure 6.15b), albeit at a lower volume. AFM images suggest that incubation with TS in DMPC bilayer had little effect on SST fibril morphology.

Figure 6.21. AFM analysis of SST-14 fibrils at 10% (w/w) with TS, topographic images of (a) SST-14 with 5 mol% TS displaying little or no fibrils and (b) SST-14 with DMPC vesicles doped with 5 mol% TS displaying neat long fibrils

Lastly, AFM analysis of SST-14 fibrils in TMS (absence of lipids) shows little or no fibrils (Figure 6.22a), which also be compared to the lower fluorescence intensity observed with ThT assay (Figure 6.14). However, SST co-incubated with DMPC bilayer doped with TMS shows long SST fibrils (Figure 6.22b) suggesting that TMS has little effect on SST fibril morphology. This, in fact, agrees with the highest final fluorescence intensity observed in the plateau region compared to SST+DMPC+TMS (Figure 6.15b).

131

Figure 6.22. AFM analysis of SST-14 fibrils at 10% (w/w) with TMS, topographic images of (a) SST-14 with 5 mol% TMS displaying little or no fibrils and (b) SST-14 with DMPC vesicles doped with 5 mol% TMS displaying neat long fibrils

6.4 Discussion RSV an active phytochemical present in many plants and wine has received much research attention due to its ability to decrease cognitive impairment and amyloid neuropathology. The exact mechanism by which RSV exerts neuroprotective effects continues to be defined. RSV has been shown to inhibit fibrillisation of toxic amyloid-forming peptides including IAPP and Aβ (3, 391, 394). In this chapter, RSV and two stilbene analogues TMS and TS, were shown to inhibit fibril formation in solution for the functional amyloid-forming peptide, SST. A significant decrease in the intensity of the fluorescence plateau was observed in the presence of all three stilbenes. Although RSV has no significant absorption in the wavelength range of ThT fluorescence, some issues have been previously observed in using ThT assays to monitor inhibition of amyloid fibril formation by inhibitors. For example, RSV has been shown to reduce the ThT intensity of pre-formed A fibrils, and it has been suggested that it might displace the bound ThT from the fibrils (403). While we have no evidence that RSV interferes with ThT binding to SST fibrils, AFM images were nevertheless used as a complementary technique to reflect deposited fibrils and in the presence of the stilbenes was significantly reduced, with little to no fibrils observed.

132

Both TMS and TS increased the lag phase, while RSV actually decreased the lag phase suggesting that nucleation of this functional amyloid occurred faster in the presence of RSV, in contrast to what has been seen for toxic amyloid formation. This suggests there may be fundamental differences between the peptide-TMS/peptide-TS interactions, and the peptide- RSV interaction. Previously RSV has been shown to interact with the toxic amyloid-forming peptide IAPP through the His-18 and Lys-1 residues (403). Interactions between aromatic side- chains and amyloid inhibitors, including polyphenols, have also been previously observed (404). Finally, π-cation interactions have been shown to stabilise globular proteins (405). SST contains a positively charged Lys residue (pKa = 10.5) at approximately neutral pH, as well as the aromatic amino acids Phe (3) and Trp all of which could interact with the stilbene inhibitors studied.

Figure 6.23. Diagrammatic illustrations of how RSV inhibits SST fibrillisation by disturbing the lipid monolayer by a dual effect approach (adapted from (29))

Addition of the stilbene inhibitor into a lipid bilayer did not result in the same trends as in solution suggesting that the interaction method may be different. In solution, the stilbenes will interfere with fibrillisation by direct interactions with the peptide, which modify the peptide- peptide interactions necessary for fibrillisation. However, in a bilayer, these molecules may affect fibrillisation in two ways. 1. By Binding directly to the peptide (similar to in solution),

133

the stilbene may prevent the peptide from interacting with the membrane (This mechanism was suggested by Evers et al. and is shown in Figure 6.23 (29) 2. By insertion into the lipid bilayer, the stilbene may change the physiochemical properties of the bilayer such that the interaction with the peptide is modified. The theoretical partition coefficients of RSV in DMPC, DPPC and DSPC vesicles have been calculated as 2.63, 3.07 and 3.11, respectively (378). The percentage of membrane-bound RSV was reported 93%, 97% and 98% in DMPC, DPPC and DSPC membranes respectively (378). The low topological polar surface areas associated with all three stilbenes indicate that all three will associate with the lipid bilayer to some extent. The tPSA is lowest for TS suggesting that this molecule may have most bilayer-association (although TMS has a slightly higher octanol-water partition coefficient). DLS data indicate that encapsulation of these molecules into the lipid bilayer did not impact the size distribution of liposomes. This is expected to affect chain packing within the lipid bilayer although this was not directly investigated during the thesis. RSV has been previously observed to act as a spacer, causing an increase in disorder lipid molecules in the bilayer (381). RSV molecule is also reported to bind near the head group region of the phospholipid, which induces an increase in surface area per lipid and subsequently decreasing membrane thickness (26).

Both TS and RSV appeared to inhibit fibrillisation in the bilayer, similar to what was observed in solution, acting against the generally promoting effect of the saturated lipid bilayers studied. This was the case, even for charged bilayers, which are strongly promoting. While the fluorescence plateau was reduced relative to that seen with the bilayer alone, it was still higher than that observed in the absence of a bilayer, suggesting that the promotion effect of the bilayer is dominant. It is possible that increasing the concentration of RSV in the bilayer may act to reduce fibrillisation even more. However, TMS actually promoted fibrillisation when encapsulated in the lipid bilayer, with higher fluorescence plateau values.

In fluid short-chain bilayers (DLPC and DMPC), the stilbenes reduced the lag time, suggesting nucleation is actually faster in the presence of these ‘inhibitors’. However, AFM images suggest that the fibrils formed in the presence of stilbene-doped DMPC bilayers can be shorter or aggregated into clusters (as for RSV doped DMPC) suggesting that the faster nucleation may have led to an off-aggregation pathway. In contrast, in gel bilayers (DPPC, DSPC), while TMS and RSV still reduce the lag time, TS increases the lag time, similar to what is observed in solution. As TS has a high logP value and a tPSA of essentially zero, it is not likely that this effect is via free TS in solution.

134

In summary, RSV, which is known as an inhibitor of toxic amyloid formation, was also shown to act as an inhibitor of fibrillisation for the functional amyloid-forming peptide SST. Two stilbene derivatives of RSV, TMS and TS, which are not polyphenols were shown to have a greater inhibiting effect in solution. The effect of the stilbenes, when encapsulated in a bilayer, was not always the same as that in solution suggesting that peptide inhibition may result from a different mechanism.

135

7. Conclusion and Future work Fundamental differences in aggregation were observed between the different peptides in solution. For the three cyclic peptides studied, SST forms typical long amyloid-like unbranched twisted fibrillar structures in solution, oxytocin formed a small number of very short fibrils, and no fibrils were detected via AFM or the ThT assay for vasopressin. It is particularly interesting to note the differences in fibrillisation between oxytocin and vasopressin, as these peptides are structurally similar, and differ only in the amino acids at positions 3 and 8. The increase in aggregation propensity for oxytocin may reflect the substitution of the Phenylalanine amino acid for Isoleucine for at position 3, as this is known to promote π – π stacking interactions, and increase the hydrophobicity of the peptide. The linear peptides Substance P and Deslorelin displayed markedly different aggregation behaviour than the cyclic peptides. Needle-like fibrillar structures were observed for Substance P, while numerous multi-laminar crystalline structures were observed for Deslorelin. Similar structures have been observed in solution for Substance P, and for LHRH, an analogue of Deslorelin and confirmed via SAXS to have internal nanotube, rather than fibril, symmetry.

For all peptides studied, both cyclic and linear, co-incubation with a lipid bilayer was observed to impact both the rate of aggregation and the morphology of fibrils formed. While there were obvious differences between the different peptides, some clear trends emerged throughout the thesis. The effect of surface charge in the lipid bilayer was dominant for all peptides studied. Strong electrostatic interactions between the typically cationic peptides, and an anionic bilayer, were found to drive the highest volume of fibrils formed in each case. In fact, AVP only formed fibrils in the presence of an anionic bilayer, with either no, or a very small volume of fibrils formed in solution, or in the presence of uncharged bilayers. For this peptide we demonstrated that the final volume of fibrils formed was proportional to the overall charge on the bilayer.

In general, lipid bilayers formed from saturated chain lipids tended to increase the rate of aggregation, and the volume of fibrils formed. In contrast, unsaturated chain lipids tended to inhibit fibrillisation, decreasing the rate of aggregation, increasing the lag time, and reducing the total volume of fibrils formed. For the peptide SST this was clearly linked to the interaction between the peptide and the lipid bilayer. Intrinsic tryptophan fluorescence studies indicated that the peptide penetrated more into the unsaturated DOPC bilayer than into the saturated chain bilayers, presumably due to the increase in lateral bilayer pressure associated with the saturated chains. We suggest that the peptide-peptide interactions necessary for fibrillisation

136

are promoted by a more peripheral attachment of the peptide to the membrane surface, but inhibited by deeper penetration into the bilayer.

The effect of chain length varied among the different peptides. For all peptides studied, no significant differences in the kagg values were observed among the different saturated chain lipids, suggesting that the actual rate of growth of the peptide fibril did not depend on the lipid environment. This is consistent with a mechanism where the lipid bilayer acts as a nucleation site for peptide-peptide interactions but does not accelerate the actual growth process. For SST, the lag period was shortest for the short-chain bilayers DLPC and DMPC, however for the other peptides studied, the lag period was typically zero suggesting no differences in the ability of the different saturated chains to promote nucleation. However, significant differences were observed between the different chain lengths in terms of the final volume of fibrils formed. For SST and Deslorelin, most fibrils were formed in the presence of the short-chain lipids DLPC and DMPC. However, for OT, AVP and SP, most fibrils were formed for the longer-chain lipids DPPC and DSPC. This suggests that the mechanism by which the lipid membrane promotes an increased volume of fibrils formed may differ depending on the peptide. This does not seem to depend on whether the peptide is linear or cyclic, or the morphology of the aggregates formed (fibrils vs nanotubes). The effect may be related to lipid solubility (the short chain lipids are typically more soluble as expected) and suggests that the membrane-mediated mechanism of fibrillisation may involve a reciprocal transfer of lipids from the bilayer to the growing fibril. It may also be related to lipid chain packing. The shorter-chain lipids DLPC and DMPC form a fluid lamellar phase at room temperature, while the longer chain lipids form a gel lamellar phase. However, we suggest that the effect of a gel vs fluid bilayer is more minor, as DOPC also forms a fluid lamellar phase at room temperature, yet is not observed to promote fibril growth for any peptide studied.

The morphology of the aggregates formed was also affected by the presence of the lipid bilayer. For SST, evidence of lipid coating of the fibrils was seen. The fibrils were more heterogeneous and had a tendency to be longer. The typical twisted morphology of the fibrils in solution was also obscured due to the lipid coating. For Des, which appears to form bundles of nanotubes in solution, addition of lipids results in a completely different aggregate morphology. We suggest that the twisted aggregates observed are twisted or helical ribbons, which are suggested to be structural intermediates during the formation of nanotubes. In this case, the lipid coating may be preventing lateral association of the protofilaments into nanotubes.

137

While the peptides studied clearly interact with the lipid bilayer in all cases, a reciprocal structural effect on the lipid bilayer was not observed for SST, in either the bulk lamellar phase or in liposomes. This is in contrast to the behaviour of toxic amyloid-forming peptides which cause major structural perturbations to the lipid bilayer, and may partially explain the reduced toxicity of functional amyloids.

In Chapter 6 we showed that the polyphenol resveratrol, which is known to inhibit fibril formation for toxic amyloid-forming peptides, also inhibited fibrillisation for the functional amyloid-forming peptide, SST. Two stilbene derivatives of resveratrol, trans-3,5,4’- trimethoxystilbene (TMS) and trans-stilbene (TS), were also shown to inhibit fibril formation in solution, despite not being polyphenols. In a lipid bilayer environment, the behaviour of these molecules, which are reasonably soluble in the lipid bilayer, was not the same as in solution. In fact, TMS, which inhibits fibrillisation in solution, consistently and strongly promoted fibrillisation in a bilayer environment. In solution the stilbene molecules presumably modify peptide-peptide interactions by direct interaction with the peptide itself. However, in a lipid bilayer, the stilbene may modify fibrillisation by changing the physicochemical properties of the bilayer such that the peptide-lipid interaction is modified.

In summary, we have shown that the lipid bilayer, and stilbene additives, have a significant effect on the fibrillisation of five different functional amyloid-forming peptides. This represents a significant increase to the body of literature in this area, as only a single experimental study has, to date, investigated this effect. We suggest that intrinsic peptide- peptide interactions are the ultimate driving force behind fibrillisation. However, the lipid membrane may promote fibrillisation by facilitating peptide-peptide interactions on the membrane surface or inhibit fibrillisation when peptide insertion into the bilayer is deep enough to preclude such interactions. The mechanism by which this occurs remains unresolved. Future work will involve direct visualisation of the peptide-lipid interactions using confocal microscopy with fluorescently tagged lipids. This could allow for direct visualisation of any reciprocal transfer of bilayer lipids into the growing fibril.

138

References

1. Arosio P, Knowles TP, Linse S. On the lag phase in amyloid fibril formation. Physical chemistry chemical physics : PCCP. 2015;17(12):7606-18. 2. Jackson MP, Hewitt EW. Why are Functional Amyloids Non-Toxic in Humans? Biomolecules. 2017;7(4). 3. Jia Y, Wang N, Liu X. Resveratrol and Amyloid-Beta: Mechanistic Insights. Nutrients. 2017;9(10):1122. 4. Dowhan W, Bogdanov M. Chapter 1 Functional roles of lipids in membranes. New Comprehensive Biochemistry. 2002;36:1-35. 5. Butterfield SM, Lashuel HA. Amyloidogenic protein-membrane interactions: mechanistic insight from model systems. Angewandte Chemie (International ed in English). 2010;49(33):5628-54. 6. Stefani M. Biochemical and biophysical features of both oligomer/fibril and cell membrane in amyloid cytotoxicity. FEBS Journal. 2010;277(22):4602-13. 7. Sasahara K, Morigaki K, Okazaki T, Hamada D. Binding of islet amyloid polypeptide to supported lipid bilayers and amyloid aggregation at the membranes. Biochemistry. 2012;51(35):6908-19. 8. Malishev R, Abbasi R, Jelinek R, Chai L. Bacterial Model Membranes Reshape Fibrillation of a Functional Amyloid Protein. Biochemistry. 2018;57(35):5230-8. 9. Drolle E, Negoda A, Hammond K, Pavlov E, Leonenko Z. Changes in lipid membranes may trigger amyloid toxicity in Alzheimer's disease. PLOS ONE. 2017;12(8):e0182194. 10. Yang M, Wang K, Lin J, Wang L, Wei F, Zhu J, et al. Gel Phase Membrane Retards Amyloid β- Peptide (1–42) Fibrillation by Restricting Slaved Diffusion of Peptides on Lipid Bilayers. Langmuir. 2018;34(28):8408-14. 11. Williams TL, Serpell LC. Membrane and surface interactions of Alzheimer’s Aβ peptide – insights into the mechanism of cytotoxicity. The FEBS journal. 2011;278(20):3905-17. 12. Milanesi L, Sheynis T, Xue W-F, Orlova EV, Hellewell AL, Jelinek R, et al. Direct three- dimensional visualization of membrane disruption by amyloid fibrils. Proceedings of the National Academy of Sciences. 2012;109(50):20455-60. 13. Bhandary S, Basu R, Manna S, Das S, Nandy P. Role of non-lamellar-forming lipid in promotion of liposomal fusion. Phase Transitions. 2009;82(3):221-7. 14. Simidjiev I, Stoylova S, Amenitsch H, Jávorfi T, Mustárdy L, Laggner P, et al. Self-assembly of large, ordered lamellae from non-bilayer lipids and integral membrane proteins in vitro. Proceedings of the National Academy of Sciences of the United States of America. 2000;97(4):1473-6. 15. Winter R, Jeworrek C. Effect of pressure on membranes. Soft Matter. 2009;5(17):3157-73. 16. Li J, Wang X, Zhang T, Wang C, Huang Z, Luo X, et al. A review on phospholipids and their main applications in drug delivery systems. Asian Journal of Pharmaceutical Sciences. 2015;10(2):81- 98. 17. Yeagle PL. Chapter 7 - Structures of Lipid Assemblies. In: Yeagle PL, editor. The Membranes of Cells (Third Edition). Boston: Academic Press; 2016. p. 115-54. 18. Escribá PV, González-Ros JM, Goñi FM, Kinnunen PKJ, Vigh L, Sánchez-Magraner L, et al. Membranes: a meeting point for lipids, proteins and therapies. Journal of Cellular and Molecular Medicine. 2008;12(3):829-75. 19. Galvagnion C, Brown JWP, Ouberai MM, Flagmeier P, Vendruscolo M, Buell AK, et al. Chemical properties of lipids strongly affect the kinetics of the membrane-induced aggregation of α- synuclein. Proceedings of the National Academy of Sciences. 2016;113(26):7065-70. 20. Grallert S, Rangel-Yagui C, Pasqualoto K, Tavares L. Polymeric micelles and molecular modeling applied to the development of radiopharmaceuticals2012. 1-16 p.

139

21. Ph DP. The Neurohypophysis: Endocrinology of Vasopressin and Oxytocin. In: De Groot LJ, Chrousos G, Dungan K, Feingold KR, Grossman A, Hershman JM, et al., editors. Endotext. South Dartmouth (MA): MDText.com, Inc.; 2000. 22. Amso Z, Cornish J, Brimble M. Short Anabolic Peptides for Bone Growth2016. n/a-n/a p. 23. Stewart AJ, Sellar R, Wilson DJ, Millar RP, Lu ZL. Identification of a novel ligand binding residue Arg38(1.35) in the human gonadotropin-releasing hormone receptor. Molecular pharmacology. 2008;73(1):75-81. 24. Valéry C, Deville-Foillard S, Lefebvre C, Taberner N, Legrand P, Meneau F, et al. Atomic view of the histidine environment stabilizing higher-pH conformations of pH-dependent proteins. Nature communications. 2015;6:7771. 25. Adamcik J, Mezzenga R. Proteins Fibrils from a Polymer Physics Perspective. Macromolecules. 2012;45(3):1137-50. 26. Fei Q, Kent D, Botello-Smith WM, Nur F, Nur S, Alsamarah A, et al. Molecular Mechanism of Resveratrol’s Lipid Membrane Protection. Scientific reports. 2018;8(1):1587. 27. Kim D-W, Kim Y-M, Kang S-D, Han Y-M, Pae H-O. Effects of Resveratrol and trans-3,5,4'- Trimethoxystilbene on Glutamate-Induced Cytotoxicity, Heme Oxygenase-1, and Sirtuin 1 in HT22 Neuronal Cells. Biomolecules & therapeutics. 2012;20(3):306-12. 28. Bonechi C, Martini S, Ciani L, Lamponi S, Rebmann H, Rossi C, et al. Using Liposomes as Carriers for Polyphenolic Compounds: The Case of Trans-Resveratrol. PLOS ONE. 2012;7(8):e41438. 29. Evers F, Jeworrek C, Tiemeyer S, Weise K, Sellin D, Paulus M, et al. Elucidating the mechanism of lipid membrane-induced IAPP fibrillogenesis and its inhibition by the red wine compound resveratrol: a synchrotron X-ray reflectivity study. J Am Chem Soc. 2009;131(27):9516-21. 30. Harrison RS, Sharpe PC, Singh Y, Fairlie DP. Amyloid peptides and proteins in review. Reviews of physiology, biochemistry and pharmacology. 2007;159:1-77. 31. Chiti F, Dobson CM. Protein misfolding, functional amyloid, and human disease. Annual review of biochemistry. 2006;75:333-66. 32. Fowler DM, Koulov AV, Balch WE, Kelly JW. Functional amyloid – from bacteria to humans. Trends in biochemical sciences. 2007;32(5):217-24. 33. Gorbenko GP, Kinnunen PK. The role of lipid-protein interactions in amyloid-type protein fibril formation. Chemistry and physics of lipids. 2006;141(1-2):72-82. 34. PEPTIDE 2.0 internet: Peptide 2.0 INC. Available from: https://www.peptide2.com/N_peptide_hydrophobicity_hydrophilicity.php. 35. Campolongo P, Ratano P, Ciotti MT, Florenzano F, Nori SL, Marolda R, et al. Systemic administration of substance P recovers beta amyloid-induced cognitive deficits in rat: involvement of Kv potassium channels. PLoS One. 2013;8(11):e78036. 36. Friedman R, Pellarin R, Caflisch A. Amyloid aggregation on lipid bilayers and its impact on membrane permeability. Journal of molecular biology. 2009;387(2):407-15. 37. Reynolds NP, Soragni A, Rabe M, Verdes D, Liverani E, Handschin S, et al. Mechanism of Membrane Interaction and Disruption by α-Synuclein. Journal of the American Chemical Society. 2011;133(48):19366-75. 38. Relini A, Marano N, Gliozzi A. Misfolding of Amyloidogenic Proteins and Their Interactions with Membranes. Biomolecules. 2014;4(1):20-55. 39. Goodchild SC, Sheynis T, Thompson R, Tipping KW, Xue W-F, Ranson NA, et al. β2- Microglobulin Amyloid Fibril-Induced Membrane Disruption Is Enhanced by Endosomal Lipids and Acidic pH. PLOS ONE. 2014;9(8):e104492. 40. Perl DP. Neuropathology of Alzheimer's disease. The Mount Sinai journal of medicine, New York. 2010;77(1):32-42. 41. Tofoleanu F, Buchete N-V. Alzheimer Aβ peptide interactions with lipid membranes: Fibrils, oligomers and polymorphic amyloid channels. Prion. 2012;6(4):339-45. 42. Lee J, Kim YH, T Arce F, Gillman AL, Jang H, Kagan BL, et al. Amyloid β Ion Channels in a Membrane Comprising Brain Total Lipid Extracts. ACS chemical neuroscience. 2017;8(6):1348-57.

140

43. Lal R, Lin H, Quist AP. Amyloid beta ion channel: 3D structure and relevance to amyloid channel paradigm. Biochimica et biophysica acta. 2007;1768(8):1966-75. 44. Davis CH, Berkowitz ML. Interaction Between Amyloid-β (1–42) Peptide and Phospholipid Bilayers: A Molecular Dynamics Study. Biophysical Journal. 2009;96(3):785-97. 45. Dobson CM. Principles of protein folding, misfolding and aggregation. Seminars in cell & developmental biology. 2004;15(1):3-16. 46. Hirschfield GM. Amyloidosis: a clinico-pathophysiological synopsis. Seminars in cell & developmental biology. 2004;15(1):39-44. 47. Chamberlain AK, MacPhee CE, Zurdo J, Morozova-Roche LA, Hill HAO, Dobson CM, et al. Ultrastructural Organization of Amyloid Fibrils byAtomic Force Microscopy. Biophysical Journal. 2000;79(6):3282-93. 48. Dai B, Li D, Xi W, Luo F, Zhang X, Zou M, et al. Tunable assembly of amyloid-forming peptides into nanosheets as a retrovirus carrier. Proceedings of the National Academy of Sciences of the United States of America. 2015;112(10):2996-3001. 49. Makin OS, Serpell LC. Structures for amyloid fibrils. The FEBS journal. 2005;272(23):5950-61. 50. Serpell LC. Alzheimer's amyloid fibrils: structure and assembly. Biochimica et biophysica acta. 2000;1502(1):16-30. 51. Dai B, Li D, Xi W, Luo F, Zhang X, Zou M, et al. Tunable assembly of amyloid-forming peptides into nanosheets as a retrovirus carrier. Proceedings of the National Academy of Sciences of the United States of America. 2015;112(10):2996-3001. 52. Pocker Y, Biswas SB. Conformational dynamics of insulin in solution. Circular dichroic studies. Biochemistry. 1980;19(22):5043-9. 53. Breydo L, Uversky VN. Structural, morphological, and functional diversity of amyloid oligomers. FEBS letters. 2015;589(19, Part A):2640-8. 54. Anoop A, Ranganathan S, Dhaked BD, Jha NN, Pratihar S, Ghosh S, et al. Elucidating the Role of Disulfide Bond on Amyloid Formation and Fibril Reversibility of Somatostatin-14: RELEVANCE TO ITS STORAGE AND SECRETION. Journal of Biological Chemistry. 2014;289(24):16884-903. 55. Morozova-Roche LA, Zurdo J, Spencer A, Noppe W, Receveur V, Archer DB, et al. Amyloid fibril formation and seeding by wild-type human lysozyme and its disease-related mutational variants. Journal of structural biology. 2000;130(2-3):339-51. 56. Krebs MR, Wilkins DK, Chung EW, Pitkeathly MC, Chamberlain AK, Zurdo J, et al. Formation and seeding of amyloid fibrils from wild-type hen lysozyme and a peptide fragment from the beta- domain. Journal of molecular biology. 2000;300(3):541-9. 57. Dovidchenko NV, Finkelstein AV, Galzitskaya OV. How to Determine the Size of Folding Nuclei of Protofibrils from the Concentration Dependence of the Rate and Lag-Time of Aggregation. I. Modeling the Amyloid Protofibril Formation. The Journal of Physical Chemistry B. 2014;118(5):1189-97. 58. Serpell LC, Sunde M, Benson MD, Tennent GA, Pepys MB, Fraser PE. The protofilament substructure of amyloid fibrils. Journal of Molecular Biology. 2000;300(5):1033-9. 59. van Grondelle W, Iglesias CL, Coll E, Artzner F, Paternostre M, Lacombe F, et al. Spontaneous fibrillation of the native neuropeptide hormone Somatostatin-14. J Struct Biol. 2007;160(2):211-23. 60. Sunde M, Blake C. The structure of amyloid fibrils by electron microscopy and X-ray diffraction. In: Wetzel R, editor. Advances in Protein Chemistry, Vol 50: Protein Misassembly. Advances in Protein Chemistry. 501997. p. 123-59. 61. Bauer HH, Aebi U, Haner M, Hermann R, Muller M, Arvinte T, et al. ARCHITECTURE AND POLYMORPHISM OF FIBRILLAR SUPRAMOLECULAR ASSEMBLIES PRODUCED BY IN-VITRO AGGREGATION OF HUMAN CALCITONIN. Journal of Structural Biology. 1995;115(1):1-15. 62. Saiki M, Honda S, Kawasaki K, Zhou DS, Kaito A, Konakahara T, et al. Higher-order molecular packing in amyloid-like fibrils constructed with linear arrangements of hydrophobic and hydrogen- bonding side-chains. Journal of Molecular Biology. 2005;348(4):983-98.

141

63. Pedersen JS, Dikov D, Flink JL, Hjuler HA, Christiansen G, Otzen DE. The changing face of glucagon fibrillation: Structural polymorphism and conformational imprinting. Journal of Molecular Biology. 2006;355(3):501-23. 64. Nilsson MR. Techniques to study amyloid fibril formation in vitro. Methods. 2004;34(1):151- 60. 65. Khurana R, Uversky VN, Nielsen L, Fink AL. Is Congo red an amyloid-specific dye? Journal of Biological Chemistry. 2001;276(25):22715-21. 66. Bousset L, Redeker V, Decottignies P, Dubois S, Le Marechal P, Melki R. Structural characterization of the fibrillar form of the yeast Saccharomyces cerevisiae prion Ure2p. Biochemistry. 2004;43(17):5022-32. 67. Chapman MR, Robinson LS, Pinkner JS, Roth R, Heuser J, Hammar M, et al. Role of Escherichia coli curli operons in directing amyloid fiber formation. Science (New York, NY). 2002;295(5556):851-5. 68. Biancalana M, Koide S. Molecular mechanism of Thioflavin-T binding to amyloid fibrils. Biochimica et biophysica acta. 2010;1804(7):1405-12. 69. Murphy RM. Kinetics of amyloid formation and membrane interaction with amyloidogenic proteins. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2007;1768(8):1923-34. 70. Khondker A, Alsop RJ, Rheinstädter MC. Membrane-Accelerated Amyloid-β Aggregation and Formation of Cross-β Sheets. Membranes. 2017;7(3):49. 71. Sengupta U, Nilson AN, Kayed R. The Role of Amyloid-β Oligomers in Toxicity, Propagation, and Immunotherapy. EBioMedicine. 2016;6:42-9. 72. Engel MFM, Khemtémourian L, Kleijer CC, Meeldijk HJD, Jacobs J, Verkleij AJ, et al. Membrane damage by human islet amyloid polypeptide through fibril growth at the membrane. Proceedings of the National Academy of Sciences. 2008;105(16):6033-8. 73. Thorn DC, Meehan S, Sunde M, Rekas A, Gras SL, MacPhee CE, et al. Amyloid fibril formation by bovine milk kappa-casein and its inhibition by the molecular chaperones alphaS- and beta-casein. Biochemistry. 2005;44(51):17027-36. 74. Stefani M, Dobson CM. Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases and biological evolution. Journal of Molecular Medicine-Jmm. 2003;81(11):678-99. 75. Uversky VN, Fink AL. Conformational constraints for amyloid fibrillation: the importance of being unfolded. Biochimica Et Biophysica Acta-Proteins and Proteomics. 2004;1698(2):131-53. 76. Furukawa Y, Nukina N. Functional diversity of protein fibrillar aggregates from physiology to RNA granules to neurodegenerative diseases. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease. 2013;1832(8):1271-8. 77. Maji SK, Perrin MH, Sawaya MR, Jessberger S, Vadodaria K, Rissman RA, et al. Functional amyloids as natural storage of peptide hormones in pituitary secretory granules. Science (New York, NY). 2009;325(5938):328-32. 78. Badtke MP, Hammer ND, Chapman MR. Functional amyloids signal their arrival. Science signaling. 2009;2(80):pe43. 79. Blanco LP, Evans ML, Smith DR, Badtke MP, Chapman MR. Diversity, biogenesis and function of microbial amyloids. Trends in microbiology. 2012;20(2):66-73. 80. Pham CL, Kwan AH, Sunde M. Functional amyloid: widespread in Nature, diverse in purpose. Essays in biochemistry. 2014;56:207-19. 81. Claessen D, Rink R, de Jong W, Siebring J, de Vreugd P, Boersma FG, et al. A novel class of secreted hydrophobic proteins is involved in aerial hyphae formation in Streptomyces coelicolor by forming amyloid-like fibrils. Genes & development. 2003;17(14):1714-26. 82. Baxa U, Cheng N, Winkler DC, Chiu TK, Davies DR, Sharma D, et al. Filaments of the Ure2p prion protein have a cross-beta core structure. Journal of structural biology. 2005;150(2):170-9. 83. Shorter J, Lindquist S. Prions as adaptive conduits of memory and inheritance. Nature Reviews Genetics. 2005;6(6):435-50.

142

84. Dos Reis S, Coulary-Salin B, Forge V, Lascu I, Bégueret J, Saupe SJ. The HET-s prion protein of the filamentous fungus Podospora anserina aggregates in vitro into amyloid-like fibrils. Journal of Biological Chemistry. 2002;277(8):5703-6. 85. Balguerie A, Dos Reis S, Ritter C, Chaignepain S, Coulary-Salin B, Forge V, et al. Domain organization and structure-function relationship of the HET-s prion protein of Podospora anserina. EMBO Journal. 2003;22(9):2071-81. 86. Ritter C, Maddelein ML, Siemer AB, Lührs T, Ernst M, Meier BH, et al. Correlation of structural elements and infectivity of the HET-s prion. Nature. 2005;435(7043):844-8. 87. Uptain SM, Lindquist S. Prions as protein-based genetic elements. Annual Review of Microbiology2002. p. 703-41. 88. Baxa U, Cheng N, Winkler DC, Chiu TK, Davies DR, Sharma D, et al. Filaments of the Ure2p prion protein have a cross-β core structure. Journal of Structural Biology. 2005;150(2):170-9. 89. King CY, Tittmann P, Gross H, Gebert R, Aebi M, Wüthrich K. Prion-inducing domain 2-114 of yeast Sup35 protein transforms in vitro into amyloid-like filaments. Proceedings of the National Academy of Sciences of the United States of America. 1997;94(13):6618-22. 90. Mackay JP, Matthews JM, Winefield RD, Mackay LG, Haverkamp RG, Templeton MD. The hydrophobin EAS is largely unstructured in solution and functions by forming amyloid-like structures. Structure. 2001;9(2):83-91. 91. Butko P, Buford JP, Goodwin JS, Stroud PA, McCormick CL, Cannon GC. Spectroscopic evidence for amyloid-like interfacial self-assembly of hydrophobin Sc3. Biochemical and biophysical research communications. 2001;280(1):212-5. 92. De Vocht ML, Reviakine I, Wosten HAB, Brisson A, Wessels JGH, Robillard GT. Structural and functional role of the disulfide bridges in the hydrophobin SC3. Journal of Biological Chemistry. 2000;275(37):28428-32. 93. Kwan AHY, Winefield RD, Sunde M, Matthews JM, Haverkamp RG, Templeton MD, et al. Structural basis for rodlet assembly in fungal hydrophobins. Proceedings of the National Academy of Sciences of the United States of America. 2006;103(10):3621-6. 94. Wosten HA, de Vocht ML. Hydrophobins, the fungal coat unravelled. Biochimica et biophysica acta. 2000;1469(2):79-86. 95. Iconomidou VA, Vriend G, Hamodrakas SJ. Amyloids protect the silkmoth oocyte and embryo. FEBS letters. 2000;479(3):141-5. 96. Kenney JM, Knight D, Wise MJ, Vollrath F. Amyloidogenic nature of spider silk. European journal of biochemistry. 2002;269(16):4159-63. 97. Fowler DM, Koulov AV, Alory-Jost C, Marks MS, Balch WE, Kelly JW. Functional amyloid formation within mammalian tissue. PLoS Biol. 2006;4(1):e6. 98. Dharmadana D, Reynolds NP, Dekiwadia C, Conn CE, Valéry C. Heparin assisted assembly of somatostatin amyloid nanofibrils results in disordered precipitates by hindrance of protofilaments interactions. Nanoscale. 2018. 99. Maji SK, Perrin MH, Sawaya MR, Jessberger S, Vadodaria K, Rissman RA, et al. Functional Amyloids as Natural Storage of Peptide Hormones in Pituitary Secretory Granules. Science (New York, NY). 2009;325(5938):328-32. 100. Gobeaux F, Wien F. Reversible Assembly of a Drug Peptide into Amyloid Fibrils: A Dynamic Circular Dichroism Study. Langmuir. 2018;34(24):7180-91. 101. Jing W, Guo F, Cheng L, Zhang J-F, Qi J-S. Arginine vasopressin prevents amyloid β protein- induced impairment of long-term potentiation in rat hippocampus in vivo. Neuroscience letters. 2009;450(3):306-10. 102. Schurmeyer T, Knuth UA, Freischem CW, Sandow J, Akhtar FB, Nieschlag E. Suppression of pituitary and testicular function in normal men by constant gonadotropin-releasing hormone agonist infusion. The Journal of clinical endocrinology and metabolism. 1984;59(1):19-24.

143

103. Casadesus G, Webber KM, Atwood CS, Pappolla MA, Perry G, Bowen RL, et al. Luteinizing hormone modulates cognition and amyloid-β deposition in Alzheimer APP transgenic mice. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease. 2006;1762(4):447-52. 104. Poujade C, Lavielle S, Chassaing G, Marquet A. Properties of substance P aggregates. Application to the synthesis and purification of substance P. Biochemical and biophysical research communications. 1983;114(3):1109-16. 105. Fowler D, Koulov A, E Balch W, Kelly J. Functional amyloid—From bacteria to humans2007. 217-24 p. 106. Lundmark K, Westermark GT, Olsen A, Westermark P. Protein fibrils in nature can enhance amyloid protein A amyloidosis in mice: Cross-seeding as a disease mechanism. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(17):6098-102. 107. Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M, et al. Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proceedings of the National Academy of Sciences of the United States of America. 1998;95(11):6448- 53. 108. McLean CA, Cherny RA, Fraser FW, Fuller SJ, Smith MJ, Beyreuther K, et al. Soluble pool of Abeta amyloid as a determinant of severity of neurodegeneration in Alzheimer's disease. Annals of neurology. 1999;46(6):860-6. 109. Selkoe DJ. Folding proteins in fatal ways. Nature. 2003;426(6968):900-4. 110. Cohlberg JA, Li J, Uversky VN, Fink AL. Heparin and Other Glycosaminoglycans Stimulate the Formation of Amyloid Fibrils from α-Synuclein in Vitro. Biochemistry. 2002;41(5):1502-11. 111. Leal SS, Botelho HM, Gomes CM. Metal ions as modulators of protein conformation and misfolding in neurodegeneration. Coordination Chemistry Reviews. 2012;256(19–20):2253-70. 112. Pfaender S, Grabrucker AM. Characterization of biometal profiles in neurological disorders. Metallomics : integrated biometal science. 2014;6(5):960-77. 113. Frederickson CJ. Neurobiology of zinc and zinc-containing neurons. International review of neurobiology. 1989;31:145-238. 114. Madsen E, Gitlin JD. Copper and iron disorders of the brain. Annual review of neuroscience. 2007;30:317-37. 115. Bush AI. Metal complexing agents as therapies for Alzheimer's disease. Neurobiology of aging. 2002;23(6):1031-8. 116. Castellani RJ, Siedlak SL, Perry G, Smith MA. Sequestration of iron by Lewy bodies in Parkinson's disease. Acta neuropathologica. 2000;100(2):111-4. 117. DeToma AS, Salamekh S, Ramamoorthy A, Lim MH. Misfolded proteins in Alzheimer's disease and type II diabetes. Chem Soc Rev. 2012;41(2):608-21. 118. Rasia RM, Bertoncini CW, Marsh D, Hoyer W, Cherny D, Zweckstetter M, et al. Structural characterization of copper(II) binding to α-synuclein: Insights into the bioinorganic chemistry of Parkinson's disease. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(12):4294-9. 119. Lee SJC, Choi TS, Lee JW, Lee HJ, Mun D-G, Akashi S, et al. Structure and assembly mechanisms of toxic human islet amyloid polypeptide oligomers associated with copper. Chemical Science. 2016;7(8):5398-406. 120. Sarell CJ, Wilkinson SR, Viles JH. Substoichiometric levels of Cu2+ ions accelerate the kinetics of fiber formation and promote cell toxicity of amyloid-{beta} from Alzheimer disease. The Journal of biological chemistry. 2010;285(53):41533-40. 121. Karr JW, Szalai VA. Cu(II) binding to monomeric, oligomeric, and fibrillar forms of the Alzheimer's disease amyloid-beta peptide. Biochemistry. 2008;47(17):5006-16. 122. Bolognin S, Messori L, Drago D, Gabbiani C, Cendron L, Zatta P. Aluminum, copper, iron and zinc differentially alter amyloid-Abeta(1-42) aggregation and toxicity. The international journal of biochemistry & cell biology. 2011;43(6):877-85.

144

123. Wallin C, Kulkarni YS, Abelein A, Jarvet J, Liao Q, Strodel B, et al. Characterization of Mn(II) ion binding to the amyloid-beta peptide in Alzheimer's disease. Journal of trace elements in medicine and biology : organ of the Society for Minerals and Trace Elements (GMS). 2016;38:183-93. 124. Lisi G, Png CYM, E Wilcox D. Thermodynamic Contributions to the Stability of the Insulin Hexamer2014. 125. Faller P, Hureau C, La Penna G. Metal ions and intrinsically disordered proteins and peptides: from Cu/Zn amyloid-beta to general principles. Accounts of chemical research. 2014;47(8):2252-9. 126. Iannuzzi C, Irace G, Sirangelo I. The effect of glycosaminoglycans (GAGs) on amyloid aggregation and toxicity. Molecules (Basel, Switzerland). 2015;20(2):2510-28. 127. Iannuzzi C, Irace G, Sirangelo I. The effect of glycosaminoglycans (GAGs) on amyloid aggregation and toxicity. Molecules. 2015;20(2):2510-28. 128. Castillo GM, Lukito W, Wight TN, Snow AD. The sulfate moieties of glycosaminoglycans are critical for the enhancement of beta-amyloid protein fibril formation. Journal of neurochemistry. 1999;72(4):1681-7. 129. Fraser PE, Darabie AA, McLaurin JA. Amyloid-beta interactions with chondroitin sulfate- derived monosaccharides and disaccharides. implications for drug development. The Journal of biological chemistry. 2001;276(9):6412-9. 130. Valle-Delgado JJ, Alfonso-Prieto M, de Groot NS, Ventura S, Samitier J, Rovira C, et al. Modulation of Abeta42 fibrillogenesis by glycosaminoglycan structure. FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 2010;24(11):4250-61. 131. Adamcik J, Sanchez-Ferrer A, Ait-Bouziad N, Reynolds NP, Lashuel HA, Mezzenga R. Microtubule-Binding R3 Fragment from Tau Self-Assembles into Giant Multistranded Amyloid Ribbons. Angewandte Chemie (International ed in English). 2016;55(2):618-22. 132. Goedert M, Jakes R, Spillantini MG, Hasegawa M, Smith MJ, Crowther RA. Assembly of microtubule-associated protein tau into Alzheimer-like filaments induced by sulphated glycosaminoglycans. Nature. 1996;383(6600):550-3. 133. Cohlberg JA, Li J, Uversky VN, Fink AL. Heparin and other glycosaminoglycans stimulate the formation of amyloid fibrils from alpha-synuclein in vitro. Biochemistry. 2002;41(5):1502-11. 134. Solomon JP, Bourgault S, Powers ET, Kelly JW. Heparin Binds 8 kDa Gelsolin Cross-β-Sheet Oligomers and Accelerates Amyloidogenesis by Hastening Fibril Extension. Biochemistry. 2011;50(13):2486-98. 135. Suk JY, Zhang F, Balch WE, Linhardt RJ, Kelly JW. Heparin accelerates gelsolin amyloidogenesis. Biochemistry. 2006;45(7):2234-42. 136. Meng F, Abedini A, Song B, Raleigh DP. Amyloid formation by pro-islet amyloid polypeptide processing intermediates: examination of the role of protein heparan sulfate interactions and implications for islet amyloid formation in type 2 diabetes. Biochemistry. 2007;46(43):12091-9. 137. Calamai M, Kumita JR, Mifsud J, Parrini C, Ramazzotti M, Ramponi G, et al. Nature and significance of the interactions between amyloid fibrils and biological polyelectrolytes. Biochemistry. 2006;45(42):12806-15. 138. Iannuzzi C, Irace G, Sirangelo I. The Effect of Glycosaminoglycans (GAGs) on Amyloid Aggregation and Toxicity. Molecules (Basel, Switzerland). 2015;20(2):2510. 139. Villaflores OB, Chen Y-J, Chen C-P, Yeh J-M, Wu T-Y. Curcuminoids and resveratrol as anti- Alzheimer agents. Taiwanese Journal of Obstetrics and Gynecology. 2012;51(4):515-25. 140. Jia Y, Liu Z, Huo X, Wang C, Meng Q, Liu Q, et al. Enhancement effect of resveratrol on the intestinal absorption of bestatin by regulating PEPT1, MDR1 and MRP2 in vivo and in vitro. International journal of pharmaceutics. 2015;495(1):588-98. 141. Rege SD, Geetha T, Griffin GD, Broderick TL, Babu JR. Neuroprotective effects of resveratrol in Alzheimer disease pathology. Frontiers in aging neuroscience. 2014;6:218. 142. Braidy N, Jugder BE, Poljak A, Jayasena T, Mansour H, Nabavi SM, et al. Resveratrol as a Potential Therapeutic Candidate for the Treatment and Management of Alzheimer's Disease. Current topics in medicinal chemistry. 2016;16(17):1951-60.

145

143. Feng Y, Wang XP, Yang SG, Wang YJ, Zhang X, Du XT, et al. Resveratrol inhibits beta-amyloid oligomeric cytotoxicity but does not prevent oligomer formation. Neurotoxicology. 2009;30(6):986- 95. 144. Song YM, Ha YM, Kim JA, Chung KW, Uehara Y, Lee KJ, et al. Synthesis of novel azo- resveratrol, azo-oxyresveratrol and their derivatives as potent tyrosinase inhibitors. Bioorganic & medicinal chemistry letters. 2012;22(24):7451-5. 145. Lamuela-Raventos RM, Romero-Perez AI, Waterhouse AL, de la Torre-Boronat MC. Direct HPLC Analysis of cis- and trans-Resveratrol and Piceid Isomers in Spanish Red Vitis vinifera Wines. Journal of Agricultural and Food Chemistry. 1995;43(2):281-3. 146. Ge J-F, Qiao J-P, Qi C-C, Wang C-W, Zhou J-N. The binding of resveratrol to monomer and fibril amyloid beta. Neurochemistry International. 2012;61(7):1192-201. 147. Singer SJ, Nicolson GL. The fluid mosaic model of the structure of cell membranes. Science (New York, NY). 1972;175(4023):720-31. 148. Spruce AE, Iwata A, White JM, Almers W. Patch clamp studies of single cell-fusion events mediated by a viral fusion protein. Nature. 1989;342(6249):555-8. 149. Seifert K, Fendler K, Bamberg E. Charge transport by ion translocating membrane proteins on solid supported membranes. Biophysical Journal. 1993;64(2):384-91. 150. Engel MFM, Yigittop H, Elgersma RC, Rijkers DTS, Liskamp RMJ, de Kruijff B, et al. Islet Amyloid Polypeptide Inserts into Phospholipid Monolayers as Monomer. Journal of Molecular Biology. 2006;356(3):783-9. 151. Zhu M, Li J, Fink AL. The association of alpha-synuclein with membranes affects bilayer structure, stability, and fibril formation. J Biol Chem. 2003;278(41):40186-97. 152. Jo E, McLaurin J, Yip CM, St George-Hyslop P, Fraser PE. alpha-Synuclein membrane interactions and lipid specificity. J Biol Chem. 2000;275(44):34328-34. 153. Necula M, Chirita CN, Kuret J. Rapid Anionic Micelle-mediated α-Synuclein Fibrillization in Vitro. Journal of Biological Chemistry. 2003;278(47):46674-80. 154. Lee HJ, Choi C, Lee SJ. Membrane-bound alpha-synuclein has a high aggregation propensity and the ability to seed the aggregation of the cytosolic form. J Biol Chem. 2002;277(1):671-8. 155. Jo E, Darabie AA, Han K, Tandon A, Fraser PE, McLaurin J. α-Synuclein–synaptosomal membrane interactions. European Journal of Biochemistry. 2004;271(15):3180-9. 156. Yip CM, Elton EA, Darabie AA, Morrison MR, McLaurin J. Cholesterol, a modulator of membrane-associated aβ-fibrillogenesis and neurotoxicity1. Journal of Molecular Biology. 2001;311(4):723-34. 157. Bokvist M, Lindstrom F, Watts A, Grobner G. Two types of Alzheimer's beta-amyloid (1-40) peptide membrane interactions: aggregation preventing transmembrane anchoring versus accelerated surface fibril formation. J Mol Biol. 2004;335(4):1039-49. 158. Choo-Smith L-Pi, Garzon-Rodriguez W, Glabe CG, Surewicz WK. Acceleration of Amyloid Fibril Formation by Specific Binding of Aβ-(1–40) Peptide to Ganglioside-containing Membrane Vesicles. Journal of Biological Chemistry. 1997;272(37):22987-90. 159. Chauhan A, Ray I, Chauhan VPS. Interaction of Amyloid Beta-Protein with Anionic Phospholipids: Possible Involvement of Lys28 and C-Terminus Aliphatic Amino Acids. Neurochemical research. 2000;25(3):423-9. 160. Chirita CN, Necula M, Kuret J. Anionic Micelles and Vesicles Induce Tau Fibrillization in Vitro. Journal of Biological Chemistry. 2003;278(28):25644-50. 161. Knight JD, Miranker AD. Phospholipid catalysis of diabetic amyloid assembly. J Mol Biol. 2004;341(5):1175-87. 162. Sparr E, Engel MF, Sakharov DV, Sprong M, Jacobs J, de Kruijff B, et al. Islet amyloid polypeptide-induced membrane leakage involves uptake of lipids by forming amyloid fibers. FEBS letters. 2004;577(1-2):117-20.

146

163. Kazlauskaite J, Sanghera N, Sylvester I, Vénien-Bryan C, Pinheiro TJT. Structural Changes of the Prion Protein in Lipid Membranes Leading to Aggregation and Fibrillization. Biochemistry. 2003;42(11):3295-304. 164. Zhao H, Tuominen EKJ, Kinnunen PKJ. Formation of Amyloid Fibers Triggered by Phosphatidylserine-Containing Membranes. Biochemistry. 2004;43(32):10302-7. 165. Wang W. Protein aggregation and its inhibition in biopharmaceutics. International journal of pharmaceutics. 2005;289(1-2):1-30. 166. Shen L, Adachi T, Vanden Bout D, Zhu XY. A Mobile Precursor Determines Amyloid-β Peptide Fibril Formation at Interfaces. Journal of the American Chemical Society. 2012;134(34):14172-8. 167. Davidson WS, Jonas A, Clayton DF, George JM. Stabilization of alpha-synuclein secondary structure upon binding to synthetic membranes. J Biol Chem. 1998;273(16):9443-9. 168. Apostolidou M, Jayasinghe SA, Langen R. Structure of α-Helical Membrane-bound Human Islet Amyloid Polypeptide and Its Implications for Membrane-mediated Misfolding. Journal of Biological Chemistry. 2008;283(25):17205-10. 169. Lee CC, Sun Y, Huang HW. How type II diabetes-related islet amyloid polypeptide damages lipid bilayers. Biophys J. 2012;102(5):1059-68. 170. Yuyama K, Yanagisawa K. Sphingomyelin accumulation provides a favorable milieu for GM1 ganglioside-induced assembly of amyloid beta-protein. Neuroscience letters. 2010;481(3):168-72. 171. Matsuzaki K. Formation of Toxic Amyloid Fibrils by Amyloid β-Protein on Ganglioside Clusters2011. 956104 p. 172. Matsubara T, Nishihara M, Yasumori H, Nakai M, Yanagisawa K, Sato T. Size and Shape of Amyloid Fibrils Induced by Ganglioside Nanoclusters: Role of Sialyl Oligosaccharide in Fibril Formation. Langmuir. 2017;33(48):13874-81. 173. Yanagisawa K. Pathological significance of ganglioside clusters in Alzheimer's disease. Journal of neurochemistry. 2011;116(5):806-12. 174. Harris JR. In vitro fibrillogenesis of the amyloid beta 1-42 peptide: cholesterol potentiation and aspirin inhibition. Micron (Oxford, England : 1993). 2002;33(7-8):609-26. 175. Matsuzaki K. Physicochemical interactions of amyloid β-peptide with lipid bilayers. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2007;1768(8):1935-42. 176. Qiu LM, Lewis A, Como J, Vaughn MW, Huang JY, Somerharju P, et al. Cholesterol Modulates the Interaction of beta-Amyloid Peptide with Lipid Bilayers. Biophys J. 2009;96(10):4299-307. 177. Curtain CC, Ali FE, Smith DG, Bush AI, Masters CL, Barnham KJ. Metal ions, pH, and cholesterol regulate the interactions of Alzheimer's disease amyloid-beta peptide with membrane lipid. Journal of Biological Chemistry. 2003;278(5):2977-82. 178. Okada T, Wakabayashi M, Ikeda K, Matsuzaki K. Formation of toxic fibrils of Alzheimer's amyloid beta-protein-(1-40) by monosialoganglioside GM1, a neuronal membrane component. J Mol Biol. 2007;371(2):481-9. 179. Matsuzaki K. Physicochemical interactions of amyloid beta-peptide with lipid bilayers. Biochimica et biophysica acta. 2007;1768(8):1935-42. 180. Kurganov B, Doh M, Arispe N. Aggregation of liposomes induced by the toxic peptides Alzheimer’s Aβs, human amylin and prion (106–126): facilitation by membrane-bound GM1 ganglioside. Peptides. 2004;25(2):217-32. 181. D'Errico G, Vitiello G, Ortona O, Tedeschi A, Ramunno A, D'Ursi AM. Interaction between Alzheimer's Aβ(25–35) peptide and phospholipid bilayers: The role of cholesterol. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2008;1778(12):2710-6. 182. Kucerka N, Perlmutter JD, Pan J, Tristram-Nagle S, Katsaras J, Sachs JN. The effect of cholesterol on short- and long-chain monounsaturated lipid bilayers as determined by molecular dynamics simulations and X-ray scattering. Biophysical journal. 2008;95(6):2792-805. 183. Williams TL, Serpell LC. Membrane and surface interactions of Alzheimer's Abeta peptide-- insights into the mechanism of cytotoxicity. The FEBS journal. 2011;278(20):3905-17.

147

184. Stefani M. Biochemical and biophysical features of both oligomer/fibril and cell membrane in amyloid cytotoxicity. The FEBS journal. 2010;277(22):4602-13. 185. Jayasinghe SA, Langen R. Lipid membranes modulate the structure of islet amyloid polypeptide. Biochemistry. 2005;44(36):12113-9. 186. Chi EY, Ege C, Winans A, Majewski J, Wu G, Kjaer K, et al. Lipid membrane templates the ordering and induces the fibrillogenesis of Alzheimer's disease amyloid-beta peptide. Proteins. 2008;72(1):1-24. 187. Bokvist M, Grobner G. Misfolding of amyloidogenic proteins at membrane surfaces: the impact of macromolecular crowding. J Am Chem Soc. 2007;129(48):14848-9. 188. Terakawa MS, Yagi H, Adachi M, Lee YH, Goto Y. Small liposomes accelerate the fibrillation of amyloid beta (1-40). J Biol Chem. 2015;290(2):815-26. 189. Wong PT, Schauerte JA, Wisser KC, Ding H, Lee EL, Steel DG, et al. Amyloid-beta membrane binding and permeabilization are distinct processes influenced separately by membrane charge and fluidity. J Mol Biol. 2009;386(1):81-96. 190. Pandey AP, Haque F, Rochet JC, Hovis JS. Clustering of alpha-synuclein on supported lipid bilayers: role of anionic lipid, protein, and divalent ion concentration. Biophys J. 2009;96(2):540-51. 191. Li X, Wan M, Gao L, Fang W. Mechanism of Inhibition of Human Islet Amyloid Polypeptide- Induced Membrane Damage by a Small Organic Fluorogen. Scientific reports. 2016;6:21614. 192. Stockl M, Fischer P, Wanker E, Herrmann A. Alpha-synuclein selectively binds to anionic phospholipids embedded in liquid-disordered domains. J Mol Biol. 2008;375(5):1394-404. 193. van Rooijen BD, Claessens MM, Subramaniam V. Membrane binding of oligomeric alpha- synuclein depends on bilayer charge and packing. FEBS letters. 2008;582(27):3788-92. 194. Ding H, Schauerte Joseph A, Steel Duncan G, Gafni A. β-Amyloid (1–40) Peptide Interactions with Supported Phospholipid Membranes: A Single-Molecule Study. Biophysical Journal. 2012;103(7):1500-9. 195. Terzi E, Holzemann G, Seelig J. Interaction of Alzheimer beta-amyloid peptide(1-40) with lipid membranes. Biochemistry. 1997;36(48):14845-52. 196. Stefanovic A, Stöckl M, Claessens M, Subramaniam V. Alpha-synuclein Oligomers Distinctively Permeabilize Complex Model Membranes2014. 197. Stefanovic AN, Stockl MT, Claessens MM, Subramaniam V. alpha-Synuclein oligomers distinctively permeabilize complex model membranes. The FEBS journal. 2014;281(12):2838-50. 198. Lopez CG, Colby RH, Cabral JT. Electrostatic and Hydrophobic Interactions in NaCMC Aqueous Solutions: Effect of Degree of Substitution. Macromolecules. 2018;51(8):3165-75. 199. McLaurin J, Chakrabartty A. Membrane disruption by Alzheimer beta-amyloid peptides mediated through specific binding to either phospholipids or gangliosides. Implications for neurotoxicity. J Biol Chem. 1996;271(43):26482-9. 200. Choo-Smith L-Pi, Surewicz WK. The interaction between Alzheimer amyloid β(1–40) peptide and ganglioside GM1‐containing membranes. FEBS letters. 1997;402(2-3):95-8. 201. Matsuzaki K, Horikiri C. Interactions of Amyloid β-Peptide (1−40) with Ganglioside- Containing Membranes. Biochemistry. 1999;38(13):4137-42. 202. McLaurin J, Franklin T, Fraser PE, Chakrabartty A. Structural transitions associated with the interaction of Alzheimer beta-amyloid peptides with gangliosides. J Biol Chem. 1998;273(8):4506-15. 203. Hane F, Drolle E, Gaikwad R, Faught E, Leonenko Z. Amyloid-beta aggregation on model lipid membranes: an atomic force microscopy study. J Alzheimers Dis. 2011;26(3):485-94. 204. Sunde M, Serpell LC, Bartlam M, Fraser PE, Pepys MB, Blake CC. Common core structure of amyloid fibrils by synchrotron X-ray diffraction. J Mol Biol. 1997;273(3):729-39. 205. Wheeler KP. The lively membranes: by R N Robertson. pp 206. Cambridge University Press. 1983. £12.50, £4.95 ISBN 0-521-23747-5 or ISBN 0-521-28202-0, paperback. Biochemical Education. 1985;13(1):40.

148

206. Kucerka N, Perlmutter JD, Pan J, Tristram-Nagle S, Katsaras J, Sachs JN. The effect of cholesterol on short- and long-chain monounsaturated lipid bilayers as determined by molecular dynamics simulations and X-ray scattering. Biophys J. 2008;95(6):2792-805. 207. Kucerka N, Nieh MP, Katsaras J. Fluid phase lipid areas and bilayer thicknesses of commonly used phosphatidylcholines as a function of temperature. Biochimica et biophysica acta. 2011;1808(11):2761-71. 208. Balleza D. Mechanical properties of lipid bilayers and regulation of mechanosensitive function: from biological to biomimetic channels. Channels (Austin, Tex). 2012;6(4):220-33. 209. Marrink SJ, Risselada J, Mark AE. Simulation of gel phase formation and melting in lipid bilayers using a coarse grained model. Chemistry and physics of lipids. 2005;135(2):223-44. 210. Mansilla MC, Cybulski LE, Albanesi D, de Mendoza D. Control of membrane lipid fluidity by molecular thermosensors. Journal of bacteriology. 2004;186(20):6681-8. 211. Chapter 2 Fundamental aspects of biological membranes. In: Ti Tien H, Ottova-Leitmannova A, editors. Membrane Science and Technology. 5: Elsevier; 2000. p. 23-82. 212. Pichler H, Emmerstorfer-Augustin A. Modification of membrane lipid compositions in single- celled organisms – From basics to applications. Methods. 2018;147:50-65. 213. Pirc K, Ulrih NP. alpha-Synuclein interactions with phospholipid model membranes: Key roles for electrostatic interactions and lipid-bilayer structure. Biochimica et biophysica acta. 2015;1848(10 Pt A):2002-12. 214. Shvadchak VV, Yushchenko DA, Pievo R, Jovin TM. The mode of alpha-synuclein binding to membranes depends on lipid composition and lipid to protein ratio. FEBS letters. 2011;585(22):3513-9. 215. Dies H, Toppozini L, Rheinstädter MC. The Interaction between Amyloid-β Peptides and Anionic Lipid Membranes Containing Cholesterol and Melatonin. PLOS ONE. 2014;9(6):e99124. 216. Marquardt D, Heberle FA, Greathouse DV, Koeppe RE, Standaert RF, Van Oosten BJ, et al. Lipid bilayer thickness determines cholesterol's location in model membranes. Soft Matter. 2016;12(47):9417-28. 217. Bhattacharya S, Haldar S. Interactions between cholesterol and lipids in bilayer membranes. Role of lipid headgroup and hydrocarbon chain–backbone linkage. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2000;1467(1):39-53. 218. Voinov Maxim A, Rivera-Rivera I, Smirnov Alex I. Surface Electrostatics of Lipid Bilayers by EPR of a pH-Sensitive Spin-Labeled Lipid. Biophysical Journal. 2013;104(1):106-16. 219. Lopez Cascales JJ, D. Oliveira Costa S, Garro A, Enriz RD. Mechanical properties of binary DPPC/DPPS bilayers2012. 11743-50 p. 220. Lemkul JA, Bevan DR. Lipid composition influences the release of Alzheimer's amyloid β- peptide from membranes. Protein science : a publication of the Protein Society. 2011;20(9):1530-45. 221. Nakazawa Y, Suzuki Y, Williamson MP, Saitô H, Asakura T. The interaction of amyloid Aβ(1– 40) with lipid bilayers and ganglioside as studied by 31P solid-state NMR. Chemistry and physics of lipids. 2009;158(1):54-60. 222. Siegel DP. Inverted micellar intermediates and the transitions between lamellar, cubic, and inverted hexagonal lipid phases. II. Implications for membrane-membrane interactions and membrane fusion. Biophysical Journal. 1986;49(6):1171-83. 223. Last NB, Rhoades E, Miranker AD. Islet amyloid polypeptide demonstrates a persistent capacity to disrupt membrane integrity. Proceedings of the National Academy of Sciences. 2011;108(23):9460-5. 224. Canale C, Torrassa S, Rispoli P, Relini A, Rolandi R, Bucciantini M, et al. Natively Folded HypF- N and Its Early Amyloid Aggregates Interact with Phospholipid Monolayers and Destabilize Supported Phospholipid Bilayers. Biophysical Journal. 2006;91(12):4575-88. 225. Sparr E, Engel MFM, Sakharov DV, Sprong M, Jacobs J, de Kruijff B, et al. Islet amyloid polypeptide-induced membrane leakage involves uptake of lipids by forming amyloid fibers. FEBS letters. 2004;577(1-2):117-20.

149

226. Relini A, Marano N, Gliozzi A. Probing the interplay between amyloidogenic proteins and membranes using lipid monolayers and bilayers. Advances in Colloid and Interface Science. 2014;207:81-92. 227. Rossi C, Chopineau J. Biomimetic tethered lipid membranes designed for membrane-protein interaction studies. European biophysics journal : EBJ. 2007;36(8):955-65. 228. Richter RP, Berat R, Brisson AR. Formation of solid-supported lipid bilayers: an integrated view. Langmuir. 2006;22(8):3497-505. 229. Devanathan S, Salamon Z, Lindblom G, Grobner G, Tollin G. Effects of sphingomyelin, cholesterol and zinc ions on the binding, insertion and aggregation of the amyloid Abeta(1-40) peptide in solid-supported lipid bilayers. The FEBS journal. 2006;273(7):1389-402. 230. van Meer G, Voelker DR, Feigenson GW. Membrane lipids: where they are and how they behave. Nat Rev Mol Cell Biol. 2008;9(2):112-24. 231. Conn C, Ces O, Mulet X, Finet S, Winter R, Seddon J, et al. Dynamics of Structural Transformations between Lamellar and Inverse Bicontinuous Cubic Lyotropic Phases2006. 108102 p. 232. Akbarzadeh A, Rezaei-Sadabady R, Davaran S, Joo SW, Zarghami N, Hanifehpour Y, et al. Liposome: classification, preparation, and applications. Nanoscale research letters. 2013;8(1):102-. 233. Sahoo SK, Labhasetwar V. Nanotech approaches to drug delivery and imaging. Drug discovery today. 2003;8(24):1112-20. 234. Allen TM. Liposomes. Opportunities in drug delivery. Drugs. 1997;54 Suppl 4:8-14. 235. Gabizon A, Goren D, Cohen R, Barenholz Y. Development of liposomal anthracyclines: from basics to clinical applications. Journal of controlled release : official journal of the Controlled Release Society. 1998;53(1-3):275-9. 236. Sharma A, Sharma US. Liposomes in drug delivery: Progress and limitations. International journal of pharmaceutics. 1997;154(2):123-40. 237. Shehata T, Ogawara K, Higaki K, Kimura T. Prolongation of residence time of liposome by surface-modification with mixture of hydrophilic polymers. International journal of pharmaceutics. 2008;359(1-2):272-9. 238. Johnston MJ, Semple SC, Klimuk SK, Ansell S, Maurer N, Cullis PR. Characterization of the drug retention and pharmacokinetic properties of liposomal nanoparticles containing dihydrosphingomyelin. Biochimica et biophysica acta. 2007;1768(5):1121-7. 239. Schroeder A, Turjeman K, Schroeder JE, Leibergall M, Barenholz Y. Using liposomes to target infection and inflammation induced by foreign body injuries or medical implants. Expert opinion on drug delivery. 2010;7(10):1175-89. 240. Tanifum EA, Ghaghada K, Vollert C, Head E, Eriksen JL, Annapragada A. A Novel liposomal nanoparticle for the imaging of amyloid plaque by MRI. Journal of Alzheimer's disease : JAD. 2016;52(2):731-45. 241. Amar-Yuli I, Adamcik J, Lara C, Bolisetty S, Vallooran JJ, Mezzenga R. Templating effects of lyotropic liquid crystals in the encapsulation of amyloid fibrils and their stimuli-responsive magnetic behavior. Soft Matter. 2011;7(7):3348-57. 242. Conn CE, Drummond CJ. Nanostructured bicontinuous cubic lipid self-assembly materials as matrices for protein encapsulation. Soft Matter. 2013;9(13):3449-64. 243. Yang M, Wang K, Lin J, Wang L, Wei F, Zhu J, et al. Gel Phase Membrane Retards Amyloid beta-Peptide (1-42) Fibrillation by Restricting Slaved Diffusion of Peptides on Lipid Bilayers. Langmuir. 2018;34(28):8408-14. 244. Ritwiset A, Krongsuk S, Johns JR. Molecular structure and dynamical properties of niosome bilayers with and without cholesterol incorporation: A molecular dynamics simulation study. Applied Surface Science. 2016;380:23-31. 245. Drolle E, Negoda A, Hammond K, Pavlov E, Leonenko Z. Changes in lipid membranes may trigger amyloid toxicity in Alzheimer's disease. PloS one. 2017;12(8):e0182194-e. 246. Dies H, Toppozini L, Rheinstadter MC. The interaction between amyloid-beta peptides and anionic lipid membranes containing cholesterol and melatonin. PLoS One. 2014;9(6):e99124.

150

247. Lamberts SWJ, Reubi JC, Krenning EP. Chapter 17 - Somatostatin. In: Bittar EE, Bittar N, editors. Principles of Medical Biology. 10: Elsevier; 1997. p. 403-19. 248. Singh PK, Maji SK. Amyloid-like fibril formation by tachykinin neuropeptides and its relevance to amyloid beta-protein aggregation and toxicity. Cell biochemistry and biophysics. 2012;64(1):29-44. 249. Anoop A, Ranganathan S, Das Dhaked B, Jha NN, Pratihar S, Ghosh S, et al. Elucidating the role of disulfide bond on amyloid formation and fibril reversibility of somatostatin-14: relevance to its storage and secretion. J Biol Chem. 2014;289(24):16884-903. 250. Hyodo S. Chapter 6 - Neurohypophysial Hormone Family. In: Takei Y, Ando H, Tsutsui K, editors. Handbook of Hormones. San Diego: Academic Press; 2016. p. 39-40. 251. Shen Y, Gong X. Chapter 1 - Experimental Tools for the Identification of Specific Genes in Autism Spectrum Disorders and Intellectual Disability. In: Sala C, Verpelli C, editors. Neuronal and Synaptic Dysfunction in Autism Spectrum Disorder and Intellectual Disability. San Diego: Academic Press; 2016. p. 3-12. 252. Grace RCR, Chandrashekar IR, Cowsik SM. Solution Structure of the Tachykinin Peptide Eledoisin. Biophysical Journal. 2003;84(1):655-64. 253. Nederpelt I, Georgi V, Schiele F, Nowak-Reppel K, Fernandez-Montalvan AE, AP IJ, et al. Characterization of 12 GnRH peptide agonists - a kinetic perspective. British journal of pharmacology. 2016;173(1):128-41. 254. Williams RW, Weaver JL. Secondary structure of substance P bound to liposomes in organic solvents and in solution from Raman and CD spectroscopy. J Biol Chem. 1990;265(5):2505-13. 255. Severini C, Petrella C, Calissano P. Substance P and Alzheimer's Disease: Emerging Novel Roles. Current Alzheimer research. 2016;13(9):964-72. 256. Gayen A, Goswami SK, Mukhopadhyay C. NMR evidence of GM1-induced conformational change of Substance P using isotropic bicelles. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2011;1808(1):127-39. 257. Tregear GW, Niall HD, Potts Jun JT, Leeman SE, Chang MM. Synthesis of Substance P. Nature New Biology. 1971;232:87. 258. Ohlen A, Lindbom L, Staines W, Hokfelt T, Cuello AC, Fischer JA, et al. Substance P and calcitonin gene-related peptide: immunohistochemical localisation and microvascular effects in rabbit skeletal muscle. Naunyn-Schmiedeberg's archives of pharmacology. 1987;336(1):87-93. 259. Holzer-Petsche U, Lembeck F, Seitz H. Contractile effects of substance P and neurokinin A on the rat stomach in vivo and in vitro. British journal of pharmacology. 1987;90(1):273-9. 260. Hasenohrl RU, Huston JP, Schuurman T. Neuropeptide substance P improves water maze performance in aged rats. Psychopharmacology. 1990;101(1):23-6. 261. Hasenohrl RU, Frisch C, Nikolaus S, Huston JP. Chronic administration of neurokinin SP improves maze performance in aged Rattus norvegicus. Behavioral and neural biology. 1994;62(2):110-20. 262. Kowall NW, Beal MF, Busciglio J, Duffy LK, Yankner BA. An in vivo model for the neurodegenerative effects of beta amyloid and protection by substance P. Proceedings of the National Academy of Sciences. 1991;88(16):7247-51. 263. Oakley AE, Perry RH, Candy JM, Perry EK. Electron microscopic appearances and implications of neuropeptide fibrillary forms. Neuropeptides. 1981;2(1):1-11. 264. Mumford EL, Squires EL, Jöchle E, Harrison LA, Nett TM, Trigg TE. Use of deslorelin short- term implants to induce ovulation in cycling mares during three consecutive estrous cycles. Animal Reproduction Science. 1995;39(2):129-40. 265. Kapoor DN, Katare OP, Kaurav H, Dhawan S. Development and optimization of in-situ forming microparticles for long term controlled delivery of deslorelin acetate2018. 059-72 p. 266. Evers F, Jeworrek C, Tiemeyer S, Weise K, Sellin D, Paulus M, et al. Elucidating the Mechanism of Lipid Membrane-Induced IAPP Fibrillogenesis and Its Inhibition by the Red Wine

151

Compound Resveratrol: A Synchrotron X-ray Reflectivity Study. Journal of the American Chemical Society. 2009;131(27):9516-21. 267. Feng Y, Wang X-p, Yang S-g, Wang Y-j, Zhang X, Du X-t, et al. Resveratrol inhibits beta- amyloid oligomeric cytotoxicity but does not prevent oligomer formation. NeuroToxicology. 2009;30(6):986-95. 268. Molinspiration [internet]: cheminformatics. Available from: https://www.molinspiration.com/. 269. Morton LA, Saludes JP, Yin H. Constant pressure-controlled extrusion method for the preparation of Nano-sized lipid vesicles. J Vis Exp. 2012(64). 270. Stetefeld J, McKenna SA, Patel TR. Dynamic light scattering: a practical guide and applications in biomedical sciences. Biophys Rev. 2016;8(4):409-27. 271. Lee C-C, Nayak A, Sethuraman A, Belfort G, McRae GJ. A Three-Stage Kinetic Model of Amyloid Fibrillation. Biophysical Journal. 2007;92(10):3448-58. 272. Meisl G, Yang X, M. Dobson C, Linse S, Knowles T. Modulation of electrostatic interactions to reveal a reaction network unifying the aggregation behaviour of the Aβ42 peptide and its variants2017. 273. Noormagi A, Primar K, Tougu V, Palumaa P. Interference of low-molecular substances with the thioflavin-T fluorescence assay of amyloid fibrils. Journal of peptide science : an official publication of the European Peptide Society. 2012;18(1):59-64. 274. Usov I, Mezzenga R. FiberApp: An Open-Source Software for Tracking and Analyzing Polymers, Filaments, Biomacromolecules, and Fibrous Objects. Macromolecules. 2015;48(5):1269- 80. 275. Nasrollahzadeh M, Atarod M, Sajjadi M, Sajadi SM, Issaabadi Z. Chapter 6 - Plant-Mediated Green Synthesis of Nanostructures: Mechanisms, Characterization, and Applications. In: Nasrollahzadeh M, Sajadi SM, Sajjadi M, Issaabadi Z, Atarod M, editors. Interface Science and Technology. 28: Elsevier; 2019. p. 199-322. 276. Khrapunov S. Circular dichroism spectroscopy has intrinsic limitations for protein secondary structure analysis. Analytical biochemistry. 2009;389(2):174-6. 277. van ‘t Hag L, Li X, Meikle TG, Hoffmann SV, Jones NC, Pedersen JS, et al. How Peptide Molecular Structure and Charge Influence the Nanostructure of Lipid Bicontinuous Cubic Mesophases: Model Synthetic WALP Peptides Provide Insights. Langmuir. 2016;32(27):6882-94. 278. Greenfield NJ. Using circular dichroism spectra to estimate protein secondary structure. Nature protocols. 2006;1(6):2876-90. 279. Whitmore L, Wallace BA. DICHROWEB, an online server for protein secondary structure analyses from circular dichroism spectroscopic data. Nucleic Acids Res. 2004;32(Web Server issue):W668-73. 280. Karran E, Mercken M, De Strooper B. The amyloid cascade hypothesis for Alzheimer's disease: an appraisal for the development of therapeutics. Nature reviews Drug discovery. 2011;10(9):698-712. 281. Kelly JW. The alternative conformations of amyloidogenic proteins and their multi-step assembly pathways. Current opinion in structural biology. 1998;8(1):101-6. 282. Dharmadana D, Reynolds NP, Conn CE, Valéry C. Molecular interactions of amyloid nanofibrils with biological aggregation modifiers: implications for cytotoxicity mechanisms and biomaterial design. Interface Focus. 2017;7(4). 283. Bowie JU. Membrane protein folding: how important are hydrogen bonds? Current opinion in structural biology. 2011;21(1):42-9. 284. Ulmer TS, Bax A, Cole NB, Nussbaum RL. Structure and dynamics of micelle-bound human alpha-synuclein. J Biol Chem. 2005;280(10):9595-603. 285. Nanga RPR, Brender JR, Vivekanandan S, Ramamoorthy A. Structure and membrane orientation of IAPP in its natively amidated form at physiological pH in a membrane environment. Biochimica et biophysica acta. 2011;1808(10):2337-42.

152

286. Skeby KK, Andersen OJ, Pogorelov TV, Tajkhorshid E, Schiøtt B. Conformational Dynamics of the Human Islet Amyloid Polypeptide in a Membrane Environment: Toward the Aggregation Prone Form. Biochemistry. 2016;55(13):2031-42. 287. Breydo L, Wu JW, Uversky VN. α-Synuclein misfolding and Parkinson's disease. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease. 2012;1822(2):261-85. 288. Delgado DA, Doherty K, Cheng Q, Kim H, Xu D, Dong H, et al. Distinct Membrane Disruption Pathways Are Induced by 40-Residue beta-Amyloid Peptides. J Biol Chem. 2016;291(23):12233-44. 289. Rodriguez Camargo DC, Korshavn KJ, Jussupow A, Raltchev K, Goricanec D, Fleisch M, et al. Stabilization and structural analysis of a membrane-associated hIAPP aggregation intermediate. eLife. 2017;6:e31226. 290. Monsellier E, Bousset L, Melki R. α-Synuclein and huntingtin exon 1 amyloid fibrils bind laterally to the cellular membrane. Scientific reports. 2016;6:19180. 291. Zhang X, St Clair JR, London E, Raleigh DP. Islet Amyloid Polypeptide Membrane Interactions: Effects of Membrane Composition. Biochemistry. 2017;56(2):376-90. 292. Sugiura Y, Ikeda K, Nakano M. High Membrane Curvature Enhances Binding, Conformational Changes, and Fibrillation of Amyloid-beta on Lipid Bilayer Surfaces. Langmuir. 2015;31(42):11549-57. 293. Li Y, Guan L, Lu T, Li H, Li Z, Li F. Interactions of the N-terminal domain of human islet amyloid polypeptide with lipid membranes: the effect of cholesterol. RSC Advances. 2016;6(99):96837-46. 294. Fusco G, Pape T, Stephens AD, Mahou P, Costa AR, Kaminski CF, et al. Structural basis of synaptic vesicle assembly promoted by alpha-synuclein. Nature communications. 2016;7:12563. 295. Fusco G, De Simone A, Gopinath T, Vostrikov V, Vendruscolo M, Dobson CM, et al. Direct observation of the three regions in alpha-synuclein that determine its membrane-bound behaviour. Nature communications. 2014;5:3827. 296. Burré J, Sharma M, Südhof TC. α-Synuclein assembles into higher-order multimers upon membrane binding to promote SNARE complex formation. Proceedings of the National Academy of Sciences of the United States of America. 2014;111(40):E4274-E83. 297. Iyer A, Roeters SJ, Schilderink N, Hommersom B, Heeren RM, Woutersen S, et al. The Impact of N-terminal Acetylation of alpha-Synuclein on Phospholipid Membrane Binding and Fibril Structure. J Biol Chem. 2016;291(40):21110-22. 298. Kikhney AG, Svergun DI. A practical guide to small angle X-ray scattering (SAXS) of flexible and intrinsically disordered proteins. FEBS letters. 2015;589(19, Part A):2570-7. 299. van Grondelle W, López-Iglesias C, Coll E, Artzner F, Paternostre M, Lacombe F, et al. Spontaneous fibrillation of the native neuropeptide hormone Somatostatin -142007. 211-23 p. 300. Korshavn KJ, Satriano C, Lin Y, Zhang R, Dulchavsky M, Bhunia A, et al. Reduced Lipid Bilayer Thickness Regulates the Aggregation and Cytotoxicity of Amyloid-β. Journal of Biological Chemistry. 2017;292(11):4638-50. 301. Seelig A, Macdonald PM. Binding of a neuropeptide, substance P, to neutral and negatively charged lipids. Biochemistry. 1989;28(6):2490-6. 302. Malishev R, Abbasi R, Jelinek R, Chai L. Bacterial Model Membranes Reshape Fibrillation of a Functional Amyloid Protein. Biochemistry. 2018. 303. Costigan SC, Booth PJ, Templer RH. Estimations of lipid bilayer geometry in fluid lamellar phases. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2000;1468(1):41-54. 304. Chen L, Yu Z, Quinn PJ. The partition of cholesterol between ordered and fluid bilayers of phosphatidylcholine: A synchrotron X-ray diffraction study. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2007;1768(11):2873-81. 305. Epand RM, Hughes DW, Sayer BG, Borochov N, Bach D, Wachtel E. Novel properties of cholesterol-dioleoylphosphatidylcholine mixtures. Biochimica et biophysica acta. 2003;1616(2):196- 208. 306. Usov I, Mezzenga R. FiberApp: An Open-Source Software for Tracking and Analyzing Polymers, Filaments, Biomacromolecules, and Fibrous Objects2015. 150210103859007 p.

153

307. Reynolds NP, Adamcik J, Berryman JT, Handschin S, Zanjani AAH, Li W, et al. Competition between crystal and fibril formation in molecular mutations of amyloidogenic peptides. Nature communications. 2017;8(1):1338. 308. Chauhan A, Ray I, Chauhan VP. Interaction of amyloid beta-protein with anionic phospholipids: possible involvement of Lys28 and C-terminus aliphatic amino acids. Neurochemical research. 2000;25(3):423-9. 309. Galvagnion C, Brown JW, Ouberai MM, Flagmeier P, Vendruscolo M, Buell AK, et al. Chemical properties of lipids strongly affect the kinetics of the membrane-induced aggregation of alpha- synuclein. Proceedings of the National Academy of Sciences of the United States of America. 2016;113(26):7065-70. 310. Meikle TG, Zabara A, Waddington LJ, Separovic F, Drummond CJ, Conn CE. Incorporation of antimicrobial peptides in nanostructured lipid membrane mimetic bilayer cubosomes. Colloids and surfaces B, Biointerfaces. 2017;152:143-51. 311. Zabara A, Meikle TG, Trenker R, Yao S, Newman J, Peat TS, et al. Lipidic Cubic Phase-Induced Membrane Protein Crystallization: Interplay Between Lipid Molecular Structure, Mesophase Structure and Properties, and Crystallogenesis. Crystal Growth & Design. 2017;17(11):5667-74. 312. Burbach JP, Luckman SM, Murphy D, Gainer H. Gene regulation in the magnocellular hypothalamo-neurohypophysial system. Physiological reviews. 2001;81(3):1197-267. 313. Sikorska E, Iłowska E, Wyrzykowski D, Kwiatkowska A. Membrane structure and interactions of peptide hormones with model lipid bilayers. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2012;1818(12):2982-93. 314. Bockaert J, Roy C, Rajerison R, Jard S. Specific binding of [3H]lysine vasopressin to pig kidney plasma membranes. Relationship of receptor occupancy to adenylate cyclase activation. Journal of Biological Chemistry. 1973;248(17):5922-31. 315. Luft FC, Steinberg H, Ganten U, Meyer D, Gless KH, Lang RE, et al. Effect of sodium chloride and sodium bicarbonate on blood pressure in stroke-prone spontaneously hypertensive rats. Clinical Science. 1988;74(6):577-85. 316. Wilkinson MF, Kasting NW. The antipyretic effects of centrally administered vasopressin at different ambient temperatures. Brain research. 1987;415(2):275-80. 317. Jard S, Gaillard RC, Guillon G, Marie J, Schoenenberg P, Muller AF, et al. Vasopressin antagonists allow demonstration of a novel type of vasopressin receptor in the rat adenohypophysis. Molecular pharmacology. 1986;30(2):171-7. 318. Akerlund M, Carlsson AM, Melin P, Trojnar J. The effect on the human uterus of two newly developed competitive inhibitors of oxytocin and vasopressin. Acta obstetricia et gynecologica Scandinavica. 1985;64(6):499-504. 319. Wheatley M, Howl J, Yarwood NJ, Hawtin SR, Davies AR, Matthews G, et al. Structure and function of neurohypophysial hormone receptors. Biochemical Society transactions. 1997;25(3):1046-51. 320. Narayanan V, Guo Y, Scarlata S. Fluorescence studies suggest a role for alpha-synuclein in the phosphatidylinositol lipid signaling pathway. Biochemistry. 2005;44(2):462-70. 321. Rodziewicz-MotowidŁo S, Sikorska E, Oleszczuk M, Czaplewski C. Conformational studies of vasopressin and mesotocin using NMR spectroscopy and molecular modelling methods. Part II: studies in the SDS micelle. Journal of Peptide Science. 2008;14(1):85-96. 322. Bhattacharjya S, Awasthi SK, Adiga PR, Balaram P. Folded conformations of antigenic peptides from riboflavin carrier protein in aqueous hexafluoroacetone. Protein Sci. 1998;7(1):123- 31. 323. Rajan R, K. Awasthi S, Bhattacharjya S, Balaram P. “Teflon-coated peptides”: Hexafluoroacetone trihydrate as a structure stabilizer for peptides1997. 125-8 p. 324. Shoffner SK, Schnell S. Estimation of the lag time in a subsequent monomer addition model for fibril elongation. Physical chemistry chemical physics : PCCP. 2016;18(31):21259-68.

154

325. Chiti F, Stefani M, Taddei N, Ramponi G, Dobson CM. Rationalization of the effects of mutations on peptide and protein aggregation rates. Nature. 2003;424(6950):805-8. 326. Adler-Abramovich L, Gazit E. The physical properties of supramolecular peptide assemblies: from building block association to technological applications. Chemical Society Reviews. 2014;43(20):6881-93. 327. Beuret N, Hasler F, Prescianotto-Baschong C, Birk J, Rutishauser J, Spiess M. Amyloid-like aggregation of provasopressin in diabetes insipidus and secretory granule sorting. BMC Biology. 2017;15(1):5. 328. Dharmadana D, Reynolds NP, Conn CE, Valéry C. Molecular interactions of amyloid nanofibrils with biological aggregation modifiers: implications for cytotoxicity mechanisms and biomaterial design. Interface focus. 2017;7(4):20160160-. 329. Roberts CJ. Non-native protein aggregation kinetics. Biotechnology and Bioengineering. 2007;98(5):927-38. 330. Fink AL. The aggregation and fibrillation of α-synuclein. Accounts of chemical research. 2006;39(9):628-34. 331. Choucair A, Chakrapani M, Chakravarthy B, Katsaras J, Johnston LJ. Preferential accumulation of Aβ(1−42) on gel phase domains of lipid bilayers: An AFM and fluorescence study. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2007;1768(1):146-54. 332. Meikle TG, Conn CE, Separovic F, Drummond CJ. Exploring the structural relationship between encapsulated antimicrobial peptides and the bilayer membrane mimetic lipidic cubic phase: studies with gramicidin A′. RSC Advances. 2016;6(73):68685-94. 333. Slusarz MJ, Slusarz R, Ciarkowski J. Molecular dynamics simulation of human neurohypophyseal hormone receptors complexed with oxytocin-modeling of an activated state. Journal of peptide science : an official publication of the European Peptide Society. 2006;12(3):171- 9. 334. Ventura S, Zurdo J, Narayanan S, Parreño M, Mangues R, Reif B, et al. Short amino acid stretches can mediate amyloid formation in globular proteins: The Src homology 3 (SH3) case. Proceedings of the National Academy of Sciences of the United States of America. 2004;101(19):7258. 335. Polymeropoulos MH, Lavedan C, Leroy E, Ide SE, Dehejia A, Dutra A, et al. Mutation in the alpha-synuclein gene identified in families with Parkinson's disease. Science (New York, NY). 1997;276(5321):2045-7. 336. Wang S, Chu C-H, Guo M, Jiang L, Nie H, Zhang W, et al. Identification of a specific α- synuclein peptide (α-Syn 29-40) capable of eliciting microglial superoxide production to damage dopaminergic neurons. Journal of neuroinflammation. 2016;13(1):158-. 337. Hili R, Yudin AK. Readily Available Unprotected Amino Aldehydes. Journal of the American Chemical Society. 2006;128(46):14772-3. 338. Kawakami T, Ohta A, Ohuchi M, Ashigai H, Murakami H, Suga H. Diverse backbone-cyclized peptides via codon reprogramming. Nat Chem Biol. 2009;5(12):888-90. 339. Choi S-j, Jeong W-j, Kang S-K, Lee M, Kim E, Ryu DY, et al. Differential Self-Assembly Behaviors of Cyclic and Linear Peptides. Biomacromolecules. 2012;13(7):1991-5. 340. Dempsey CE, Mason PE, Jungwirth P. Complex Ion Effects on Polypeptide Conformational Stability: Chloride and Sulfate Salts of Guanidinium and Tetrapropylammonium. Journal of the American Chemical Society. 2011;133(19):7300-3. 341. Convert O, Ploux O, Lavielle S, Cotrait M, Chassaing G. Analysis of tachykinin-binding site interactions using NMR and energy calculation data of potent cyclic analogues of substance P. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology. 1988;954:287- 302. 342. Kowall NW, Beal MF, Busciglio J, Duffy LK, Yankner BA. An in vivo model for the neurodegenerative effects of beta amyloid and protection by substance P. Proceedings of the National Academy of Sciences of the United States of America. 1991;88(16):7247-51.

155

343. Williams RW, Weaver JL. Secondary structure of substance P bound to liposomes in organic solvents and in solution from Raman and CD spectroscopy. Journal of Biological Chemistry. 1990;265(5):2505-13. 344. Choo LP, Jackson M, Mantsch HH. Conformation and self-association of the peptide hormone substance P: Fourier-transform infrared spectroscopic study. The Biochemical journal. 1994;301 ( Pt 3):667-70. 345. Chassaing G, Convert O, Lavielle S. Preferential conformation of substance P in solution. Eur J Biochem. 1986;154(1):77-85. 346. Perry EK, Oakley AE, Candy JM, Perry RH. Properties and possible significance of substance P and insulin fibrils. Neuroscience letters. 1981;25(3):321-5. 347. Wu C-SC, Yang JT. Instability of terminal amino acid residues in oligopeptides in sodium dodecyl sulfate solution. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology. 1983;746(1):72-80. 348. Rueger M, Bienert M, Mehlis B, Gast K, Zirwer D, Behlke J. Self-association of the neuroregulatory peptide substance P and Its C-terminal sequences. Biopolymers. 1984;23(4):747-58. 349. Choo LP, Jackson M, Mantsch HH. Conformation and self-association of the peptide hormone substance P: Fourier-transform infrared spectroscopic study. Biochemical Journal. 1994;301(3):667. 350. Jackson M, Mantsch HH, Spencer JH. Conformation of magainin-2 and related peptides in aqueous solution and membrane environments probed by Fourier transform infrared spectroscopy. Biochemistry. 1992;31(32):7289-93. 351. Chassaing G, Convert O, Lavielle S. Conformation analogy between substance P and physalaemin. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology. 1986;873(3):397-404. 352. Whitehead TL, McNair SD, Hadden CE, Young JK, Hicks RP. Membrane-induced secondary structures of neuropeptides: a comparison of the solution conformations adopted by agonists and antagonists of the mammalian tachykinin NK1 receptor. Journal of medicinal chemistry. 1998;41(9):1497-506. 353. Coy DH, Coy EJ, Vilchez-Martinez JA, de la Cruz A, Arimura A, Schally AV. Suppression of gonadotropin release and ovulation in animals by inhibitory analogs of luteinizing hormone-releasing hormone. Current topics in molecular endocrinology. 1976;3:339-54. 354. Ott TR, Troskie BE, Roeske RW, Illing N, Flanagan CA, Millar RP. Two mutations in extracellular loop 2 of the human GnRH receptor convert an antagonist to an agonist. Molecular endocrinology (Baltimore, Md). 2002;16(5):1079-88. 355. Sundaram S, Durairaj C, Kadam R, Kompella UB. Luteinizing hormone-releasing hormone receptor-targeted deslorelin-docetaxel conjugate enhances efficacy of docetaxel in prostate cancer therapy. Molecular cancer therapeutics. 2009;8(6):1655-65. 356. Herbert CA, Trigg TE, Renfree MB, Shaw G, Eckery DC, Cooper DW. Effects of a Gonadotropin-Releasing Hormone Agonist Implant on Reproduction in a Male Marsupial, Macropus eugenii1. Biology of Reproduction. 2004;70(6):1836-42. 357. B. Patel A, Srivastava S, S. Phadke R. Substance P (free acid) Adopts Different Conformation than Native Peptide in DMSO, Water and DPPC Bilayers2001. 129-38 p. 358. Duplaa H, Convert O, Sautereau AM, Tocanne JF, Chassaing G. Binding of substance P to monolayers and vesicles made of phosphatidylcholine and/or phosphatidylserine. Biochimica et biophysica acta. 1992;1107(1):12-22. 359. Bignardi C. Substance P Self-Aggregation Revised: A Chromatographic and Mass Spectrometry Analysis2012. 360. Jaspers H, HorvÁTh A, MezÖ I, KÉRi G, Binst GV. Conformational study of a series of somatostatin analogues with antitumor and/or GH inhibitory activity. International Journal of Peptide and Protein Research. 1994;43(3):271-6.

156

361. Diao J, Burre J, Vivona S, Cipriano DJ, Sharma M, Kyoung M, et al. Native alpha-synuclein induces clustering of synaptic-vesicle mimics via binding to phospholipids and synaptobrevin- 2/VAMP2. eLife. 2013;2:e00592. 362. Wu CSC, Hachimori A, Yang JT. Lipid-induced ordered conformation of some peptide hormones and bioactive oligopeptides: predominance of the helix over the .beta.-form. Biochemistry. 1982;21(19):4556-62. 363. Chandra S, Chen X, Rizo J, Jahn R, Südhof TC. A Broken α-Helix in Folded α-Synuclein. Journal of Biological Chemistry. 2003;278(17):15313-8. 364. Shahzadi Z, Mukhopadhyay C. Interaction Between Luteinizing Hormone-Releasing Hormone and GM1-Doped Cholesterol/Sphingomyelin Vesicles: A Spectroscopic Study. The Journal of membrane biology. 2017;250(6):617-27. 365. Keire DA, Fletcher TG. The conformation of substance P in lipid environments. Biophys J. 1996;70(4):1716-27. 366. Mika JT, Moiset G, Cirac AD, Feliu L, Bardají E, Planas M, et al. Structural basis for the enhanced activity of cyclic antimicrobial peptides: The case of BPC194. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2011;1808(9):2197-205. 367. Shi J, He M, Cao J, Wang H, Ding J, Jiao Y, et al. The comparative analysis of the potential relationship between resveratrol and stilbene synthase gene family in the development stages of grapes (Vitis quinquangularis and Vitis vinifera). Plant physiology and biochemistry : PPB. 2014;74:24-32. 368. Baur JA, Sinclair DA. Therapeutic potential of resveratrol: the in vivo evidence. Nature Reviews Drug Discovery. 2006;5:493. 369. Jang M, Cai L, Udeani GO, Slowing KV, Thomas CF, Beecher CWW, et al. Cancer Chemopreventive Activity of Resveratrol, a Natural Product Derived from Grapes. Science (New York, NY). 1997;275(5297):218. 370. Milder IE, Arts IC, van de Putte B, Venema DP, Hollman PC. Lignan contents of Dutch plant foods: a database including lariciresinol, pinoresinol, secoisolariciresinol and matairesinol. The British journal of nutrition. 2005;93(3):393-402. 371. Lu Z, Zhang Y, Liu H, Yuan J, Zheng Z, Zou G. Transport of a cancer chemopreventive polyphenol, resveratrol: interaction with serum albumin and hemoglobin. Journal of fluorescence. 2007;17(5):580-7. 372. Fahr A, Hoogevest Pv, May S, Bergstrand N, S. Leigh ML. Transfer of lipophilic drugs between liposomal membranes and biological interfaces: Consequences for drug delivery. European Journal of Pharmaceutical Sciences. 2005;26(3):251-65. 373. Tsuchiya H. Stereospecificity in membrane effects of catechins. Chemico-Biological Interactions. 2001;134(1):41-54. 374. Tsuchiya H, Tanaka T, Nagayama M. Antiproliferative Effects Associated with Membrane Lipid Interaction of Green Tea Catechins. Journal of Health Science. 2008;54(5):576-80. 375. Neves AR, Nunes C, Amenitsch H, Reis S. Effects of resveratrol on the structure and fluidity of lipid bilayers: a membrane biophysical study. Soft Matter. 2016;12(7):2118-26. 376. Gulaboski R, Borges F, Pereira CM, Cordeiro MN, Garrido J, Silva AF. Voltammetric insights in the transfer of ionizable drugs across biomimetic membranes: recent achievements. Combinatorial chemistry & high throughput screening. 2007;10(7):514-26. 377. Giaginis C, Tsantili-Kakoulidou A. Alternative measures of lipophilicity: from octanol-water partitioning to IAM retention. Journal of pharmaceutical sciences. 2008;97(8):2984-3004. 378. Wesołowska O, Kużdżał M, Štrancar J, Michalak K. Interaction of the chemopreventive agent resveratrol and its metabolite, piceatannol, with model membranes. Biochimica et Biophysica Acta (BBA) - Biomembranes. 2009;1788(9):1851-60. 379. Fabris S, Momo F, Ravagnan G, Stevanato R. Antioxidant properties of resveratrol and piceid on lipid peroxidation in micelles and monolamellar liposomes. Biophysical chemistry. 2008;135(1- 3):76-83.

157

380. Wisniewska A, Draus J, Subczynski WK. Is a fluid-mosaic model of biological membranes fully relevant? Studies on lipid organization in model and biological membranes. Cellular & molecular biology letters. 2003;8(1):147-59. 381. Brittes J, Lucio M, Nunes C, Lima JL, Reis S. Effects of resveratrol on membrane biophysical properties: relevance for its pharmacological effects. Chemistry and physics of lipids. 2010;163(8):747-54. 382. de Ghellinck A, Shen C, Fragneto G, Klosgen B. Probing the position of resveratrol in lipid bilayers: A neutron reflectivity study. Colloids and surfaces B, Biointerfaces. 2015;134:65-72. 383. Wesolowska O, Kuzdzal M, Strancar J, Michalak K. Interaction of the chemopreventive agent resveratrol and its metabolite, piceatannol, with model membranes. Biochimica et biophysica acta. 2009;1788(9):1851-60. 384. Selvaraj S, Mohan A, Narayanan S, Sethuraman S, Krishnan UM. Dose-dependent interaction of trans-resveratrol with biomembranes: effects on antioxidant property. Journal of medicinal chemistry. 2013;56(3):970-81. 385. Longo E, Ciuchi F, Guzzi R, Rizzuti B, Bartucci R. Resveratrol induces chain interdigitation in DPPC cell membrane model systems. Colloids and surfaces B, Biointerfaces. 2016;148:615-21. 386. Koukoulitsa C, Durdagi S, Siapi E, Villalonga-Barber C, Alexi X, Steele BR, et al. Comparison of thermal effects of stilbenoid analogs in lipid bilayers using differential scanning calorimetry and molecular dynamics: correlation of thermal effects and topographical position with antioxidant activity. European biophysics journal : EBJ. 2011;40(7):865-75. 387. Sarpietro MG, Spatafora C, Tringali C, Micieli D, Castelli F. Interaction of Resveratrol and Its Trimethyl and Triacetyl Derivatives with Biomembrane Models Studied by Differential Scanning Calorimetry. Journal of Agricultural and Food Chemistry. 2007;55(9):3720-8. 388. Neves AR, Nunes C, Reis S. Resveratrol induces ordered domains formation in biomembranes: Implication for its pleiotropic action. Biochimica et biophysica acta. 2016;1858(1):12-8. 389. Neves AR, Nunes C, Reis S. New Insights on the Biophysical Interaction of Resveratrol with Biomembrane Models: Relevance for Its Biological Effects. The journal of physical chemistry B. 2015;119(35):11664-72. 390. Garcia-Garcia J, Micol V, de Godos A, Gomez-Fernandez JC. The cancer chemopreventive agent resveratrol is incorporated into model membranes and inhibits protein kinase C alpha activity. Archives of biochemistry and biophysics. 1999;372(2):382-8. 391. Tu LH, Young LM, Wong AG, Ashcroft AE, Radford SE, Raleigh DP. Mutational analysis of the ability of resveratrol to inhibit amyloid formation by islet amyloid polypeptide: critical evaluation of the importance of aromatic-inhibitor and histidine-inhibitor interactions. Biochemistry. 2015;54(3):666-76. 392. Willem M, Garratt AN, Novak B, Citron M, Kaufmann S, Rittger A, et al. Control of peripheral nerve myelination by the beta-secretase BACE1. Science (New York, NY). 2006;314(5799):664-6. 393. Zhang W, Xiong H, Callaghan D, Liu H, Jones A, Pei K, et al. Blood-brain barrier transport of amyloid beta peptides in efflux pump knock-out animals evaluated by in vivo optical imaging. Fluids and barriers of the CNS. 2013;10(1):13. 394. Mishra R, Sellin D, Radovan D, Gohlke A, Winter R. Inhibiting islet amyloid polypeptide fibril formation by the red wine compound resveratrol. Chembiochem : a European journal of chemical biology. 2009;10(3):445-9. 395. Radovan D, Opitz N, Winter R. Fluorescence microscopy studies on islet amyloid polypeptide fibrillation at heterogeneous and cellular membrane interfaces and its inhibition by resveratrol. FEBS letters. 2009;583(9):1439-45. 396. Wei L, Jiang P, Xu W, Li H, Zhang H, Yan L, et al. The molecular basis of distinct aggregation pathways of islet amyloid polypeptide. J Biol Chem. 2011;286(8):6291-300. 397. Molinspiration Cheminformatics [cited 2019 16-01]. Available from: https://www.molinspiration.com/.

158

398. Belleri M, Ribatti D, Nicoli S, Cotelli F, Forti L, Vannini V, et al. Antiangiogenic and vascular- targeting activity of the microtubule-destabilizing trans-resveratrol derivative 3,5,4'- trimethoxystilbene. Molecular pharmacology. 2005;67(5):1451-9. 399. Pan MH, Gao JH, Lai CS, Wang YJ, Chen WM, Lo CY, et al. Antitumor activity of 3,5,4'- trimethoxystilbene in COLO 205 cells and xenografts in SCID mice. Molecular carcinogenesis. 2008;47(3):184-96. 400. Yang YT, Weng CJ, Ho CT, Yen GC. Resveratrol analog-3,5,4'-trimethoxy-trans-stilbene inhibits invasion of human lung adenocarcinoma cells by suppressing the MAPK pathway and decreasing matrix metalloproteinase-2 expression. Molecular nutrition & food research. 2009;53(3):407-16. 401. Hasiah AH, Ghazali AR, Weber JF, Velu S, Thomas NF, Inayat Hussain SH. Cytotoxic and antioxidant effects of methoxylated stilbene analogues on HepG2 hepatoma and Chang liver cells: Implications for structure activity relationship. Human & experimental toxicology. 2011;30(2):138- 44. 402. Hernandez B, Carelli C, Coic YM, De Coninck J, Ghomi M. Vibrational analysis of amino acids and short peptides in aqueous media. V. The effect of the disulfide bridge on the structural features of the peptide hormone somatostatin-14. The journal of physical chemistry B. 2009;113(38):12796- 803. 403. Tu L-H, Young LM, Wong AG, Ashcroft AE, Radford SE, Raleigh DP. Mutational analysis of the ability of resveratrol to inhibit amyloid formation by islet amyloid polypeptide: critical evaluation of the importance of aromatic-inhibitor and histidine-inhibitor interactions. Biochemistry. 2015;54(3):666-76. 404. Porat Y, Mazor Y, Efrat S, Gazit E. Inhibition of islet amyloid polypeptide fibril formation: a potential role for heteroaromatic interactions. Biochemistry. 2004;43(45):14454-62. 405. S Prajapati R, Sirajuddin M, Durani V, Sreeramulu S, Varadarajan R. Contribution of Cation−π Interactions to Protein Stability †2006. 15000-10 p.

159

Appendix Table S8.1. Analysis of hydrophobicity, net charge, and % charged amino acids of SST-14 using PEPTIDE 2.0 (34)

Amino Mw Net Hydrophobic Peptides Structure acid Charged amino (kDa) charge residue% residues acids %

Somatostatin 1.60 14 +2 36 Cyclic 14

Vasopressin 1 9 +2 22 Cyclic 33

Oxytocin 1 9 +1 22 Cyclic 22

Substance P 1.3 11 +3 55 Linear 45

Deslorelin 1.3 10 +1 50 Linear 44

Figure S8.1. Shows the DLS curves for liposomes doped with various lipid compositions

160

Table S8.2. Shows the hydrodynamic size of liposomes doped with various compositions

Sample Avg size-nm DOPC 130 DOPC+Chol 148 DOPC+DOPS 135 DMPC 135 DMPC+Chol 125 DMPC+DMPS 125 DLPC 128 DPPC 89 DSPC 108

Table S8.3. Quantitative analysis of the CD spectra for SST incubated in liposomes

Liposomes Wavelength (nm) After 3 hours DOPC 205 65DOPC+35DOPS After 72 hours DOPC 204 DMPC weak intensity 65DOPC+35DOPS 210

161

Table S8.4. Shows the effect of concentration on kagg, half-time and lag phase for SST in NaCl

Lag phase t50 or (t1/2) -1 2 % (w/w) kagg (s ) (hours) (hours)3 0.01 absent 0 0

0.05 absent 0 0

0.1 absent 0 0

0.5 absent 0 0

1 absent 0 0

2.5 58 112 0.0036712

5 12 51 0.017882

Plateau followed by one-phase association2 and half-time3

Table S8.5. Quantitative analysis of the CD spectra for SP incubated in liposomes after 72- hours

Liposomes Wavelength (nm) After 72 hours DOPC Weak intensity

162

Table S8.6. Shows the hydrodynamic size of saturated lipid vesicles with the presence and absence of stilbenes at 5 mol%

Sample Avg size-nm DLPC 128 DLPC+RSV 122 DLPC+TS 164 DLPC+TMS 91 DMPC 135 DMPC+RSV 148 DMPC+TS 164 DMPC+TMS 85 DPPC 89 DPPC+RSV 106 DPPC+TS 97 DPPC+TMS 72 DSPC 108 DSPC+RSV 121 DSPC+TS 122 DSPC+TMS 95

Table S8.7. Shows the hydrodynamic size of DMPC+Chol vesicles with the presence and absence of stilbenes at 5 mol%

Sample Avg size-nm DMPC+Chol 125 DMPC+Chol+RSV 122 DMPC+Chol+TS 5mol% 118 DMPC+Chol+TMS 5mol% 115

163

Table S8.8. Shows the hydrodynamic size of anionic charged vesicles with the presence and absence of stilbenes at 5 mol%

Sample Avg size-nm DOPC+DOPS 135 DOPC+DOPS+RSV 122 DOPC+DOPS+TS 115 DOPC+DOPS+TMS 98 DMPC+DMPS 125 DMPC+DMPS+RSV 91 DMPC+DMPS+TS 89 DMPC+DMPS+TMS 105

3000000 e

c 2000000

n e

c -1

s Kagg = 0.01857 s

r o

u 1000000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.2. ThT curve fit for SST (5% w/w) in NaCl

164

1500000 e

c 1000000

e c

s -1

e Kagg = 0.02384 s

r o

u 500000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.3. ThT curve fit for SST (5% w/w) in DOPC

4×106

n e

c -1

s Kagg = 0.05889 s e

r 6

o 2×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.4. ThT curve fit for SST (5% w/w) in DLPC

165

8×106

e 6×10

e c

s 4×106 -1

e Kagg = 0.03803 s

u l

F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.5. ThT curve fit for SST (5% w/w) in DMPC

6×106 e

c 6

n 4×10

e -1

c Kagg = 0.06305 s

e r

o 6

u 2×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.6. ThT curve fit for SST (5% w/w) in DPPC

166

5000000

4000000

e c

n -1

e 3000000 Kagg = 0.05110 s

e r

o 2000000

l F 1000000

0 0 50 100 150 200 Time (hours)

Figure S8.7. ThT curve fit for SST (5% w/w) in DSPC

4000000

e 3000000

e c

s 2000000 -1

e Kagg = 0.02871 s

u l

F 1000000

0 0 50 100 150 200 Time (hours)

Figure S8.8. ThT curve fit for SST (5% w/w) in DOPC+Chol

167

1×107

n e

c -1

s Kagg = 0.03023 s e

r 5×106

l F

0 0 50 100 150 200 Time (hours)

Figure S8.9. ThT curve fit for SST (5% w/w) in DMPC+Chol

6×106

c n

e 6

c 4×10 -1

s Kagg = 0.3253 s

o u

l 2×106 F

0 0 50 100 150 200 Time (hours)

Figure S8.10. ThT curve fit for SST (5% w/w) in DOPC+DOPS

168

1.5×107 e

c 7 n 1×10 -1

e Kagg = 0.2193 s

e r

o 6

u 5×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.11. ThT curve fit for SST (5% w/w) in DMPC+DMPS

8×106

e 6×10

c n

e -1 c Kagg = 0.1099 s

s 4×106

u l

F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.12. ThT curve fit for OT (0.1% w/w) in NaCl

169

1.5×107 e

c 7

n 1×10

e Kagg = 0.1396

e r

o 6

u 5×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.13. ThT curve fit for OT (0.5% w/w) in NaCl

2.5×107

2×107

e c

n 7

e 1.5×10 c

s -1

e Kagg = 0.6874 s

r 7

o 1×10

l F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.14. ThT curve fit for OT (5% w/w) in NaCl

170

2×107

1e .5×10

e c

s 1×107

r o

u -1 l Kagg = 0.2833 s F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.15. ThT curve fit for OT (0.5% w/w) in DOPC

2×107

1e .5×10

e c

s 1×107 e

r -1

o Kagg = 0.2485 s

u l

F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.16. ThT curve fit for OT (0.5% w/w) in DOPC+Chol

171

3×107 e

c 7

n 2×10

e r

o 7 u 1×10 -1

l Kagg = 0.2583 s F

0 0 50 100 150 200 Time (hours)

Figure S8.17. ThT curve fit for OT (0.5% w/w) in DOPC+DOPS

2.5×107

2×107

e c

n 7 e 1.5×10 c -1

s Kagg = 0.3466 s e

r 7

o 1×10

l F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.18. ThT curve fit for OT (0.5% w/w) in DLPC

172

2.5×107

2×107

e c

n 7

e 1.5×10

c s

e -1

r 7 Kagg = 0.3739 s

o 1×10

l F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.19. ThT curve fit for OT (0.5% w/w) in DMPC

3×107 e

c 7

n 2×10

e c

s -1

e Kagg = 0.2177 s r

o 7

u 1×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.20. ThT curve fit for OT (0.5% w/w) in DMPC+Chol

173

2.5×107

2×107

e c

n 7 e 1.5×10 c -1

s Kagg = 0.3592 s e

r 7

o 1×10

l F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.21. ThT curve fit for OT (0.5% w/w) in DMPC+DMPS

3×107 e

c 7

n 2×10

e -1

c Kagg = 0.3317 s

e r

o 7

u 1×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.22. ThT curve fit for OT (0.5% w/w) in DPPC

174

3×107 e

c 7

n 2×10

e -1

c Kagg = 0.3760 s

e r

o 7

u 1×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.23. ThT curve fit for OT (0.5% w/w) in DSPC

60000 e

c 40000

e c

s -1

e Kagg = 0.02002 s

r o

u 20000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.24. ThT curve fit for AVP (5% w/w) in 95DMPC+5DMPS

175

80000

e 60000

n e

c -1

s 40000 Kagg = 0.02208 s

u l

F 20000

0 0 50 100 150 200 Time (hours)

Figure S8.25. ThT curve fit for AVP (5% w/w) in 85DMPC+15DMPS

150000 e

c 100000

s e

r -1

o Kagg = 0.02806 s

u 50000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.26. ThT curve fit for AVP (5% w/w) in 75DMPC+25DMPS

176

200000

e 150000

e c

s 100000 e -1

r Kagg = 0.03245 s

u l

F 50000

0 0 50 100 150 200 Time (hours)

Figure S8.27. ThT curve fit for AVP (5% w/w) in 65DMPC+35DMPS

100000

80000

c n

e 60000

e r

o 40000 -1

u Kagg = 0.003315 s

l F 20000

0 0 50 100 150 200 Time (hours)

Figure S8.28. ThT curve fit for Des (10% w/w) in DOPC

177

100000

80000

c n

e 60000

s e

r -1

o 40000 Kagg = 0.01523 s

l F 20000

0 0 50 100 150 200 Time (hours)

Figure S8.29. ThT curve fit for Des (10% w/w) in DOPC+Chol

200000

e 150000

e c

s 100000

o u l -1 F 50000 Kagg = 0.008910 s

0 0 50 100 150 200 Time (hours)

Figure S8.30. ThT curve fit for Des (10% w/w) in DOPC+DOPS

178

500000

400000

c n

e 300000

c -1

s Kagg = 0.03367 s

e r

o 200000

l F 100000

0 0 50 100 150 200 Time (hours)

Figure S8.31. ThT curve fit for Des (10% w/w) in DLPC

350000

e 300000

n e

c -1

s 250000 Kagg = 0.02301 s

u l

F 200000

150000 0 50 100 150 200 Time (hours)

Figure S8.32. ThT curve fit for Des (10% w/w) in DMPC

179

140000

e 120000

e c

s 100000

o u l Kagg = 0.02456 s-1 F 80000

60000 0 50 100 150 200 Time (hours)

Figure S8.33. ThT curve fit for Des (10% w/w) in DPPC

150000 e

c 100000

e c

s -1

e Kagg = 0.02491 s

r o

u 50000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.34. ThT curve fit for Des (10% w/w) in DSPC

180

200000

e 150000

e c

s 100000 e

r -1

o Kagg = 0.02911 s

u l

F 50000

0 0 50 100 150 200 Time (hours)

Figure S8.35. ThT curve fit for Des (10% w/w) in DMPC+Chol

250000

200000

c n

e 150000

s e

r -1

o 100000 Kagg = 0.03673 s

l F 50000

0 0 50 100 150 200 Time (hours)

Figure S8.36. ThT curve fit for SP (10% w/w) in DLPC

181

200000

e 150000

e c

s 100000

e r

o -1 u

l Kagg = 0.01754 s

F 50000

0 0 50 100 150 200 Time (hours)

Figure S8.37. ThT curve fit for SP (10% w/w) in DMPC

800000

e 600000

e c

s 400000 e

r -1

o Kagg = 0.02004 s

u l

F 200000

0 0 50 100 150 200 Time (hours)

Figure S8.38. ThT curve fit for SP (10% w/w) in DPPC

182

600000

e 500000

e c

s 400000 -1

e Kagg = 0.02631 s

u l

F 300000

200000 0 50 100 150 200 Time (hours)

Figure S8.39. ThT curve fit for SP (10% w/w) in DSPC

400000

e 300000

e c

s 200000 e

r -1

o Kagg = 0.02247 s

u l

F 100000

0 0 50 100 150 200 Time (hours)

Figure S8.40. ThT curve fit for SP (10% w/w) in DOPC+Chol

183

650000

600000

c n

e 550000

s e

r -1

o 500000 Kagg = 0.1574 s

l F 450000

400000 0 50 100 150 200 Time (hours)

Figure S8.41. ThT curve fit for SP (10% w/w) in DMPC+DMPS

1500000 e

c 1000000

e c

s -1

e Kagg = 0.04019 s

r o

u 500000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.42. ThT curve fit for SST (5% w/w) in resveratrol (RSV)

184

3000000 e

c 2000000

r o

u 1000000

l Kagg = 0.00341 s-1 F

0 0 50 100 150 200 Time (hours)

Figure S8.43. ThT curve fit for SST (5% w/w) in trans stilbene (TS)

2000000

e 1500000

e c

s 1000000 e

r -1

o Kagg = 0.02777 s

u l

F 500000

0 0 50 100 150 200 Time (hours)

Figure S8.44. ThT curve fit for SST (5% w/w) in trimethoxy resveratrol (TMS)

185

2500000

2000000

c n

e 1500000 -1

c Kagg = 0.2229 s

e r

o 1000000

l F 500000

0 0 50 100 150 200 Time (hours)

Figure S8.45. ThT curve fit for SST (5% w/w) in DLPC+RSV

3000000 e

c 2000000

s e

r -1

o Kagg = 0.03194 s

u 1000000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.46. ThT curve fit for SST (5% w/w) in DLPC+TMS

186

600000 e

c 400000

s e

r -1

o Kagg = 0.06603 s

u 200000

l F

0 0 50 100 150 200 Time (hours)

Figure S8.47. ThT curve fit for SST (5% w/w) in DLPC+TS

2500000

2000000

c n

e 1500000

e r

o 1000000

-1 u

l Kagg = 0.1354 s F 500000

0 0 50 100 150 200 Time (hours)

Figure S8.48. ThT curve fit for SST (5% w/w) in DMPC+RSV

187

1×107

8×106

e c

n 6

e 6×10

s e

r 6 -1

o 4×10 Kagg = 0.01584 s

l F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.49. ThT curve fit for SST (5% w/w) in DMPC+TS

2.5×107

2×107

e c

n 7

e 1.5×10

s e r 7 -1

o 1×10 Kagg = 0.02076 s

l F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.50. ThT curve fit for SST (5% w/w) in DMPC+TMS

188

6×106 e

c 6

n 4×10

e r

o -1 6

u 2×10 Kagg = 0.02006 s

l F

0 0 50 100 150 200 Time (hours)

Figure S8.51. ThT curve fit for SST (5% w/w) in DMPC+Chol+RSV

1.5×107 e

c 7

n 1×10

e r

o 6 -1 u 5×10

l Kagg = 0.03882 s F

0 0 50 100 150 200 Time (hours)

Figure S8.52. ThT curve fit for SST (5% w/w) in DMPC+Chol+TS

189

2×107

1e .5×10

e c

s 1×107 e r -1

o Kagg = 0.06057 s

u l

F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.53. ThT curve fit for SST (5% w/w) in DMPC+Chol+TMS

1×107

8×106

e c

n 6 e 6×10 -1

c Kagg = 0.4770 s

s e

r 6

o 4×10

l F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.54. ThT curve fit for SST (5% w/w) in DMPC+DMPS+RSV

190

1.5×107 e

c 7

n 1×10

e c

s -1 e

r Kagg = 0.4213 s o 6

u 5×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.55. ThT curve fit for SST (5% w/w) in DMPC+DMPS+TS

2.5×107

2×107

e c

n 7 e 1.5×10

c -1

s Kagg = 0.3329 s e

r 7

o 1×10

l F 5×106

0 0 50 100 150 200 Time (hours)

Figure S8.56. ThT curve fit for SST (5% w/w) in DMPC+DMPS+TMS

191

1×107

8×106

e c

n 6

e 6×10

s e

r 6

o 4×10 u

l Kagg = 0.02506 s-1 F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.57. ThT curve fit for SST (5% w/w) in DOPC+DOPS+TS

1.5×107 e

c 7

n 1×10

e r

o -1 6 Kagg = 0.2180 s

u 5×10

l F

0 0 50 100 150 200 Time (hours)

Figure S8.58. ThT curve fit for SST (5% w/w) in DOPC+DOPS+TMS

192

2500000

2000000

c n

e 1500000

e r

o 1000000

-1 u

l Kagg = 0.02100 s F 500000

0 0 50 100 150 200 Time (hours)

Figure S8.59. ThT curve fit for SST (5% w/w) in DPPC+RSV

4000000

e 3000000

e c

s 2000000

e r

o -1 u

l Kagg = 0.02764 s

F 1000000

0 0 50 100 150 200 Time (hours)

Figure S8.60. ThT curve fit for SST (5% w/w) in DPPC+TS

193

1×107

8×106

e c

n 6

e 6×10 c

s -1

e Kagg = 0.04871 s

r 6

o 4×10

l F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.61. ThT curve fit for SST (5% w/w) in DPPC+TMS

3000000 e

c 2000000

e r

o -1

u 1000000 Kagg = 0.04248 s

l F

0 0 50 100 150 200 Time (hours)

Figure S8.62. ThT curve fit for SST (5% w/w) in DSPC+RSV

194

5000000

4000000

c n

e 3000000

s e

r -1

o 2000000 Kagg = 0.02799 s

l F 1000000

0 0 50 100 150 200 Time (hours)

Figure S8.63. ThT curve fit for SST (5% w/w) in DSPC+TS

8×106

e 6×10

e c

s 4×106 e

r -1

o Kagg = 0.03053 s

u l

F 2×106

0 0 50 100 150 200 Time (hours)

Figure S8.64. ThT curve fit for SST (5% w/w) in DSPC+TMS

195