Mass Spectrometry Method Development to Identify Binding Ligands Against A2AR Nanodisc Complex

Home , Preladenant

Author Ma, Jun

Published 2017-11

Thesis Type Thesis (PhD Doctorate)

School School of Environment and Sc

DOI https://doi.org/10.25904/1912/1816

Downloaded from http://hdl.handle.net/10072/380580

Griffith Research Online https://research-repository.griffith.edu.au

Mass Spectrometry Method Development to Identify

Binding Ligands Against A2AR Nanodisc Complex

Jun Ma, BSc (Hons)

Griffith Institute for Drug Discovery,

School of Natural Science, Griffith University

Submitted in fulfilment of the requirements of the degree of Doctor of Philosophy

November 2017

Abstract

Protein is essential for human physiological processes and signalling pathways. Mass spectrometry (MS) is an important tool for ligand identification against protein target.

This project aims to establish an MS-based ligand identification method towards neurodegenerative disease-related protein targets, including cytosolic LRRK2 protein, adenosine A2A receptor (A2AR), and α2-adrenergic receptor (α2AR). The LRRK2 subdomains (Roc/GTPase, COR, and MAPKKK/kinase) and GPCRs (A2AR, A2AR-GFP, and α2AR) were cloned, sequenced, expressed and purified for MS assay. The proteins were solubilised in different types of detergents (such as Triton X-100 and n-dodecyl-β-

D-maltoside). Moreover, the A2AR, A2AR-GFP, and α2AR were reconstituted into self- assembled phospholipid bilayers (or nanodisc) to improve the solubility and stability.

Because LRRK2 subdomains and α2AR had solubility and yield problems that would limit the MS analysis, the A2AR nanodisc was consequently demonstrated as a suitable protein target for MS method development. Two MS methods, native ESI MS and ultrafiltration- based affinity LC/MS, were attempted for ligand identification against A2AR nanodisc.

The ultrafiltration-based affinity LC/MS was successfully developed to assay the interactions (binding affinities) between A2AR nanodiscs and 15-ligand mixture (eight known ligands with seven unrelated negative ligands). The ultrafiltration-based affinity

LC/MS allowed identification of ligands to a relatively stable unliganded/apo GPCR in nanodisc environment.

Statement of originality

This work has not previously been submitted for a degree or diploma in any university.

To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.

(Signed)__

Jun Ma

Table of contents

Abstract ...... i Statement of originality ...... ii Table of contents ...... iii List of figures ...... vii List of tables ...... xi Acknowledgements ...... xii Thesis-related publications ...... xiii List of abbreviations ...... xiv

Chapter 1 LRRK2 and A2AR are Parkinson’s disease-related proteins ...... 1 1.1 Cytosolic LRRK2 protein ...... 3 1.1.1 LRRK2 enzymatic domains of MAPKKK/kinase and Roc/GTPase ...... 6 1.1.2 Ligands targeting the G2019S mutation of LRRK2 ...... 11 1.2 A class of pharmaceutically relevant membrane proteins – G-protein coupled receptors (GPCRs) ...... 13 1.2.1 GPCR intracellular signalling pathway ...... 16 1.2.2 Structure-based knowledge of GPCRs ...... 22

1.2.3 Two neurological disease-related GPCRs – A2AR and α2AR ...... 28 1.3 Phospholipid environment for GPCR structure and function ...... 31 1.3.1 Self-assembled nanodisc for GPCRs ...... 37 1.3.2 Nanodisc reconstitution by different lipids and MSPs ...... 44 1.3.3 Recent advances of nanodisc technique ...... 48 1.4 Mass spectrometry (MS) ...... 50 1.4.1 Native MS and affinity LC/MS ...... 54 1.4.2 Ion source - The electrospray ionisation (ESI) ...... 61 1.4.3 Mass analysers - Fourier transform ion cyclotron resonance (FT-ICR) and time-of-flight (TOF) ...... 65 1.4.4 Recent MS advances ...... 68 1.5 Project proposal ...... 70 1.6 Aims ...... 71 Chapter 2 Cloning, expression and purification of LRRK2 subdomains (Roc/GTPase, COR and MAPKKK/kinase) ...... 72 iii

2.1 Abstract ...... 72 2.2 Results and Discussion ...... 72 2.2.1 Physicochemical property prediction by bioinformatics ...... 72 2.2.2 Cloning of Roc/GTPase and MAPKKK/kinase domains using traditional restriction enzyme-based digestion and ligation method ...... 75 2.2.3 Colony PCR and sequencing reaction ...... 81 2.2.4 Expression and purification of Roc/GTPase, COR and MAPKKK/kinase domains ...... 84 2.3 Methods ...... 88 2.3.1 Online bioinformatics analysis ...... 88 2.3.2 Polymerase chain reaction (PCR) ...... 88 2.3.3 Cloning by traditional restriction enzyme-based digestion and ligation method ...... 89 2.3.4 Recombinant plasmid verification by colony PCR and DNA sequencing ...... 89 2.3.5 Protein expression and purification ...... 89

Chapter 3 Expression and purification of A2AR, A2AR-GFP, and α2AR...... 92 3.1 Abstract ...... 92 3.2 Results and discussion ...... 92 3.2.1 Bioinformatics analysis for secondary GPCR structures ...... 92

3.2.2 Expression and purification of A2AR and α2AR ...... 94 3.2.3 Quantification using analytical size exclusion chromatography (aSEC) ...... 96 3.2.4 CPM thermostability assay ...... 98 3.3 Methods ...... 102 3.3.1 Secondary structure of GPCR proteins analysis by GPCRdb ...... 102 3.3.2 GPCR expressions and purification ...... 102 3.3.3 aSEC ...... 104 3.3.4 CPM thermostability assay ...... 105

Chapter 4 Reconstitution of A2AR, A2AR-GFP, and α2AR into nanodiscs ...... 106 4.1 Abstract ...... 106 4.2 Results and discussion ...... 106

4.2.1 Membrane scaffold protein (MSP) cloning by Overlap Phusion® PCR ...... 106 4.2.2 Expression and purification of MSP and TEV ...... 108 4.2.3 Nanodisc reconstitution ...... 110

4.2.4 A2AR nanodisc characterisation by DLS, negative staining EM, and CPM ...... 115 4.3 Methods ...... 122 4.3.1 MSP cloning by quick Overlap Phusion® PCR and mutagenesis ...... 122 4.3.2 GPCR nanodisc reconstitution and optimisation ...... 124 4.3.3 DLS ...... 125 4.3.4 Negative-staining electron microscopy ...... 125

Chapter 5 Ligand identification against A2AR nanodisc using native MS ...... 127 5.1 Abstract ...... 127 5.2 Results and discussion ...... 128 5.2.1 Parameters optimisation for ESI-FT-ICR MS ...... 128

5.2.2 Purified MSP1D1 (△His6) and MSP2N2 (△His6) soluble membrane scaffold proteins detection by native ESI-FT-ICR MS ...... 130 5.2.3 Empty nanodisc analysis by native ESI-FT-ICR MS and ESI-Q-ToF MS ...... 132 5.2.4 ESI-ICR-FT MS optimizations for nanodisc/detergent/protein mixture ...... 138 5.3 Methods ...... 142 5.3.1 Sample preparation for native ESI-FT-ICR MS ...... 142 5.3.2 Ligand detection by native ESI-FT-ICR MS ...... 143 5.3.3 Data analysis for native ESI-FT-ICR MS ...... 143

Chapter 6 Ligand identiﬁcation of adenosine A2A receptor in self-assembled nanodiscs by ultrafiltration-based aﬃnity LC/MS ...... 144 6.1 Abstract ...... 145 6.2 Introduction ...... 145 6.3 Materials and methods ...... 147 6.3.1 Constructs and GPCR expressions ...... 147 6.3.2 Purification and NuPAGE electrophoresis ...... 148 6.3.3 Analytical size exclusion chromatography (aSEC) ...... 149 v

6.3.4 Membrane scaﬀold protein expression and purification ...... 149

6.3.5 Reconstitution of the A2AR into the nanodisc ...... 150 6.3.6 Ligand identification by aﬃnity MS ...... 150 6.4 Results and discussion ...... 151

6.4.1 MSP1D1 and A2AR purifications ...... 152

6.4.2 A2AR nanodisc reconstitution ...... 153 6.5 Conclusions ...... 158 Chapter 7 General conclusions and future directions...... 159 Bibliography ...... 161 Appendix ...... 219

List of figures

Figure 1.1: A representation of LRRK2 domains and genetic variations...... 5

Figure 1.2: Primary amino acid sequences and tertiary structures of MAPKKK domain

(residues 1800 - 2138)...... 7

Figure 1.3: The X-ray crystal structure of human Roc/GTPase dimer (PDB ID: 2ZEJ). .. 8

Figure 1.4: Intracellular signals generated by GTPase cycling (A) and the proposed GTP-

dependent signalling pathway of LRRK2 (B)...... 9

Figure 1.5: Hypothetical LRRK2 regulation pathway in the dopaminergic neuron...... 10

Figure 1.6: Strategies for developing LRRK2 binding ligands...... 12

Figure 1.7: The phylogenetic tree of human GPCR superfamily...... 15

Figure 1.8: Examples of GPCR responses for five different types of ligands...... 17

Figure 1.9: A schematic diagram of GPCR signalling pathways in human...... 19

Figure 1.10: Advances in GPCR structural determination...... 24

Figure 1.11: GPCR architecture...... 27

Figure 1.12: A2AR and α2AR distributions and physiological effects in the human body.

...... 31

Figure 1.13: Structures and shapes of different phospholipids and cholesterol...... 32

Figure 1.14: The apoA-I and MSP structures...... 39

Figure 1.15: Nanodisc reconstitution procedures and membrane protein structures solved

in nanodiscs...... 41

Figure 1.16: General features and schematics of principal components in different MS.

...... 56

Figure 1.17: The schematic diagram of direct native ESI MS and ultrafiltration-based

affinity LC/MS workflow...... 57

Figure 1.18: An example of MSP1D1 mass spectrum resolved by native MS (ESI-FT-ICR

vii

MS and ESI-Q-ToF MS)...... 58

Figure 1.19: Identification of protein-ligand binding by affinity LC/MS from extracted ion

chromatography (EIC)...... 60

Figure 1.20: The schematic illustration of controlled-current ESI...... 63

Figure 1.21: The principle of FT-ICR mass analyser and the outline of ESI-FT-ICR MS.

...... 66

Figure 1.22: Schematic diagram of Q-ToF MS...... 67

Figure 2.1: The instability index and GRAVY values of Roc/GTPase, COR and

MAPKKK/kinase domains were calculated by ProtParam...... 74

Figure 2.2: Schematic diagram of pET-22b(+) expression vector, PCR products and

expressed proteins...... 76

Figure 2.3: The optimisation of template DNA concentration and annealing temperature.

...... 77

Figure 2.4: Optimisation of digestion time and PCR product purification...... 79

Figure 2.5: The quantity and quality of DNA were checked by Nanodrop...... 80

Figure 2.6: The colony PCR products of MAPKKK/kinase and Roc/GTPase ligated

products...... 81

Figure 2.7: The restriction digestion and protein expression of selected colonies...... 83

Figure 2.8: Expression of Roc/GTPase, COR and MAPKKK/kinase domains...... 85

Figure 2.9: Optimisation of lysis conditions for Roc, MAPKKK/kinase, and COR domains.

...... 86

Figure 3.1: Snake plots of A2AR (GenBank sequence ID: NP_000666.2) and α2AR

(GenBank sequence ID: NP_000672.3) generated from GPCRdb (http://gpcrdb.org/).

...... 93

Figure 3.2: The NuPAGE electrophoresis of α2AR, A2AR and A2AR-GFP purifications .. 95

viii

Figure 3.3: The relationships between BSA amount and chromatographic peak area (A)

or peak height (B)...... 96

Figure 3.4: The aSEC profiles of purified A2AR, A2AR-GFP and α2AR (A) and

corresponding normalised aSEC profiles (B)...... 97

Figure 3.5: The CPM thermostability profiles of A2AR DDM with different binding

ligands (A) and the comparison of melting temperature (Tm) with apo A2AR (B). . 99

Figure 3.6: Stability of apo A2AR in DDM micelle at 4 ºC over storage period of four

weeks...... 100

Figure 4.1: Schematic diagram of overlap Phusion® PCR for MSP cloning...... 107

Figure 4.2: Small-scale purification of MSP variants...... 108

Figure 4.3: The TEV purification (A) and MSP1D1 digestion optimisation (B)...... 110

Figure 4.4: A2AR nanodisc reconstitution ratio was optimised...... 111

Figure 4.5: The optimisation of A2AR-GFP (A and B) and α2AR (C and D) nanodisc

reconstitution...... 113

Figure 4.6: An empty nanodisc reconstituted by MSP2N2 with DMPC in different

MSP2N2: DMPC ratios of 1:300, 1: 350, and 1: 400...... 114

Figure 4.7: The DLS results of A2AR nanodisc comparing to A2AR in DDM detergent.

...... 115

Figure 4.8: Analysis of apo A2AR nanodisc stability by NuPAGE gel electrophoresis (A)

and SEC (B)...... 116

Figure 4.9: Aggregation and monomeric A2AR nanodisc species were applied to negative-

staining EM...... 117

Figure 4.10: CPM thermostability assays...... 118

Figure 5.1: Important ESI-FT-ICR MS parameters for optimizing a standard protein of 10

µM bCA (m/z 1200 - 3000)...... 129

Figure 5.2: MS spectra of MSP1D1 (m/z 1680 - 2500) analysed by ESI-FT-ICR MS. 131

Figure 5.3: MS spectra of MSP2N2 analysed by ESI-FT-ICR MS (m/z 2600 - 3280). 132

Figure 5.4: The full spectra of MSP1D1-encircled empty nanodisc acquired by native ESI-

FT-ICR MS (A) and Q-ToF MS (B)...... 133

Figure 5.5: The MS specta of empty nanodisc (m/z 980 - 1630) analysed by native ESI-

FT-ICR MS...... 134

Figure 5.6: MS spectra of empty nanodisc (m/z 900-1400) analysed by ESI-Q-ToF MS.

...... 136

Figure 5.7: The optimization of skimmer voltage (ESI-FT-ICR MS) for the mixture of 10

µM TB protein and 1.1% OG detergent...... 138

Figure 5.8: The optimisation of in-source CID energy (ESI-FT-ICR MS) for the mixture

of 10 µM TB protein and 1.1% OG detergent...... 139

Figure 5.9: The MS spectrum of the mimic mixture (10 µM TB protein/empty nanodisc

and 1.1% OG detergent) obtained at the skimmer voltage of 120 V and in-source CID

energy of -30 eV (ESI-FT-ICR MS)...... 141

Figure 6.1: Schematic diagram of affinity MS ligand identification for the A2AR nanodisc..

...... 152

Figure 6.2: NuPAGE electrophoresis analysis and aSEC profile of the A2AR...... 153

Figure 6.3: The optimisation of nanodisc reconstitution ratios...... 154

Figure 6.4: (A) Total ion chromatograms (TIC) acquired through compounds from the

mixture alone (green), A2AR nanodisc (red) and negative control (blue). (B) Extracted

ion chromatograms (EIC) of the compounds from the A2AR nanodisc (solid line) and

the negative control (dashed line), each chromatogram (from left to right) represents

a compound with strong, weak, or nonspecific binding to the A2AR...... 155

List of tables

Table 1.1: Examples of PD risk factors...... 2

Table 1.2: Comparison of micelle, bicelle, liposome, nanodisc and SMALP...... 34

Table 1.3: Different membrane scaffold proteins...... 45

Table 1.4: Properties of some phospholipids used in nanodisc reconstitution...... 47

Table 1.5: Biophysical (NMR, MS, SPR, thermal shift assay and X-ray crystallography),

biochemical (HTS, FBDD, and ITC) and in silico SAR modelling methods for

protein-ligand interaction studies...... 52

Table 3.1: Quantification of A2AR, A2AR-GFP and α2AR yield in Figure 3.4 based on linear

regression relationship in Figure 3.3...... 97

Table 4.1. Primers for overlap Phusion® PCR and QuickChange site-directed mutagenesis.

...... 122

Table 6.1: Identification and binding index determination for each compound in a 15-

compound mixture by ultrafiltration-based affinity LC/MS analysis...... 157

Acknowledgements

I would like to convey my sincerest appreciation to my supervisors, Professor Ronald Quinn,

Professor Zhi-Jie Liu, and Dr Wendy Loa-Kum-Cheung, for the help and guidance through my

PhD study. I would also like to extend my thanks to Professor Wenqing Shui, Professor Dong

Wu, Professor Qingtao Shui, and Professor George Mellick, who provide support, supervision and help with extreme patience throughout my thesis. Without the support of Professor Zhi-Jie

Liu, Professor Ronald Quinn and Dr Wendy Loa-Kum-Cheung, I would not have the invaluable research experience and knowledge from this project.

Second, I thank Professor Raymond Stevens and the cloning, cell expression, and protein puriﬁcation facilities of iHuman Institute (ShanghaiTech University) for their support. I thank the MS facility provider in National Centre for Protein Science Shanghai. I also thank the financial support from the Griffith Postgraduate Scholarship. Without the above efforts, this work would not have been possible. It was also a pleasure to work with all the members of iHuman institute (ShanghaiTech University) and Griffith Institute for Drug Discovery (Griffith

University).

Last but not least, I would like to thank my parents for their continued support through all my studies. I would not have attempted this endeavour without their help.

xii

Thesis-related publications

 (Chapter 3, 4 and 6) Ma, J., Lu, Y., Wu, D., Peng, Y., Loa-Kum-Cheung, W., Peng, C., Quinn

R. J., Shui, W., Liu, Z. J., (2017). Ligand Identification of Adenosine A2AR Receptor in Self-

Assembled Nanodisc by Affinity Mass Spectrometry, Analytical Methods, 9(40).

doi:10.1039/C7AY01891F;

 (The same ultrafiltration affinity LC/MS method was also used for ligand identification

towards human 5-HT2C GPCR receptor) Peng, Y., Zhao, S., Wu, Y., Cao, H., Xu, Y., Liu, X.,

Shui, W., Cheng, J., Zhao, S., Shen, L., Ma, J., Quinn, R. J., Stevens, R., Zhong, G., Liu, Z.

J., (2018) Identification of Natural Products as Novel Ligands for the Human 5-HT2C

Receptor, Biophysics Reports. https://doi.org/10.1007/s41048-018-0047-1;

 (Poster presentation) Ma, J., Lu, Y., Wu, D., Quinn R. J., Shui, W., Liu, Z. J., (5th/Dec/2016),

Ligand Identification of Adenosine A2AR Receptor in Self-Assembled Nanodisc, iHuman

4th Forum, ShanghaiTech University, Shanghai;

 (Poster presentation and rapid fire talk) Ma, J., Lu, Y., Wu, D., Peng, Y., Loa-Kum-Cheung,

W., Peng, C., Quinn R. J., Shui, W., Liu, Z. J., (2nd/Nov/2017), Ligand Identification of

nd Adenosine A2AR Receptor in Self-Assembled Nanodisc by Affinity Mass Spectrometry, 2

Queensland Mass Spectrometry Symposium, QIMR, Brisbane.

xiii

List of abbreviations a.a. – Amino acids

AA – Arachidonic acid

ABC transporter – ATP binding cassette transporter

AC – Adenylyl cyclase

ADP – Adenosine diphosphate

AFM – Atomic force microscopy

ANK – Ankyrin

API – Atmospheric pressure ionisation apoA-I – Apolipoprotein A-1

APP – Amyloid precursor protein

ARM – Armadillo

ATP – Adenosine triphosphate bp – Base pair bR – Bacteriorhodopsin

BRIL – Apocytochrome b562RIL

BSA – Bovine serum albumin

β2AR – β2 adrenergic receptor cAMP – Cyclic AMP

CHAPS – 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate chr – Chromosome

CL – Cardiolipin

CMC – Critical micelle concentration

COR – C-terminal of Roc

CPM – Thiol-speciﬁc ﬂuorochrome N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl]maleimide

xiv

CPR – Cytochrome P450 reductase

CV – Column volume

CYP3A4 – Cytochrome P450 3A4

DLS – Dynamic light scattering

DM – n-Decyl-β-ᴅ-maltopyranoside

DDM – n-Dodecyl-β-ᴅ-maltoside

DMPC – 1,2-dimyristoyl-sn-glycero-3-phosphocholine

DMPG – 1,2-dimyristoyl-sn-glycero-3-phospho-(1′-rac-glycerol)

DOB – (−)-1–2,5-dimethoxy-4-bromophenyl-2-aminopropane

DOPC – 1,2-dioleoyl-sn-glycero-3-phosphocholine

DOPE – 1,2-dioleoyl-sn-glycero-3- phosphoethanolamine

DOPG – 1,2-dioleoyl-sn-glycero-3-phospho-(1′-rac-glycerol)

DRD1 – Dopamine receptor D1

ECD – Extracellular domain

EM – Electron microscopy

ERK2 – Extracellular signal-regulated kinase 2

ES – Electrospray

ESI – Electrospray ionisation

ESI-MS – Liquid chromatography-electrospray ionisation mass spectrometry

FAB – Fast atom bombardment

FDA – Food and drug administration

FhuA – Ferric hydroxamate uptake protein

FI – Fluorescence intensity

FP – Fluorescence polarization

FPLC – Fast protein liquid chromatography

FRET – Fluorescence resonance energy transfer

FT-ICR – Fourier transform ion cyclotron resonance

FT-ICR MS – Fourier transform ion cyclotron resonance mass spectrometer

FT-MS – Fourier transform mass spectrometer

4E-BP – Eukaryotic translation initiation factor 4E-binding protein

GAP – GTPase-activating protein

GDP – Guanosine diphosphate

GEF – Guanine nucleotide exchange factor

Gi – Inhibitory regulative G-protein

GPCR – G protein-coupled receptor

GRK – G protein-coupled receptor kinase

Gs – Stimulative regulative G-protein

GST – Glutathione S-transferase

GtCT – G protein transducing C-terminal peptide

GTP – Guanosine triphosphate

HEK293T – Human embryonic kidney 293T cell line

His6 – Hexahistidine hLRRK2 – Human leucine-rich repeat kinase 2

HAc – Acetic acid hONS – Human olfactory neurosphere-derived

HPLC – High performance liquid chromatography

HTS – High-throughput screen

IC50 – Concentration required to inhibite 50% of enzyme activity

ICD – Intracellular domain

ICR – Ion cyclotron resonance

xvi

IMAC – Immobilised metal affinity chromatography

IP – Inositol phosphates

IP3 – Inositol 1,4,5-trisphosphate iPD – Idiopathic Parkinson’s disease

IPTG – isopropyl-β-ᴅ-thiogalactoside

IR – Infrared

IUPHAR – International Union of Basic and Clinical Pharmacology kDa – Kilodalton kb – Kilobase pair

L-dopa – Laevodihydroxyphenylalanine

LB – Luria Bertani broth

LBs – Lewy bodies

LCP – Lipidic cubic phase

LeuT – Leucine transporter

LMNG – Lauryl maltose neopentyl glycol

LRR – Leucine-rich repeat

LRRK2 – Leucine-rich repeat kinase 2

LSD – d-lysergic acid diethylamide

MALDI – Matrix-assisted laser desorption ionisation

MAP – Mitogen-activated protein

MAPK – MAP kinase

MAPKKK – MAP kinase kinase kinase

MBP – Myelin basic protein

MCS – Multiple cloning sites mGluR2 – Metabotropic glutamate receptor 2

xvii

MLC – Myosin light chain

MLK – Mixed-lineage kinase mRNA – Messenger RNA

MP – Maximum-parsimony

MPNL – Membrane protein nanodisc library

MS – Mass spectrometry

MSP – Membrane scaffold protein

NH3 – Ammonia

NH4Ac – Ammonium acetate

NJ – Neighbour-joining

Nrf-2 – Nuclear factor erythroid 2-related factor-2

NMR – Nuclear magnetic resonance

NTS1 – Neurotensin receptor type 1

OD – Optical density

OMIM – Online Mendelian Inheritance in Man

OmpA – Outer membrane proteins A

OmpX – Outer membrane protein X

P loop – Phosphate-binding loop

PBS – Phosphate buffered saline

PC – phosphocholine

PG – phosphoglycerol

PLE – polar lipid extract

PCR – Polymerase chain reaction

PD – Parkinson’s disease

PDB – Protein data bank

xviii

PE – phosphoethanolamine pI – Isoelectric point

PINK1 – Phosphatase and tensin (PTEN)-induced kinase 1

PKA – protein kinase A

PKC – protein kinase C

PLA – phospholipase A

PLC – phospholipase C

PMSF – Phenylmethanesulfonyl fluoride

POPC – 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

POPG – 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1′-rac-glycerol)

POPS – 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine ppm – Parts per million

Sc PrP – Prion of Scrapie

PTH1R – Parathyroid hormone 1 receptor

RI – Ribonuclease inhibitor

Roc – Ras of complex protein

RIPK – Receptor-interacting protein kinases

SAXS – Small angle X-ray scattering

SDS – Sodium dodecyl sulphate

SDS-PAGE – Sodium dodecyl sulfate polyacrylamide gel electrophoresis

SEC – Size-exclusion chromatography

SMA – Co-polymer

SNARE – Soluble NSF attachment protein receptors

SNpc – Substantia nigra pars compacta

SPA – Scintillation procimity assay

xix

TFMPP – m-trifluoromethylphenylpiperazine

TGF-α – Transforming growth factor α

TKL – Tyrosine kinase-like

TOF – Time-of-ﬂight

TRF – Time-resolved fluorescence

TR-FRET – Time-resolved fluorescence resonance energy transfer

TRK – Tropomyosin kinase

TRPV1 – Transient receptor potential cation channel subfamily V member 1

TSP – Thermospray ionisation

UCH-L1 – Ubiquitin COOH-terminal hydrolase L1

UF – Uranyl formate

Chapter 1 LRRK2 and A2AR are Parkinson’s disease-related proteins

Parkinson’s disease (PD) is the second most common neurodegenerative disease with a prevalence of more than 1% after 60 years of age 1. Globally, it is estimated that 6.9 million people suffer from PD and the number will double to 14.2 million by 2040 2. In

Australia, there are more than 70,000 people living with PD in 2015 3. Males are more often affected by PD than females at a ratio of 3:2 4. The clinical symptoms of PD were first characterised by James Parkinson in 1817 as resting tremor, bradykinesia, muscular rigidity, and postural instability 5. The pathological hallmark is the selective death of dopaminergic neuron in substantia nigra pars compacta (SNpc) of midbrain, which causes neurotransmission to be disordered 6. The other neuropathological symptoms include ubiquitin-proteasome protein degradation, mitochondrial dysfunction, oxidative stress, Lewy bodies (LBs) in the cytoplasm of dopaminergic neurons, fewer neuronal projections towards ganglia and synapses in nigrostriatal pathway 6-10. The average rate of dopaminergic neuron loss in SNpc is 0.1 – 0.2% per year, but PD patient has an accelerated rate of neuron loss for more than 70% – 80% per year 11.

PD is widely recognised as a multifactorial disease. The epidemiology and etiology studies suggest that PD is caused by a complex interplay between genetic and environmental factors 12-14. Ageing, gender, ethnicity, pesticide exposure, cumulative oxidative stress, impaired mitochondrial function, genetic mutations, protein misfolding, neuroinflammation are the risk factors of PD 4,15,16. Specifically, PD is reported to correlate with genetic mutations and aberrant protein expressions of PARK1 to PARK20, nuclear factor erythroid 2-related factor 2 (Nrf2), adenosine A2A receptor (A2AR), serotonin receptors (5-HT receptors), postsynaptic mGluR5 receptor, dopaminergic D2 receptor (Table 1.1). The detailed biological mechanisms of LRRK2 and A2AR will be introduced in section 1.1 and 1.2.

There is no cure for PD currently, except for some alleviation treatments such as administering Laevodihydroxyphenylalanine (L-dopa) and dopamine agonists. Recently, deep brain stimulation surgeries have also been used to rehabilitate electrical signal problems in neural circuits of PD patients 17.

Table 1.1: Examples of PD risk factors.

Genes Descriptions

A single base pair substitution at position 53 (A53T) in the long arm of PARK1 (SNCA chromosome 4 (4q21-4q22) promotes α-synuclein aggregation in Lewy or α-synuclein) bodies (LBs) and cause autosomal dominant early-onset PD 18.

One or more deletions in the long arm of chromosome 6 (6q25.2-q27) are PARK2 (parkin) associated with autosomal recessive early-onset PD 19.

PARK5 (UCH- A missense mutation of I93M in the fourth exon (4q14) contributes to L1) protein aggregation in neurons 20.

Missense mutations in MAPKKK/kinase domain of LRRK2 (chr12p 11.2- PARK8 (LRRK2) q13.1) are associated with autosomal dominant late-onset PD 21.

Nrf2-mediated antioxidant insufficiency contributes to oxidative stress and Nrf2 correlate with PD onset 22.

Overexpressed A2AR protein inhibits glutamate receptor activity and causes A2AR decreased level of nigrostriatal dopamine 23,24.

Dopaminergic D2 Loss of dopaminergic D2 receptor reduces the midbrain dopamine level and receptor impairs the corticostriatal synaptic transmission 25.

Serotonin 5-HT 5-HT receptors facilitate dopaminergic release in the dopamine-depleted receptors parkinsonian brain 26.

Glutamate Postsynaptic glutamate mGluR5 receptor regulates dopaminergic mGluR5 receptor transmitter release in PD patient 27.

1.1 Cytosolic LRRK2 protein

The LRRK2 gene (GenBank accession number: AAV63975) contains 51 exons that span

144 kb of the chromosomal region. The mRNA transcript of LRRK2 is 9 kb, which translates into a 286 kDa multi-domain protein with 2,527 amino acids. The LRRK2 protein is expressed in various tissues including blood, kidneys, lungs, spleen and brain and critical for metabolism and immunity 28-35. Marín (2006) confirms that the LRRK2 has a general structure of LRR-Roc-COR-MAPKKK 36. The N-terminus and C-terminus of LRRK2 are important for its stability and functions 37,38.

The genetic segregation phenomenon of LRRK2 indicates that the missense mutations in

LRRK2 are a causative factor of PD 28,29,39,40. In contrast to the mutations of parkin 19, DJ-

1 41 and PINK1 42 that contribute to early-onset PD 28, LRRK2 mutations account for late- onset autosomal-dominant and sporadic PD, with a mutation frequency from 2% to 40% in different populations, countries, and ethnicities 43-47. Because LRRK2 is responsible for late-onset PD, it indicates that LRRK2 may affect late neuronal development progressively. The LRRK2 mutations were first discovered in a Sagamihara PD family and several European PD families in 2004 21,28. However, the pathological symptoms were polymorphic for these families. The Sagamihara PD family with inherited I2020T mutation showed nigral degeneration without LB formation, dementia, cognitive dysfunction, or depression 48. However, the European PD families with inherited Y1699C or R1441C mutation demonstrated pathologic features of dementia, progressive supranuclear palsy (PSP), ubiquitin-positive neuronal inclusions or nigral degeneration

29. Therefore, patients with different LRRK2 mutations may present varied symptom combinations, including tauopathies (microtubule-associated tau protein aggregation), α- synucleinopathies (α-synuclein aggregation), ubiquitin deposits, and nigral neuronal loss in the SNpc 49.

Some common LRRK2 mutations, including R1441C/G/H (4321C>T/G or 4322G>A),

Y1699C (5096A>G), G2019S (6055G>A), and I2020T (6059T>C), are located within the LRR-Roc-COR-MAPKKK construct 50-52 (Figure 1.1). The R1441C/G/H mutations are at the same position at the end of α3 helix of GTPase/Roc domain, while the G2019S and I2020T mutations are located in the Mg2+-binding loop of MAPKKK/kinase domain.

These mutations (R1441C/G/H, Y1699C, G2019S and I2020T) are proposed to influence substrate accessibility and affinity to MAPKKK/kinase and Roc/GTPase domains 53. In

PD patients, G2019S substitution is reported as the most common LRRK2 mutation, accounting for 2.4% of familial PD cases and 0.3% of sporadic PD cases in Australia 44.

Furthermore, the penetrance of G2019S mutation is age-dependent, increasing from 17% at age of 50 to 85% at age of 70 54. Functional studies suggest that the G2019S mutation increases the LRRK2 phosphorylation activity for around three folds comparing to the wild-type LRRK2 36,55-57. Some LRRK2 mutations cause downstream protein (such as α- synuclein and tau) hyperphosphorylation, misfolding and aggregation 28,29,58. The overexpression of LRRK2 protein with G2019S mutation in neuronal cells causes a dramatic reduction in neurite length, mitochondria dysfunction, increased apoptosis, and enhanced oxidative stress 59-61. In mouse model, G2019S mutation leads to the degenerations of nigral dopaminergic neurons and primary cortical neurons 58,62,63. Some studies proposed that the structural LRR, COR, and WD40 domains possibly act as a hinge for LRRK2 dimerisation, which may activate and phosphorylate the downstream

PD-related proteins 32,37. However, the direct evidence on LRRK2 regulatory mechnism is still missing.

Figure 1.1: A representation of LRRK2 domains and genetic variations.LRRK2 protein has 2527 amino acids, encoded by 51 exons. It contains domains of armadillo (ARM), ankyrin repeat (ANK), leucine-rich repeat (LRR), Ras of complex protein (Roc), C- terminal of Roc (COR), mitogen-activated kinase kinase kinase (MAPKKK), and WD40.

The boundaries of domains are indicated by the numbers below the protein sequence. The reported common PD pathogenic mutations are indicated by the numbers on the top of gene sequence. This figure is modified from Li, Tan & Yu, 2014.

Phylogenetic analysis shows that the human LRRK1 and LRRK2 fall into the Roco family with a characteristic LRR-Roc-COR-MAPKKK domains in the structure 36. The

Roc/GTPase and MAPKKK/kinase domains are important for regulating cellular processes, signal transduction, protein-protein interaction, and membrane trafficking 64-

66. The Roc domain of LRRK1 controls the activity of its own MAPKKK domain and subsequently regulate downstream MAP kinase (MAPK) activity 67. Hence, the presence of both Roc and MAPKKK domains in LRRK2 raises the possibility of a combination of intermolecular and intramolecular controlling systems, which the LRRK2 Roc/GTPase domain can directly regulate the LRRK2 MAPKKK/kinase activity.

1.1.1 LRRK2 enzymatic domains of MAPKKK/kinase and Roc/GTPase

Phylogenetic analysis (SI 1 in Appendix) shows that the human LRRK1 and LRRK2 are the only homologous kinase proteins that belong to receptor-interacting protein kinase

(RIPK) and the mixed lineage kinase (MLK) families, with serine/threonine and tyrosine kinase activities. The kinases from the TKL and the RIPK families are known to activate the mitogen-activated protein kinase (MAPK) signalling cascade and subsequently regulate the signal transduction, cell cycles, cytoskeletal rearrangement, cell apoptosis, and differentiation 68-70. The MAPKKK/kinase domain of LRRK2 consists of several conserved features, including ATP-binding cleft, catalytic loop, Mg2+ binding loop, and activation loop (Figure 1.2 (A)).

Since the X-ray structure of LRRK2 MAPKKK domain is not solved to date, its homology model of extracellular signal-regulated kinase 2 (ERK2) is used to gain insight of the LRRK2 MAPKKK structure 71-74 (Figure 1.2 (B)). This homology model shows that the ATP-binding cleft (residue 1800 – 2016) at N-terminus contains a glycine-rich loop (residues 1885 – 1982) and a catalytic loop (residues 1989 – 2002), which can stabilise the phosphate group of ATP and transfer the phosphate group from ATP to a substrate 75. The MAPKKK domain retains the kinase activity through the residue of aspartate (D1994) that is after a conserved arginine (R1993) residue in the catalytic loop.

The activation loop (residue 2017 – 2045) is located in the centre of MAPKKK and after the ATP-binding cleft. The activation loop contains a flexible Mg2+ binding loop

(including a conserved DYG tripeptide motif), P loop and P+1 loop (including a conserved APE tripeptide motif). The MAPKKK activity is controlled by the phosphorylation of activation loop. The DYG motif rotates out of the hydrophobic activation loop and occupies the ATP-binding cleft in the inactive state of MAPKKK

(inactive DYG-out form). Once the activation loop is phosphorylated, the activation loop will move towards to solvent and the ATP-binding cleft become exposed and activated 6

(active DYG-in form). It is found that the D2017F/Y mutations change the active DYG- in form to inactive DYG-out form, while the G2019S and I2020T mutations stabilise the active DYG-in form and increase the LRRK2 kinase activity 75-77.

Figure 1.2: Primary amino acid sequences and tertiary structures of MAPKKK domain

(residues 1800 - 2138).(A) The amino acid sequences of MAPKKK domain. The conserved regions (bold letters and grey boxes) include catalytic loop (between β6 and

β7), Mg2+ binding loop (between β8 and β9), activation loop (P loop and P+1 loop),

R1993, catalytic core of D1994, DYG and APE tripeptide motifs. (B) The homologous

ERK2 protein model of active DYG-in MAPKKK (PDB ID: 5U6I) with conserved ATP- binding cleft (yellow), activation loop (red), catalytic loop (cyan), and G2019/I2020

(magenta) between ATP-binding cleft and activation loop. 7

Roc is a small GTPase domain with a molecular mass of 20 - 25 kDa. It is the only human

LRRK2 subdomain with a solved X-ray structure (Figure 1.3). Roc/GTPase forms a dimer when expressed in vitro 32. The Roc dimer includes a GTPase catalytic core region, and a guanine nucleotide phosphate-binding loop (i.e. activation loop or P loop) that is common for ATP- or GTP-binding proteins 78. The P loop stabilises the binding with

GDP (inactive form) or GTP (active form).

Figure 1.3: The X-ray crystal structure of human Roc/GTPase dimer (PDB ID:

2ZEJ).Two individual Roc monomers are shown separately by green and red ribbon representations. Each Roc domain consists of five α-helices (α1 - α5) and six β strands

(β1 - β6). The binding ligands of GDP-Mg2+ are shown as ball-and-stick representations

(yellow sticks of GDP and blue balls of Mg2+). The activation loop (or P loop) that binds to GDP/GTP is between the α1 helix and β1 strand. This figure is modified from Deng et al. 2008.

The Roc/GTPase domain plays an important role in intracellular signalling transduction

66,79-81. An upstream signal can stimulate the GDP dissociation from the GDP-bound inactive Roc domain, which is followed by the GTP binding (Figure 1.4 (A)). This step is stimulated by the guanine nucleotide exchange factor (GEF) and leads to conformational changes. The GTPase activating protein (GAP) helps to convert the GTP- bound active form to the GDP-bound inactive form. This cycling elicits the downstream effects and regulates the cellular processes.

Figure 1.4: Intracellular signals generated by GTPase cycling (A) and the proposed GTP- dependent signalling pathway of LRRK2 (B).

It is known that Roc/GTPase domain can regulate MAPKKK/kinase activity in the human

LRRK1 protein, which is homologous to LRRK2 67. Similarly, a GTP-dependent LRRK2 reciprocal feedback model (Figure 1.4 (B)) is proposed in which GTP binds to the Roc domain and regulates the phosphorylation activity of MAPKKK domain 67,81,82. The evidence includes that the K1347A and T1348N mutations in Roc/GTPase domain prevent the GTP accessibility to the activation loop and subsequently abolish the kinase activity in neuronal cells 66,81,83,84. Additionally, the T2031A and S2032A mutations in the Roc activation loop diminish 80 - 90% of the autophosphorylation activity of

MAPKKK/kinase domain and form a kinase-dead LRRK2 variant 56.

This reciprocal feedback loop (Figure 1.5) may be important for functions of tubulins, microtubules, protein trafficking, and neurite outgrowth 85-88. Numerous downstream effectors and interaction partners of LRRK2 have been identified, but it is still lack of the understanding of endogenous signalling pathway, direct physiological substrates, and

LRRK2 mechanisms in the PD pathogenesis.

Figure 1.5: Hypothetical LRRK2 regulation pathway in the dopaminergic neuron.

1.1.2 Ligands targeting the G2019S mutation of LRRK2

Nowadays, the LRRK2-associated PD can be readily diagnosed by genome sequencing and intervened by neuroprotective drugs 89,90. The dopamine deficiency in PD patients is currently treated and relieved by administering dopamine precursors, such as laevodihydroxyphenylalanine (L-dopa). Nevertheless, this treatment causes many side effects such as long-term central gastrointestinal and cardiovascular complications and progressive drug efficacy loss 91. Many PD therapeutic drugs also target the kinase domain, because phosphorylation is an important regulatory mechanism in cells 92. The characterisation of downstream signalling molecules of LRRK2 MAPKKK/kinase domain delineates the LRRK2 regulatory mechanism in the neurons 69. As the G2019S mutant of LRRK2 has an increased kinase activity and substrate specificity, inhibitors against LRRK2 kinase domain activity have been developed. The inhibition of kinase activity shows a relief from the neurotoxicity in PD models of Caenorhabditis elegans and Drosophila with G2019S mutation 93. In consequence, the kinase inhibitors targeting the G2019S mutant of LRRK2 may elicit neuroprotective activities against PD 94,95.

There are three types of LRRK2 kinase inhibitors, namely type I, type II, and type III kinase inhibitors. Most LRRK2 kinase inhibitors are known as type I kinase inhibitors that directly compete with ATP in the ATP-binding site during its active state. More than

40 type I kinase inhibitors have been discovered against LRRK2 wild-type and G2019S mutant by 2014 (SI 4 in Appendix). However, the ATP-binding site is highly conserved across the human kinome (around 519 kinases) and the physiological concentration of

ATP in the human body is maintained at a high and constant level (1 - 5 mM in

68,96,97 concentration) . As a result, type I kinase inhibitors usually have a high dosage (IC50) and less specificity. The type II inhibitors bind to both the ATP-binding site and an adjacent allosteric region of the kinase. Besides targeting the ATP-binding site of LRRK2 by type I and II kinase inhibitors, the Roc/GTPase domain and protein-protein interaction 11

domains (ARM, LRR, COR, and WD40) can also be used for developing kinase inhibitors

(Figure 1.6). These domains are reported to be responsible for dimerisation, allosteric activation, downstream signalling pathway, and protein localisation 32,37,50,98-102. These type of inhibitors, binding exclusively to the adjacent allosteric pocket around the ATP- binding site, are called type III kinase inhibitors 103. The type II and III kinase inhibitors may disrupt the hydrophobic interaction and LRRK2 dimerisation and subsequently reduce neuronal toxicity 50. Since the allosteric sites are not so highly conserved as the

ATP binding site, the type II and III kinase inhibitors are more specific with lower off- target effects than type I kinase inhibitors. For example, a bidentate-binding dual inhibitor

(SR9444), binding to the allosteric pocket outside of ATP binding sites of LRRK2-

G2019S mutant, can improve mitochondrial functions in neuronal cells with a reasonable selectivity and potency 104.

Figure 1.6: Strategies for developing LRRK2 binding ligands.This figure is modified from Ho et al., 2014.

As the direct downstream substrate of LRRK2 MAPKKK/kinase is not known, many functional studies use LRRK2 autophosphorylation, generic kinase substrates (myelin basic protein (MBP) and myosin light chain (MLC)), and putative kinase substrates (such

as moesin, ezrin, radixin, eukaryotic translation initiation factor 4E-binding protein (4E-

BP), tubulin-β-2C, MAPKK, synthesised peptides of LRRKtide and Nictide) to estimate the kinase activity 37,58,105-108. It is important to note that the kinase activity measured by generic, synthetic substrate or autophosphorylation may not directly correlate with the real kinase activity in vivo. Another challenge is the lack of structural information of the human LRRK2 MAPKKK domain. The structure-based drug design for LRRK2

MAPKKK domain is based on the homologous kinases from Dictyostelium discoideum

(PDB ID: 4F0F) and ERK2 protein (PDB ID: 5U6I) 71-74. The elucidation of LRRK2 enzymatic activities, downstream substrate and protein-protein interaction will facilitate a better understanding of the relationship between LRRK2 and PD pathogenesis.

1.2 A class of pharmaceutically relevant membrane proteins – G-protein coupled receptors (GPCRs)

Cells, as the primary structural and functional units of human, are responsible for communicating with the extracellular environment and neighbour cells via a cell membrane barrier. Membrane proteins (MPs) are critical players in processing foreign signals (such as taste, light, odour, hormones, lipids and chemicals), regulating chemical transportation, and maintaining cellular homoeostasis. G-protein coupled receptors

(GPCRs) are the largest and one of the most influential MPs in humans. There are 826

GPCRs in the human genome, which comprise 3% of the proteins in the human proteome

109. GPCRs are pleiotropic membrane proteins mediating the transduction of diverse arrays of extracellular stimuli into the cells. In human, the signalling molecules range from small chemicals (hormones, neurotransmitters, metabolites, odorants, tastants, and lipids) to larger peptides (chemokines and incretins). The signalling molecules exert specific downstream effects through GPCRs and subsequently regulate immune response, hormone secretion, neurotransmitter release, smooth muscle relaxation/contraction, cell apoptosis, chemotaxis, cell aggregation, nociception, neuroplasticity, sleep-wakefulness 13

cycles, food intake, learning, memory, and behaviour 109-112. It is estimated that around

80% signals of hormones, neurotransmitters, neuromodulators, autocrines and paracrines transmit across the cell membrane through GPCRs 113.

GPCRs are one of the most important pharmaceutical targets and there are 33% of marketed drugs target GPCRs 114. These receptors are related to embryonic development, stem cell maintenance 115, and disease progressions including cardiovascular diseases

(CVDs) 116-118, inflammation 119-121, analgesia 122-124, cancer 125-128, metabolism disorders

(type II diabetes and obesity) 129-131, and neurologic disorders (multiple sclerosis,

Parkinson’s disease and Alzheimer’s disease) 132-136. GPCRs are non-uniformly expressed in various tissues, organs and cell types and provide tissue-selective functions 137. For instance, the rhodopsin-like receptors are expressed in the sensory system to mediate physiological functions of vision, gustatory, odorant detections, and renin-angiotensin- aldosterone system 138,139. The chemokine receptors are expressed in the immune system

119,140,141, while the aminergic receptors (including adrenergic, dopaminergic, muscarinic, histaminergic and serotoninergic receptors) are expressed in the central/peripheral nervous system 142-144. Additionally, the secretin receptor of glucagon-like peptide-1 receptor (GLP1R) in the pancreas is necessary for maintaining blood glucose level 145,146, while the cannabinoid type 1 receptor (CB1) in the CNS is essential for inhibiting glutamate release during neurotransmission 147-150.

The human GPCRs are divided into five major families according to the GRAFS acronym classification system that are rhodopsin-like family (719 members), secretin family (16 members), glutamate family (22 members), adhesion family (33 members) and frizzled/taste2 family (36 members) 151. Besides the human GPCR families, another two

GPCR families are the fungal mating pheromone receptor family in fungus and the cyclic

AMP receptor family in Dictyostelium discoideum (amoeba). To date, the crystal

structures of 36 rhodopsin-like receptors, four secretin receptors, two glutamate receptors, and one frizzled receptors have been published (Figure 1.7). Among the 826 GPCRs, the endogenous ligands of around 140 orphan GPCRs have not yet been identified 114.

Figure 1.7: The phylogenetic tree of human GPCR superfamily.The receptors with asterisks are the GPCRs with solved structures to date. The figure is adapted from Lv et al., 2016 and updates until March/2018.

1.2.1 GPCR intracellular signalling pathway

Ligands include ions, hormones, chemokines, organic compounds, odorants, biogenic amines (neurotransmitters), peptides, proteins, lipids, nucleotides and photons 151. The endogenous ligands are the common signalling molecules for GPCRs in the human body.

GPCR ligands can bind to the extracellular domain, within the cavity of the helix bundle in the outer half transmembrane region, the orthosteric binding pocket or an allosteric site surrounding the binding pocket 152. The small-molecule ligands (i.e. compounds) bind deeper in the transmembrane domain crevice than the larger ligands (i.e. peptides) 153.

The GPCR ligands are important in drug discovery and tools for understanding neurotransmission, enzymatic catalysis, signal transduction, and immune response.

In the absence of ligand interaction, GPCRs usually appear as a heterogeneous population with a mixture of different states and conformations. At the apo/unliganded state, the

GPCR physiological activity is called constitutive or basal activity. The inactive and active GPCR conformations are in constant equilibrium 154,155. The apo GPCR shows different structural arrangements and functions in response to different types of ligands.

The general terms to describe GPCR ligands encompass strong/full agonist, partial agonist, neutral competitive antagonist, inverse agonist, and partial inverse agonist

(Figure 1.8). Different types of ligands have different selectivity and elicit different biological activities. An agonist will increase the receptor activity and downstream G- protein response. A strong/full agonist binds preferentially to the active conformation of

GPCR and drives the equilibrium to the fully active state, such as fully activated β1 adrenergic receptor by noradrenaline. A partial agonist still has a higher affinity for active conformation over inactive form, but elicit a partial response and less selectivity when compared to a strong (full) agonist. The ligand, binding selectively to the inactive conformation and dampening the receptor activity below the basal/constitutive level, is called a strong inverse agonist. The strong inverse agonist of GPCR drives the equilibrium 16

towards an inactive state, resulting in a decreased GPCR activity. A partial inverse agonist has a higher affinity to the inactive form of GPCR than the active form, but less selectivity than the strong inverse agonist. A neutral competitive antagonist blocks/inhibits the agonist or inverse agonist binding to the receptor (avoiding their pharmacological functions) and maintains the receptor at basal activity. The neutral antagonist has an equal affinity to both active and inactive conformations so it does not alter the equilibrium 156,157.

Based on the sources, the GPCR ligands are also divided into two different categories, i.e. endogenous ligands and exogenous ligands. The endogenous ligands are natural ligands that originate from the physiological system, while exogenous ligands are added to a physiological system as drug or external chemicals by researchers.

Figure 1.8: Examples of GPCR responses for five different types of ligands.

Upon ligand binding, GPCRs can make a conformational change at the cytoplasmic side and cause the accessibility changes to the downstream effectors, including canonical heterotrimeric G-protein, non-canonical β-arrestin, or other messengers (Figure 1.9). The heterotrimeric G-proteins belong to the GTPase family. They can undergo GTPase cycling and transduce the signalling to the downstream effectors (Figure 1.4 (A)). Gα is one of the subunits of heterotrimeric G-protein, which interacts with GDP at an inactive state or GTP at an active state. In the absence of extracellular signal or at apo/unliganded 17

state, Gα forms an inactive complex with GDP and Gβγ (inactive GPCR/Gα/GDP/Gβγ complex). Upon receptor activation, the receptor undergoes a conformational change and

158 the GDP is exchanged to GTP . The Gα dissociates from GPCR and Gβγ, and forms an active complex with GTP (active Gα/GTP complex) (Figure 1.9). Two subunits of

Gα/GTP and Gβγ remain within the cell membrane and diffuse laterally. These subunits

159 will subsequently interact with other downstream messengers . The activated Gα/GTP proteins relay different signals to the downstream messengers depending on four different

160 Gα families, i.e. Gi/o, Gs, Gq and G12/13 . These four Gα families can stimulate or inhibit the activity of downstream messengers including adenylyl cyclase (AC) and phospholipase C (PLC). The AC and PLC are enzymes that associate with the intracellular side of cell membrane and catalyse the reactions in the cytoplasm. The signalling networks in the cytoplasm recruit various intracellular messengers, such as protein kinase C (PKC), protein kinase A (PKA), cyclic AMP (cAMP), 1,2-diacylglycerol

2+ 110 (DAG), 1,4,5-trisphosphate (IP3) and intracellular Ca . For example, the activated Gα initiates the catalysis of ATP to cAMP by AC enzyme and phosphorylation of phosphatidylinositol (PI) to DAG and IP3 by PLC enzyme 161. These signalling networks are critical for a variety of body processes, such as blood clotting by thrombin receptor in platelets, inflammation-induced airway contraction by β2AR, and exocrine secretion by a purinergic P2Y11 receptor on pancreatic duct epithelial cells 162-164. After GTP is hydrolysed to GDP, the heterotrimeric G-protein reassociates with the GPCR again. By this activation/deactivation cycle, GPCRs act as a fine-tuning switch on the cell surface and initiate a broad range of physiological responses. The downstream messengers are considered as a biomarker for GPCR activity and can be used for screening assays.

Alternatively, the GPCRs can also undergo phosphorylation at the carboxyl terminus by

GPCR kinase (GRK) 165,166. The phosphorylated GPCR will recruit a multifunctional adaptor protein of β-arrestin. The β-arrestin blocks G-protein coupling with GPCR 18

through a steric hindrance mechanism. This process is referred as desensitisation (Figure

1.9). The β-arrestin signalling pathway mediates clathrin-dependent endocytosis as well as several independent downstream effectors, such as β-arrestins scaffold mitogen- activated protein kinases (MAPKs), extracellular signal-regulated kinase (ERK), c-Jun

N-terminal kinase (JNK), tyrosine kinases and E3 ubiquitin ligases 167-170. The G-protein signalling pathway and β-arresting signalling pathway are important for sensations, hormone production, and neurotransmission.

Figure 1.9: A schematic diagram of GPCR signalling pathways in human. G-protein signalling pathway: Ligand binding induces a GPCR conformational change and activate downstream signalling pathway. The activated GPCR, coupling to heterotrimeric

G-protein (Gα/Gβγ) and GTPase, is dissociated. The activated G-protein signalling transduces the signal to second messengers including cAMP (levels can increase or decrease in response to different GPCR activation signals), IP3, DAG and Ca2+. β- arrestin signalling pathway: The phosphorylated GPCR recruits β-arrestin and blocks

G-protein signalling pathway (desensitisation). The β-arrestin mediates endocytosis of activated GPCRs and transduces β-arrestin signalling to the downstream messengers, including MAPKs, tyrosine kinases and E3 ubiquitin ligases. 19

GPCR biased agonism (functional selectivity) was also discovered in the past decades

171,172. The biased agonism proposes that ligand(s) can activate one or two parallel signalling pathways (G-protein or β-arrestin-mediated signalling pathway) to different level. As a result, an agonist in one signalling pathway can also play a role as a neutral antagonist or inverse agonist in another signalling pathway in different cell types and functional status. The biased agonism (functional selectivity) may be resulted from a diversity of coupled proteins (G-proteins or β-arrestin), different ligand scaffolds, preferred downstream signalling molecules, ligand/receptor oligomerisation, and allosteric modulation 173-175. Therefore, the biased agonism theory proposes that a ligand is classified explicitly by their individual effect in certain cells, instead of simply being classified as an agonist or antagonist. A ligand, initiating GPCR signalling pathway through either G-protein or β-arrestin selectively, is named as G-protein-biased ligand or

β-arrestin-biased ligand correspondingly. The notable examples of biased agonism are 5-

HT2 serotonin receptors, µ-opioid receptors, β1 and β2 adrenergic receptors, V2 vasopressin receptors, and some dopaminergic receptors 155,176-178. For instance, serotonin is the primary endogenous ligand of 5-HT2 receptors (including 5-HT2A, 5-HT2B, and 5-

HT2C receptors). The 5-HT2 receptors are coupled to both phospholipase C (PLC) and

179 phospholipase A2 (PLA2) enzymes . The 5-HT2 can activate PLC by Gq/11 and lead to the stimulation of secondary messengers of DAG, IP3 and Ca2+ (PLC-IP pathway) (Figure

1.9). The 5-HT2 can also stimulate PLA2 and result in the release of arachidonic acid or

180,181 AA (PLA2-AA pathway) .

It is found that serotonergic agonists (bufotenin, d-lysergic acid diethylamide (LSD), (−)-

1–2,5-dimethoxy-4-bromophenyl-2-aminopropane (DOB), quipazine, TFMPP, mCPP, psilocybin, mescaline, lisuride and ergotamine) can bind 5-HT2 receptor with high affinity but activate receptor differentially, which exhibit functional selectivity at 5-HT2 receptors 178,182-184. For example, DOB and LSD are hallucinogenic drugs, whereas 20

lisuride and ergotamine have no hallucinogenic effects 185,186. Thus, DOB and LSD might regulate a different signalling pathway from lisuride and ergotamine in neural cells.

Furthermore, DOB and LSD show pleiotropic hallucinogenic potential in human 179,187.

For 5-HT2 receptor-mediated PLC-IP3 signalling pathway, LSD demonstrated little stimulation activity (22%) relative to that of DOB (80%) 178,188. Therefore, the activations of 5-HT2 receptor by different agonists stimulate the secondary messengers (IP3, DAG,

AA, Ca2+ and 2-arachidonoylglycerol) differently 189. It is also an explanation why some hallucinogenic phenethylamine derivatives of 5-HT2 agonists have psychedelic side effects (such as DOB), whereas other serotonergic compounds do not. The biased agonism of 5-HT2 receptor or functional selectivity suggests some novel antipsychotic drug development approaches towards schizophrenia and depression with no or less side effect.

The intracellular secondary messengers of GPCR signalling pathway (Figure 1.9) can be used to monitor ligand binding and signalling transduction by technologies of radioactive ligand binding assay, calcium (Ca2+) flux assay, GTP/cAMP level modulation, and β- arrestin recruitment assay. The calcium ion is naturally produced upon activating the

GPCR coupled Gq — one of the four most important Gα protein families (Figure 1.9). The

2+ genetic modification allows intracellular Ca coupling to all four Gα protein types, i.e.

2+ Gi/o, Gs, Gq, and G12/13. Fluorescent Ca can be used to visualise small-molecule binding to GPCRs. The Ca2+ flux assay together with fluorescent imaging plate reader (FLPR) has been used to identify the endogenous ligand of GPR14 (or urotensin-II receptor), which is important for vasoconstriction and cardiovascular homeostasis 190,191. Similarly, the cAMP, one of the most important secondary messenger and biomarker in GPCR signalling pathways, is coupled with luminescent tags (such as pigment from frog skin cell) and used for cell-based high-throughput screenings assay towards small molecules

192-196 . The Gs/Gq-coupled GPCRs stimulate cAMP production, whereas Gi/Go-coupled

GPCRs inhibit cAMP production. The intracellular cAMP level can be directly reflected by luminescent intensity in different subcellular localisations. It has developed into a broadly applicable screening assay based on intracellular cAMP level.

The binding of β-arrestin to activated GPCR causes receptor desensitisation (receptor- arrestin complex endocytosis) and inhibit the GPCR signalling pathway. During the desensitisation, the ligand is removed and the receptor is recycled back to the cell membrane (Figure 1.9). β-Arrestin linked with a green fluorescent protein (GFP) tag can be used in genetically engineered cell lines. The β-arrestin movement can be monitored during the recycling process 197,198. Lastly, another technique of bioluminescence resonance energy transfer (BRET) measures changes in light emission based on the interaction between an electron donor attached to β-arrestin and an electron acceptor expressed with target GPCR 199,200.

GPCR is one of the most heavily investigated protein targets, with an exponentially increased number of publications 201,202. The structural determinations of GPCRs allow better understandings of GPCR functions.

1.2.2 Structure-based knowledge of GPCRs

GPCR structures facilitate the structure-based drug discovery towards many diseases and contribute better understanding of the underlying physiological roles of GPCRs. Three methods currently used for GPCR structural determinations are cryogenic electron microscopy (cryo-EM), nuclear magnetic resonance (NMR), and X-ray crystallography.

The GPCR structural determinations were convoluted and torturous in the early stage.

The GPCRs are typically heavily post-translationally modified by glycosylation, acylation, and phosphorylation, herein these modifications lead to heterogeneous receptors on the cell membrane 203. The other difficulties for GPCR structural 22

determination include intrinsic protein instability, low natural abundance, toxicity during overexpression, and difficulties in purification and crystallography.

The bacteriorhodopsin, a proton-pumping membrane protein, was used to gain the first insight into membrane protein structure by cryo-EM, since it has an organised pattern in two-dimensional crystals. The first bR structure gave a resolution of 7 Å in 1975 204. Since then, a better resolution closer to 3 Å has been attempted. The first three-dimensional atomic resolution structure of purple bacterium Rhodopseudomonas viridis was determined in 1985 205. The landmark in the GPCR research was the structure of bovine rhodopsin by X-ray crystallography in 2000 at a resolution of 2.8 Å 206. It was the first

GPCR crystal structure obtained. It took another seven years to develop more advanced technologies and systematic methods for human β2 adrenergic receptor (β2AR) structure determination in 2007 207,208. It was the first human high-resolution GPCR crystal structure with a diffusible bound ligand. This structure showed GPCR binding with a partial inverse agonist and illustrated how GPCR was activated.

Among the five GPCR classes, rhodopsin-like receptors are the most heavily studied to date (Figure 1.10 (A)) and these structures provide a basis for the homology modelling, drug design, and functional studies for other classes of GPCRs 202. Despite the fact that

GPCRs are such important therapeutic targets, only 48 unique GPCR structures (totally

220 available GPCR structures) from four GPCR families are solved (Figure 1.10 (A)).

This is a small component of the 741 unique MP structures (totally 1,337 available MP structures including GPCR structures) and accounts for only 1% protein structures deposited in PDB (totally 137,322 available protein structures) thus far 209-212. The challenge is because of the extreme difficulty in obtaining good quality and quantity of

GPCR crystals from the cell expression systems.

Figure 1.10: Advances in GPCR structural determination.(A) The ladder type of phylogenetic tree of solved human GPCRs is generated by GPCRdb (http://docs.gpcrdb. org). The alignment is based on amino acid sequences of seven transmembrane domains

(TM1 - TM7). The phylogenetic distance is calculated by neighbour-joining method. The branch lengths are even. The annotations of “R”, “L” and “C” on the top of the coloured vertical stripes represent “receptor family”, “ligand type” and “class” respectively; (B)

The total number of published GPCR structures each year since 2000. 24

There has been an exponential growth of solved GPCR crystal structures since 2007

(Figure 1.10 (B)). This fast growth is due to a number of technological advances since

2007, such as stabilising mutagenesis 213, protein engineering 214-216, expression hosts development, conformation-selective nanobody development, stabilising ligands screening, lipidic cubic phase (LCP) crystallisation 217, new mild detergents, X-ray free- electron laser (XFEL), single particle electron microscopy (EM) analysis and microfocus synchrotron beamlines 215.

The amphipathic GPCRs reside in an insulated phospholipid bilayer, and have a prominent seven-helical transmembrane (7TM) structure (Figure 1.11 (A)). Even though

GPCRs may have little sequence similarity, they all have well-conserved 7TM architecture 112. The GPCRs contain several conserved motifs that are important in receptor structures and functions. For instance, each α-helix or transmembrane (TM) region consists of 25 - 35 consecutive hydrophobic residues 218. These seven α-helices

(TM1 - TM7) form a receptor in a counter-clockwise manner, starting from extracellular

N-terminus until intracellular C-terminus (Figure 1.11 (A)). The 7TM bundles are connected by hydrophilic extracellular and intracellular loops (ECLs or ICLs), i.e. three extracellular loops (ECL1 - ECL3), three intracellular loops (ICL1 - ICL3) and followed by a short cytosolic helix (eighth helix or H8) resting in parallel to the cell membrane.

The other conserved GPCR features include a disulphide bond between ECL2 and TM3, a D[E]RY motif at the interface of TM3 and ICL2, a hydrogen bond (or ionic lock) between D[E]RY motif in TM3 and D/E residue in TM6, a CWxP motif in TM6, and a

NPxxY motif in TM7 219,220.

Even though most solved GPCR structures have a similar barrel conformation, the seven

α-helices of GPCRs tend to have different tilts and rotations to form different binding pockets, sizes, ECL shapes, electrostatic forces of bound ligands 202. The loops and transmembrane regions are flexible and display different conformations at active and inactive states. For example, the superposition of active and inactive A2AR structures in

Figure 1.11 (B) suggests that the transmembrane helix has a propensity to bend upon activation. In general, the distinctions of GPCR structures are noticeable on the extracellular domain (ECD) half side (including ECL and extracellular N-terminus), intracellular C-terminus, and intracellular loop three (ICL3). These distinctions are not only identified across different GPCR families but also within the same GPCR family 221.

The ECL and ECD regions show pronounced diversities for their secondary structures and disulphide crosslinking patterns, which are evolved for different accessibilities to different ligands 221,222. On the contrary, the ICL-TM half side of GPCR is more conserved than ECL-TM half (except for ICL3). The ICL and ICD regions interact with downstream effectors, such as heterotrimeric G-protein, β-arrestin or other downstream signalling messengers 151,222.

Except for several ligand-free GPCR structures, 223-227, most solved GPCR structures are in complex with diffusible ligands of antagonists, agonists or inverse agonists which are important to stabilise the GPCR. The inactive state of the receptor is considered to be more stable and less conformationally heterogeneous than receptor at active state 154.

Therefore GPCR structures are usually solved at the inactive states with antagonists

149,150 binding first, followed by the structures at agonist-bound active states, such as CB1 ,

228-231 207,232,233 A2AR , and β2AR . The structures of both states provide a snapshot of the conformational changes between inactive form and active form.

Figure 1.11: GPCR architecture.(A) The A2AR crystal structure (PDB ID: 4EIY). It depicts conserved GPCR features of extracellular domain (ECD), transmembrane domains (TM1 - TM7), intracellular domain (ICD), extracellular loops (ECL1 - ECL3), intracellular loops (ICL1 - ICL3), and sequences of D[E]RY, CWxP, and NPxxY; (B)

The ECL-TM half side of A2AR makes more dramatic movement than ICL-TM half side from inactive state (red, PDB ID: 3EML) to active state (grey, PDB ID: 3QAK). 27

The structural study of GPCR provides a solid base to understand GPCR physiological mechanisms and behaviours. The information about the subtle differences between these closely resembled GPCRs is critical. It helps to understand ligand recognition, receptor activation, allosteric modulation and dimerisation of GPCRs. The new GPCR structures propel the pharmacological developments for the GPCR-related diseases and reveal the overall signalling transduction mechanisms. Class A GPCRs are typically activated by a small-molecule ligand, while class B GPCRs usually bind with larger peptide hormones, which are used for treating diabetes, bone diseases and obesity. It is imperative to have more structure information on non-rhodopsin-like GPCR family. To understand the

GPCR signalling pathway at cellular level, it is also important to determine the complex structure of GPCR coupling with intracellular effectors (such as G-protein and β-arrestin).

There are only several complex structures available 234-243. Additionally, the structural determination of GPCR dimer or oligomer is also important for understanding GPCR signaling transduction mechanism and functions. Last but not least, the stabilising ligand of GPCRs helps crystallisation, GPCR deorphanisation, and drug discovery. As a result, the development of ligand identification methods towards GPCR is an important goal for the academic society and pharmaceutical industry.

1.2.3 Two neurological disease-related GPCRs – A2AR and α2AR

In the human body, the purine nucleoside of adenosine is an important modulator that maintains physiological functions in heart 244,245, kidney 246,247, lung 248, bone marrow 249, liver 250,251, immune system 252-254, and nervous system 255. Four types of adenosine

GPCRs are discovered in the mammalians, which are receptors of A1, A2A, A2B and A3

256. These receptors use adenosine as preferred endogenous agonist. The activation of adenosine receptors will transduce the signal through different families of G-proteins (Gs,

Gi and Golf). The classification of A1, A2 and A3 is based on the effect of inhibiting (Gi) or stimulating (Gs) the adenylyl cyclase (AC) release. The A1 and A3 receptors are usually 28

coupled with Gi, while A2 is coupled with Gs and Golf. For secondary messengers (Figure

1.9), both adenosine A2A receptor (A2AR) and adenosine A2B receptor (A2BR) stimulate the cAMP formation and mobilisation of intracellular Ca2+ 257. Among these four adenosine receptors, A2AR is the most well-studied receptor, which is widely distributed across blood vessels, smooth muscle, endothelium, lymphatics, sympathetic and parasympathetic nervous systems 258. It involves in many physiological and pathophysiological conditions. For example, it is responsible for immune-mediated inflammation diseases 244,248,259-268, Parkinson’s disease 23,24, epilepsies, sleep disorders, pain, drug addictions 269-272, and neuroprotections from ischemia, stroke, post-traumatic

247,251,255,273-284 brain injury, and excitotoxicity . The A2AR knockout mice show slower pain stimuli but increased platelet aggregation, heart rate and blood pressure 285. It indicates that A2AR is important in regulating glutamate and dopamine release in the brain, and responsible for increasing the blood flow to the myocardium by causing the vasodilation of coronary arteries 286. Some epidemiological evidences and behaviour studies also support a correlation between A2AR and neurodegenerative disease of PD.

For example, theophylline and caffeine are well-known A2AR antagonists. The tea and coffee drinkers show a lower risk of PD 287-290. Additionally, alcoholism is reported to

291 increase the risk of PD . While A2AR level shows a correlation with alcohol drinking behaviour 292.

A2AR is co-expressed with dopamine D2 receptor in striatopallidal neurons or basal ganglia where these two receptors form a heterodimeric complex and functionally oppose

293,294 to each other . For instance, the stimulation of A2AR activity will induce the decrease of D2 receptor affinity to dopamine. The primary pathology of PD is known to be the loss of nigrostriatal dopamine and lead to the reduction of dopamine D2 receptor activity

295,296 . Hence, A2AR selective antagonist of 8-[(E)-2-(3,4-dimethoxyphenyl)ethenyl]-1,3- diethyl-7-methyl-purine-2,6-dione (also known as KW-6002 or Istradefylline) 29

inactivating the A2AR activity emerges as a potential non-dopaminergic therapy for PD

295,297-300 . An increasing numbers of A2AR antagonists are also tested on PD animal models and even progressed into clinical phase II trials 287,301.

Another GPCR related to neurologic disorder is adrenergic receptors, which include α1,

302-304 α2, β1 and β2 adrenergic receptors . Based on different human chromosomal localisations, α2 adrenergic receptor subdivides into α2A, α2B, and α2C adrenergic receptors, which are mostly associated with Gi heterotrimeric G-protein (G-protein signalling

305-308 pathway) or phosphorylated by GRK2 (β-arrestin signalling pathway) . The α2A adrenergic receptor (α2AR) modulates the signalling transduction via cAMP and synaptic neurotransmitters (such as catecholamines: dopamine, noradrenaline/norepinephrine and

309-312 epinephrine/adrenaline) in central and peripheral nervous systems . α2AR is important for the proliferation of intestinal epithelial cells 313,314, potassium channel activation 315, blood pressure 316, and some cognitive behaviour disorders (such as attention-deficit, impaired memory, hyperactivity, depression, stress, analgesia and anxiety) 309,317-324. In conclusion, these two GPCRs are therapeutically relevant to inflammation, psychiatric disorders and neurodegenerative diseases (Huntington’s disease, Alzheimer’s disease, Parkinson’s disease, multiple sclerosis) 325-327 (Figure 1.12).

Figure 1.12: A2AR and α2AR distributions and physiological effects in the human body.

The common implications of both A2AR and α2AR are indicated in the centre.

1.3 Phospholipid environment for GPCR structure and function

The natural cell membrane is a continuous lamellar bilayer, which serves as a protective barrier, regulates molecular trafficking, performs secretion, endocytosis, and signal transduction. It is composed of phospholipids, glycolipids, cholesterols, peripheral membrane proteins and transmembrane proteins. The traditional cell membrane model is called a fluid mosaic model 328. This model is supported by the lateral diffusions of MPs within the cell membrane. Phospholipid is an amphipathic molecule with a hydrophilic polar head and a hydrophobic acyl chain (Figure 1.13). There are four major classes and one minor class of phospholipids in mammalian cell membrane, i.e. phosphatidylcholine

(PC), phosphatidylethanolamine (PE), sphingomyelin (SPH), phosphatidylserine (PS), and phosphatidylinositol (PI, an important signalling molecule such as PIP2 and PIP3)

329,330. 31

Figure 1.13: Structures and shapes of different phospholipids and cholesterol.CxHy and

CnHm correspond to acyl chains. (A) Cholesterol and common phospholipids in the cell membrane; (B) Positive curvature formed by inverted conical lipids (such as PE and PS);

(C) Lamellar bilayer formed by cylindrical lipids (such as PC); (D) Negative curvature formed by conical lipids (such as oleic acid and cholesterol hemisuccinate); (E) Lamellar bilayer formed by inverted conical lipids and conical lipids.

The phospholipids are asymmetrically distributed across the membrane bilayer, where

PC and SPH are located in the outer leaflet, while PE, PS and PI in the inner leaflet 331.

The PS and PI are anionic lipids and result in a net negative charge on the cytoplasmic face of cell membrane 332. On the contrary, PC and SPH are zwitterionic in charge.

Cholesterols are also dispersed throughout the phospholipid bilayers. The cholesterols maintain the integrity, fluidity, rigidity and function of cell membrane and play an important role as lipid rafts and signalling molecules 333.

Besides the low cellular expression and intrinsically labile problems, GPCRs are also difficult to solubilise after removing from the cell membrane. GPCR is usually solubilised by surface-active surfactant or detergent 334. Cells are physically disrupted by hypotonic lysis buffer, followed by repeated dounce homogenisation. Then GPCRs are removed from their natural phospholipid bilayer by detergent solubilisation. To maintain the integrity of GPCR structures, mild detergent is used for solubilisation such as N-dodecyl-

335 β-D-maltoside (DDM), DM, or LMNG . Most GPCRs are labile during the solubilisation, and even slight disturbance may lead to structural misfolding, denaturation, aggregation, and sometimes degradation by proteases 217. Another major challenge for

GPCR purification is to maintain the receptor in a functionally active, soluble, monodisperse, and homogeneous state. Therefore, other than directly solubilising GPCRs in detergent micelle, several techniques to maintain GPCRs at a native-like environment such as bicelle, amphipol, liposome, nanodisc, and styrene-maleic acid co-polymer lipid particle (SMALP) have been developed. These techniques facilitate the structural and functional studies (Table 1.2).

Table 1.2: Comparison of micelle, bicelle, liposome, nanodisc and SMALP.

Description Advantages Disadvantages

 GPCR destabilisation;

 Need to screen expensive

detergents;

 Short storage period; A hydrophobic  Alter conformation and environment thermodynamic which cause Detergent  Convenient; applied for GPCR GPCR misfolding and micelle  Easy for use purification and function loss; crystallisation  Maintain the detergent

concentration above CMC in

experiment;

 Heterogeneous micelles in

size.

 Large;

 Difficult to control size and

lipid stoichiometry;

 Heterogeneous diameters (30

A sphere-shaped nm - 2.5 μm);

lipid vesicle  More stable than detergent;  Large-size liposomes cause Liposome containing one or  Non-toxic and flexible undesirable light scattering

more lipid bilayers. and complicate optical

spectroscopy

measurements;

 Preclude ligand access;

 Low solubility;

 Due to high curvature

surface, liposomes are prone

to aggregation when osmosis

suddenly changes.

A mixture of long  Lipid mixture recipe is  Native-like environment; chain and short limited in availability;  Small size; chain lipids with a  Need to maintain lipid molar Bicelle specific ratio. The  Can be used for NMR; ratio in experiments; short chain lipids  Can be used for  Temperature changes cause contribute high- crystallisation phase change. curvature.

 Good stability;  Larger than detergent  Preserve the GPCR A complex of micelle; oligomer; Amphipol amphipathic  Can’t resemble the lipid  Low critical aggregation polymer environment. concentration;

 Detergent-free

 Mimic native-like  Real cell membrane has an

phospholipid asymmetrical charge/lipid

A native-like environment; distribution;

phospholipid  Better stability and  Need to optimise the lipids: Nanodisc bilayer encircled by activity than in liposome MSP: GPCR ratio;

two scaffold  Detergent-free  Need to optimise lipid and

proteins (MSPs).  Improved GPCR stability; MSP species;

 Homogeneous particles  MSP (lipoprotein) has higher

with controlled sizes; electron density than SMA in

 Oligomerization can EM application;

happen;  MSP influences the

 Can be used for NMR, functional/structural studies

EM, SPR, cell-free from the lateral side of

expression, and other GPCRs;

techniques;

 Adjustable lipid

compositions.

 Ligand access from top

and bottom;

 Nanodisc structures are

already defined.

 Phospholipid  Magnesium concentration A complex with environment; interferes the SMALP SMALP styrene maleic acid  Can be used for EM; reconstitution; (lipodisq) co-polymer (SMA)  Homogeneity in particle  Lack of structural and lipid size; characterisation

 New technique without saposin-  Promising in high- A complex of widely applied yet; lipoprotein resolution cryo-EM saposin scaffold  The thermostability and nanoparticle studies on membrane protein and lipids suitable for applying on MS (Salipro) proteins is unknown

Among these models, the nanodisc is an advantageous, well-established and versatile tool

to maintains GPCR in a native-like phospholipid bilayer environment. Nanodisc not only

controls the physical size by varying the MSP sizes (macro-environment), but also allows 36

the adjustment of lipid types and stoichiometry (micro-environment) for specific protein.

Comparing to detergent micelle and liposome, nanodisc is a better and more powerful investigation tool for GPCR biochemical and biophysical studies.

1.3.1 Self-assembled nanodisc for GPCRs

MPs can be denatured or inactivated if it is extracted from the native cell membrane directly. In the 2000s, Sligar’s group first developed nanolipoprotein particle (NLP) or nanodisc technique in order to study the native architecture of a membrane protein of microsomal cytochrome P450 reductase in hepatic endoplasmic reticulum 336-339. The prototype of nanodisc is originated from human high-density lipoprotein (HDL) particle, which is formed by apolipoprotein A-1 or apoA-I 338,340,341. The HDLs are lipoprotein particles with variable sizes and rich in phospholipids, but poor in cholesterols 342-344. In human serum, the HDL particles involve in reverse cholesterol transport and highly relate to the risks of coronary artery diseases 345. ApoA-I is a 28 kDa protein with 243 amino acids, comprising 70% of HDL mass (Figure 1.14 (A)). ApoA-I is released from liver and intestine and enters into plasma. It consists of an N-terminal globular domain with 43 residues and an amphipathic helix with ten tandem repeating units (eight segments of 22- mer amino acid repeats and two segments of 11-mer repeats). Each repeat is punctuated by proline and glycine residues. The long repetitive helical structure in apoA-I is suggested to be evolved from internal gene duplication and function in lipid binding 346-

348. Furthermore, the lipid-free crystal structures of C-terminal apoA-I helices (PDB IDs:

1AV1, 2N5E and 3R2P), full-length apoA-I (PDB: 2A01), and computational studies suggest a double belt conformation of apoA-I 349-359. It is also found that the deletion of

N-terminal globular domain of apoA-I decreases the cholesterol contents in HDL but not change the apoA-I spherical structure (Figure 1.14 (B)). The removal of helix-1, 9 or 10 impedes the HDL functions 360,361. Based on the knowledge of apoA-I, Stephen Sligar and

his colleagues developed a genetically engineered membrane scaffold protein (MSP) with deleted apoA-I N-terminal domains 338,340,341.

Nanodisc, as its name suggested, is a discoidal self-assembled lipid bilayer and can be solubilised in detergent-free aqueous solution. The GPCR is encircled by two amphipathic helical MSP proteins (Figure 1.14 (B)). The interior side of MSP interacts and encircles the acyl tails of phospholipids and shields them from exterior aqueous solution. The MSP length determines the nanodisc circumference and could be adjusted by selective addition or truncation of MSP α-helix segment (Table 1.3). The adjustable property of MSP is necessary for GPCR oligomerisation. The average diameters of GPCR transmembrane domains are around 3 - 5 nm 341,362,363 and the sizes of nanodiscs range from 8 - 17 nm in diameter 364-366. Two popular MSP constructs for nanodisc reconstitutions are MSP1D1 with 9.5 - 9.7 nm in diameter, and MSP1E3D1 with 12.9 nm in diameter 367. The larger size of nanodisc has the potential to accommodate more GPCR transmembrane segments during the nanodisc reconstitution, such as rhodopsin and mGluR2 dimers in MSP1E3D1-encircled nanodiscs 368-370. The structure and compositions of nanodiscs has been successfully studied by gel filtration 371-373, small- angle X-ray scattering (SAXS) 373-376, atomic force microscopy (AFM) 371, radiolabeled lipid titration 367, molecular dynamics simulation 377, and ESI-MS 378.

Figure 1.14: The apoA-I and MSP structures.(A) The full-length apoA-I structure (PDB

ID: 2A01) displays a compact N-terminal four-helix bundle and two C-terminal α-helices;

(B) The truncated apoA-I structure (Δ1-54, Δ121-142) with lipid-binding acyl tail (PDB

ID: 2N5EB) forms a two antiparallel double belt. Schematic diagram of side-view (C) and top-view (D) of MSP1D1-encircled empty nanodisc.

The procedures for nanodisc reconstitution include (1) GPCRs overexpression in insect or mammalian cells, (2) solubilisation of cell membrane components in detergent, (3) purification of GPCR and MSP (ΔHis6) by immobilised metal ion affinity chromatography (IMAC), (4) incubation of membrane proteins (solubilised in detergent), phospholipids (solubilised in sodium cholate, SDS, DDM, or CHAPS), and MSPs in a stoichiometric-dependent manner, (5) initiation of nanodisc self-assembly by detergent removal. The detergent removal methods include dialysis or physical adsorption by hydrophobic adsorbents (Bio-Beads SM-2TM or Amberlite XAD hydrophobic beads) at optimal temperature, (6) protein and lipid aggregates removal by size exclusion chromatography (SEC) coupled to fast protein liquid chromatography (FPLC) system, (7) purification of GPCR containing nanodiscs by IMAC. Since hexahistidine (His6) tag is only present in GPCR, the empty nanodiscs will be removed in a wash step; (8) examination of final GPCR nanodisc yield and purity by NuPAGE gel electrophoresis

(Figure 1.15 (A)). The homogeneity of final GPCR nanodisc particles can be analysed by dynamic light scattering (DLS) (Figure 4.7). The DLS analyses the fluctuations in the intensity of scattered light arising from the Brownian motion and diffusion of the particles of interest 379. It is a straightforward, non-invasive and rapid technique for characterising the particle size (spherical diameter) distributions and particle size variations in small quantities.

Figure 1.15: Nanodisc reconstitution procedures and membrane protein structures solved in nanodiscs.(A) Nanodisc reconstitution procedures; (B) Examples of solved MP structures in nanodisc: TRVP1 (PDB ID: 5IRZ) by cryo-EM, OmpX (PDB ID: 2M06) by

NMR, and bR (PDB ID: 5BR2) by X-ray crystallography. 41

Even though the mechanism by which GPCRs are incorporated into self-assembled nanodisc remains elusive, the nanodisc has been successfully applied to many GPCR studies. It provides a native-like environment for proper GPCR conformation, function, and ligand interaction 380-382. The nanodisc also helps with the study of GPCR oligomerisation by adjusting MSP length and GPCR/lipid/MSP stoichiometry. Certain

GPCRs exist as dimers or higher-order oligomers in living cells, such as rhodopsin, β1AR,

383-385 380,386 369 β2AR, and GABAB . The stable dimers of β2AR , rhodopsin and NT1 receptor 387 have been reconstituted into nanodiscs, even though it is found that the oligomeric state of receptors interacted with G-proteins less efﬁciently than monomers.

For example, the external stimuli of photons trigger the isomerisation of 11-cis-retinal

(chromophore) to all-trans-retinal in rhodopsin. This retinal isomerisation induces a conformational change of rhodopsin and activates the downstream heterotrimeric G- protein (Gt). The rhodopsin monomer and dimer are reconstituted into nanodisc separately.

The stoichiometry for reconstituteing rhodospin dimer is 1:1:90 (rhodopsin: MSP1E3D1: phospholipids) with a diameter of 12.8 nm 366,369,388. The studies show that the rhodopsin in nanodisc (either rhodopsin monomer or dimer) is phosphorylated by GRK1 and bound to β-arrestin in a 1: 1 ratio, thus one rhodopsin is efficient for β-arrestin binding and signal transduction 368,389-391.

What is more, the lipid compositions of nanodisc can also be adjusted. Lipid content is known to be critical for GPCR functions 392-396. The synthetic, natural and lipid extract mixture are available and applied to membrane protein activity studies. For example, the activities of MsbA (ABC transporter from E. coli) and proteorhodopsin (microbial proton pump) showed a correlation with lipid compositions in nanodisc. The MsbA in nanodiscs with different lipids (DMPC, DMPG, DLPC or DOPC) result in different lipid-dependant

ATPase activities 397. Likewise, the proteorhodopsins in nanodiscs with POPC or DMPC show different photoactivities 366. The nanodisc containing long and unsaturated POPC 42

contributes to faster photocycle than nanodisc containing shorter and saturated DMPC. It is assumed that the long and unsaturated POPC well encloses the hydrophobic transmembrane domain of proteorhodopsin than DMPC. The other studies on CYP3A4-

CPR complex, respiratory complex II, and bacteriorhodopsin also show a lipid-dependant activities. For example, the CYP3A4-CPR complex in nanodisc containing anionic lipids shows an accelerated proton-coupled electron transfer rate 398-401. The anionic cardiolipin lipid is important for respiratory complex II activity 402, and the anionic DOPG/DMPG mixutre important for bacteriorhodopsin activity.

Additionally, some studies demonstrate that nanodisc can preserve MP in a native state and assist MP refolding. For instance, the MPs can be solubilised and denatured in ionic detergent-like SDS buffer. Intriguingly, the denatured MPs are observed to undergo refolding in self-assembled nanodisc 334,403,404. Similarly, both folding and native unfolding intermediates of bacteriorhodopsin (bR) are observed in DMPC-containing nanodiscs by single-molecule atomic force microscopy (AFM) study, which suggests nanodisc can promote protein refolding process 405.

Last but not least, nanodisc has a better ligand accessibility than detergent micelle and liposome in native-like environment. The scintillation proximity assay (SPA) is a sensitive technique to detect the presence of proximal radioligand to the protein target attached on scintillant beads within a distance of 8 µm 406,407. The combination of nanodisc and SPA technique is used to study the ligand binding affinity to the membrane proteins of LeuT and FhuA. The SPA results demonstrate that the LeuT and FhuA have higher bioactivities and ligand affinities in nanodiscs than that in detergent micelle 406,408.

These studies emphasise the advantages of incorporating membrane proteins or GPCRs in native-like and detergent-free nanodisc. Nanodisc is an stable, monodisperse and homogeneous biological membrane that contributes to minimal precipitation, degradation, aggregation, and structural misfolding under the experimental procedures (including buffer wash, buffer exchange, and long-time incubation with ligands) 409-414. Nanodisc also does not perturb the ligand accessibility to MP as that in detergent and liposome.

1.3.2 Nanodisc reconstitution by different lipids and MSPs

The MSP/lipid compositions of nanodiscs can be investigated by LC-MS/MS and 1H

NMR 415. The hydrophobic side of MSP makes contact with the acyl chain of phospholipid, and the hydrophilic side of MSP orients outward to the solvent (Figure

1.14). The number of lipids per nanodisc (NL) and the number of amino acids in MSP

2 (NMSP) has a relationship of NLS = (0.423NMSP - 9.75) , where S represents the mean surface area (measured in Å2) per lipid used to form the nanodisc. The quadratic relationship between the number of lipid molecules per nanodisc and the MSP length ensure the formation of ﬂat two-dimensional nanodisc particle. The unsaturated POPC has a larger surface area than saturated DPPC, thus POPC containing nanodisc is more loosely packed than DPPC containing nanodiscs 338,365. In MSP1-encircled nanodisc, it is consisted of 130 POPC phospholipid molecules compared to 160 DPPC phospholipid molecules 365. The MSP1 and MSP2 are the first two MSP constructs introduced by

Sligar’s lab 338 and some other MSPs are shown in Table 1.3.

Table 1.3: Different membrane scaffold proteins. The MSP diameters are determined by size exclusion chromatography and small angle X-ray scattering (SAXS).

Molecular Scaffold Diameter Composition Weight variant (Å) (kDa)

MSP1 FX-H1-H2-H3-H4-H5-H6-H7-H8-H9-H10 24.6 97 - 98

FX-H1-H2-H3-H4-H5-H6-H7-H8-H9-H10-GT-H1-H2- MSP2 48 95 H3-H4-H5-H6-H7-H8-H9-H10

MSP1E1 FX-H1-H2-H3-H4-H4-H5-H6-H7-H8-H9-H10 27.5 104 - 106

MSP1E2 FX-H1-H2-H3-H4-H5-H4-H5-H6-H7-H8-H9-H10 30 111 - 119

MSP1E3 FX-H1-H2-H3-H4-H5-H6-H4-H5-H6-H7-H8-H9-H10 32.5 121 - 129

MSP1D1 TEV-H1D (1-11)-H2-H3-H4-H5-H6-H7-H8-H9-H10 24.6 95 - 97

MSP1D2 TEV-H2-H3-H4-H5-H6-H7-H8-H9-H10 N/A N/A

MSP2N1 TEV-H1D (1-11)-H2-H3-H4-H5-H6-H7-H8-H9-H10- N/A N/A GT-H1D (1-11)-H2-H3-H4-H5-H6-H7-H8-H9-H10

MSP2N2 TEV-H1D (1-11)-H2-H3-H4-H5-H6-H7-H8-H9-H10 45.5 150 - 165 -GT-H2-H3-H4-H5-H6-H7-H8-H9-H10

MSP2N3 TEV-H1D (1-11)-H2-H3-H4-H5-H6-H7-H8-H9-H10- 46 152 - 170 GT-H1D (1-17)-H2-H3-H4-H5-H6-H7-H8-H9-H10 MSP1E3 TEV-H1D (1-11)-H2-H3-H4-H5-H6-H4-H5-H6-H7- 30 128 D1 H8-H9-H10

Note: FX denotes the N-terminal sequence of MGHHHHHHIEGR; TEV denotes the N- terminal sequence of MGHHHHHHHDYDIPTTENLYFQG; H1D (1-11) denotes the helix I amino acid followed by deleted amino acid sequence in brackets.

The phospholipids used for nanodisc reconstitution include zwitterionic phospholipids

(e.g. POPC, DMPC, DPPC, and DOPE), anionic phospholipids (e.g. POPS, DOPC,

POPG, DOPG, DMPA, or DMPG, POPE), a multicomponent phospholipid mixture, natural soybean lipid extract, E. coli lipid extract, chicken egg PC lipid extract, and porcine polar brain lipid extract (Table 1.4). The considerations for phospholipid selection include natural lipid environment for particular MP, phase transition temperature, and lipid shapes. Different head groups and acyl chains determine the corresponding geometrical shapes (Figure 1.13). The ubiquitous PC lipids with large

(methylated) headgroups form a cylindrical shape and self-assemble into a lamellar bilayer. This feature is similar to the cell membrane structure and maintain nanodisc sample optically transparent. However the non-bilayer lipids with much larger or smaller headgroups than acyl chains form conical or inverted conical shapes (Figure 1.13). The non-bilayer lipids are essential for junction formation, activities, curvature, and at a specific region 416,417. The lipid compositions not only influence membrane fluidity and surface charge, but also GPCR function and structure. Cell membranes from different species have different lipid compositions that are essential for proper protein structures and functions. For example, a phospholipid mixture of PE/PG/PC/CL is important for maintaining bovine cytochrome c oxidase structure, while only PC is identified to be important for paracoccus denitrificans cytochrome c oxidase structure 418. The presence of negatively charged lipids in zwitterionic lipids is found to be important for nanodisc stability and up to 100% negatively charged lipids have be used in experiments 419-421.

The nanodisc possesses a similar phenomenon of phase transition as native membranes.

The lipid thermotropic phases describe the lipid dynamics, which include ordered/disordered phases and fluid/gel/solid phases. For example, phospholipids are organised into ordered gel phase at low temperature, but disordered fluid phase or liquid crystalline phase at higher temperature. It is found that the phase transition of nanodisc 46

from gel phase to liquid crystalline phase is similar to that of native cell membrane 376.

Notably, liquid crystalline phase is important for nanodisc self-assembly. The nanodisc

409 reconstitution is essential to perform above lipid phase transition temperature or Tm .

Table 1.4: Properties of some phospholipids used in nanodisc reconstitution.

T Name Charge m Shape Applications (ºC)

POPC -2.6 cylinder 366,374,422-436 364,366,405,422,432,434,437- DMPC 23 cylinder 439 zwitterionic DPPC 41 cylinder 364,440 inverted DOPE -16 364,404,422,436 cone inverted POPS 14 431,432,441-443 cone DOPC -17 cylinder 422 POPG -4 cylinder 364,374,434 DOPG -18 cylinder 444 anionic inverted DMPA 52 445,446 cone DMPG 23 cylinder 364,422,432,434,444,446-448 inverted POPE 25 436,443 cone Soybean polar lipid 449-452 extract E. coli lipid extract 422,452-456 natural Chicken egg PC N/A N/A extract 452,457,458 phospholipids extracts Porcine polar brain 459 lipid extract

The molar ratio of GPCRs: MSPs: lipids is key to generating reproducible, stable and homogenous nanodisc (Figure 1.15 (A)). A high percentage of lipids in MSP/lipid/GPCR mixture contributes to an increased lipid aggregation during nanodisc self-assembly

(Figure 4.4). Therefore, it is necessary to optimise the molar ratio prior to large-scale nanodisc reconstitution 338,371. Besides that, if detergent removes too fast during the incubation, it also contributes to increased amount of GPCR aggregation.

1.3.3 Recent advances of nanodisc technique

A wide range of MPs have been investigated in a nanodisc model with different topologies and oligomeric states, including cytochrome P450s 336,339,460, cytochrome P450 reductase (CPY3A4) 337,400, monomeric and trimetric bacteriorhodopsin (bR) 371,374,405,461-

463, OmpA 464, OmpX 462, TRPV1 450,451, SNARE 413,465, and GPCRs of NK1R 362, µ- opioid receptor 459, NTS1 362,369,380,415,459,466, monomeric and dimeric bovine rhodopsin receptor 340,369,389,390,467, monomeric and dimeric mGluR2 370, PTH1R 468, mouse olfactory

469 470 362,380,466 362 receptor , endothelin A receptor, endothelin B receptor , β2AR , DRD1 , and CCR5 471. Besides the incorporation of single MP species in nanodisc, an MP nanodisc library (MPNL) from bacteria 472, yeast 473 and mammalian cells 474 are also incorporated into nanodisc directly. Recent applications on nanodisc have been involved in many fields such as cell-free MP expression 422,470,475,476, MS (will be discussed in section 1.4.4), EM 413,450,451,465,477-479, AFM 371,405, surface plasmon resonance (SPR)

397,441,480,481, NMR 445,482-486. The most recent advances are direct crystallisations of integral membrane proteins in nanodiscs 444,487.

The structure determination for membrane protein remains to be challenging. Nanodisc makes it is possible to observe the MP structure directly by negative staining EM (2D image) and single particle cryo-EM (3D image) 477,488-490. The EM is known to have an observation difficulty for asymmetrical proteins with MW smaller than 64 kDa in theory

and smaller than 100 – 150 kDa in practice 477,491-495. Owing to the resolution revolution of cryo-EM in the past decade, some membrane proteins were resolved by cryo-EM after reconstitution into nanodisc, such as transient receptor potential cation channel subfamily

V member 1 (TRPV1) 450,451, soluble NSF attachment protein receptors (SNARE)

413,465,478, drug discharge outer membrane protein OprM 488, rabbit ryanodine receptor

489 (RyR1), pore-forming TcdA1 toxin subunit , low resolution of apo mouse 5-HT3 GPCR receptor 479. Cryo-EM has been used to determine the atomic structures of GPCR complexes in detergents, such as A2AR (in complex with NECA agonist, mini-Gs, Gβγ subunits and nanobody Nb35) 240, class B GPCR of GLP-1 receptor (in complex with

237,241 agonists, Gs, Gβγ subunits and nanobody Nb35) , class B GPCR of calcitonin

234 receptor (in complex with peptide ligand Gs, Gβγ subunits) , and β2AR complexes at low resolutions 242,243. However, the non-routine sample preparation, low-resolution, slow data collection, large data size, radiation damage and image motion induced by electron beam limit the EM application on structural solving.

Other recent advances include the cell-free expression system used for membrane protein structural and functional studies. The MPs are expressed and pre-assembled into nanodiscs or co-expressed with nanodisc components 496-500. In comparison with other membrane mimetics (detergents micelle, bicelle, and liposome), nanodisc enhances the yield, solubility and stability of membrane proteins. The cell-free expression in nanodisc decreases the potential cytotoxicity and low MP yield. It also avoids the use of detergent for MP solubilisation. The combination of cell-free expression system and nanodisc has been applied to receptor tyrosine-protein kinase ErbB3, the voltage-sensing channel of

KvAP, bR 497, bacterial proton pump proteorhodopsin, multidrug resistance transporters of RmrE and SugE, phospho-MurNAc-pentapeptide translocase MraY 422, and human endothelin A and B receptors 470.

The SPR technique is also applied to ligand identification against MPs in nanodiscs

441,480,481 . The SPR is used to measure the association/dissociation constant (Kon/Koff) and binding affinity (Kd) between ligand and protein target. The binding partners (such as protein target) is immobilised on the surface of SPR chip and the solubilised ligands in solution flow towards the immobilised protein target. The binding event can be inferred by determining the refractive index change on the surface of SPR chip. In practice, SPR has already been used to measure the Kd of endothelin A and B in nanodisc against b-ET-

470 1 and b-ET-3 peptide ligands . Recently A2AR-GFP nanodisc was also studied by SPR, which showed a higher binding activity (i.e. 80% in nanodisc compared to 52% in

501 detergent), faster association rate (Kon), and better stability than in detergent . These results suggest that nanodisc maintains the GPCR in a stable and active state.

In conclusion, the benefits of nanodisc are: (1) the increase of protein solubility and stability; (2) protein at native-like state; (3) more biological relevance; (4) control the size and composition precisely and better homogeneity; (5) the extracellular and cytoplasmic sides of nanodisc remain to be unrestrictedly accessible to surrounding environment and ligand binding. These advantages are essential for the physiological, biochemical, structural, and ligand identification analysis of membrane protein (including GPCRs).

1.4 Mass spectrometry (MS)

Ligand can bind to a particular protein target as a "lock and key" model 502. The ligand molecules vary in sizes and classes, such as atoms, compounds, ions, nucleic acids, carbohydrates, vitamins or other proteins. A specific ligand binding to the active site of proteins can induce protein conformational change and regulate protein bioactivity 503.

Various biophysical and biochemical approaches have been developed to study the protein-ligand interactions, including high-throughput screening (HTS), structure-

activity relationship (SAR), structure-based design (SBD), fragment-based drug design

(FBDD), nuclear magnetic resonance (NMR) and mass spectrometry (MS) (Table 1.5).

The development of an automation system for HTS makes it possible to screen a large number of compounds (more than 100,000 compounds) against a biological target in one day 504,505. The ligands can be initially discovered by HTS, and the specific biological functions (such as synaptic transmission, enzyme catalysis, signal transduction and immune response) can be elucidated later on 506. However, there are several limitations for HTS, including fluorescence quenching by inner-filter effect in the radiochemical/fluorescence-based screening system, unknown fluorescent compounds, buffer composition influence, large compound library required, detection sensitivity limitation, mass resolution accuracy and quality of the sample 507-509. Therefore, it is recommended to conduct multiple HTS assays to validate the results and overcome the limitations 510.

The structure-activity relationship (SAR) statistically analyses the relationship between three-dimensional (3D) chemical structures and their biological activities. It can be achieved by principal component analysis (PCA), discriminant analysis (DA), or correspondence factor analysis (CFA). The analysis of SAR cluster the initial HTS hits with similar bioactivity and bioaffinity, and determine the common chemical group responsible for the biological activity 511-513. Thus, SAR improves the potency and selectivity of drug design. The structure-based design (SBD) is in complementary to HTS,

FBDD and SAR techniques. SBD relies on the knowledge on protein three-dimensional structure and it can work in parallel with other ligand identification methods. SBD includes virtual molecular docking by DOCK, Schrödinger or ZINC15 514-518 and molecular dynamics (MD) simulation 519,520.

Table 1.5: Biophysical (NMR, MS, SPR, thermal shift assay and X-ray crystallography),

biochemical (HTS, FBDD, and ITC) and in silico SAR modelling methods for protein-

ligand interaction studies.

Methods Advantages Disadvantages

 Poor chemical space coverage and very diverse compounds;  Fast and immediate (automation system  Extensive HTS compound library; can screen 105 compounds with various  Validation of positive target-ligand MWs in one day); interaction from false positives hits;  Not limited to target types (protein, cells  Expensive automating system; or microorganisms) and compound  Protocol optimisation for reproducible information; HTS data;  Highly sensitive detection methods;  Pure and consistent samples (target and  Biochemical reaction volume and compounds) each time; sample consumption can be minimized  Development of analysis software and (e.g. 1,536-well microtitre plate); big data size;  Identification of high-affinity  Fluorescence-based HTS assay with compounds (Kd in µM range) better sensitivity than radiochemical- based assay, but increased interference

 Only modest to low affinity hits (100  Smaller compound library (thousands) µM – 1 mM) are identified; and smaller molecular mass (< 300 Da);  Subsequent structural optimisation to  Better coverage for physicochemical increase the binding affinity (medicinal FBDD properties and chemical space; chemistry);  Major scaffold structure is acquired from  Due to low affinity, medium to large the common fragment binding quantity of target and fragment used for structures. HTS and further confirmation

 In complementary to HTS;  Predictions rely on computational  Conclude a large set of chemical hits and model and software programming; classify into a smaller group; SAR  Structurally unrelated compound may  Predictions based on hits information; be ruled out from predictions, so  Identification of important functional different classes of the compounds with groups on the chemicals

diverse physicochemical properties are required;  False correlation may arise from biased computational analysis

 Some proteins are difficult and time-  Highly accurate; consuming for crystallisation; SBD based on  Determination of structure and  No dynamic information from one X-ray interaction with ligands at high atomic crystal structure; crystallography resolution;  Need enough crystal sample;  No molecular weight limit.  Expensive and low speed

Target-based  Molecular mass limit (<30, 000 Da); nuclear  Detect biomolecules configuration under  Large amount of sample required; magnetic normal physiological conditions;  Difficult and time-consuming to analyse resonance  Observe dynamic information; the spectrum; (NMR)  Non-destructive method.  Difficult for insoluble protein

spectroscopy

 Some buffer composition can influence  Monitor bindings in real-time; the kinetic analysis; Surface  Radio label-free (tagged);  Need stable sample for immobilisation plasmon  Very sensitive; (no conformational change or disruption resonance  Reusable SPR chip; during experiments) (SPR)  Automotive injection and measurement.  Heterogeneous surface condition can influence the binding and results

 Fast and direct measurement (label-free)

without immobilisation;  Large quantity and high concentration;  Determined thermodynamic binding Isothermal  Non-specific binding also generate profile (binding enthalpy); titration small binding enthalpy;  Reaction in native buffer system; calorimetry  Slow and not high throughput;  Non-destructive; (ITC)  Not suitable for low-soluble and  Result acquired in single experiment, unstable samples; with increased ligand concentration;  No MW limit

 Real-time monitoring fluorescent  Need reactive functional group in the Radiochemical/ signals; biological sample;

fluorescence-  Simple machine setup (only need  Interferences from storage buffer based assay suitable excitation and emission (DMSO) and fluorescent/coloured (thermal shift wavelengths); ligands; assay)  Most dyes are commercially available;  Need to avoid light;  Good sensitivity;  Safety for radioactive substances;  Can be high throughput (384/1536-well  Fitting can be biased (linear relationship plate) in certain range).

 Fast;  Little structure information; ESI-MS  Able to detect small amount of sample;  Destructive method;  Able to observe protein in native state  Difficult for insoluble proteins.

1.4.1 Native MS and affinity LC/MS

MS is a primary technique for nucleic acid characterisation 521, protein analysis 522 and

ligand identification 523. MS analysis is usually complementary to the HTS and other

biophysical methods such as NMR, IR, Raman spectroscopy and analytical

ultracentrifugation. It can be used to study the covalent or non-covalent interactions

between protein and ligands. MS instruments are different in detection speed, resolution,

sensitivity and mass range. Most MS instruments are comprised of five principal

components: sample inlet, ion source, mass analyser, detector and data system (Figure

1.16). The sample solution is introduced into the MS and evaporated into a gaseous phase

through the inlet system either at atmospheric pressure or under vacuum. The gaseous

droplets are ionised into cations by losing electrons (Figure 1.20 (A)). Different ions are

separated by accelerating and focusing into a beam. Then this beam of ions is bent by an

external magnetic field and separated based on the mass-to-charge ratio (m/z) in a mass

analyser. Finally, they can be propelled towards the detector for measurement and

analysing on computers. Over the past years, the improvement of ionisation and mass

analyser techniques were the major driving force for MS development.

The assembly and dissociation of non-covalent protein-ligand complex is important for accomplishing protein functions in the human body. The non-covalent binding complex can be preserved by using a soft ESI source and transferred into various mass analysers at native state. These mass analysers seperate ions based on different principles, including quadruple electric fields (quadrupole MS), time-of-flight (Q-TOF MS), or selective ejection of ions from a three-dimensional trapping field (ion trap or Fourier transform ion cyclotron resonance MS) (Figure 1.16) 524. The conventional quadrupole analyser is the simplest instrument for mass-to-charge (m/z) determination. There are more analyser types with better resolution power are available nowadays, such as magnetic sector, ion trap, time-of-flight (ToF) and Fourier transformed ion cyclotron (FT-ICR). Nowadays,

FT-ICR MS, Q-TOF MS, and Orbitrap MS are the most commonly used mass analyser.

Figure 1.16: General features and schematics of principal components in different MS.

(A) five common components for most MS instruments (B) 12Tesla FT-ICR MS, (C)

MaXis Q-TOF MS, and (D) Orbitrap MS. 56

There are two complementary MS-based ligand identification methods that are direct ligand detection by native ESI MS and indirect ligand detection by affinity LC/MS

(Figure 1.17).

Figure 1.17: The schematic diagram of direct native ESI MS and ultrafiltration-based affinity LC/MS workflow.

The advantage of using direct MS detection method is that it can observe the binding event between protein and ligands straightforward. The native MS directly acquire the information about protein topology, protein dynamic information and non-covalent ligand-protein binding information, such as directly analysing the non-covalent binding complex by ESI-FT-ICR MS 525. The native ESI MS determines the protein mass-to- charge ratio (m/z) rapidly and accurately. The mass and charge information can be calculated from the isotopic peaks by formulae: z = 1/ΔM, z is the protein charge, and

ΔM is the difference between two isotopic m/z peaks (Figure 1.18). Subsequently, the molecular mass can be further determined by using average m/z multiplied by the charge.

Figure 1.18: An example of MSP1D1 mass spectrum resolved by native MS (ESI-FT-ICR

MS and ESI-Q-ToF MS).(A) MSP1D1 isotopic peaks at charge state of +10 (ESI-FT-ICR

MS); (B) at charge state of +20 (ESI-Q-ToF MS). The insert figures are the whole

MSP1D1 protein spectrum with different charge states. 58

For instrument optimisation, it requires to tune the acceleration voltage and gas pressure to give an optimal ion desolvation and preserve the non-covalent complex during transmission 526. The protein samples for native ESI MS do not need any treatment or immobilisation procedures. The ESI-FT-ICR MS is a high-resolution MS instrument that can analyse large biomolecules including proteins, amino acids, and saccharides 527-529.

Recently, some challenging molecules larger than 150 kDa are also solved by ESI-FT-

ICR MS 378,530. The MS allows the detection of protein-ligand complex in nanomolar or micromolar concentration. The non-covalent binding interactions, which include ionic bonds, hydrogen bonds, dipole-dipole bonds and van der Waals forces, are usually reversible and intrinsically fragile. It is common in biological processes such as transportation pathways, metabolism, and cell signalling pathway. Therefore the screening of reversible non-covalent binding ligand, which affects the protein bioactivity, is essential for drug development 503. The native ESI-MS is able preserve the non-covalent interaction within the instrument and detects the binding event. This technique has been used for studying the membrane proteins (including GPCRs) in recent years.

However, the major limitation for native ESI-MS is the protein sample preparation and buffer selection, which is the key to high-quality mass spectra. The stability and solubility of the protein are also critical. As a result, most native ESI MS screenings still focus on soluble proteins since the sample homogeneity, solubility, and stability are better than

MPs. Besides that, some protein storage buffers consist of large amount of non-volatile salts (such as NaCl, KCl, and MgCl2), anti-frozen agents, and surfactants. These reagents are non-volatile and incompatible with electrospray ionisation (ESI) technique. Thus the protein storage buffers are normally exchanged to a volatile ammonium acetate or ammonium bicarbonate at physiological pH for native ESI MS. However, these procedures highly rely on the stability and solubility of proteins. For unstable proteins, they may undergo partial or complete denaturation, unfolding, oligomerisation and 59

aggregation during buffer exchange. Therefore, it is not readily applicable to hydrophobic and unstable proteins. Moreover, native MS is still a low throughput screening method

(several minutes per sample for data acquisition) until the invention of Nanomate®

(Triversa). Nanomate® can be coupled with ESI MS and allowed for medium throughput capability. It allows the screening of 384 compounds or complex daily and consumption of only 50 - 200 pmol of proteins.

Instead of measuring the protein-ligand complex by native ESI-MS, another MS-based ligand identification method by indirect affinity LC/MS is available (Figure 1.17). The affinity LC/MS provides a method to detect the binding ligands in a stable, sensitive and high throughput format. It has been successfully applied to ligand screening against various soluble protein targets 531,532.

Figure 1.19: Identification of protein-ligand binding by affinity LC/MS from extracted ion chromatography (EIC).The binding index (BI) is calculated by integrating chromatography from each target versus control.

As shown in Figure 1.19, binding index (BI) is the ratio between MS relative intensity of ligand binding to protein target comparing with that binding to negative control. The higher BI will indicate a relatively stronger binding affinity. The properties of GPCR instability and insolubility limit the screening efficiency and capability of native ESI MS.

Our studies (in Chapter 5) indicated that affinity LC/MS could be applied to ligand

screening against GPCR nanodisc complex. The combination of nanodisc and affinity

LC/MS techniques makes it possible to screen the ligands against unstable membrane proteins in a high-throughput manner.

1.4.2 Ion source - The electrospray ionisation (ESI)

The ion source is an essential component of MS where the ionisation (compounds, proteins or nucleic acids) occurs. The types of ion sources include thermospray ionisation

(TSP), fast atom bombardment (FAB) electrospray ionisation (ESI), and matrix-assisted laser desorption ionisation (MALDI) 533. The MALDI and ESI are two common ion sources and have been used to detect protein-protein interaction and protein-lipid interactions. The ESI is an example of atmospheric pressure ionisation (API) source which ionises the sample at atmospheric pressure, followed by introducing the ions into a mass analyser in vacuum. The invention of ESI enables the production of intact heavy molecular ions (larger than 100 kDa) from a liquid phase 534-536.

The highly charged ions are transferred from solution and intermediate highly charged droplets to gas phase by the assistance of electrical energy, dry gas flow, and high temperature in ESI source (Figure 1.20 (A)). There are three steps involved in ESI source: nebulization of sample solution into electrically charged droplets, droplet ﬁssion, and production of desolvated ions 537-539. In detail, the end of the microcapillary is a small- diameter needle, which maintains a high electrical potential (typically 2 to 6 kV) between capillary and counter-electrode nozzle 534,537,540. A fine spray of highly charged droplets with the same polarity is generated from the tip of needle 541. As sample moves towards the counter electrode, a countercurrent flow of drying nebulising gas (e.g. N2) is introduced to ESI source. The nebulising gas with high temperature makes the solvent of charged sample droplets evaporate and makes the droplet sizes reduce. The drying gas continuously segregates the large ionic droplets into smaller particles. At a critical point

(Rayleigh limit) when the Coulomb repulsive force is greater than the cohesive force, the fission (Coulombic explosion) occurs. Lastly, the desolvated and multiple charged gaseous ions are produced 539,542. During the electrospray ionization, the positive/negative ions ([M+H]+) are formed by the addition of protons (H+), alkali cations (Na+ and K+), ammonium cations (NH4+), or proton abstraction ([M-H]-). For example, the molecule

(M) in positive ion mode undergoes M + nH+ → [M+nH]n+ or M + Na+ → [M+Na]+, while the molecule (M) in negative ion mode undergoes M - nH+ → [M-nH]n-. Normally, positive mode is used for analysing samples that have function group readily accept protons (such as amide and amino groups in protein or peptide), and the negative mode is used for analysing samples that have function group readily lose protons (such as carboxylic acids and hydroxyls in nucleic acids and saccharides). The continuous stream of gaseous ions is transported from the region of atmospheric pressure ionisation source through the orifice and further into the higher vacuum chambers (quadrupoles, hexapoles, collision cells, and mass analyser) in the MS (Figure 1.20 (B)). From ESI source, a distribution of different charged molecular ions is emitted and sampled by a skimmer cone.

Figure 1.20: The schematic illustration of controlled-current ESI.(A) The mechanism of

ESI. The figure is adapted from Ho et al., 2003; (B) A view of an electrospray source with the interface between the ion source in atmospheric pressure and the mass analyser in vacuum. The figure is adapted from de Hoffmann & Stroobant, 2007.

ESI source is suitable for determining the protein structural information 543-545 and the binding of small molecules 546, protein-protein 547, protein-nucleic acid 548-550, DNA-drug

539,551,552, antibody-antigen 525,534, protein-metal ion 553, and drug-DNA-protein complexes

554. For analytical researches, the ESI source brings many benefits. Firstly, ESI part can 63

be readily connected to liquid chromatography (LC) system to separate complex mixtures

537,555,556. ESI MS can also analyse the samples in solution at nearly physiological conditions of pH, concentration, and temperature 534. Due to the softness of ESI technique, weak non-covalent interactions can survive when transferring the complex from solution to gaseous phase without fragmentation happening 503,525,555,557-561. The observed protein- ligand binding event in ESI MS is close to actual solution phase in most cases 525. The intact non-covalent bound complex may be fragmented when introducing the in-source collision energy/voltage. The nonspecific binding or low energetic aggregation can be reduced by designing a negative control with nonspecific binding affinity background

525,562,563.

However, the major limitation of using ESI is the requirement of using desalted samples.

The ESI source is a constant-current electrochemical cell (Figure 1.20 (B)). During the formation of cations in the spray, electrons move from the sample solution to the electrical circuit and the ESI source maintains as electrical neutrality. If there are many inorganic ions (e.g. Na+ or K+) present in the solution, adducts will be formed due to the combination of neutral molecules with ionised inorganic ions other than protons. For example: M + iNa+ + jK+ → (M + iNa + jK)(i+j)+ instead of exclusive M + H+→ (M + H)+.

Therefore, the high concentration of non-volatile salt inhibits the sample forming protonated ions and causing the signal suppression and poor signal-noise ratio 564,565. The ammonium acetate (NH4Ac) is a type of volatile salts, which produce gaseous forms of ammonia (NH3) and acetic acid (HAC) at high temperature. The buffer exchange to ammonium acetate will reduce the adduct formation in the capillary and improve the mass accuracy 566. Of note, non-volatile salt or surfactant is important for stability and solubility of membrane proteins (such as GPCRs). The narrower ESI emitter tips (less than 0.5 µm) will better resolve the signals of protein in non-volatile buffer 567,568. The development of low solution flow rate ESI sources of “micro electrospray” (nano-ESI ionisation source) 64

has dramatically improved the sensitivity of MS 523,569,570. The diameter of nano-ESI source can be as small as 0.2 µm comparing to normal ESI of 1 µm in diameter.

1.4.3 Mass analysers - Fourier transform ion cyclotron resonance (FT-ICR) and

time-of-flight (TOF)

FT-ICR mass analyser determines the mass-to-charge ratio (m/z) of ion based on ion cyclotron resonance frequency (ωc) in a fixed magnetic field (B). When charged particles pass through a constant magnetic ﬁeld (B), the particles will be under a magnetic force or

Lorentz force (FM) perpendicular to the direction of magnetic field and make an circular trajectory motion according to the right-hand rule (Figure 1.21 (A and B)). To transform the ion cyclotron resonance frequency (ωc) to m/z (Figure 1.21 (D)), the Fourier transformation equation is m/z = 571,572. For the same charge state ions, the radius of the circular trajectory is inversely proportional to the mass of ion. The heavier ions, the less deflected than lighter ions. Thus the ions with different masses can be focused on the detector plate with varying the strength of the magnetic field. The ion with a spiral motion passes through a series of pumping stages with increasingly vacuum pressures. Finally, the ion falls into the ion trap which is known as magnetic field ICR cell (Figure 1.21 (C)).

There is a trapping plate at the end of the ion trap, which prevents the ions from escaping.

The ion trap is located inside a spatial uniform static superconducting high field magnet, which is cooled by liquid helium.

Figure 1.21: The principle of FT-ICR mass analyser and the outline of ESI-FT-ICR MS.(A)

According to the right-hand rule, when a charged particle passes through a constant magnetic ﬁeld, denoted as B, the particles will be under a magnetic force (FM) perpendicular to the direction of magnetic field; (B) The constant magnetic force (FM) in

ICR cell will deflect the ion motion as a circular trajectory; (C) The ESI-FT-ICR MS is equipped with an external ESI ion source, differential pumping stages, ion transfer optics and an ICR cell. The ICR cell is located inside the superconducting magnet; (D) The detected oscillation frequency is converted to m/z ratio by Fourier transformation (FT).

The figures are adapted from Gross & Roepstorff, 2011.

Time-of-flight (ToF) mass analyser is also extensively used in MS because of their convenience of use and calibration. ToF MS determines the mass-to-charge ratio (m/z) by measuring the ion’s “time-of-flight” when travel through a field-free ToF tube (Figure

1.22). The ions are accelerated by an electric field (U) with known voltage after entering.

The velocity (ν) of an ion is inversely proportional to m/z. For same charge ions, the heavier ions, the slower in speed. Therefore the ions with different m/z are separated in

ToF MS. The ToF mass analyser with an addition of a quadrupole (Q) analyser is called

Q-ToF mass analyser. The Q-ToF MS permits MS/MS analysis, collision induced dissociation (CID) and provides better accuracy for complex molecules.

Figure 1.22: Schematic diagram of Q-ToF MS.It includes an ESI source, multiple quadrupoles (collision cell), a field-free flight tube and ion detector. This figure is adapted from Silvia & Amadeo, 2006.

Resolution and accuracy are important properties in defining the analytical capabilities of a mass analyser. The resolution of MS refers to the ability to differentiate two adjacent signals. For example, the resolution of two adjacent mass peaks, m and m+Δm, is m/∆m.

The larger molecule, the greater number of charge states can be found from MS spectrum due to the larger molecular surface area. However, the spectrum at higher charge state will have a smaller spacing between two adjacent m/z peaks (i.e. Δm/z). In this case, the signals of large molecules need to be better resolved by a high-resolution MS 573 or dissociation into smaller fragments by collision energy (CID or CAD) in collision cells.

In another way, buffers can be adjusted to higher pH, which makes proteins at lower charge states (z) and larger spacing between adjacent m/z peaks (i.e. Δm/z) 574,575.

FT-ICR MS is one of the most sensitive methods for ion detection with a resolution above

107 and accuracy as low as 0.1 ppm 571,576-580. The high resolution MS is also important for determining the mutation and disulphide bond formation in the protein. For example, the high resolution MS can differentiate between Asp and Asn, Leu and Asn, Ile and Asn,

Glu and Gln, sulfhydryl groups (R-SH) and disulphide bond (R-S-S-R’) in protein samples 581. The Q-ToF MS is also a sensitive and high speed acquisition instrument that permits accuracy within 5ppm (1% corresponds to 10,000 ppm) 582.

1.4.4 Recent MS advances

Nowadays, MS plays a major role in biological research, pharmaceutical discoveries, and clinic applications. It provides an accurate and sensitive analysis method to investigate a subtle mass change within a complex sample (cell lysate, crude protein samples, protein- protein complex, protein-ligand complex) during biological processes. Protein is an important drug target because it is responsible for modulating biological processes and functions in the human body. MS is an ideal technology for studying proteins, including protein-ligand identification, protein post-translational modification, and protein

quantification. For protein studies, MS is in complementary to nuclear magnetic resonance (NMR) spectroscopy, circular dichroism (CD), light scattering, X-ray crystallography, and electron microscopy (EM). The advantages and disadvantages of these technologies for identifying protein-ligand binding have been discussed and compared in Table 1.5.

The most advantageous aspect of MS is less consumption of samples than other biophysical techniques due to its high sensitivity. It can recognise the protein molecular weight (native MS), regulatory binding sites (tandem MS/MS), complex compositions

(native MS), stoichiometry, protein topology/folding/conformational changes (HDX MS)

583,584, protein post-translational modifications (native MS), ligand/drug/cofactor identifications (native MS) 525, protein thermostability (denaturing MS), protein-protein interactions (native MS) 525 and inter-protein energetics during forming complex (native

MS) 585 with small amount of sample. The MS revolution not only pushes the determined

MW of protein over the limit of 150 kDa 378,530, but also study some complicated protein complexes. For example, nanodisc complex can contain and stabilise the membrane proteins for MS investigation. Nanodisc highlights the possibility of future membrane protein studies, such as subunit identification, stoichiometry chemistry, and interactions between lipids and MPs. Marty et al. (2012, 2014, 2016) used native ESI-MS to determine the molecular weight and polydispersity of intact empty nanodiscs and show that the intact ‘empty’ nanodiscs can be preserved in the gas phase 378,586,587. Han and his colleagues used direct proxy ligand ESI MS and nanodisc techniques to investigate the binding interactions (association constant, Ka) between a glycolipid (the lipid reconstituted in empty nanodisc) and a soluble protein called cholera toxin B subunit

588 homopentamer (CTB5) .

The affinity LC/MS is an emerging application on MS methods, especially on drug discovery, ligand discovery, and clinical identifications (proteomics and metabolomics).

The combination of sensitive MS and quantitative assay (affinity LC/MS) makes it possible to measure the ligand-protein binding affinity and specificity with low abundance of protein samples. The affinity LC/MS quantifies compounds by integrating intensities of the peaks in MS spectrum. The protein-ligand complex can be enriched by

SPE/HLB column or microbeads (immunoprecipitation). The ligand preparation is done by methanol evaporation and re-solubilisation in a proper MS-compatible buffer.

1.5 Project proposal

Proteins are important pharmaceutical targets in biological research. Specifically, LRRK2,

α2AR, and A2AR are implicated for neurodegenerative diseases, such as Parkinson’s disease and Alzheimer’s disease. The studies on these protein structures, functions and drug developments require the discovery of their binding ligands. MS is an important ligand identification method. However, the MS development for target proteins is limited by protein stability and solubility. For example, GPCRs are insoluble in MS-compatible salt buffer and require detergent for solubilisation and preservation. Therefore, α2AR and

A2AR will be stabilised in nanodiscs to further improve their solubility and stability. The suitability of these protein targets (LRRK2 subdomains, α2AR nanodisc, and A2AR nanodisc) will be evaluated focusing on their stability, solubility, and homogeneity. The desired protein target(s) with good stability, solubility, and homogeneity will be applied to MS-based ligand identification method development. The ligand discoveries against

LRRK2 and PD-related GPCRs will help with signalling transduction elucidation, GPCR crystallisation, GPCR deorphanisation, and drug development.

1.6 Aims

① To clone, express and purify LRRK2 subdomains (Roc/GTPase, COR, and

MAPKKK/kinase), A2AR, A2AR-GFP and α2AR (Chapter 2 and 3);

② To evaluate the solubility, stability, homogeneity and yield of LRRK2 subdomains

(Roc/GTPase, COR, and MAPKKK/kinase), A2AR, A2AR-GFP and α2AR (Chapter 2

and 3);

③ To reconstitute A2AR, A2AR-GFP and α2AR into nanodiscs (Chapter 4);

④ To evaluate solubility, stability, homogeneity and yield of nanodisc complexes

(Chapter 4);

⑤ To develop an MS-based ligand identification assay (native MS or ultrafiltration-

based affinity LC/MS) against soluble and stable protein targets (Chapter 5).

Chapter 2 Cloning, expression and purification of LRRK2 subdomains

(Roc/GTPase, COR and MAPKKK/kinase)

2.1 Abstract

PD is widely recognised as a multifactorial disease with risk factors of ageing, gender, ethnicity, pesticide exposure, protein misfolding, and genetic trait 4,15,16. LRRK2 is an important PD-causative factor and a relatively large protein (286 kDa) with multiple enzymatic domains. LRRK2 was discovered in 2004 and it is the only known homologous protein to LRRK1 in human. LRRK1 is reported to have an intermolecular controlling mechanism between the Roc/GTPase domain and the MAPKKK/kinase domain, and reported to also regulate the intramolecular protein target 67. LRRK1 and LRRK2 have similar protein domain organisations, thus LRRK2 is also proposed to have a similar reciprocal feedback loop that is responsible for PD pathogenesis. As there is no direct evidence on LRRK2 intracellular signalling pathway and no effective LRRK2-targeting drug, we aim to develop a MS-based ligand identification method towards LRRK2 subdomains (Roc/GTPase, COR and MAPKKK/kinase) and use the ligand tool to elucidate LRRK2 signalling pathway and develop potential drug candidates against

LRRK2-linked PD. In this chapter, we will introduce the cloning, expression and purification of three LRRK2 subdomains (Roc/GTPase, COR and MAPKKK/kinase) and evaluate the suitability of these LRRK2 subdomains for native MS.

2.2 Results and discussion

2.2.1 Physicochemical property prediction by bioinformatics

The full-length human LRRK2 cDNA and protein sequences were obtained from

GenBank, National Centre for Biotechnology Information (GenBank IDs: BC117180.1 and NP_940980.3). The boundaries of Roc/GTPase, COR and MAPKKK/kinase domains

were defined according to the literature 589-591. The stability, solubility and other physicochemical properties of each LRRK2 subdomain were predicted by an online bioinformatics tool ProtParam (http://www.expasy.org/tools/protparam.html). Based on the LRRK2 amino acid sequence, physicochemical properties of molecular weight, isoelectric point (PI), aliphatic index, amino acid percentage, instability index, and grand average of hydropathy (GRAVY) were computed by ProtParam.

The instability index and GRAVY value were calculated as important indicators for protein stability and solubility. The instability index represents the protein stability in a test tube. A stable protein has an instability index smaller than 40, while an unstable protein has an instability index above 40 592. The GRAVY value represents the protein hydropathy. When pH equals 7, the hydrophilic amino acids include arginine, lysine, glycine, asparagine, serine, glutamine and aspartate, while the hydrophobic amino acids include phenylalanine, isoleucine, tryptophan, leucine, valine, methionine, tyrosine, cysteine and alanine 593. The GRAVY sums hydropathy of all the amino acids at pH 7.0 and is divided by the number of amino acids 594. More negative GRAVY values indicate the proteins are more hydrophilic. Based on instability index and GRAVY value, the boundaries of Roc/GTPase, COR and MAPKKK/kinase domains can be further optimised (Figure 2.1).

Figure 2.1: The instability index and GRAVY values of Roc/GTPase, COR and

MAPKKK/kinase domains were calculated by ProtParam.The numbers below X-axis denotes constructs with different boundaries (start-end). The values next to Y-axis denote the instability index or GRAVY value.

As Roc amino acid sequence increased from 1333 – 1500 (168 a.a.) to 1333 – 1878 (545 a.a.), the protein became less stable and less hydrophilic (Figure 2.1 (A1) and (A2)). The

Roc construct of 1333 – 1509 was predicted to be more stable than the literature recommended sequence of 1333 – 1510 (178 a.a.) 589-591. Both of these two constructs 74

(1333 – 1509 (177 a.a.) and 1333 – 1510 (178 a.a.)) were predicted as hydrophilic proteins.

Therefore, the Roc sequence of 1333 – 1509 (177 a.a.) was selected as the cloning construct which was rich in lysine (10.2%), leucine (9.7%) and alanine (7.0%) but contained less cysteine (0.5%), tryptophan (1.6%) and tyrosine (2.2%).

As the COR amino acid sequence increased from 1502 – 1878 (368 a.a.) to 1550 – 1878

(331 a.a.), these constructs were all predicted to be unstable but hydrophilic (Figure 2.1

(B1) and (B2)). The increasing amino acid length of COR domain caused the protein to be slightly unstable, but did not significantly change the hydrophobicity. The COR domain of 1511 – 1878 (368 a.a.) was the most stable construct and this construct was consequently chosen for COR expression. The chosen COR construct was rich in leucine

(14.7%), glutamate (9.3%) and isoleucine (8.0%) but contained less threonine (1.1%), tryptophan (1.6%) and cysteine (2.1%).

The MAPKKK/kinase constructs from 1879 – 2091 (213 a.a.) to 1879 – 2149 (271 a.a.) were predicted as stable and hydrophilic proteins (Figure 2.1 (C1) and (C2)). The increasing length of MAPKKK construct did not dramatically change the instability index but slightly affected its hydrophobicity. The MAPKKK sequence of 1879 – 2138 (260 a.a.) was used for MAPKKK expression. This construct was rich in leucine (13.4%), alanine (8.2%) and glycine (6.7%) but contained less tryptophan (0.4%), cysteine (1.9%) and methionine (2.2%).

2.2.2 Cloning of Roc/GTPase and MAPKKK/kinase domains using traditional

restriction enzyme-based digestion and ligation method

Three boundaries of adjacent LRRK2 subdomains (Roc/GTPase, COR, and

MAPKKK/kinase) were determined by ProtParam. The plasmid strain of pET-22b(+)

(Novagen) was used as the expression vector, to express the recombinant proteins of

Roc/GTPase, COR, and MAPKKK/kinase under the control of T7 RNA polymerase promoter and lac operon (Figure 2.2).

Figure 2.2: Schematic diagram of pET-22b(+) expression vector, PCR products and expressed proteins.(A) Some important features of pET-22b(+), including plasmid size, restriction sites, lac promoter (lacI), multiple cloning sites (MCS), and ampicillin resistance (Ap); (B) The PCR product included a start codon (ATG), NdeI/XhoI restriction sites, and 6× His tag. The upstream T72 forward primer and downstream T7 reverse primer were used for colony PCR and DNA sequencing; (C) The expressed Roc,

COR and MAPKKK proteins included a lysine linker between protein and 6×His tag.

The COR domain was amplified and cloned in pET-22b(+) by Dr Christopher Love (my works on COR expression and purification will be covered later). The DNA sequences of

Roc/GTPase and MAPKKK/kinase were first amplified by polymerase chain reaction

(PCR) from the full-length human LRRK2 cDNA (MGC: 150789, IMAGE: 40125731).

The DNA sizes of Roc/GTPase and MAPKKK/kinase domains were shown as 800 bp and 580 bp respectively on DNA agarose gel. The PCR conditions, template DNA amount and annealing temperature, were optimised according to agarose gel electrophoresis results (Figure 2.3). The sharp and intense bands indicated pure and large quantities of

PCR products amplified in an optimal PCR condition.

Figure 2.3: The optimisation of template DNA concentration and annealing temperature.(A) The PCR products in different lanes were amplified with different amount of template LRRK2 DNA; (B) The lanes were PCR products amplified at different optimised annealing temperatures.

Figure 2.3 (A) indicated that 5 ng of template DNA amount produced the highest amount of PCR product. Additionally, the optimal annealing temperature to amplify Roc/GTPase and MAPKKK/kinase ranged from 50 °C to 52 °C (Figure 2.3 (B)). Because higher annealing temperature contributes to a higher PCR specificity, the annealing temperature 77

of 52 °C was chosen for PCR amplification. As a result, the Roc/GTPase and

MAPKKK/kinase were amplified from 5 ng full-length LRRK2 cDNA at an annealing temperature of 52 °C.

The amplified PCR products and pET-22b(+) vector were double digested with

XhoI/NdeI (NEB). The DNA sequence of Roc or MAPKKK was subsequently ligated and subcloned into pET-22b(+) vector between XhoI and NdeI restriction sites (Figure

2.2 (A)). However, NdeI restriction enzyme, which produces only two-base overhang, was previously reported to have lower digestion and ligation efficiency than XhoI 595.

Therefore, the complete digestion of PCR products and pET-22b(+) vector by NdeI would increase the ligation success rate and prevent self-ligation. The NdeI digestion time was optimised from one to four hours, and sequentially digested by XhoI for another four hours (Figure 2.4 (A)). The digested products were separated by agarose gel electrophoresis.

Figure 2.4: Optimisation of digestion time and PCR product purification.(A) Sequential digestion of pET-22b(+) with NdeI followed by XhoI. Lane 1, GeneRulerTM 1 kb DNA ladder; Lane 2, 1 µg of undigested pET-22b(+); Lane 3 to lane 6, 1 µg of pET-22b(+) was digested by 20 units of NdeI from 1 hr to 4 hrs; Lane 7 to lane 10, after 1 µg of pET-

22b(+) was completely digested by 20 units of NdeI, pET-22b(+) was subsequently digested with 20 units of XhoI from another 1 hr to 4 hrs; (B) Lanes 1, 3 and 5 showed pET-22b(+), MAPKKK/kinase and Roc PCR products before XhoI/NdeI digestion. Lanes

2, 4 and 6 were purified pET-22b(+), MAPKKK/kinase and Roc PCR products after

XhoI/NdeI digestion. 79

The sequential digestion (Figure 2.4 (A)) showed that 20 units of NdeI can completely digest 1 µg of pET-22b(+) vector after four hours. The restriction enzyme buffers, PCR primers, and polymerase enzymes were removed using Isolate® II PCR purification and gel extraction kit (Bioline) prior to the ligation reaction. The quality and quantity of purified DNA were checked on agarose gel (Figure 2.4 (B)) and by Nanodrop® (Figure

2.5). The Nanodrop® measured the ratios of absorbance at 260 and 280 (260/280), which was used to assess the purify of DNA sample. The pure DNA typically yields a 260/280 ratio of 1.8 596.

Figure 2.5: The quantity and quality of DNA were checked by Nanodrop.The samples included purified pET-22b(+) vector after double digestion, purified PCR products

(MAPKKK/kinase and Roc/GTPase), purified PCR products after double digestion

(MAPKKK/kinase and Roc/GTPase).

The Roc/GTPase or MAPKKK/kinase was subcloned into pET-22b(+) vector and transformed into E. coli DH5αTM competent cells. The bacteria containing recombinant plasmids were grown as colonies on agar plates with 100 µg/mL ampicillin and screened by colony PCR.

2.2.3 Colony PCR and sequencing reaction

Colony PCR was used to screen the E. coli DH5α colonies that contained the correct recombinant plasmids with Roc/GTPase or MAPKKK/kinase DNA sequence. The successful insertion of Roc/GTPase or MAPKKK/kinase sequence between XhoI and

NdeI would generate a colony PCR product of 710 bp or 960 bp respectively. The E. coli

DH5α colonies containing empty pET-22b(+) vectors would produce a colony PCR product of 275 bp. The T72 forward primer and T7 reverse primer were used to amplify the DNA sequence between XhoI and NdeI (Figure 2.6).

Figure 2.6: The colony PCR products of MAPKKK/kinase and Roc/GTPase ligated products.The ligations were under different pET-22b(+) vector: PCR insert ratios (1: 3, 1:

1 and 3: 1). The lane 49 (A2) and lane 51 (B2) were the negative controls containing empty pET-22b(+) vectors. 81

Most colony PCR products had a size of 275 bp, which was the same as negative controls of colony-49 (Figure 2.6 (A2)) and colony-51 (Figure 2.6 (B2)). These colonies contained empty pET-22b(+) vectors without Roc/GTPase or MAPKKK/kinase inserts. This result indicates a low digestion/ligation efficiency of NdeI that was previously reported 595. The colony-21 (vector: insert = 1: 1) produced colony PCR product of 960 bp which was consistent with the DNA size of MAPKKK/kinase (Figure 2.6 (A1)). The colony-17

(vector: insert = 1: 3), colony-38 (vector: insert = 1: 1), and colony-48 (vector: insert =1:

1) generated colony PCR fragment sizes of 710 bp, 600 bp and 300 bp respectively

(Figure 2.6 (B1) and (B2)). The size of PCR product from colony-17 was consistent with the DNA size of Roc/GTPase.

The colony-5 and colony-21 containing MAPKKK/kinase ligated products and colonies-

17, colony-38 and colony-48 containing Roc/GTPase ligated products were further examined by restriction digestion with XhoI/NdeI (NEB) and protein expression (Figure

2.7).

Figure 2.7: The restriction digestion and protein expression of selected colonies.(A) The lasmids from colony-5 and colony-21 (MAPKKK/kinase ligated plasmids) were double digested by XhoI/NdeI; (B) The plasmid from colony-21 with correct insert DNA size was further checked by protein expression; (C) The plasmids from colony-17, -38 and -

48 (Roc/GTPase domain) were double digested with XhoI/NdeI; (D) The plasmids from colony-17 and -38 with correct insert DNA sizes were further checked by protein expression. "-" denotes the plasmids before XhoI/NdeI double digestion; "+" denotes the plasmids after XhoI/NdeI digestion.

The recombinant plasmids were extracted and purified using PureYieldTM plasmid miniprep system (Promega). The restriction digestion of recombinant plasmid from colony-21 generated two DNA fragments, 5.4 kb (pET-22b vector) and 780 bp

(MAPKKK/kinase insertion). The molecular mass of expressed protein by colony-21 was

30 kDa, which was the same as MAPKKK/kinase protein (http://www.expasy.org/ tools/protparam.html). As a result, colony-21 might contain the correct recombinant plasmids with MAPKKK/kinase sequence (Figure 2.7 (A) and (B)). Similarly, double digestion of recombinant plasmid from colony-17 generated two fragments of 5.4 kb

(pET-22b vector) and 530 bp (Roc/GTPase insertion). The molecular mass of expressed protein by colony-17 was 21.4 kDa, which was consistent with the molecular mass of

Roc/GTPase protein (Figure 2.7 (C) and (D)).

As a result, the plasmids were extracted from colony-21 (MAPKKK/kinase) and colony-

17 (Roc/GTPase) and amplified by Prism BigDye terminator cycle sequencing 3.1 kit

(ABI) according to the protocol from Griffith DNA sequencing facility

(http://www.griffith.edu.au/science-aviation/dna-sequencing-facility). The prepared

DNA was submitted to the Griffith DNA sequencing facility and the sequencing results were confirmed with published full-length LRRK2 cDNA sequence (GenBank IDs:

BC117180.1) by BioEdit software (SI 2 in Appendix).

2.2.4 Expression and purification of Roc/GTPase, COR and MAPKKK/kinase

domains

The E. coli BL21: DE3 strain is deficient in lon and ompT proteases and under the control of Lac promoter of pET-22b 597. It was used to express Roc/GTPase (21.4 kDa), COR

(43.9 kDa) and MAPKKK/kinase (30.2 kDa) domains. The protein was expressed in E. coli BL21 after induction with 0.4 mM IPTG at 37 °C for 4 hrs. The whole cell lysate before and after the induction was separated by SDS-PAGE and visualised by Coomassie 84

blue staining. The relative migration of each protein was compared in parallel with the

PageRuler Plus pre-stained protein ladder (Thermo) (Figure 2.8).

Figure 2.8: Expression of Roc/GTPase, COR and MAPKKK/kinase domains.Each protein was induced with 0.4 mM IPTG for up to four hours. The proteins were separated on 12.5% SDS-PAGE gel and visualised by Coomassie blue staining.

The results (Figure 2.8) showed that the recombinant proteins of Roc/GTPase, COR or

MAPKKK/kinase were induced and expressed with 0.4 mM IPTG after four hours. The whole cell was applied to lysis and chromatography purification. To increase the protein solubility, the pH of lysis buffer was required to be around one unit away from the isoelectric points (PIs) of the protein. Because the theoretical PI values of Roc/GTPase,

COR, and MAPKKK/kinase domains were 9.74, 5.79 and 7.74 respectively, the pH of lysis buffer was determined as 7.0.

Around 3.5 g of wet bacterial pellet was collected from 1 L bacterial culture and the pellet was resuspended in 40 mL lysis buffer. The lysis buffers contained 50 mM Tris-HCl, pH

7.0, NaCl, glycerol, 1 mg/ml lysozyme PMSF, and TritonX-100 (Figure 2.9 (D)). The concentrations of NaCl salt, Triton X-100, and glycerol were optimised to enhance the protein solubility. The soluble and insoluble fractions were separated by centrifuging at

20,000× g for 30 mins.

Figure 2.9: Optimisation of lysis conditions for Roc, MAPKKK/kinase, and COR domains.The “-IPTG” and “+IPTG” indicate the whole cell lysate without induction and whole cell lysate with recombinant protein induced. The “S” and “I” indicate the soluble and insoluble fractions after cell lysis. 86

However, when expressing the proteins in large-scales, a significant amount of each protein was observed in insoluble fractions under different lysis conditions (Figure 2.9).

Since satisfactory solubility is an important requirement for native ESI-MS samples, these domains with low solubility were expected to experience difficulty for native ESI MS detection.

The full-length LRRK2 was previously expressed in mammalian cells, and found to be unstable and insoluble. Various expression organisms had been reported to express

LRRK2 subdomain previously, including E. coli, yeast and mammalian cells 32,38,74,598,599.

In this experiment, the LRRK2 subdomains (Roc/GTPase, COR, and MAPKKK/kinase) were expressed at 37°C in E. coli BL21. Even though the Roc/GTPase, COR, and

MAPKKK/kinase were predicted to have a reasonable solubility, the experimental results still showed these LRRK2 subdomains were low in solubilities. In other studies, the

LRRK2 homologous proteins, including Roco4 kinase domain from Dictyostelium discoideum (PDB ID: 4F0F) and ERK2 protein from Rattus norvegicus (PDB ID: 5U6I)

71-74, were also used to understand LRRK2 mechanisms. Based on existing proteins, genetic modification techniques can be used to solve the solubility problem in the future, such as addition of GST-tag, lowering the culturing temperature, and the use of different detergents in lysis buffer. In conclusion, the LRRK2 solubility problem would preclude these protein constructs from native MS studies.

2.3 Methods

2.3.1 Online bioinformatics analysis

The stability, solubility and other physicochemical properties of each LRRK2 subdomain were predicted by an online bioinformatics tool called ProtParam

(http://www.expasy.org/tools/protparam.html). Based on the amino acid sequence of each LRRK2 subdomain, the ProtParam computed the protein physicochemical properties, such as amino acid percentage, protein molecular mass, isoelectric point (PI), instability index, aliphatic index, and grand average of hydropathy (GRAVY).

2.3.2 Polymerase chain reaction (PCR)

For Roc domain, forward and reverse primers were: 5’-TAT TAA CAT ATG AAC CGA

ATG AAA CTT ATG ATT G-3’ and 5’- TAC TCG AGT TAA TGG TGG TGA TGG

TGA TGC TTA AGG CTC TCG TTT ATG ATG GT-3’. For COR domain, forward and reverse primers were: 5’-TAT TAA CAT ATG AAG ATC CGA GAT CAG CTT G-3’ and 5’- TAC TCG AGT TAA TGG TGG TGA TGG TGA TGC TTT TCA AAT TCC

AAC TCA TCA T -3’. For MAPKKK domain, forward and reverse primers were: 5’-

TAT TAA CAT ATG CAA GCT CCA GAG TTT CTC CTA G-3’ and 5’-TAC TCG

AGT TAA TGG TGG TGA TGG TGA TGC TTG ACT AAT TCA GCT GAA TTC A-3’.

All forward primers contained NdeI restriction sites (underlined) and all reverse primers consisted of XhoI restriction sites (underlined) and stop codons (in bold). Additionally, all reverse primers included exopeptidase-cleavable C-terminal hexahistidine tags (in italics). The reactions were performed by adding 0.5 - 5 ng template plasmid, 1.25 units

AmpliTaq Gold® 360 DNA polymerase (Invitrogen), 1× AmpliTaq Gold® 360 Buffer, 1.5 mM MgCl2, 0.2 mM dNTP and 0.4 µM forward and reverse primers in a volume of 50

µL. The thermal cycle was: first denaturation for 10 mins at 95 °C followed 30 cycles of denaturation (30 s at 95 °C), annealing (30 s at 52 °C) and elongation (1 min/kb at 72 °C).

A final elongation was performed for 10 mins at 72 °C. 88

2.3.3 Cloning by traditional restriction enzyme-based digestion and ligation method

The DNA fragments in agarose gel were purified using an Isolate® PCR and gel kit

(Bioline). Ten femtomole (30 ng) of double digested pET-22b(+) was ligated with DNA fragment of Roc, COR, and MAPKKK in the presence of a 1× ligase buffer (NEB) and five units of T4 DNA ligase (Fermentas). The ligation reaction was initiated by mixing purified double digested PCR products, purified double digested pET-22b(+) vectors, T4 ligase buffer (Fermentas), and T4 ligase (Fermentas) overnight at 25 °C. The molar ratio of vector: insert was optimised from 1: 3, 1: 1 to 3: 1 in a volume of 10 µL. The amount of vector and PCR product was calculated using NEBioCalculatorTM (v1.3.7).

The resulting recombinant plasmids of pET-22b-Roc, pET-22b-COR and pET-22b-

MAPKKK were diluted and transformed into E. coli DH5αTM competent cells

(Invitrogen). The E. coli were grown as colonies on agar plates with 100 µg/mL ampicillin.

The colonies containing different cloned plasmids were checked using colony PCR.

2.3.4 Recombinant plasmid verification by colony PCR and DNA sequencing

The colony PCR had the same thermal cycle and conditions as previous PCR for cloning each LRRK2 subdomains. The bacterial colony was removed from the agar plate with a pipette tip and resuspended in water. The resuspended colony was added to the PCR tube.

2.3.5 Protein expression and purification

Proteins were induced for expression by a chemical inducer called isopropyl-β-ᴅ- thiogalactoside (IPTG). When optical density at 595 nm (OD595) reached 0.6 to 0.8, a concentration of 0.4 mM IPTG was added to the bacterial culture to induce the recombinant protein expression. The bacterial cultures were grown at 37 °C for another four hours after induction. The cultures were taken out every one hour for SDS-PAGE analysis and OD595 measurement. After four hours induction, the bacteria were pelleted by centrifuging at 4,000× rpm for 40 mins at 4 °C. About 3.5 g of wet weight E. coli pellet 89

was harvested from one litre of bacterial culture. The pellets were stored at -20 °C or -

80 °C until use.

For each gram of cell pellet, 10 mL of lysis buffer (50 mM Tris-HCl, pH 7.0, 500 mM

NaCl, 0.1 mM EDTA, 0.1 mM DTT, TritonX-100, 1 mM PMSF, and 1 mg/mL lysozyme) was used to resuspend the cells. The cell suspension was subjected to five cycles of freezing and thawing, followed by five rounds of 30 s of insonation by a 25 kHz ultra- sonicator. The insonated samples were centrifuged at 20,000× g for 30 mins at 4 °C. The supernatant soluble fraction was stored temporarily at 4 °C until being subjected to affinity chromatography purification.

A column volume (CV) of 3 mL nickel Chelating Sepharose fast flow (GE Healthcare) was poured into a 10mL chromatography column (BioRad). Before protein purification, the chromatography column was washed sequentially by 2× CV of 1 M sodium hydroxide,

5× CV of sterilised distilled water, 2× CV of 0.2 M Nickel sulphate, 2× CV sterilised distilled water and 2× CV lysis buffer (50 mM Tris-HCl, 0.5 M NaCl, 20% sucrose, pH7.2). The bacteria lysate was loaded on the top of the column and collected in 50 mL

Falcon tubes. Then, the column was washed with 2× CV of lysis buffer (50 mM Tris-HCl,

0.5 M NaCl, 20% sucrose, pH 7.2) to remove any unbound proteins.

The lysis buffers containing increasing imidazole concentrations (i.e. 10 mM, 50 mM,

100 mM, 200 mM, 300 mM and 500 mM) were also washed through the chromatography column. Each fraction was analysed by SDS-PAGE. The fractions with purified eluted protein were pooled together and dialysed against buffer of 50 mM Tris-HCl, 0.5 M NaCl,

20% sucrose, pH7.2, 0.1 mM EDTA, 1 mM DTT and 1 mM PMSF.

The protein sizes and qualities were estimated by SDS-PAGE. The pellets were resuspended in 1× SDS sample buffer (50 mM Tris-HCl pH 6.8, 10% (v/v) glycerol, 2%

(w/v) SDS and 0.01% (w/v) bromophenol blue) and incubated at 98 °C for 10 mins. The samples were loaded onto 12.5% polyacrylamide resolving gel (1.5mm thick) with 5% polyacrylamide stacking gel (BioRad). All samples were loaded against PageRulerTM

Plus prestained protein ladder (Thermo). The gels were run in a Mini-PROTEAN® cell system (BioRad) with 1× Tris-glycine running buffer at a constant voltage of 100 V.

Upon the completion of gel running, the gels were removed from cassettes and rinsed once in distilled water for five minutes with rocking. The gels were stained for three hours or overnight in staining solution (40% methanol, 10% glacial acetic acid and 2.5 g/L

Coomassie Brilliant Blue R-250) with rocking, followed by destaining in solution of 40% methanol and 10% glacial acetic acid overnight.

Chapter 3 Expression and purification of A2AR, A2AR-GFP, and α2AR

3.1 Abstract

Apart from the LRRK2 subdomains with solubility problems, another two neurodegenerative disease-related GPCR proteins (A2AR and α2AR) were also expressed for MS-based ligand identification method development. Unlike cytosolic LRRK2 protein, the A2AR and α2AR are GPCR proteins with seven transmembrane domains located in the cell membrane. This class of cell membrane proteins are important drug targets because they control the communication with the extracellular environment and transduce the signals to downstream effectors. The ligand identification against GPCRs are important for GPCR deorphanisation, GPCR signalling transduction elucidation,

GPCR stabilisation and GPCR crystallisation. Many isolation and preservation techniques to maintain GPCRs solubility and stability in vitro have been developed, such as detergent micelle and nanodisc. In this Chapter, A2AR A2AR-GFP, and α2AR were expressed, purified, and stored in DDM detergent buffers for a short period. The solubility and protein yield were assessed by NuPAGE electrophoresis and analytical size exclusion chromatography (aSEC). The stabilities of GPCRs were estimated by CPM thermostability assay and NuPAGE electrophoresis. The receptors (A2AR, A2AR-GFP, or

α2AR), which had reasonable solubility, stability, yield, and homogeneity, were then used for MS-based ligand identification method development.

3.2 Results and discussion

3.2.1 Bioinformatics analysis for secondary GPCR structures

GPCRs have a seven-transmembrane (7TM) structure, with hydrophobic transmembrane domains and hydrophilic extracellular loops or intracellular loops. The A2AR and α2AR belong to the rhodopsin-like GPCR family. Many A2AR crystal structures have been solved so far, yet no α2AR crystal structure is available. The complete amino acid sequences of A2AR (UniProt ID: P29274) and α2AR (UniProt ID: P08913) from Homo 92

sapiens were retrieved from an online bioinformatics tool called GPCRdb

(http://gpcrdb.org/). In the delineation of secondary structures of A2AR and α2AR, the residue diagrams (or snake plots) were generated by GPCRdb. The snake plots of secondary structures showed conserved GPCR features, such as ionic lock, toggle switch

(CWxP) and a conserved NPxxY sequence (Figure 3.1).

Figure 3.1: Snake plots of A2AR (GenBank sequence ID: NP_000666.2) and α2AR

(GenBank sequence ID: NP_000672.3) generated from GPCRdb (http://gpcrdb.org/). 93

600 The A2AR and α2AR were genetically engineered GPCRs with C/N truncations, ICL3 truncations, BRIL fusion partner replacement, and mutagenesis. These modifications made the highly dynamic GPCRs less flexible and more thermostable. The two- dimensional topologies of A2AR and α2AR well resembled the general structure of rhodopsin-like GPCRs.

3.2.2 Expression and purification of A2AR and α2AR

The A2AR, A2AR-GFP and α2AR sequences were cloned into pFastBac1 vectors

(Invitrogen), which were used to generate recombinant baculoviruses by Bac-to-Bac technique (Invitrogen). The baculovirus containing GPCR DNA sequences was used to infect the Spodoptera frugiperda (Sf9) insect cells and overexpress the GPCRs on insect cell membrane. The GPCRs were extracted from cell membrane by dounce homogenisation, ultracentrifugation, and detergent solubilisation. Because A2AR, A2AR-

GFP and α2AR contained 840 Da C-terminal hexahistidine tags, these tags facilitated protein purification by immobilised metal affinity chromatography (IMAC). The wash

2+ and elution fractions from Co -NTA chromatography of A2AR, A2AR-GFP and α2AR were examined by NuPAGE gel electrophoresis (Figure 3.2).

Monomer Monomer

Aggregation Aggregation

Figure 3.2: The NuPAGE electrophoresis of α2AR, A2AR and A2AR-GFP purifications (A,

B and C). “W” and “E” represented “wash fraction” and “elution fraction”; (D) The aSEC profiles of purified A2AR with addition of different ligands. The aSEC height value

(mAU280) indicated the amount of aggregation and monomeric species; (E) The normalised aSEC profiles of A2AR purifications with addition of different binding ligands.

The data points were normalised against highest aSEC peak (mAU280) value and converted into percentage (%). It indicated the percentage of aggregation relative to the monomeric species.

The NuPAGE gel electrophoresis results showed that A2AR and α2AR protein were present at 40 kDa and A2AR-GFP was at 65 kDa. The protein aggregations were shown at a higher molecular weight between 70 and 130 kDa. The target protein bands were slightly wider than the soluble protein bands due to glycosylation (Figure 3.2 (A), (B) and (C)). 95

Moreover, aSEC was used to analyse the stabilising effect of detergent and ligands to the receptors (Figure 3.2 (D) and (E)). A2AR was purified with the addition of different types of stabilising ligands, i.e. a high-affinity ligand of ZM-241385, a low-affinity endogenous ligand of theophylline, or with no ligand added. The addition of two different stabilising antagonists (ZM-241385 and theophylline) did not significantly improve the A2AR homogeneity, nor did it reduce aggregation. Of note, the apo-A2AR (without ligand added) showed an excellent stability and homogeneity. Because the presence of a binding ligand in protein target will sterically hinder the access of other ligands, the apo-state of A2AR was selected as an ideal screening target for MS-based ligand identification.

3.2.3 Quantification using analytical size exclusion chromatography (aSEC) aSEC was used to check the protein homogeneity, differentiate aggregation and monomeric species, and quantify the protein yield. The relationship between peak height

(or peak area) and amount of standard protein (bovine serum albumin, BSA) was plotted to assess linearity (Figure 3.3).

Figure 3.3: The relationships between BSA amount and chromatographic peak area (A) or peak height (B). 96

The peak height (or peak area) of the protein chromatography showed a proportional correlation to the protein amount. The R-squared values (R2) of both aSEC standard curves were close to one (1), which indicated a good fit for linear regression. Moreover, the protein amounts calculated by peak height and peak area resulted in a similar value, thus the linear equations can be used to calculate relative ratio between the aggregation and monomeric species.

(A) (B)

Figure 3.4: The aSEC profiles of purified A2AR, A2AR-GFP and α2AR (A) and corresponding normalised aSEC profiles (B).

Table 3.1: Quantification of A2AR, A2AR-GFP and α2AR yield in Figure 3.4 based on linear regression relationship in Figure 3.3.

Height of monomer Amount (µg) for Total yield (µg) for

peak (mAU280) 0.5 μL 1 L culture

α2AR 108.86 35.89 1435.92

A2AR 256.79 89.70 3588.00

A2AR-GFP 62.57 19.06 1143.87

The aSEC profiles (Figure 3.4 (A)) showed that A2AR (3.5 mg/1 L culture) had the highest yield than the yields of α2AR (1.4 mg/1 L culture) and A2AR-GFP (1.1 mg/1 L culture).

Because the molecular mass of A2AR-GFP is slightly heavier than the other two GPCRs, the A2AR-GFP (in green colour) showed a slight shift in the aSEC profiles. The normalised aSEC profiles (Figure 3.4 (B)) indicated that the apo A2AR and apo A2AR-

GFP had higher percentage of monomeric species compared to the apo α2AR.

As a result, the apo A2AR has the highest yield and best homogeneity among these three apo-state GPCRs (A2AR, A2AR-GFP, and α2AR). It may be due to the fact that the A2AR construct is well-studied and well-optimised with many solved crystal structure, but the

α2AR construct is still under optimisation for better homogeneity. The homogeneous apo

A2AR can be directly used for CPM thermostability assay.

3.2.4 CPM thermostability assay

Cysteines are embedded in the interior of GPCRs to form disulphide bonds 601. During the thermal cycle (from 20 ºC to 90 ºC), GPCRs are denatured and interior disulphide bonds of cys-cys are broken. N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl] maleimide or CPM dye will react with exposed free cysteines and form covalent bonds with free thiol side chains, which generate a conjugate with fluorescent excitation/emission wavelength of 384/463 nm. The melting temperature (Tm) is the temperature at which it is half of the maximum fluorescence. CPM thermostability assay was performed to calculate the Tm of A2AR in DDM detergent and compare the stabilising effect of different known binding ligands (Figure 3.5).

Figure 3.5: The CPM thermostability profiles of A2AR DDM with different binding ligands (A) and the comparison of melting temperature (Tm) with apo A2AR (B).The addition of A2AR unrelated ligand (AM-6538) was used as negative control. “ns” denoted

“not significant”, **** denoted the p-value < 0.0001, n = 4.

The Tm of apo A2AR in DDM detergent micelle was determined as 51.06 ±2.61 °C (n =

4), which was not significantly different from negative control of unrelated AM-6538 ligand (52.07 ±0.96 °C, n = 4). However, the addition of stabilising binding ligands resulted in higher melting temperatures than apo A2AR (Figure 3.5 (B)). In general, the

A2AR melting temperatures were dramatically increased by at least 10 ºC after adding ligands ZM-241385 (63.27 ±0.94 °C, n = 4), UK-432097 (71.15 ±1.39 °C, n = 4), SCH-

442416 (60.88 ±0.42 °C, n = 4), and SCH-58261 (61.38 ±0.33 °C, n = 4). The CPM assay compared the stabilising effects of different ligands to apo A2AR in detergent environment

(HEPES/NaCl/DDM/CHS).

The aSEC was also used to estimate how fast the apo A2AR aggregated over the storage period of four weeks at 4 ºC (Figure 3.6).

Monomer

Monomer Aggregation

Aggregation

Figure 3.6: Stability of apo A2AR in DDM micelle at 4 ºC over storage period of four weeks.(A) The aSEC profiles of A2AR in DDM stored for up to four weeks; (B) The normalised aSEC profile.

The aSEC results indicated that the detergent maintained the apo A2AR as relatively homogeneous at 4 ºC for up to two weeks (Figure 3.6). A significant aggregation appeared after storing for two weeks at 4 ºC. The percentage of A2AR aggregation accounted for

6.90% of the total A2AR sample in the first week, compared to the percentage of aggregation after two weeks (16.58%) and four weeks (29.39%).

The apo-state of A2AR, A2AR-GFP and α2AR were successfully expressed and purified in sf9 insect cells. This human A2AR construct was the same as the construct used for X-ray

600 structure of agonist-bound A2AR . This construct consisted of a fusion protein of apocytochrome b562RIL (BRIL) in the third intracellular loop (ICL3). The C-terminus was truncated to residue A317 to stabilise the receptor and minimise the flexibility.

602-606 Additionally, A2AR is a popular prototype for GPCR assay development . However, the apo α2AR showed a significant percentage of aggregation after purification, which indicated a necessity to add a stabilising ligand during purification. Because a high

100

percentage of aggregation in apo α2AR would influence CPM reacting with exposed

601 cysteine side chains and infer a false melting temperature, the homogeneous A2AR was used for CPM thermostability assay.

The Tm of A2AR was measured and used to compare the stabilising effect of different ligands. Intriguingly, the addition of agonist of UK-432097 significantly increased the Tm of A2AR for around 20 ºC. As a result, this agonist-bound A2AR construct was likely to be stable in the active state. This CPM thermostability assay was easily set up for non- fluorescently tagged ligands. However, the CPM assay has some limitations. For example, a fluorescent ligand can influence the absorption of UV at 384 - 463 nm during the detection. The CPM assay also lacked the binding information between ligand and protein.

It did not exclude the possibility of unspecific ligand binding against GPCRs. Therefore, the establishment of MS-based ligand identification assay would be an important assay to understand the GPCR-ligand binding.

The apo A2AR and apo A2AR-GFP showed the best stability and homogeneity compared to the other two GPCRs (A2AR-GFP and α2AR) in both CPM and aSEC assays. The apo

A2AR can maintain its homogeneity in the DDM detergent micelle for less than two weeks at 4 °C (Figure 3.6). The A2AR, A2AR-GFP and α2AR will be further reconstituted into a phospholipid-like nanodisc to improve the stability and solubility. The thermostability of apo A2AR in detergent micelle will be compared with that in nanodisc environment.

101

3.3 Methods

3.3.1 Secondary structure of GPCR proteins analysis by GPCRdb

GPCRdb, an open-access GPCR database, was used to present the information about

GPCR crystal structure in Protein Data Bank (PDB), single-point mutagenesis prediction, known binding ligands of receptors, and structure-based sequence alignment. The predicted mutagenesis and homologous sequences were used for optimising GPCR structure towards better thermostability. In our study, it predicted secondary structure and phylogenetic analysis of GPCRs. The residue diagrams (snake plots) were exported from “protein search” (http://gpcrdb.org/protein/) and the 2D topologies of receptors were demonstrated with full termini, loops, transmembrane domains, and conserved residues. The residues and positions were coloured and overstruck by clicking. The residue positions were indicated by both generic Ballesteros-Weinstein numbering scheme (with a conserved X.50 position in each transmembrane) and receptor-specific amino acid numbering scheme 607. The “phylogenetic trees” (http://gpcrdb.org/ phylogenetic_trees/targetselection) was generated based on GPCR transmembrane domain sequences and displayed by ladder representations 608. The interpretation included several coloured strips next to the receptor names, which were GPCR family, bound ligand type, and GPCR class respectively.

3.3.2 GPCR expressions and purification

600 The genetically engineered human A2AR and α2AR constructs contained an N-terminal

HA signal sequence followed by a FLAG tag, and a C-terminal 10× His tag. The A2AR and α2AR also consisted of a fusion protein of apocytochrome b562RIL (BRIL) in the third intracellular loop (ICL3) to enhance the thermostability and limit the ICL3 flexibility. The A2AR was truncated to residue A317 to further stabilise the A2AR and minimise the flexibility at C-terminus. Based on this A2AR construct, a green fluorescent

® protein (GFP) tag was cloned to the C-terminus of A2AR construct by overlap phusion 102

PCR. The A2AR, A2AR-GFP, and α2AR were cloned and expressed in a pFastBac1 vector

(Invitrogen). The recombinant baculoviruses were generated by Bac-to-Bac system

(Invitrogen) and infected into Spodoptera frugiperda (Sf9) insect cells. The passage 0 baculoviruses were used to infect 5 mL of Sf9 insect cells at a density of 2 – 3 ×

106 cells/mL. The cells were scaled up to 1 L and grown at 27 °C for 48 hrs prior to being harvested.

For large scale purification of the receptors, one litre of frozen Sf9 cell was thawed and dounce homogenised twice in the hypotonic lysis buffer that contained 20 mM HEPES pH 7.5, 10 mM NaCl, 10 mM MgCl2, 20 mM KCl, and protease inhibitor cocktail tablets

(Roche). The disrupted cells were collected by centrifuging at 40,000 rpm for 35 mins.

The dounce homogenisation was repeated another three times with the addition of 1 M

NaCl in the hypotonic lysis buffer to remove the unwanted soluble and membrane- associated proteins. The disrupted cell membrane pellet was resuspended in the buffer containing 50 mM HEPES pH 7.5, 800 mM NaCl, 2 mg/mL iodoacetamide. The cell membrane proteins were solubilised with 0.5% (w/v) DDM (Affymetrix) and 0.1% (w/v) cholesteryl hemisuccinate (CHS, Sigma) for another four hours. The cell membrane debris was separated from solublised cell membrane proteins through centrifuging at

35,000 rpm for 40 mins, and the supernatant containing the GPCRs was collected.

The A2AR, A2AR-GFP and α2AR contained 840 Da hexahistidine-tags at the C-termini, which facilitated protein purification by IMAC. In this case, the cobalt (Co2+)-NTA resin was used for affinity chromatography, because cobalt (Co2+)-NTA resin showed higher specificity than nickel (Ni2+)-NTA resin (but lower binding strength than Ni2+-NTA resin).

As a result, a lower concentration of imidazole was used to elute the GPCRs from Co2+-

NTA resin. The solubilised GPCRs were incubated with TALON® Cobalt IMAC resin

(Clontech) overnight in the presence of 20 mM imidazole. Then the resin was applied to

103

a gravity column (Poly-Prep, Bio-Rad) and washed by 20 column volumes (CVs) of wash buffer containing 50 mM HEPES pH 7.5, 800 mM NaCl, 0.05% (w/v) DDM, 0.01% (w/v)

CHS, and 20 mM imidazole. The receptors were finally eluted out with buffer containing

50 mM HEPES pH 7.5, 800 mM NaCl, 0.05% (w/v) DDM, 0.01% (w/v) CHS, and 220 mM imidazole.

Because SDS can induce the denaturation and aggregation of hydrophobic GPCR proteins

403, NuPAGE Bis-Tris gel, which does not contain SDS, was used for gel electrophoresis.

The protein samples were mixed with 4× LDS sample buffer (Novex) and separated on

10% NuPAGE Bis-Tris gels. After Coomassie blue staining, the fractions containing the receptor were pooled together and concentrated using a 100 kDa cutoff concentrator

(Sartorius). The final protein concentration was determined by colourimetric Bradford assay. All purification buffers and procedures were carried out at 4 °C.

3.3.3 aSEC

The purified GPCRs were applied to a Sepax Nanofilm SEC-250 column. The column was connected to an Agilent model 2000 HPLC system with a flow rate of 0.5 mL/min and signal detection set to 280 nm. The column was pre-equilibrated with HPLC buffer containing 50 mM HEPES pH 7.5, 500 mM NaCl, 2% (v/v) glycerol, 0.05% (w/v) DDM

(Affymetrix) and 0.01% (w/v) CHS (Sigma). The samples and column were maintained at 4 °C throughout the analysis. To avoid the protein concentration out of the range of linear regression (Figure 3.3), the concentrated GPCR samples (A2AR, α2AR and A2AR-

GFP) were diluted 80 times by HPLC compatible buffer before injecting into HPLC for quantification.

104

3.3.4 CPM thermostability assay

The N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl]maleimide or CPM dye

(Invitrogen, Carlsbad, CA) was a thiol-specific fluorochrome probe. The GPCR final concentration used for CPM assay was 0.015 mg/mL. The CPM dye was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 4 mg/mL (stock concentration). The

CPM final concentration was 0.007 mg/mL. The concentrations of all added ligands were

100 µM. The incubation temperature was increased from 20 ºC to 90 ºC at a rate of 1

ºC/min. The experiments were repeated four times (n = 4), and the significances of adding different ligands were statistically analysed by one-way ANOVA.

105

Chapter 4 Reconstitution of A2AR, A2AR-GFP, and α2AR into nanodiscs

4.1 Abstract

The apo/unliganded A2AR, A2AR-GFP, and α2AR were expressed, extracted, and purified in detergent micelle. The results in Chapter 3 showed that GPCRs maintained homogeneity in detergent micelle for less than two weeks at 4 ºC and the Tm of apo A2AR was around 51.06 ± 2.61 °C (n = 4). In ESI MS, protein analyte is subjected to high temperature dry gas flow, which assists in protein droplet evaporation in the ESI source before entering the quadrupole and mass analyser. As a result, the detergent-solubilised apo-state GPCRs are likely to aggregate and disrupt at high temperature in the ESI source.

To overcome this challenge, a phospholipid-like nanodisc complex was prepared to contain the GPCR in a more stable and homogeneous state. The thermostability and homogeneity of GPCR in nanodisc were measured by CPM thermostability assay and size exclusion chromatography (SEC), and compared with that in DDM detergent micelle.

The homogeneous and stable apo GPCR nanodisc complex with good yield was used to develop MS-based ligand identification method.

4.2 Results and discussion

4.2.1 Membrane scaffold protein (MSP) cloning by Overlap Phusion® PCR

The PCR primers for overlap Phusion® PCR and mutagenesis PCR were designed as shown in Table 4.1. The schematics of experimental procedures are shown in Figure 4.1.

The sequences of MSP1E2D1 (150 bp), MSP1E3D1 (250 bp), and MSP1 (750 bp) were amplified by PCR and separated on agarose gel (Figure 4.1 (B)). The 5’- and 3’-ends of

PCR products (or megaprimers) contained sequences complementary to the cloning site in MSP1D1 cloning vector. The bands of PCR products at the correct sizes were excised and purified for further overlap Phusion® PCR (step A).

106

The megaprimers and MSP1D1 cloning vector were mixed for overlap Phusion® PCR

(step B). The cloned plasmids were treated by DpnI enzyme digestion and selected by antibiotic resistance (step C). The plasmids were confirmed by DNA sequencing (SI 3 in

Appendix).

Figure 4.1: Schematic diagram of overlap Phusion® PCR for MSP cloning.Figure (A) was adapted from Bryksin & Matsumura, 2010. Figure (B) is the experimental illustration of overlap Phusion® PCR.

The overlap Phusion® PCR cloning method (Figure 4.1) was used to create genetically engineered plasmids expressing MSP1, MSP1D1, MSP1E2D1 and MSP1E3D1. The

QuickChangeTM site-directed mutagenesis was also used to mutate certain sequences within the pET-28a-MSP1D1 cloning vectors and create the plasmids expressing

MSP1D2, MSP2N2, MSP2N3 and MSP1E1D1.

107

4.2.2 Expression and purification of MSP and TEV

For small-scale expressions, the tobacco etch virus (TEV) protease and cloned MSP plasmid variants, MSP1 (24 kDa), MSP1D2 (24 kDa), MSP2N2 (45.5 kDa), MSP2N3

(46.2 kDa), MSP1E1D1 (27.5 kDa), MSP1E2D1 (30 kDa), and MSP1E3D1 (32 kDa), were transformed into E. coli BL21 competent cells and induced with 0.4 mM IPTG

(Figure 4.2).

MSP1E3D1

MSP1D2 MSP1

MSP2N2 MSP2N3

MSP1E1D1 MSP1E2D1

Figure 4.2: Small-scale purification of MSP variants.The “F”, “S” and “I” denote the

“flow-through fraction”, “soluble fraction”, and “insoluble fraction” respectively.

108

Gel electrophoresis (Figure 4.2) showed that different MSP variants were cloned into pET-28a-MSP1D1 vector and expressed in E. coli BL21 competent cells. MSP1D1 is one of the most commonly used MSP with a diameter of 9.5 - 9.7 nm and it can contain most

GPCRs with diameter of 3 - 5 nm in average 341,362,363. Therefore, MSP1D1 expression was scaled up and purified for reconstitution of GPCRs in nanodiscs. In addition,

MSP2N2 (second largest MSP with a diameter of 15 - 16.5 nm) was also scaled up and further purified.

The MSP1D1 and TEV protease proteins with N-terminal His6 tags facilitated the affinity chromatography purification by Ni-NTA (Figure 4.3 (A)). The His6 tag of MSP1D1 was cleaved by purified TEV protease. The digestion ratio (w/w) between TEV protease and

MSP1D1 (TEV:MSP1D1) was optimised from 1: 10 to 1: 100 and the MSP1D1 amount was maintained at 50 µg (Figure 4.3 (B)). The TEV and MSP1D1 mixtures were incubated at 4 ºC overnight (around 16 hrs). The digested MSP1D1 was purified by Ni-

NTA affinity chromatography again. The completely digested MSP1D1 without His6 tag

(ΔHis6) was present in the wash fractions containing low concentration of imidazole (0 -

20 mM). The incompletely digested MSP1D1 with His6 tag (+His6) was in the elution fractions containing high concentration of imidazole (400 mM).

109

Figure 4.3: The TEV purification (A) and MSP1D1 digestion optimisation (B).The “S”,

“I” and “W” denote the “soluble fraction”, “insoluble fraction”, and “wash fraction” respectively. The “MSP1D1(ΔHis6)” and “MSP1D1(+His6)” denote the “digested

MSP1D1 without His6 tag” and “incompletely digested MSP1D1 with His6 tag”.

The digestion optimisation results (Figure 4.3 (B)) indicated the TEV: MSP1D1 (w/w) ratios of 1: 10 and 1: 25 was sufficient to digest MSP1D1 at 4 ºC overnight (around 16 hrs). While, the TEV: MSP1D1 (w/w) ratios of 1: 50 and 1: 100 still showed incompletely digested MSP1D1 (+His6) at slightly higher molecular mass. After TEV digestion optimisation, the MSP1D1 (ΔHis6) was purified in large-scale for GPCRs reconstitution in nanodiscs.

4.2.3 Nanodisc reconstitution

For A2AR nanodisc reconstitution, the detergent-solubilised phospholipid mixture

(POPC/POPS), A2AR in detergent and MSP1D1 were mixed at physiological pH 7.5 at 4

ºC. The reconstitution ratio of GPCR: MSP1D1: lipids was optimised as 1: 2: 25, 1: 2: 50,

1: 2: 100, 1: 5: 100: 250, 1: 5: 500, 1: 8: 200, and 1: 8: 400 (Figure 4.4). The A2AR nanodisc self-assembly was initiated after removing detergents by Bio-BeadsTM (Bio-Rad) from the mixture of A2AR, POPC/POPS, and MSP1D1.

110

Aggregation Monomer Monomer Aggregation

Monomer

Aggregation

Figure 4.4: A2AR nanodisc reconstitution ratio was optimised.(A) The reconstitution optimisation was analysed by SEC in small-scale; (B) The SEC profile was normalised to the highest peak value in the curve. The highest peak values were normalised to 100% and lowest values normalised to 0%; (C) The SEC profile of large-scale A2AR nanodisc reconstitution; (D) The purified A2AR in DDM. The “W” and “E” denote “wash fractions” and “elution fractions”; (E) The homogeneous A2AR nanodisc monomeric peak was purified by Ni-NTA chromatography. The “W” and “E” denote “wash fractions” and

“elution fractions”. 111

The reconstituted A2AR nanodisc, aggregations, and free MSP1D1 were detected and separated by size exclusion chromatography (SEC) as shown in Figure 4.4 (A and B).

When applying the reconstituted A2AR nanodisc sample to FPLC, the aggregations of phospholipids/A2AR were present in the void volume of the size exclusion column at around 9 mL, and A2AR homogenous peak was at 13 mL preceding free MSP/empty nanodisc at 14 - 16 mL. A small reaction volume of 100 µL was used for nanodisc reconstitution optimisation. The SEC profiles of nanodisc were obtained and normalised against the highest peak value in the curve. The optimal A2AR: MSP1D1: lipid ratio was determined as 1: 5: 250. The SEC profiles (Figure 4.4 (A and B)) also indicated that the increase of lipid in the reconstitution mixture promoted the formation of aggregation and free MSP/empty nanodisc, such as at the ratios of 1: 2: 100, 1: 5: 500, and 1: 8: 400.

The optimised reconstitution ratio was applied to large-scale A2AR nanodisc reconstitution (Figure 4.4 (C)). The homogeneous A2AR nanodisc was purified by SEC and Ni-NTA chromatography. The empty nanodisc contained only lipids and MSP1D1

(ΔHis6) but A2AR contained C-terminal His tag, hence the A2AR containing nanodisc was separated from empty nanodisc by Ni-NTA chromatography (Figure 4.4 (E)). The A2AR nanodisc was eluted by a higher concentration of 400 mM imidazole and stored in buffer of 50 mM HEPES PH7.5, 150 mM NaCl. The nanodisc sample was finally concentrated to10 mg/mL.

112

Figure 4.5: The optimisation of A2AR-GFP (A and B) and α2AR (C and D) nanodisc reconstitution.The “W” and “E” denote the “wash fractions” and “elution fractions”. 113

Similarly, the optimal reconstitution ratios for A2AR-GFP nanodisc and α2AR nanodisc were determined as 1: 8: 250 and 1: 10: 700 respectively (Figure 4.5). The A2AR-GFP with a green fluorescent tag was used to track protein loss and denaturation during the process and evaluate the nanodisc reconstitution method. However, the α2AR nanodisc showed a low reconstitution yield (mAU280 ~32) compared to A2AR nanodisc (mAU280

~190) and A2AR-GFP nanodisc (mAU280 ~420) (Figure 4.5 (D). It may indicate a significant loss/denaturation during the reconstitution process. Because A2AR nanodisc showed the best homogeneity and reconstitution yield, it was further characterised in section 4.2.4 before being applied to MS screening.

Besides the optimisation for A2AR, A2AR-GFP and α2AR nanodisc reconstitution, another empty nanodisc composed of a single type phospholipid of DMPC and membrane scaffold protein of MSP2N2 (150 - 165 nm in diameter) was also generated (Figure 4.6).

Figure 4.6: An empty nanodisc reconstituted by MSP2N2 with DMPC in different

MSP2N2: DMPC ratios of 1:300, 1: 350, and 1: 400.

114

The SEC profile of MSP2N2/DMPC encircled nanodisc with DMPC lipids was different from A2AR, A2AR-GFP and α2AR nanodisc, which were constructed by POPC/POPS and

MSP1D1 (Figure 4.4 and Figure 4.5). The MSP2N2/DMPC reconstituted nanodisc showed more aggregation and free MSP2N2 (Figure 4.6). Of note, this profile was consistent with the SEC profile from the published literature which reconstituted tetrameric TRPV1 membrane protein into one nanodisc 450. This MSP2N2/DMPC encircled empty nanodisc can be applied to GPCR oligomerisation study in the future.

4.2.4 A2AR nanodisc characterisation by DLS, negative staining EM, and CPM

The DLS was used to characterise the homogeneity of nanodisc particles (spherical diameter) in a small quantity. The homogeneity of A2AR nanodisc was compared to that of A2AR DDM based on their spherical particle radius (Figure 4.7).

Figure 4.7: The DLS results of A2AR nanodisc comparing to A2AR in DDM detergent. 115

The size distribution (“average radius”) of A2AR nanodisc indicated that nanodisc had an average radius of 4.92 ± 0.12 nm or diameter of 9.84 ±0.24 nm. Moreover, the A2AR nanodisc (4.92 ± 0.12 nm) was smaller in radius than A2AR DDM (5.64 ± 1.37 nm). The

“% number” was used to report the relative amount of main species in the total sample solution. The A2AR nanodisc was more homogenous than A2AR DDM based on the radius

±SEM (100.0% ±0% compared to 99.4% ± 0.2%).

The homogenous A2AR nanodisc was concentrated to 10 mg/mL and stored at 4 ºC for four weeks. The NuPAGE gel electrophoresis and FPLC were used to analyse the A2AR nanodisc stability over the four weeks (Figure 4.8).

Monomeric A2AR nanodisc

Figure 4.8: Analysis of apo A2AR nanodisc stability by NuPAGE gel electrophoresis (A) and SEC (B).(A) The number on each lane of the NuPAGE gel indicated the day(s) of nanodisc samples stored at 4 ºC. The apo-A2AR nanodisc was stored up to four weeks (28 days); (B) The normalised SEC profile of each A2AR nanodisc samples stored at 4 ºC for up to four weeks.

The SEC profiles of A2AR nanodisc showed no significant change in its homogeneity and monodispersity over the four-week storage period at 4 ºC. There was also no aggregation appearing on NuPAGE gel during storage at 4 ºC. The results indicated that nanodisc maintained the apo A2AR in a stable state for at least four weeks at 4 ºC. This result 116

demonstrated that self-assembled phospholipid nanodisc kept A2AR in a better stability than in DDM detergent micelle (Figure 3.6). The negative-stained electron microscopy

(EM) was also used to qualitatively examine homogeneity of the A2AR nanodisc (Figure

4.9).

Figure 4.9: Aggregation and monomeric A2AR nanodisc species were applied to negative- staining EM.(A) The SEC profile of A2AR nanodisc reconstitution; (B) The aggregation fraction examined by negative-staining EM; (C) The homogeneous A2AR nanodisc fraction examined by negative-staining EM. 117

The aggregation species were observed as large stacked disc clusters (Figure 4.9 (B)).

Increasing lipid ratio in the reconstitution mixture led to more aggregated discs and larger particles as shown in Figure 4.9 (B). The monomeric fractions at 13 mL showed a more homogeneous and monodisperse A2AR nanodisc particle that oriented randomly on the staining grid.

The CPM thermostability assay was used to examine the thermostability (Tm) of A2AR in nanodisc environment compared to A2AR in the DDM detergent micelle (Figure 3.5). The incubation temperature increased from 20 ºC to 90 ºC with speed of 1 ºC/min.

Figure 4.10: CPM thermostability assays.(A) A2AR nanodisc thermostability assay

(including addition of different ligands and unrelated negative control of AM-6538); (B)

The comparison of thermostability of A2AR in nanodisc and DDM detergent micelle. The statistical analysis was one-way ANOVA. The “ns”, “**”, and “***” denote “not significant”, “p-value ≤ 0.01”, and “p-value ≤ 0.001”. The sample numbers of thermostability assay was n = 4 for A2AR DDM, and n = 6 for A2AR nanodisc.

118

The melting temperature of A2AR in nanodisc was measured by CPM thermostability assay. The melting temperature for apo A2AR nanodisc was 57.33 ± 1.18 ºC (n = 6). The addition of antagonists (ZM-241385, SCH-442416, and SCH-58261) contributed to a significant improvement of A2AR nanodisc melting temperature (Figure 4.10 (B)).

Specifically, the addition of ZM-241385 to A2AR nanodisc showed a 6 ºC increase in the melting temperature (63.40 ± 2.51 ºC, n = 6). The other two antagonists of SCH-442416

(62.03 ± 2.17 ºC, n = 6) and SCH-58261 (62.42 ± 2.32 ºC) showed around 5 ºC increase of the melting temperatures. Furthermore, the negative control ligand of AM-6538 (57.93

±0.367 ºC) did not improve the melting temperature of A2AR nanodisc. Intriguingly, the apo A2AR nanodisc showed that the melting temperature significantly increased by 6.5 °C in comparison to apo A2AR in DDM detergent. However, addition of the same stabilising ligands (ZM-241385, SCH-442416, or SCH-58261) to A2AR nanodisc and A2AR DDM resulted in the same melting temperature. The CPM assay, in general, was applicable to

GPCR nanodisc sample and Tm reflected the stabilising effect of each ligand.

This reconstitution study underscored the importance of GPCR: MSP1D1: lipids ratios for efﬁcient GPCR nanodisc reconstitution. Because the average diameter of GPCR transmembrane domain is 30 - 50 Å 341,362,363, the most common MSP construct of

MSP1D1 (diameter of 95 - 97 Å) is enough to contain the monomeric and dimeric GPCRs in nanodisc without constraint 467. The lengths of MSPs determine the sizes of the nanodiscs and thus affect the MSP: lipid ratios. For example, the second largest MSP of

MSP2N2 (diameter of 150 - 165 Å) constructed nanodisc contains more lipids than

MSP1D1 constructed nanodisc. Additionally, a change in the lipid composition (such as

POPC/POPS or DMPC) also affects the MSP: lipid ratio because the areas occupied by the lipid polar head groups are different. For example, the MSP1D1-encircled empty nanodisc can contain 125 unsaturated POPC or 160 DMPC per disc, and the MSP1E3D1-

119

encircled empty nanodisc (diameter of 128 Å) consists of 260 POPC or 320 DMPC per disc 367,373,376,411.

Moreover, the incorporation of GPCR into nanodisc displaces a certain amount of lipid molecules in the centre. The incorporation of bacteriorhodopsin (bR) into nanodisc replaces approximately 37 DMPC phospholipids 461, and rhodopsin displaces approximately 50 POPC phospholipids 369. From empirical experience, an extra number of lipids is usually added to the nanodisc reconstitution mixture. Thus, different MSP-to- lipid ratios are optimised for new combinations of MSP, lipid and GPCRs. In our experiment, the optimal A2AR: MSP1D1: POPC/POPS lipids ratio was determined as 1:

5: 250, A2AR-GFP: MSP1D1: POPC/POPS lipids ratio was 1: 8: 250, and α2AR: MSP1D1:

POPC/POPS lipids ratio was 1: 10: 700 (Figure 4.4 and Figure 4.5).

The GPCR nanodisc reconstitution is initiated by removing detergents, and the excess phospholipids and GPCRs will aggregate. The aggregation was formed in a significant amount if the ratio of GPCR: MSP: lipid was away from the optimal ratio. For α2AR,

A2AR and A2AR-GFP nanodisc reconstitutions, the experimentally determined ratios of

GPCR: MSP1D1: phospholipids were 1: 10: 700, 1: 8: 250 and 1: 5: 250 respectively.

Theoretically, every nanodisc was encircled by two MSP1D1. The A2AR: MSP1D1 ratio of 1: 2 showed lower yield than the ratio of 1: 5 in the optimisation results (Figure 4.4 (A and B)). It indicated that the increasing number of MSP amount favoured the assembly of monomeric GPCR in nanodisc. In contrast, the lower amount of MSP was reported to favour the oligomerisation of GPCRs 369,390,397,465,482,609,610. For example, the rhodopsin dimer was reconstituted into nanodisc at a rhodopsin: MSP ratio of 1: 1 369.

Furthermore, the natural phospholipids in cell membrane are asymmetrically distributed, such that the phosphatidylserine (PS) and phosphatidylethanolamine (PE) are primarily in the cytoplasmic leaflet, while the phosphatidylcholine (PC) and sphingomyelin (SPH) 120

are primarily in the outer leaflet 611,612. The cholesterol is also added to PC/PS phospholipid in nanodisc and help GPCR to fulfil its functions 490. In our study, the reconstituted α2AR, A2AR and A2AR-GFP nanodiscs contained two phospholipid mixtures

— zwitterionic POPC (70%) and negatively charged POPS (30%) to mimic the environment in the plasma membrane.

The A2AR, A2AR-GFP and α2AR were successfully reconstituted into nanodisc. The A2AR-

GFP with a green fluorescence was used to monitor the protein loss and reconstitution reaction during the experiment. The α2AR nanodisc showed a low yield level (mAU280

~32) compared to A2AR nanodisc (mAU280 ~190) and A2AR-GFP nanodisc (mAU280 ~420)

(Figure 4.5 (D). Therefore, A2AR nanodisc with good yield and stability was used as protein target for further MS development. The A2AR nanodisc was concentrated to 10 mg/mL in a detergent-free buffer (HEPES/NaCl buffer), which showed a good solubility.

The apo A2AR nanodisc also showed a good stability comparing to apo A2AR DDM. The nanodisc increased the melting temperature (Tm) of apo A2AR DDM by 6 °C and stabilised the receptor at 4 °C for at least four weeks. Notably, the nanodisc not only enhanced A2AR stability and kept receptor in a soluble sate, but also maintained the ligand accessibility to the receptors. The ligands could bind and stabilise the A2AR.

In conclusion, the nanodiscs were successfully reconstituted by different lengths of MSPs

(MSP1D1 or MSP2N2) and different lipid compositions (POPC/POPS mixture or

DMPC). We generated a workflow to reconstitute different GPCRs (such as A2AR, A2AR-

GFP and α2AR) into covalently circularised nanodiscs with high homogeneity. The A2AR in nanodisc had a better thermostability comparing to that in DDM detergent. The A2AR nanodisc in detergent-free buffer will facilitate the use of NMR and cryo-EM to study the structure and function of GPCRs. More importantly, the A2AR nanodisc will be used to develop MS-based ligand identification method.

121

4.3 Methods

4.3.1 MSP cloning by quick Overlap Phusion® PCR and mutagenesis

The pET-28a-MSP1D1 (plasmid number: 20061) was purchased from Addgene and used

as vector for overlap Phusion® PCR cloning. The mechanism of overlap Phusion® PCR

was similar to QuickChangeTM site-directed mutagenesis 613. The megaprimers had two

overlapping upstream and downstream ends that were complementary to the vector

(Figure 4.1). The optimised overlapping region between megaprimer and vector was

around 13 – 17 bp and the total length of megaprimer was around 45 bp. The primers

were designed by bioinformatics tools of BioEdit or Primer-BLAST (Table 4.1).

Table 4.1. Primers for overlap Phusion® PCR and QuickChange site-directed mutagenesis.

MSP variants Forward Primers Reverse Primers

5'-CGCACGCAGCGGTTCCACCTTTTGAC 5'-AATTATATCGTCAAAAGGTGGAACC MSP1E1D1 GATATAATTCCATCTCTTCCTGCCATTTT ATATCTCGATGACTTTCAGAAAAAATGG (mutagenesis from TTCTGAAAGTCATCGAGATATGGTTCCAC CAGGAAGAGATGGAATTATATCGTCAAA MSP1D1) CTTTTGACGATATAATT-3' AGGTGGAACCGCTGCGTGCG-3'

MSP1E2D1 (overlap 5’-ACTCCATGAGCTCCAAGAGAAGCTCA 5’-GCGATCGCGCATTTCTTCGCCTAAT PCR) GCCCATATCTCGATGACTTTCAGA-3’ GGGCTGAGCTTCTCTTGGAGCTC-3’

MSP1E3D1 (overlap 5’-TGTTGATGCACTCCGGACTCATTTGG 5’-CGCTGGCGAAGTTCATCCGAATACG PCR) CGCCATATCTCGATGACTTTCAGA-3’ GCGCCAAATGAGTCCGGAGTGC-3’ MSP1D2 5'-TGTCCCAGAATTCCTGCGTCACGGGA 5'-TACTGAGAATTTGTATTTTCAGGGT (mutagenesis from CCCTGAAAATACAAATTCTCAGTA-3' CCCGTGACGCAGGAATTCTGGGACA-3' MSP1D1) 5’-GGAGATATACCATGGGTCATCATCAT 5’- ACGGGGCCCAGTTGTTCGCGAAGT MSP1 (overlap PCR) CATCATCATATTGAGGGACGTC-3’ TTACTGAAGGTAGACGTAACAGA-3’ MSP2N1 5’-ATATACTAAAAAGCTGAATACCCAGT 5’-GTGCTCGAGTGCGGCCGCAAGCTTA (mutagenesis from CTACCTTCAGTAAACTTCGCGAAC-3’ CTGGGTATTCAGCTTTTTAGTATAT-3’ MSP1D1)

122

MSP2N2 5’-ATATACTAAAAAGCTGAATACCCAGC 5’- GTGCTCGAGTGCGGCCGCAAGCTT (mutagenesis from CCGTGACGCAGGAATTCTGGGACA-3’ ACTGGGTATTCAGCTTTTTAGTATA-3’ MSP1D1) MSP2N3 5'-TACCCAGGGTACCCGCGAACAACTGG 5'-CCTGCGTCACGGGGCCCAGTTGTTC (mutagenesis from GCCCCGTGACGCAGG-3' GCGGGTACCCTGGGTA-3' MSP2N2) MSP1E2 5’-ACTCCATGAGCTCCAAGAGAAGCTCA 5’-CGATCGCGCATTTCTTCGCCTAATG (mutagenesis from GCCCATATCTCGATGACTTTCAGAAAAAA GGCTGAGCTTCTCTTGGAGCTCATGGAG MSP1) T-3’ T-3’

The megaprimers contained MSP sequence and upstream/downstream complementary

regions to pET-28a-MSP1D1 vectors. The megaprimers were first generated by Phusion®

DNA polymerase PCR (Figure 4.1 (step A)). The annealing temperature for PCR was 58

ºC and could be optimised if necessary. The megaprimers were separated on agarose gel

and the megaprimers of correct sizes were excised, extracted and purified from agarose

gel by QIAquick® gel extraction kit (Qiagen) and silica-column purifications. The

overlapping extension PCR was initiated by mixing purified megaprimers, vector (pET-

28a-MSP1D1) and other PCR reagents in a volume of 50 µL (Figure 4.1 (step B)). The

PCR were 18 – 22 cycles.

The pET-28a-MSP1D1 parental vector (without MSP insertion) was multiplied and

methylated in E. coli DH5α (dam+). It was recognised and digested by DpnI restriction

enzyme at 37 °C for two hours (Figure 4.1 (step C)). The overlap Phusion® PCR products

(with MSP insertion) were retained after DpnI digestion and transformed into E. coli

DH5α competent cells. The transformed E. coli were plated out on agar plates with

kanamycin resistance. The plasmids were extracted and sequenced.

123

4.3.2 GPCR nanodisc reconstitution and optimisation

The GPCRs were temporarily stored in mild DDM detergent at 4 ºC for less than two weeks before nanodisc reconstitution. 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS) were purchased from Sigma-Aldrich. The POPC and POPS stocks were prepared by solubilising in 200 mM sodium cholate. To optimise the nanodisc reconstitution, each receptor was mixed with MSP and phospholipid mixture (POPC/POPS) in different ratios such as 1: 2: 25, 1: 2: 50, 1: 2: 100, 1: 5: 100, 1: 5: 250, 1: 5: 500, 1: 8: 200, and 1: 8: 400.

For phospholipid mixture (POPC/POPS), the ratio of POPC: POPS was maintained at 7:

3. To initiate the nanodisc reconstitution, the pre-treated Bio-Beads SM-2TM (Bio-Rad) was added and incubated overnight at 4 °C. The pre-treated Bio-Beads gradually adsorbed the detergents to aid in nanodiscs self-assembly. The Bio-Beads were separated from solution by centrifugation at high-speed centrifugation at 14,000× g.

The GPCR nanodisc complex was isolated and purified by a fast protein liquid chromatography (FPLC) system that connected with a Superdex 200, 10/300 GL gel filtration column at a flow rate of 0.4 mL/min. The reconstituted GPCR nanodisc peak

(mAU280) appeared at around 12.5 mL, which was after an aggregation peak (at around 9 mL) and preceding an MSP peak (at around 17 mL).

The optimised reconstitution condition showed the least aggregation and the most significant amount of homogenous species. The homogeneous fractions were further purified by immobilised metal ion affinity chromatography (IMAC) to separate empty nanodisc from GPCR-containing nanodisc. The Ni-NTA resin was first pre-equilibrated with 50 mM HEPES pH 7.5, 150 mM NaCl and 25 mM imidazole. Because hexahistidine

(His6) tag was only present in the GPCR, MSP1D1 (ΔHis6) encircled empty nanodiscs were removed in the wash step with 20 CVs of 50 mM HEPES pH 7.5, 150 mM NaCl,

124

50 mM imidazole. The GPCR containing nanodisc was finally eluted by 50 mM HEPES pH 7.5, 150 mM NaCl, 400 mM imidazole. The eluted fractions were assessed by

NuPAGE gel electrophoresis to evaluate the GPCR quantity and purity. The fractions containing GPCR nanodiscs were pooled, concentrated, and buffer exchanged into a final concentration of 10 mg/mL in buffer of 50 mM HEPES pH 7.5, 150 mM NaCl by a 30 kDa cut-off Amicon® centrifugal filter.

4.3.3 DLS

The nanodisc samples were diluted to 1 mg/mL (0.5 - 10 mg/ml in general) and centrifuged at 14,000× g or filtration to minimise the contaminations (such as dust or precipitations) prior to DLS analysis. The sample was loaded in a quartz cuvette and measured by a DynaPro nanostar instrument (Wyatt Technology) at room temperature.

The same sample was measured ten times (replicates) and three different batches of nanodisc samples were measured by DynaPro nanostar instrument in total. Data was analysed by DYNAMICS software (Wyatt Technology) to estimate the radius and molecular weight of the nanodisc particles.

4.3.4 Negative-staining electron microscopy

A volume of 100 ml 1% (w/v) uranyl formate (UF) solution was prepared in a dark room.

The prepared solution was covered with aluminium foil and stored at -80 °C until use.

The negative-staining EM grids (thin carbon-coated copper grids) were previously prepared by Professor Qingtao Shen. The grids were removed from a grid box and glow- discharged for 30 s. The discharged grids were placed on a clean filter paper in a petri dish with cover.

The discharged grid was held with tweezers at a 45° angle and 5 μL of the nanodisc sample (0.1 mg/mL nanodisc sample diluted with water to avoid salt background) was dispensed on the carbon film side of grid and incubated for one minute. The excess 125

nanodisc solution was removed by gently touching the edge of the grid with filter paper

(Whatman). The grid was briefly washed by 35 μL of deionised water on Parafilm for one minute. The excess solution was removed by filter paper. The grid was blotted by 1%

UF for one minute then the excess solution was removed by filter paper. The wash and staining steps were repeated three times before being imaged by a Tecnai T12 electron microscope fitted with a Tungsten filament operating at 120 kV. The images were recorded at a magnification of 67,000×.

126

Chapter 5 Ligand identification against A2AR nanodisc using native MS

5.1 Abstract

Among LRRK2, A2AR nanodisc, and α2AR nanodisc targets, the A2AR nanodisc with a good stability and solubility was selected as a suitable protein target for MS-based ligand identification method development. GPCRs are important drug-target and A2AR ligand identification can lead to a non-dopaminergic therapeutic target for PD.

Two MS methods were optimised for ligand identification against A2AR nanodisc, which are native MS (Chapter 5) and ultrafiltration-based affinity LC/MS (in Chapter 6). For native MS, the A2AR nanodisc sample were directly injected into ESI-FT-ICR MS and

ESI-Q-ToF MS. The molecular mass of protein sample and binding ligand can be directly calculated from MS spectrum. To date, there is only one report on the direct observation of ADP/ATP (endogenous ligand) binding with P2Y1R (GPCR) in DDM detergent buffer using an “in-house modified Orbitrap MS” 614, and there is no report on the identification of GPCR ligands using native ESI-FT-ICR MS or native ESI-Q-ToF MS in nanodisc environment. The GPCR samples need to travel at a high temperature from the ESI source towards quadrupoles and mass analyser during native MS analysis, thus native ESI-FT-

ICR MS is strictly limited by solubility and stability of the GPCR. The phospholipids of nanodisc need to be removed before A2AR enters the mass analyser in order to acquire mass information of the GPCR. The instrumental parameters of skimmer energy and in- source collision-induced fragmentation (CID) can be adjusted to help with phospholipid removal and liberate the GPCRs. It has been reported that the addition of octyl-β- glucoside (OG) detergent to nanodisc sample may help with the nanodisc dissociation 615.

As a result, native MS parameters were optimised for the mixture of A2AR nanodisc and

OG detergent. However, it was unsuccessful in developing native MS method to identify ligands against A2AR nanodisc after numourous optimisations.

127

Alternatively, the ultrafiltration-based affinity LC/MS method was used to identify ligands against A2AR nanodisc. The ligand-A2AR nanodisc complex was first enriched by centrifugal filtration, followed by methanol dissociation, methanol evaporation, and ligand resolubilisation. Ligand binding affinity was quantified using LC/MS system. The advantage of using affinity LC/MS was that the incubation was done at 4 ºC and it did not require the injection of the protein or nanodisc complex into MS.

To date, even though the resolving power of MS has been dramatically improved, the application of MS towards GPCR ligand identification is still challenging with few publication. The development of MS-based ligand identification method towards GPCRs will help with GPCR stabilisation, GPCR crystallisation, GPCR deorphanisation, and drug discovery. This development will also propel the application of MS application in the challenging GPCR research.

5.2 Results and discussion

5.2.1 Parameters optimisation for ESI-FT-ICR MS

Bovine carbonic anhydrase (bCA) was used as a standard protein for tuning and optimising the ESI-FT-ICR MS parameters. The heated gas and protein ions travelled across an electrical potential (voltage), between a heated capillary and a skimmer, therefore, the instrumental parameters of capillary, skimmer, and ion funnel, which are located in front of the MS instrument between ESI source and quadrupole, were first tuned systematically (Figure 1.16). The influential parameters, contributing to a significant effect on signalling-to-noise ratio and signal intensity, would be used to optimise A2AR signals afterwards (Figure 5.1 and more details in SI 5 in Appendix).

128

Figure 5.1: Important ESI-FT-ICR MS parameters for optimizing a standard protein of 10

µM bCA (m/z 1200 - 3000).More details in SI 5 in Appendix.

The bCA spectrum was first acquired at optimised MS parameters for high m/z at around

10 kDa (SI 5 (A) in Appendix). The electrical potential (voltages) at the capillary, skimmer, and ion funnel determined the transmission efficiencies of protein ions and MS sensitivity. Specifically, the decrease of capillary exit voltage from 220 V to 100 V reduced the signal intensity to half, but the increase of capillary exit voltage to 300V did not change the signal intensity (SI 5 (B1 and B2) in Appendix). When “ESI capillary voltage” or “skimmer voltage” was increased, there was also an improvement on bCA transmission efficiency and signal intensity (SI 5 (D1, D2, E1, and E2) in Appendix).

129

The ion funnel, at the interface between ion source and mass spectrometer detector, was responsible for the bCA ion transmission and focusing. The decrease of the ion funnel voltage led to a reduction in bCA signal intensity but more bCA ion peaks at lower charge state were observed (SI 5 (C1 and C2) in Appendix). It indicates that the protein ions were less focused and more ion species with different charge states were transferred into the mass analyser. The parameter of “ion funnel voltage” would be useful when requiring wider range of ion peaks albeit with weaker ion signals. These MS technical parameters that contribute to ion transmission efficiency and sensitivity would be tuned intensively when analysing the A2AR nanodisc by native MS.

5.2.2 Purified MSP1D1 (His6) and MSP2N2 (His6) soluble membrane scaffold

proteins detection by native ESI-FT-ICR MS

MSP1D1 and MSP2N2 are the scaffold proteins encircling the nanodisc and purified as soluble proteins. The purified MSP1D1 and MSP2N2 proteins were applied to ESI-FT-

ICR MS. The MSP protein spectrum will be the background peaks detected in A2AR nanodisc and empty nanodisc samples. The basis of charge and molecular mass determination were introduced in Figure 1.18.

130

m/z

Figure 5.2: MS spectra of MSP1D1 (m/z 1680 - 2500) analysed by ESI-FT-ICR MS.

The skimmer voltage was optimised to improve the signal intensity of MSP1D1 (ΔHis6) spectra. The signal intensity of +10 charge state was peaked at 2.5 × 107 when skimmer

TM voltage at 100 V. The MS spectrum of MSP1D1 (ΔHis6) was analysed by DataAnalysis software. The charge states of MSP1D1 were determined by the “charge state ruler” function in the software and manually calculated from each monoisotopic mass peaks.

The average molecular mass of MSP1D1 (ΔHis6) protein was determined as 22.085 kDa from the MS spectra. The MS spectra of another scaffold protein, MSP2N2, was also determined and identified by ESI-FT-ICR MS (Figure 5.3).

131

m/z

Figure 5.3: MS spectra of MSP2N2 analysed by ESI-FT-ICR MS (m/z 2600 - 3280).

Similarly, the charge states of MSP2N2 were also determined using the DataAnalysisTM software and manual calculation. The average molecular mass of MSP2N2 was determined as 45079.41549 kDa (Figure 5.3). The MSP results (Figure 5.2 and Figure

5.3) demonstrated that the increase of skimmer voltage at the ESI source improved the signal-to-noise ratio with better “clearness”. However, decreased signal intensity was observed when skimmer voltage increased above 120 V (Figure 5.2). It may be the result of protein disruption or breakdown at higher skimmer voltages at the ESI source and reduced protein species towards the mass analyser.

5.2.3 Empty nanodisc analysis by native ESI-FT-ICR MS and ESI-Q-ToF MS

After the molecular mass and MS spectra of membrane scaffold proteins (MSP1D1 and

MSP2N2) were determined, the MS spectra of empty nanodisc were also acquired by

ESI-FT-ICR MS and ESI-Q-TOF MS. 132

Figure 5.4: The full spectra of MSP1D1-encircled empty nanodisc acquired by native

ESI-FT-ICR MS (A) and Q-ToF MS (B). 133

The full spectra of MSP1D1-encircled empty nanodisc acquired by ESI-FT-ICR MS and

ESI-Q-ToF MS showed overwhelmed phospholipid signals at m/z of 1520 (2× POPC) and 1542 (POPC+POPS) and possible MSP1D1 signals at m/z 800 - 1,400 (Figure 5.4).

The weak MSP1D1 signals from the empty nanodisc spectra were further analysed as shown in Figure 5.5 and Figure 5.6.

m/z

Figure 5.5: The MS specta of empty nanodisc (m/z 980 - 1630) analysed by native ESI-

FT-ICR MS.(A) full empty nanodisc spectrum; (B) MSP1D1 signals (m/z 980 - 1430) in empty nanodisc spectrum at skimmer voltage of 25 V; (C) MSP1D1 signals (m/z 980 -

1430) in empty nanodisc spectrum at skimmer voltage of 50 V. 134

The full MS spectrum showed a dominating phospholipid signals over MSP1D1 signals at skimmer voltage of 25 V (Figure 5.5 (A)). The phospholipid signal intensity was up to

4.5 × 109, which was 100-fold higher than MSP1D1 signal intensity of 3.2 × 107 (Figure

5.5 (B)). The MSP1D1 signal intensity (3.5 × 108) was improved when the skimmer voltage was increased to 50 V (Figure 5.5 (C)). However, the lipids signals still dominated in the empty nanodisc. In a similar way, the MS spectra of empty nanodisc were also acquired by ESI-Q-ToF MS (Figure 5.6).

135

Figure 5.6: MS spectra of empty nanodisc (m/z 900-1400) analysed by ESI-Q-ToF MS. 136

As native ESI-FT-ICR MS (Figure 5.5), the native ESI-Q-ToF MS for empty nanodisc

(Figure 5.6) also showed charge states of MSP1D1 from +16 to +24. Another parameter, in-source collision-induced dissociation (CID) energy, was also applied to the nanodisc sample and successfully disrupted the empty nanodisc and released the MSP1D1. When in-source CID ranged from -10 – -20 eV, there were some signals of lipid-bound MSP1D1 in the MS spectra (Figure 5.6 (B and C)). When in-source CID energy changed to -30 eV, no lipid-bound MSP1D1 was found in MS spectrum (Figure 5.6 (D)). It suggested that the empty nanodisc was partially disrupted when in-source CID was between -10 eV and

-20 eV, and the empty nanodisc was completely disrupted at in-source CID of -30 eV.

The ESI-FT-ICR MS and ESI-Q-ToF MS results concluded the acquisition conditions for

MSP1D1 signals (m/z 800 - 1,400) and the full spectrum of MSP1D1 with overwhelmed phospholipid signals at m/z of 1520 (2× POPC) and 1542 (POPC+POPS) (Figure 5.4,

Figure 5.5 and Figure 5.6). The skimmer voltage helped to dissociate the salts and impurities in the empty nanodisc sample and contributed to a better signal-to-noise in the spectra. Moreover, the in-source CID energy showed a similar effect as skimmer voltage and preserved the internal non-covalent bonds between POPC lipid and MSP1D1.

Furthermore, the MSP1D1 signals detected by ESI-FT-ICR MS (3.5 × 108) were significantly higher than MSP1D1 signals detected by ESI-Q-ToF MS (4.5 × 104).

Therefore, the ESI-FT-ICR MS was used for further A2AR nanodisc analysis. It was reported that the addition of octyl-β-glucoside (OG) detergent into the nanodisc sample might help with the nanodisc dissociation 615. A mixture of control soluble TB protein

(provided by Professor Ron Quinn), empty nanodisc, and OG detergent was injected into

ESI-FT-ICR MS to evaluate the effects of OG detergent, and optimise the instrument parameters.

137

5.2.4 ESI-ICR-FT MS optimizations for nanodisc/detergent/protein mixture

A 40 kDa soluble TB protein, with a similar molecular mass as A2AR (45 kDa), was used as an A2AR mimic. A concentration of 1.1% OG (2× CMC) detergent was first mixed with TB protein and used to optimise the skimmer voltage and in-source CID energy systematically (Figure 5.7 and Figure 5.8).

Figure 5.7: The optimization of skimmer voltage (ESI-FT-ICR MS) for the mixture of 10

µM TB protein and 1.1% OG detergent.The shaded areas of MS spectra (m/z 2200 to

2800) contain the TB protein signals. The shaded areas are zoomed in and the TB protein charge states (+15 to +18) are displayed in figure inlet.

138

The obtained mass spectrum showed the detergent dimer (m/z 602) and detergent trimer

(m/z 894) signals, which suppressed the TB proteins signals (m/z 2200 - 2800). The highest TB protein signal intensity with best signal-to-noise ratio was obtained with a skimmer voltage of 120 V. Therefore, the skimmer was maintained at 120 V during the following in-source CID optimisation experiments.

Figure 5.8: The optimisation of in-source CID energy (ESI-FT-ICR MS) for the mixture of 10 µM TB protein and 1.1% OG detergent. The shaded areas of MS spectra (m/z 2200 to 3200) contain the TB protein signals. The shaded areas are zoomed in and the TB protein charge states (+13 to +19) are displayed in figure inlet.

139

When skimmer voltage kept at 120 V, the in-source CID energy was optimised from -10 eV to -30 eV. The results showed a dramatic decrease of OG detergent signal intensity as

CID energy was changed to -30 eV. As a result, the instrument parameters of skimmer voltage of 120 V combined with a CID energy of -30 eV contributed to the highest TB protein signal intensity.

A concentration of 10 µM empty nanodisc was added and mixed with 10 µM of TB protein and 1.1% OG detergent to mimic the mixture of A2AR nanodisc and 1.1% OG detergent. The optimised instrument parameters (skimmer voltage of 120 V and in-source

CID energy of -30 eV) were applied to the mixture of 10 µM TB protein/empty nanodisc and 1.1% OG detergent.

140

Figure 5.9: The MS spectrum of the mimic mixture (10 µM TB protein/empty nanodisc and 1.1% OG detergent) obtained at the skimmer voltage of 120 V and in-source CID energy of -30 eV (ESI-FT-ICR MS).The shaded areas of the MS spectrum (m/z 2400 to

3200) with TB protein signals at different charge states (+13 to +17) are displayed in the figure inlet.

The result (Figure 5.9) indicated that the optimised skimmer and CID improved the signal intensities of soluble TB protein (40 kDa) and signal-to-noise ratio in a mixture (10 µM

TB protein/empty nanodisc and 1.1% OG detergent). The instrument was optimised and tuned for analysing the native protein at molecular mass of 40 kDa, but it did not observe the A2AR (45 kDa) by native ESI MS, despite further parameter optimisation attempted. 141

This is likely due to the fact that A2AR has a much lower stability and solubility than TB protein, and contributed a lower response. Moreover, the soluble TB protein (40 kDa) was not located in the interior hydrophobic centre of a nanodisc; therefore, the mimic mixture (10 µM TB protein/empty nanodisc and 1.1% OG detergent) did not exactly resemble the A2AR nanodisc with OG detergent in native MS analysis.

In conclusion, the native ESI-FT-ICR MS was successfully used to detect different compositions of A2AR nanodisc such as POPC/POPS lipids and MSP1D1/MSP2N2.

However, the available instrument setup is more suitable to detect various soluble proteins and ligands instead of insoluble membrane proteins or GPCRs. The most recent achievement by Robinson's lab is to directly identify the endogenous ligand of ADP

614 binding to P2Y1R by nESI-MS . The mass spectrometer used for direct GPCR-ligand identification is internally modified and it was not feasible to reproduce in currently MS facility 616,617. The lipid signals from the GPCR nanodisc suppressed the GPCR signals, which are low in MS response (due to its low solubility). As a result, ultrafiltration-based affinity LC/MS method were also investigated for ligand identification against A2AR nanodisc complex.

5.3 Methods

5.3.1 Sample preparation for native ESI-FT-ICR MS

All protein samples (A2AR nanodisc, empty nanodisc, MSP1D1, and MSP2N2) were buffer exchanged against 200 mM ammonium acetate pH 7.5 and 2× CMC DDM by using 0.5 mL 10 kDa or 100 kDa molecular weight cut-off centrifugal filter (Merk,

Amicon, AU). The buffer for all samples was exchanged three times, which was equivalent to a 1,000× dilution of original HEPES/NaCl buffer. The centrifugation speed was at 2,500× g at 4 °C for 30 mins at each time. The concentration of the protein was estimated by NanodropTM 1000 and further measured by colourimetric Bradford protein

142

assay. The final protein concentration was diluted into 100 µM as a stock solution. The injected sample concentration ranged from 10 µM to 50 µM.

5.3.2 Ligand detection by native ESI-FT-ICR MS

To produce intact gas phase ions, the source was operated at a pressure of ~3.5 mbar for

ESI-FT-ICR MS and Q-TOF MS. The typical spray concentration of the proteins was

10 μM. The mass spectra were recorded with a capillary voltage of 1.2 kV and a cone voltage of 150 V. The pressures were 8.67 × 10-8 mbar in the time-of-flight (TOF) analyser or 5.45 × 10-10 mbar in the ICR analyser.

5.3.3 Data analysis for native ESI-FT-ICR MS

The mass spectra from Q-TOF MS (Bruker) and FT-ICR MS (Bruker) were extracted and analysed by DataAnalysisTM (Bruker). The molecular mass of protein was manually calculated from isotopic peaks (Figure 1.18) or acquired automatically from the deconvoluted spectrum.

143

Chapter 6 Ligand identiﬁcation of adenosine A2A receptor in self-assembled

nanodiscs by ultrafiltration-based aﬃnity LC/MS

(Reference: Anal. Methods, 2017, 9, 5851–5858)

144

6.1 Abstract

A nanodisc is designed as a vehicle to contain GPCRs in a native-like phospholipid environment. It maintains GPCRs in a better thermostability and homogeneity state in a detergent-free buffer compared to a detergent micelle environment. Here, we established an affinity mass spectrometry-based approach to assay the interaction between the adenosine A2A receptor in the nanodisc and eight known ligands. The affinity mass spectrometry results are consistent with the previously published data for binding affinities. The combination of nanodisc and affinity MS techniques allows the identification of the GPCR ligands in a relatively stable, unliganded/apo and native state.

6.2 Introduction

Membrane proteins are key players in processing foreign signals (such as taste, light, odour and chemicals) and maintaining cellular homoeostasis. They can trigger downstream signalling cascades and immune responses. A large number of membrane proteins belong to the superfamily of G-protein coupled receptors (GPCRs), which is

109 predicted to have a number of 826 GPCRs in the human genome . The adenosine A2A receptor (A2AR) is one of the most important GPCRs in the human body.

There are four types of adenosine receptors discovered in mammalians, which are A1,

256 A2A, A2B and A3 receptors . These receptors use adenosine as the endogenous ligand, which is a crucial neuromodulator regulating a broad range of physiological functions in

252,618 the human central and peripheral nervous system . The A2AR plays a critical role in many physiological functions. For example, the knockout mice show a slower response to pain stimuli but increased platelet aggregation, heart rate and blood pressure 285. It is implicated that the A2AR can cause vasodilation of coronary arteries and increase the blood flow to the myocardium. The A2AR is also responsible for some pathophysiological conditions, such as inflammatory diseases, Parkinson's disease, epilepsies, sleep disorders,

145

pain, and drug addictions 23,259,269-272. These properties indicate the therapeutic relevance of the A2AR in treating inflammation, psychiatric disorders and neurodegenerative diseases.

The GPCRs with solved structures usually contain bound ligands to stabilise the receptors in a certain state. The endogenous ligands include ions, organic compounds, odorants, amines, peptides, proteins, lipids, nucleotides, and photons 151,619. These ligands tend to bind at the extracellular domain, within the cavity of the GPCR binding pocket or in the outer half transmembrane region 620. These regulatory sites are readily available to circulating small ligands, which may be potential drug candidates. Approximately 33% of marketed small-molecule drugs are GPCR targeted 621. These ligands help to regulate cellular processes through GPCRs such as synaptic transmission, enzyme catalysis, signal transduction and immune response.

Ligand screening methods against protein targets include thermal shift assays, surface plasmon resonance (SPR), ligand-observed NMR spectroscopy, mass spectrometry (MS), radioactive ligand binding assay and X-ray crystallography. In the past decade, MS has emerged as a powerful and high-performance technique for ligand detection and screening in complex biological systems 622-626. For the direct capture of GPCR-ligand binding by MS, the major challenges are the sample preparation and instrument performance. The detergent signals can potentially overwhelm the GPCR signals significantly. Also, it needs well-tuned collisional energy to dissociate the GPCR from the local micelle environment. The most recent achievement by Robinson's lab is to

614 directly identify the endogenous ligand of ADP binding to P2Y1R by nESI-MS . The mass spectrometer used for direct GPCR-ligand identification is internally modified and may not be feasible for some other MS labs 616,617.

146

In the past, aﬃnity MS has been applied to ligand identification towards various soluble

531,532,627,628 proteins and certain purified GPCRs such as human vasopressin (V2) and

629 muscarinic type 1 (M1) receptors that are solubilised in detergent-containing buﬀer .

Since the environment of detergent micelles tend to destabilise specific GPCRs during purification 335, an engineered nanodisc technology has been developed for GPCR studies

338,340. Nanodiscs constitute self-assembled disc-like phospholipid bilayers that are encircled by two amphipathic helical scaﬀold proteins, called membrane scaﬀold proteins

(MSPs). The sizes of lipid bilayers normally range from 8.6-16 nm diameters, depending on MSP variants 367. From our dynamic light scattering (DLS) results, the average

MSP1D1 diameter is estimated to be 9.8 nm. Nanodiscs can maintain GPCRs at a nearly native state with better thermostability and homogeneity in detergent-free buffer (such as phosphate-buffered saline and Tris/NaCl buffer). Our previous CPM thermostability assay (results not shown) indicates that the melting temperature of A2AR nanodiscs is 5 °C higher than it in DDM detergent and A2AR nanodisc is stable at 4 °C for at least one month.

Moreover, nanodiscs will not influence the accessibility of ligands into the ligand binding pocket from extracellular and intracellular interfaces 341,630. Here we developed a method to apply the challenging GPCR nanodisc on aﬃnity LC/MS with low interference and

600 less false positives. In our study, we reconstitute the A2AR into a nanodisc and utilise the ultrafiltration-based LC/MS to measure the ligand binding aﬃnity to the A2AR nanodisc.

6.3 Materials and methods

6.3.1 Constructs and GPCR expressions

600 The genetically engineered A2AR construct contained an HA signal sequence and a

FLAG tag at the N-terminus and a 10× His-tag at the C-terminus. For the crystallisation purpose, the receptor also consisted of a fusion protein, apocytochrome b562RIL (BRIL), in the third intracellular loop and C-terminal truncation to residue A317 to further stabilise 147

the receptor and minimise the flexibility. This A2AR construct was cloned into a pFastBac1 vector (Invitrogen). The recombinant baculoviruses were generated by the

Bac-to-Bac system (Invitrogen) and infected into Spodoptera frugiperda (Sf9) cells. The passage 0 baculovirus was used to infect 5 mL of Sf9 cells at a density of 2-3 × 106 cells per mL. Cells were grown at 27 °C for 48 hrs prior to harvest.

6.3.2 Purification and NuPAGE electrophoresis

For large scale purification of the A2AR, 1 L frozen Sf9 cells were thawed and dounce homogenised twice in the hypotonic lysis buﬀer containing 20 mM HEPES pH 7.5, 10 mM NaCl, 10 mM MgCl2, 20 mM KCl, and protease inhibitor cocktail tablets (Roche).

The disrupted cells were collected by centrifuging at 40,000 rpm for 35 mins. To remove the unwanted soluble and membrane-associated proteins, dounce homogenisation was repeated another three times with the addition of 1 M NaCl in the hypotonic lysis buﬀer.

The disrupted cell membrane pellet containing the A2AR was resuspended in the buﬀer containing 50 mM HEPES pH 7.5, 800 mM NaCl and 2 mg/mL iodoacetamide and rocked for 30 min, followed by solubilisation with 0.5% (w/v) DDM (Aﬀymetrix) and

0.1% (w/v) cholesteryl hemisuccinate (CHS, Sigma) for four hours. The solubilised membranes were isolated by centrifuging at 35 000 rpm for 40 mins, and the supernatant containing the A2AR was incubated with the TALON® IMAC resin (Clontech) overnight in the presence of 20 mM imidazole. The resin was applied to a gravity column (Poly-

Prep, Bio-Rad) and washed using 20 column volumes (CVs) buﬀer containing 50 mM

HEPES pH 7.5, 800 mM NaCl, 0.05% (w/v) DDM, 0.01% (w/v) CHS and 20 mM imidazole. The A2AR was eluted with buﬀer containing 50 mM HEPES pH 7.5, 800 mM

NaCl, 0.05% (w/v) DDM, 0.01% (w/v) CHS, and 220 mM imidazole.

The protein samples were mixed with 4× LDS sample buﬀer (Novex) and separated on

10% NuPAGE bis-Tris gels. After Coomassie blue staining, the fractions containing the

148

A2AR were pooled together and concentrated by using a 100 kDa cutoﬀ concentrator

(Sartorius). The final protein concentration was determined by colourimetric Bradford assay. All purification buﬀers and purification procedures were carried out at 4 °C.

6.3.3 Analytical size exclusion chromatography (aSEC)

Purified GPCRs were applied to a Sepax Nanofilm SEC-250 column using an Agilent model 2000 HPLC system at a flow rate of 0.5 mL and the signal detection set to 280 nm.

The column was pre-equilibrated with 50 mM HEPES pH 7.5, 500 mM NaCl, 2% (v/v) glycerol, 0.05% (w/v) DDM (Aﬀymetrix) and 0.01% (w/v) CHS (Sigma). Samples and columns were maintained at 4 °C throughout the analysis.

6.3.4 Membrane scaﬀold protein expression and purification

MSP1D1 was expressed by using a pET-28a vector system (Novagen) with the BL21-

Gold (DE3) strain (Stratagene) as a host. A single colony was inoculated into 5 mL Luria-

Bertani broth (LB) containing 50 mg/mL kanamycin as a starting culture and further transferred into 1 L sterilised LB. When OD600 reaches 0.8-1.2, the culture was induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) for another four hours. The bacteria were harvested by centrifugation at 4000g for 30 mins. The pellet from the 1 L culture was resuspended in 40 mL 1× phosphate-buﬀered saline (PBS) and lysed by repeated sonication. The lysate was clarified by centrifugation at 20,000g for 30 mins.

The lysate was loaded onto the nickel column under gravity flow. The column was washed with buffer containing 1× PBS, pH 7.5 and 50 mM imidazole for 20 CVs, then the MSP1D1 was eluted with buffer containing 1PBS pH 7.5 and 0.4 M imidazole. The purity and quantity of the fractions were analysed by electrophoresis. The fractions containing MSP1D1 were pooled together and buffer exchanged to 20 mM Tris/HCl, 0.5

M NaCl, 0.5 mM EDTA and pH 7.5 at 4 °C. The final concentration for MSP1D1 was 10 mg/mL. 149

6.3.5 Reconstitution of the A2AR into the nanodisc

1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl- sn-glycero-3-phospho-L-serine (POPS) were purchased from Sigma-Aldrich. The POPC and POPS stocks were solubilised by using sodium cholate. To optimise the ratio for nanodisc reconstitution, the A2AR was mixed with MSP1D1 and phospholipids (a mixture of POPC/POPS) with diﬀerent ratios, i.e. 1 : 2 : 200, 1 : 5 : 100, 1 : 5 : 250. For the phospholipids, the ratio of POPC : POPS was maintained at 7 : 3. The mixture was incubated with pre-treated Biobeads (Bio-Rad) overnight at 4 °C, which adsorbed the detergents to aid in nanodisc formation.

To isolate the homogeneous nanodisc, the product was run through a fast protein liquid chromatography (FPLC) system with the Superdex 200, 10/300 GL column. A reconstituted nanodisc peak appeared at around 12.5 mL, after an aggregation peak and preceding a MSP1D1 peak. The fractions were further purified by immobilised metal ion aﬃnity chromatography (IMAC) to separate the empty nanodisc from the A2AR nanodisc.

6.3.6 Ligand identification by aﬃnity MS

The A2AR nanodisc was incubated with a compound mixture consisting of eight known ligands (SCH-442416, ZM-241385, UK-432097, SCH-58261, NECA, adenosine, theophylline, and caﬀeine) and seven unrelated compounds at a final concentration of 2

µM (A2AR nanodisc) and 0.4 µM (ligand) at 4 °C for two hours. The incubated sample was diluted with 150 mM ammonium acetate and filtered through a 100 kDa MW cutoﬀ concentrator (Sartorius) by centrifugation at 5000× g for 5 min at 4 °C. After washing with the dilution buﬀer twice, the A2AR nanodisc proteins retained on the ultrafiltration membrane were collected into a new centrifugal tube from which the protein-ligand complexes were isolated. The bound compounds were released from proteins by methanol dissociation, followed by centrifugation at 13,000g for 20 min at 4 °C. The supernatant

150

was passed through the OstroTM 96-well plate (Waters) to remove protein and phospholipid interferents. The eluent was dried in a speed vacuum and re-dissolved in 50% methanol. The empty nanodisc was used as a negative control under the same procedures.

Three experimental replicates were prepared for each pair of the target and the negative control. Samples were analysed on an Agilent 6530 TOF coupled to an Agilent 1260

HPLC system. The compounds were eluted from the Eclipse Plus C18 column (2.1 mm

× 100 mm, 3.5 mm, Agilent, USA) at a flow rate of 0.4 mL/min, with the mobile phases of water/0.1% formic acid (A) and acetonitrile/0.1% formic acid (B). The LC gradient was as follows: 0–2 min, 5% B; 2–2.1 min, 5–20% B; 2.1–10 min, 20–35% B; 10–13 min,

35–60% B; 13–13.5 min, 60–90% B; 13.5–16.5 min, 90% B and re-equilibration for 4 min. Full-scan mass spectra were acquired in the range of m/z 100-1000 on an Agilent

6530 TOF with ESI source settings: voltage 3000 V, gas temperature 350 C and fragmentor 150 V. For each pair of target and control samples, all replicates were injected after the reference sample (compound alone). LC/MS chromatograms for specific ligands were extracted using MassHunter software (Agilent, USA) based on the accurate mass measurement with a deviation of 10 ppm and also an RT tolerance of 0.2 min compared with the reference compound. MS responses were represented by the integrated peak areas of the corresponding extracted ion chromatograms. The MS binding index (BI) referred to the ratio of MS response of a specific ligand detected in the A2AR nanodisc incubation sample relative to the control. From our past aﬃnity MS screening results, the positive hits were selected based on an average BI value >2 and relative standard deviation (RSD) <30% 532.

6.4 Results and discussion

Here we introduce an aﬃnity MS screening method for the A2AR nanodisc. An overview of the aﬃnity MS ligand screening workflow for the A2AR nanodisc is shown in Figure

6.1. 151

Figure 6.1: Schematic diagram of affinity MS ligand identification for the A2AR nanodisc.

The A2AR nanodisc preparation steps include: (1) A2AR purification; (2) A2AR purified proteins, phospholipids and MSP1D1 were mixed and incubated; (3) the A2AR nanodisc formed after detergent removal by adding Biobeads™. The affinity MS steps include: (1) the A2AR nanodisc incubated with the compound mixture; (2) unspecific binding washed off by ultrafiltration; (3) methanol dissociation; (4) lipids and protein removal by solid phase extraction; (5) binding index (BI) determination by LC/MS.

6.4.1 MSP1D1 and A2AR purifications

To validate the eﬀectiveness of the MS-based screening technique to the GPCR nanodisc, we chose the human A2AR nanodisc as a target. The MSP1D1 with a measured diameter of 9.8 nm was used in nanodisc reconstitution. The A2AR and MSP1D1 were purified as

600,630 previously described . The A2AR and MSP1D1 were examined by NuPAGE electrophoresis (Figure 6.2 (A)). The predicted molecular weights of A2AR and MSP1D1 were 50 kDa and 22.5 kDa, respectively. The elution fractions with the most proteins were concentrated and the final concentrations were determined by Bradford assay.

Usually, 1.5-2 mg of A2AR were purified from 1 L biomass. Further analysis of A2AR homogeneity was done by aSEC (Figure 6.2(B)). The aSEC result showed that 95% of

152

A2ARs were present as monomeric peaks compared to 5% of aggregation peaks. There was no compound added during the purification procedures.

Figure 6.2: (A) NuPAGE electrophoresis analysis of A2AR and MSP1D1-∆His6 purified proteins. M, W1, W2, W3, E1, E2 and E3 represent marker, 1st wash fraction, 2nd wash fraction, 3rd wash fraction, 1st elution fraction, 2nd elution fraction, and 3rd elution fraction,

nd respectively. Most A2AR proteins were eluted out in the 2 elution fraction; (B) aSEC profile of the A2AR in DDM buffer shows a homogeneous sample preparation.

6.4.2 A2AR nanodisc reconstitution

Several models are used to contain the GPCRs and facilitate the structural and functional elucidations, such as detergent micelles, mixed detergent/lipids micelles, bicelles, liposomes and nanodiscs 334. Among these models, nanodiscs have many advantages and are a versatile tool to maintain the GPCR in a native-like phospholipid bilayer environment. It not only controls the physical size of the nanodisc (macro-environment) by varying the lipid stoichiometry and MSP variant but also allows the adjustment of nanodisc lipid types (micro-environment). We chose a nanodisc constituted by a lipid mixture of POPC/POPS 382,431,432. Since POPS is a negatively charged lipid and POPC is a neutral zwitterionic lipid, a mixture of POPC/POPS was used to mimic the environment of the plasma membrane. The molar ratio of POPC and POPS was maintained as 7 : 3. 153

Figure 6.3: (A) The optimisation of reconstitution ratios of A2AR : MSP1D1-∆His6 : phospholipids. FPLC analyses the homogeneity and monodispersity of the reconstituted

A2AR nanodisc. The chromatographic peaks at 9 mL represent aggregation formed during nanodisc reconstitution, and the peaks at 12.5 mL represent the monomeric A2AR successfully reconstituted into the nanodisc; (B) the reconstituted A2AR nanodisc is examined by NuPAGE analysis. The two bands correspond to A2AR and MSP1D1-∆His6.

To incorporate the A2AR into the nanodisc, the ratio of A2AR, MSP1D1-∆His6 and lipids

(POPC/POPS) was systematically optimised. The empirical ratio of MSP1D1 : lipids for the empty nanodisc was 1 : 70 in previous experiments. However, the incorporation of the A2AR will lead to the replacement with some amount of lipids in the nanodisc.

Therefore the experimentally determined ratio of A2AR : MSP1D1 : lipids was 1 : 5 : 250 for generating a monodisperse and homogeneous nanodisc sample (Figure 6.3 (A)). The

NuPAGE gel indicated the successful incorporation of A2AR into the nanodisc (Figure

6.3 (B)). Ultrafiltration-based affinity mass spectrometry analysis for A2AR nanodisc ligand identification Ultrafiltration-based affinity LC/MS has been employed to screen small molecule ligands against various soluble protein targets 531,532,627,628 yet never

154

reported to be applied to any GPCR nanodisc. We have found that the variation of the detergent concentration in the sample preparation of purified GPCR proteins would affect ligand recovery from the ultrafiltration device (data not shown). In contrast, nanodiscs under the detergent-free conditions can better suit the need for ligand detection by ultrafiltration-based affinity LC/MS analysis.

Figure 6.4: (A) Total ion chromatograms (TIC) acquired through compounds from the mixture alone (green), A2AR nanodisc (red) and negative control (blue). (B) Extracted ion chromatograms (EIC) of the compounds from the A2AR nanodisc (solid line) and the negative control (dashed line), each chromatogram (from left to right) represents a compound with strong, weak, or nonspecific binding to the A2AR.

We prepared a 15-compound mixture containing ZM-241385, UK-432097, SCH-58261,

SCH-442416, NECA, adenosine, theophylline, caffeine and other seven unrelated compounds in 50% methanol. A final concentration of 0.4 mM of each compound was

155

mixed with 2 mM of the A2AR nanodisc to assay the ligand binding affinity to A2AR nanodisc by ultrafiltration-based affinity LC/MS. The inhibitory constants (Ki) and types of compounds were extracted from ChEMBL (https://www.ebi.ac.uk/chembl/compound) and BindingDB database (https://www.bindingdb.org/bind/index.jsp). The Ki values were used as an indicator for protein–ligand interaction activities. The binding index (BI) referring to the ratio of MS response of a given ligand detected in the A2AR nanodisc incubation sample versus the control (empty nanodisc) was used to assess the specific enrichment of the ligand associated with the A2AR (Figure 6.4). Previous studies have shown that ligands with BI values>2 are positive bindings to the target protein 532. As expected, all eight known ligands of the A2AR in the 15-compound mixture were verified in this assay and the rest of seven compounds unrelated to the A2AR all had BI values below the threshold (Table 6.1). More importantly, BI values for SCH-442416 (15.9 ±

4.3), UK-432097 (47.8 ± 4.9), SCH-58261 (24.8 ± 4.8), NECA (30.5 ± 3.9) and adenosine

(27.7 ± 4.9) with relatively strong affinity (Ki < 1 mM) were much higher than those of two weakly bound ligands theophylline (3.7 ± 1.0) and caffeine (3.25 ± 1.2) in the compound mixture (Ki > 1 mM). Although the ranking order of BI values is in general agreement with the binding affinity of different ligands, the exception does exist in some cases, such as for ZM-241385 (BI = 3.7 ± 0.1). Its low BI value might be related to the kinetics of the ligand association/dissociation with the receptor. Nevertheless, the current affinity LC/MS indicates that the nanodisc system coupled to the ultrafiltration-based affinity LC/MS technique is feasible and promising for ligand identification from compound mixtures towards specific GPCR targets in a high-throughput manner.

156

Table 6.1: Identification and binding index determination for each compound in a 15- compound mixture by ultrafiltration-based affinity LC/MS analysis.

BIa BIa BIa Retention Mass A2AR Ki Compound Rep-1 Rep-2 Rep-3 time accuracy ligandb (nM)b (min) (ppm) SCH-442416 15.1 12.1 20.6 14.9 0.64 Antagonist 0.05

ZM-241385 3.70 3.84 3.65 8.62 0.12 Antagonist 1.6

UK-432097 50.7 50.5 42.2 10.6 -1.19 Agonist 4.0

SCH-58261 19.4 26.3 28.6 14.3 -1.61 Antagonist 5.0

NECA 34.8 27.4 29.2 3.62 -0.99 Agonist 20

Adenosine 26.1 23.7 33.2 1.05 -0.88 Agonist 700

Theophylline 3.27 2.95 4.78 4.74 -1.82 Antagonist 1584

Caffeine 3.39 2.02 4.34 5.51 -0.90 Antagonist 2000

Negative-1 1.14 1.24 1.20 11.3 -0.64 NO NA

Negative-2 0.86 1.11 0.87 11.5 1.76 NO NA

Negative-3 0.83 1.26 1.39 13.0 1.03 NO NA

Negative-4 1.00 0.86 1.08 16.2 -0.50 NO NA

Negative-5 0.00 0.00 0.00 11.5 -0.32 NO NA

Negative-6 0.95 0.91 1.15 16.0 -0.17 NO NA

Negative-7 1.42 1.30 1.18 7.23 0.21 NO NA

aBI determined in the affinity MS assay from three independent replicates. bData from

ChEMBL (https://www.ebi.ac.uk/chembl/compound) and Binding DB databases

(https://www.bindingdb.org/bind/index.jsp).

157

6.5 Conclusions

The affinity LC/MS has been successfully applied to ligand screening against both soluble proteins and purified GPCR dissolved in detergents 531,532,627-629. This is the first report of combining nanodisc and affinity LC/MS techniques to identify ligands against unstable

GPCRs. We provide a protocol for a robust, quantitative and reliable ligand identification method by combining GPCR nanodisc and affinity LC/MS techniques. The reconstitution of GPCRs into nanodiscs would potentiate ligand screening efficiency in an automated affinity LC/MS platform due to better thermostability of the protein target. The affinity

LC/MS technique allows for sensitive and convenient detection of ligands in a wide range of affinities to GPCR targets, and can be readily set up without much instrumental modification for different GPCR target screening.

158

Chapter 7 General conclusions and future direction

We investigated two different ligand identification methods, i.e. native ESI-FT-ICR MS and ultrafiltration-based affinity LC/MS. The native ESI-FT-ICR MS was successfully used to detect different compositions of A2AR nanodisc such as POPC/POPS lipids and

MSP1D1/MSP2N2. However, the available instrument setup is more suitable to detect various soluble proteins or ligands instead of insoluble membrane proteins or GPCRs.

The most recent achievement by Robinson's lab is to directly identify the endogenous

614 ligand of ADP binding to P2Y1R by nESI-MS . The mass spectrometer used for direct

GPCR-ligand identification is internally modified and it was not feasible to reproduce in currently MS facility 616,617. The lipid signals from the GPCR nanodisc suppressed the

GPCR signals which are low in MS response (due to its low solubility).

Ultraﬁltration-based affinity LC/MS has been successfully applied to ligand screening against soluble proteins and this is the first report of combining the nanodisc technique with affinity LC/MS technique to screen ligands against unstable GPCRs. This affinity

LC/MS protocol provides a robust, quantitative and reliable ligand identification method by combining GPCR nanodisc and aﬃnity LC/MS techniques. The reconstitution of

GPCRs into nanodiscs would enable ligand screening in an automated aﬃnity LC/MS platform due to better thermostability of the protein target. Nanodisc maintains the GPCR in a near-native phospholipid bilayer environment. Nanodisc not only controls the physical size (macro-environment) by varying the MSP variants, but also allows the adjustment of lipids types (micro-environment). Nanodisc did not influence the accessibility of ligands into the ligand binding pocket from extracellular and intracellular interfaces.

The aﬃnity LC/MS technique allows for sensitive and convenient detection of ligands in a wide range of aﬃnities to GPCR targets and could be readily set up without much

159

instrumental modification for diﬀerent GPCR target screening. The MS-based ligand identification technique against GPCRs will facilitate the GPCR crystallisation, GPCR deorphanisation, and drug discovery in the future.

160

Bibliography

1. de Lau LML, Breteler MMB. Epidemiology of Parkinson's disease. Lancet Neurology 2006;5(6):525-35. 2. Dorsey ER, Bloem BR. The Parkinson pandemic - a call to action. Journal of the American Medical Association Neurology 2018;75(1):9-10. 3. Living with Parkinson's disease: An updated economic analysis. Deloitte Australia. July 2015. 4. Kalia LV, Lang AE. Parkinson's disease. Lancet 2015;386(9996):896-912. 5. Parkinson J. An essay on the shaking palsy. Journal of Neuropsychiatry and Clinical Neurosciences 2002;14(2):223-36. 6. Dauer W, Przedborski S. Parkinson's disease: Mechanisms and models. Neuron 2003;39(6):889-909. 7. Fearnley JM, Lees AJ. Ageing and Parkinson's disease: substantia nigra regional selectivity. Brain 1991;114(5):2283-301. 8. Gelb DJ, Oliver E, Gilman S. Diagnostic criteria for Parkinson disease. Archives of Neurology 1999;56(1):33-9. 9. Esteves AR, Swerdlow RH, Cardoso SM. LRRK2, a puzzling protein: insights into Parkinson's disease pathogenesis. Experimental Neurology 2014;261:206-16. 10. Liu G, Aliaga L, Cai H. α-synuclein, LRRK2 and their interplay in Parkinson's disease. Future Neurology 2012;7(2):145-53. 11. Covy JP, Giasson BI. α-synuclein, leucine-rich repeat kinase-2, and manganese in the pathogenesis of Parkinson disease. Neurotoxicology 2011;32(5):622-9. 12. Schapira AH. Future directions in the treatment of Parkinson's disease. Movement Disorders 2007;22(S17):S385-S91. 13. Simon-Sanchez J, Schulte C, Bras JM, Sharma M, Gibbs JR, Berg D, et al. Genome-wide association study reveals genetic risk underlying Parkinson's disease. Nature Genetics 2009;41(12):1308-12. 14. Pankratz N, Wilk JB, Latourelle JC, DeStefano AL, Halter C, Pugh EW, et al. Genomewide Association Study for Susceptibility Genes Contributing to Familial Parkinson Disease. Human Genetics 2009;124(6):593-605. 15. Collier TJ, Kanaan NM, Kordower JH. Ageing as a primary risk factor for Parkinson's disease: evidence from studies of non-human primates. Nature Reviews Neuroscience 2011;12(6):359-66. 16. Beal MF. Mitochondria, oxidative damage, and inflammation in Parkinson's

161

disease. Annals of the New York Academy of Sciences 2003;991(1):120-31. 17. Kochanski RB, Bus S, Pal G, Metman LV, Sani S. Optimization of microelectrode recording in deep brain stimulation surgery using intraoperative computed tomography. World Neurosurgery 2017;103(103):168-73. 18. Spillantini MG, Schmidt ML, Lee VM, Trojanowski JQ, Jakes R, Goedert M. α- synuclein in Lewy bodies. Nature 1997;388(6645):839-40. 19. Kitada T, Asakawa S, Hattori N, Matsumine H, Yamamura Y, Minoshima S, et al. Mutations in the parkin gene cause autosomal recessive juvenile parkinsonism. Nature 1998;392(6676):605-8. 20. Leroy E, Boyer R, Auburger G, Leube B, Ulm G, Mezey E, et al. The ubiquitin pathway in Parkinson's disease. Nature 1998;395(6701):451-2. 21. Funayama M, Hasegawa K, Ohta E, Kawashima N, Komiyama M, Kowa H, et al. An LRRK2 mutation as a cause for the Parkinsonism in the original PARK8 family. Annals of Neurology 2005;57(6):918-21. 22. Todorovic M, Wood SA, Mellick GD. Nrf2: A Modulator of Parkinson's Disease? Journal of Neural Transmission 2016;123(6):611-9. 23. Morelli M, Di Paolo T, Wardas J, Calon F, Xiao D, Schwarzschild MA. Role of

adenosine A2A receptors in parkinsonian motor impairment and L-DOPA-induced motor complications. Progress in Neurobiology 2007;83(5):293-309.

24. Ekonomou A, Poulou PD, Matsokis N, Angelatou F. Stimulation of adenosine A2A receptors elicits zif/268 and NMDA ϵ2 subunit mRNA expression in cortex and striatum of the “weaver” mutant mouse, a genetic model of nigrostriatal dopamine deficiency. Neuroscience 2004;123(4):1025-36. 25. Seeman P, Niznik HB. Dopamine receptors and transporters in Parkinson's disease and schizophrenia. The Federation of American Societies for Experimental Biology Journal 1990;4(10):2737-44. 26. Fox SH, Chuang R, Brotchie JM. Serotonin and Parkinson's disease: on movement, mood, and madness. Movement Disorders 2009;24(9):1255-66. 27. Zhu A, Kil K-E, Kura S, Gong C, Zhang Z, Choi J-K, et al. Modulation of presynaptic (mGluR4) and postsynaptic (mGluR5) metabotropic glutamate receptors in 6-OHDA induced Parkinson’s disease model. Journal of Nuclear Medicine 2013;54(S2):524-30. 28. Paisán-Ru1 z C, Jain S, Evans EW, Gilks WP, Simón J, van der Brug M, et al. Cloning of the gene containing mutations that cause PARK8-linked Parkinson's disease. Neuron 2004;44(4):595-600. 162

29. Zimprich A, Biskup S, Leitner P, Lichtner P, Farrer M, Lincoln S, et al. Mutations in LRRK2 cause autosomal-dominant Parkinsonism with pleomorphic pathology. Neuron 2004;44(4):601-7. 30. Galter D, Westerlund M, Carmine A, Lindqvist E, Sydow O, Olson L. LRRK2 expression linked to dopamine-innervated areas. Annals of Neurology 2006;59(4):714-9. 31. Thévenet J, Pescini Gobert R, Hooft van Huijsduijnen R, Wiessner C, Sagot YJ. Regulation of LRRK2 expression points to a functional role in human monocyte maturation. Public Library of Science One 2011;6(6):e21519-e32. 32. Deng J, Lewis PA, Greggio E, Sluch E, Beilina A, Cookson MR. Structure of the ROC domain from the Parkinson's disease-associated leucine-rich repeat kinase 2 reveals a dimeric GTPase. Proceedings of the National Academy of Sciences 2008;105(5):1499-504. 33. Herzig MC, Kolly C, Persohn E, Theil D, Schweizer T, Hafner T, et al. LRRK2 protein levels are determined by kinase function and are crucial for kidney and lung homeostasis in mice. Human Molecular Genetics 2011;20(21):4209-23. 34. Ness D, Ren Z, Gardai S, Sharpnack D, Johnson VJ, Brennan RJ, et al. Leucine- rich repeat kinase 2 (LRRK2)-deficient rats exhibit renal tubule injury and perturbations in metabolic and immunological homeostasis. Public Library of Science One 2013;8(6):e66164-e75. 35. Tong Y, Giaime E, Yamaguchi H, Ichimura T, Liu Y, Si H, et al. Loss of leucine- rich repeat kinase 2 causes age-dependent bi-phasic alterations of the autophagy pathway. Molecular Neurodegeneration 2012;7(1):2-17. 36. Marín I. The Parkinson Disease Gene LRRK2: Evolutionary and Structural Insights. Molecular Biology and Evolution 2006;23(12):2423-33. 37. Greggio E, Zambrano I, Kaganovich A, Beilina A, Taymans JM, Daniels V, et al. The Parkinson disease-associated leucine-rich repeat kinase 2 (LRRK2) is a dimer that undergoes intramolecular autophosphorylation. Journal of Biological Chemistry 2008;283(24):16906-14. 38. Piccoli G, Onofri F, Cirnaru MD, Kaiser CJO, Jagtap P, Kastenmüller A, et al. Leucine-Rich Repeat Kinase 2 Binds to Neuronal Vesicles through Protein Interactions Mediated by Its C-Terminal WD40 Domain. Molecular and Cellular Biology 2014;34(12):2147-61. 39. Gilks WP, Abou-Sleiman PM, Gandhi S, Jain S, Singleton A, Lees AJ, et al. A common LRRK2 mutation in idiopathic Parkinson's disease. Lancet 163

2005;365(9457):415-6. 40. Di Fonzo A, Rohe CF, Ferreira J, Chien HF, Vacca L, Stocchi F, et al. A frequent LRRK2 gene mutation associated with autosomal dominant Parkinson's disease. Lancet 2005;365(9457):412-5. 41. Bonifati V, Rizzu P, van Baren MJ, Schaap O, Breedveld GJ, Krieger E, et al. Mutations in the DJ-1 gene associated with autosomal recessive early-onset Parkinsonism. Science 2003;299(5604):256-9. 42. Valente EM, Abou-Sleiman PM, Caputo V, Muqit MM, Harvey K, Gispert S, et al. Hereditary early-onset Parkinson's disease caused by mutations in PINK1. Science 2004;304(5674):1158-60. 43. Satake W, Nakabayashi Y, Mizuta I, Hirota Y, Ito C, Kubo M, et al. Genome-wide association study identifies common variants at four loci as genetic risk factors for Parkinson's disease. Nature Genetics 2009;41(12):1303-7. 44. Huang Y, Halliday GM, Vandebona H, Mellick GD, Mastaglia F, Stevens J, et al. Prevalence and clinical features of common LRRK2 mutations in Australians with Parkinson's Disease. Movement Disorders 2007;22(7):982-9. 45. Nuytemans K, Theuns J, Cruts M, Van Broeckhoven C. Genetic etiology of Parkinson disease associated with mutations in the SNCA, PARK2, PINK1, PARK7, and LRRK2 genes: a mutation update. Human Mutation 2010;31(7):763- 80. 46. Thaler A, Ash E, Gan-Or Z, Orr-Urtreger A, Giladi N. The LRRK2 G2019S mutation as the cause of Parkinson's disease in Ashkenazi Jews. Journal of Neural Transmission 2009;116(11):1473-82. 47. Klein C, Westenberger A. Genetics of Parkinson’s disease. Cold Spring Harbor Perspectives in Medicine 2012;2(1):a008888-a903. 48. Hasegawa K, Stoessl AJ, Yokoyama T, Kowa H, Wszolek ZK, Yagishita S. Familial parkinsonism: study of original Sagamihara PARK8 (I2020T) kindred with variable clinicopathologic outcomes. Parkinsonism & Related Disorders 2009;15(4):300-6. 49. Wszolek ZK, Pfeiffer RF, Tsuboi Y, Uitti RJ, McComb RD, Stoessl AJ, et al. Autosomal dominant parkinsonism associated with variable synuclein and tau pathology. Neurology 2004;62(9):1619-22. 50. Daniëls V, Vancraenenbroeck R, Law BMH, Greggio E, Lobbestael E, Gao F, et al. Insight Into The Mode of Action of The LRRK2 Y1699C Pathogenic Mutant. Journal of Neurochemistry 2011;116(2):304-15. 164

51. Li JQ, Tan L, Yu JT. The role of the LRRK2 gene in parkinsonism. Molecular Neurodegeneration 2014;9(1):47-63. 52. Bardien S, Lesage S, Brice A, Carr J. Genetic characteristics of leucine-rich repeat kinase 2 (LRRK2) associated Parkinson’s disease. Parkinsonism & Related Disorders 2011;17(7):501-8. 53. Nolen B, Taylor S, Ghosh G. Regulation of protein kinases: controlling activity through activation segment conformation. Molecular Cell 2004;15(5):661-75. 54. Kachergus J, Mata IF, Hulihan M, Taylor JP, Lincoln S, Aasly J, et al. Identification of a novel LRRK2 mutation linked to autosomal dominant parkinsonism: evidence of a common founder across European populations. American Journal of Human Genetics 2005;76(4):672-80. 55. West AB, Moore DJ, Biskup S, Bugayenko A, Smith WW, Ross CA, et al. Parkinson's disease-associated mutations in leucine-rich repeat kinase 2 augment kinase activity. Proceedings of the National Academy of Sciences 2005;102(46):16842-7. 56. Luzón-Toro B, de la Torre ER, Delgado A, Pérez-Tur J, Hilfiker S. Mechanistic insight into the dominant mode of the Parkinson's disease-associated G2019S LRRK2 mutation. Human Molecular Genetics 2007;16(17):2031-9. 57. Anand VS, Reichling LJ, Lipinski K, Stochaj W, Duan W, Kelleher K, et al. Investigation of leucine-rich repeat kinase 2: enzymological properties and novel assays. The Federation of European Biochemical Societies Journal 2009;276(2):466-78. 58. MacLeod D, Dowman J, Hammond R, Leete T, Inoue K, Abeliovich A. The familial Parkinsonism gene LRRK2 regulates neurite process morphology. Neuron 2006;52(4):587-93. 59. Heo HY, Park JM, Kim CH, Han BS, Kim KS, Seol W. LRRK2 enhances oxidative stress-induced neurotoxicity via its kinase activity. Experimental Cell Research 2010;316(4):649-56. 60. Iaccarino C, Crosio C, Vitale C, Sanna G, Carri MT, Barone P. Apoptotic mechanisms in mutant LRRK2-mediated cell death. Human Molecular Genetics 2007;16(11):1319-26. 61. Plowey ED, Cherra SJ, 3rd, Liu YJ, Chu CT. Role of autophagy in G2019S- LRRK2-associated neurite shortening in differentiated SH-SY5Y cells. Journal of Neurochemistry 2008;105(3):1048-56. 62. Smith WW, Pei Z, Jiang H, Moore DJ, Liang Y, West AB, et al. Leucine-rich 165

repeat kinase 2 (LRRK2) interacts with parkin, and mutant LRRK2 induces neuronal degeneration. Proceedings of the National Academy of Sciences 2005;102(51):18676-81. 63. Dusonchet J, Kochubey O, Stafa K, Young SM, Jr., Zufferey R, Moore DJ, et al. A rat model of progressive nigral neurodegeneration induced by the Parkinson's disease-associated G2019S mutation in LRRK2. Journal of Neuroscience 2011;31(3):907-12. 64. Hatano T, Kubo S, Imai S, Maeda M, Ishikawa K, Mizuno Y, et al. Leucine-rich repeat kinase 2 associates with lipid rafts. Human Molecular Genetics 2007;16(6):678-90. 65. McKay MM, Morrison DK. Integrating signals from RTKs to ERK/MAPK. Oncogene 2007;26(22):3113-21. 66. West AB, Moore DJ, Choi C, Andrabi SA, Li X, Dikeman D, et al. Parkinson's disease-associated mutations in LRRK2 link enhanced GTP-binding and kinase activities to neuronal toxicity. Human Molecular Genetics 2007;16(2):223-32. 67. Korr D, Toschi L, Donner P, Pohlenz HD, Kreft B, Weiss B. LRRK1 protein kinase activity is stimulated upon binding of GTP to its Roc domain. Cellular Signalling 2006;18(6):910-20. 68. Manning G, Whyte DB, Martinez R, Hunter T, Sudarsanam S. The protein kinase complement of the human genome. Science 2002;298(5600):1912-34. 69. Guo L, Wang W, Chen SG. Leucine-rich repeat kinase 2: relevance to Parkinson's disease. International Journal of Biochemistry & Cell Biology 2006;38(9):1469- 75. 70. Johnson GL, Lapadat R. Mitogen-activated protein kinase pathways mediated by ERK, JNK, and p38 protein kinases. Science 2002;298(5600):1911-2. 71. Scott JD, DeMong DE, Greshock TJ, Basu K, Dai X, Harris J, et al. Discovery of a 3-(4-Pyrimidinyl) Indazole (MLi-2), an Orally Available and Selective Leucine- Rich Repeat Kinase 2 (LRRK2) Inhibitor that Reduces Brain Kinase Activity. Journal of Medicinal Chemistry 2017;60(7):2983-92. 72. Gilsbach BK, Kortholt A. Structural biology of the LRRK2 GTPase and kinase domains: implications for regulation. Frontiers in Molecular Neuroscience 2014;7(1):32-40. 73. Gilsbach BK, Messias AC, Ito G, Sattler M, Alessi DR, Wittinghofer A, et al. Structural characterization of LRRK2 inhibitors. Journal of Medicinal Chemistry 2015;58(9):3751-6. 166

74. Gilsbach BK, Ho FY, Vetter IR, van Haastert PJM, Wittinghofer A, Kortholt A. Roco kinase structures give insights into the mechanism of Parkinson disease- related leucine-rich-repeat kinase 2 mutations. Proceedings of the National Academy of Sciences 2012;109(26):10322-7. 75. Ray S, Liu M. Current understanding of LRRK2 in Parkinson's disease: biochemical and structural features and inhibitor design. Future Medicinal Chemistry 2012;4(13):1701-13. 76. Ray S, Bender S, Kang S, Lin R, Glicksman MA, Liu M. The Parkinson Disease- linked LRRK2 Protein Mutation I2020T Stabilizes an Active State Conformation Leading to Increased Kinase Activity. Journal of Biological Chemistry 2014;289(19):13042-53. 77. Liu M, Bender SA, Cuny GD, Sherman W, Glicksman M, Ray SS. Type II kinase inhibitors show an unexpected inhibition mode against Parkinson’s disease-linked LRRK2 mutant G2019S. Biochemistry 2013;52(10):1725-36. 78. Saraste M, Sibbald PR, Wittinghofer A. The P-loop — a common motif in ATP- and GTP-binding proteins. Trends in Biochemical Sciences 1990;15(11):430-4. 79. Vigil D, Cherfils J, Rossman KL, Der CJ. Ras superfamily GEFs and GAPs: validated and tractable targets for cancer therapy? Nature Reviews Cancer 2010;10(12):842-57. 80. Takai Y, Sasaki T, Matozaki T. Small GTP-binding proteins. Physiological Reviews 2001;81(1):153-208. 81. Ito G, Okai T, Fujino G, Takeda K, Ichijo H, Katada T, et al. GTP binding is essential to the protein kinase activity of LRRK2, a causative gene product for familial Parkinson's disease. Biochemistry 2007;46(5):1380-8. 82. Boon JY, Dusonchet J, Trengrove C, Wolozin B. Interaction of LRRK2 with kinase and GTPase signaling cascades. Frontiers in Molecular Neuroscience 2014;7(64):1-8. 83. Smith WW, Pei Z, Jiang H, Dawson VL, Dawson TM, Ross CA. Kinase activity of mutant LRRK2 mediates neuronal toxicity. Nature Neuroscience 2006;9(10):1231-3. 84. Stafa K, Trancikova A, Webber PJ, Glauser L, West AB, Moore DJ. GTPase activity and neuronal toxicity of Parkinson's disease-associated LRRK2 is regulated by ArfGAP1. Public Library of Science Genetics 2012;8(2):e1002526- e50. 85. Hsu CH, Chan D, Greggio E, Saha S, Guillily MD, Ferree A, et al. MKK6 binds 167

and regulates expression of Parkinson's disease-related protein LRRK2. Journal of Neurochemistry 2010;112(6):1593-604. 86. Caesar M, Zach S, Carlson CB, Brockmann K, Gasser T, Gillardon F. Leucine- rich repeat kinase 2 functionally interacts with microtubules and kinase- dependently modulates cell migration. Neurobiology of Disease 2013;54(1):280- 8. 87. Chan D, Citro A, Cordy JM, Shen GC, Wolozin B. Rac1 protein rescues Neurite retraction caused by G2019S leucine-rich repeat kinase 2 (LRRK2). Journal of Biological Chemistry 2011;286(18):16140-9. 88. Law BM, Spain VA, Leinster VH, Chia R, Beilina A, Cho HJ, et al. A direct interaction between leucine-rich repeat kinase 2 and specific β-tubulin isoforms regulates tubulin acetylation. Journal of Biological Chemistry 2014;289(2):895- 908. 89. Jasinska-Myga B, Kachergus J, Vilarino-Guell C, Wider C, Soto-Ortolaza AI, Kefi M, et al. Comprehensive sequencing of the LRRK2 gene in patients with familial Parkinson's disease from North Africa. Movement Disorders 2010;25(13):2052-8. 90. Rubio JP, Topp S, Warren L, St Jean PL, Wegmann D, Kessner D, et al. Deep sequencing of the LRRK2 gene in 14,002 individuals reveals evidence of purifying selection and independent origin of the p.Arg1628Pro mutation in Europe. Human Mutation 2012;33(7):1087-98. 91. Di Stefano A, Sozio P, Cerasa LS, Iannitelli A. L-Dopa prodrugs: an overview of trends for improving Parkinson's disease treatment. Current Pharmaceutical Design 2011;17(32):3482-93. 92. Vancraenenbroeck R, Lobbestael E, Maeyer MD, Baekelandt V, Taymans JM. Kinases as targets for Parkinson's disease: from genetics to therapy. CNS & Neurological Disorders Drug Targets 2011;10(6):724-40. 93. Liu Z, Hamamichi S, Dae Lee B, Yang D, Ray A, Caldwell GA, et al. Inhibitors of LRRK2 kinase attenuate neurodegeneration and Parkinson-like phenotypes in Caenorhabditis elegans and Drosophila Parkinson's disease models. Human Molecular Genetics 2011;20(20):3933-42. 94. Ian W, Herbie N. Do molecularly targeted agents in oncology have reduced attrition rates? Nature Reviews Drug Discovery 2008;8(1):15-6. 95. Calabresi P, Di Filippo M, Gallina A, Wang Y, Stankowski JN, Picconi B, et al. New synaptic and molecular targets for neuroprotection in Parkinson’s disease. Movement Disorders 2013;28(1):51-60. 168

96. Knight ZA, Shokat KM. Features of selective kinase inhibitors. Chemistry & Biology 2005;12(6):621-37. 97. Spitaler M, Villunger A, Grunicke H, Überall F. Unique structural and functional properties of the ATP-binding domain of atypical protein kinase C-ι. Journal of Biological Chemistry 2000;275(43):33289-96. 98. Klein CL, Rovelli G, Springer W, Schall C, Gasser T, Kahle PJ. Homo- and heterodimerization of ROCO kinases: LRRK2 kinase inhibition by the LRRK2 ROCO fragment. Journal of Neurochemistry 2009;111(3):703-15. 99. Dachsel JC, Taylor JP, Mok SS, Ross OA, Hinkle KM, Bailey RM, et al. Identification of potential protein interactors of LRRK2. Parkinsonism & Related Disorders 2007;13(7):382-5. 100. Ho FY, Rosenbusch KE, Kortholt A. The Potential of Targeting LRRK2 in Parkinson’s Disease. In: Rana AQ, editor. A Synopsis of Parkinson's Disease: InTech; 2014. p. 230. 101. Guaitoli G, Raimondi F, Gilsbach BK, Gómez-Llorente Y, Deyaert E, Renzi F, et al. Structural model of the dimeric Parkinson’s protein LRRK2 reveals a compact architecture involving distant interdomain contacts. Proceedings of the National Academy of Sciences 2016;113(30):e4357-e66. 102. Rudenko I, Chia R, Cookson M. Is inhibition of kinase activity the only therapeutic strategy for LRRK2-associated Parkinson's disease? BioMed Central Medicine 2012;10(1):20-8. 103. Garuti L, Roberti M, Bottegoni G. Non-ATP competitive protein kinase inhibitors. Current Medicinal Chemistry 2010;17(25):2804-21. 104. Feng Y, Chambers JW, Iqbal S, Koenig M, Park H, Cherry L, et al. A small molecule bidentate-binding dual inhibitor probe of the LRRK2 and JNK kinases. American Chemical Society Chemical Biology 2013;8(8):1747-54. 105. Jaleel M, Nichols RJ, Deak M, Campbell DG, Gillardon F, Knebel A, et al. LRRK2 phosphorylates moesin at threonine-558: characterization of how Parkinson's disease mutants affect kinase activity. The Biochemical Journal 2007;405(2):307-17. 106. Imai Y, Gehrke S, Wang HQ, Takahashi R, Hasegawa K, Oota E, et al. Phosphorylation of 4E-BP by LRRK2 affects the maintenance of dopaminergic neurons in Drosophila. The European Molecular Biology Organization Journal 2008;27(18):2432-43. 107. Gillardon F. Leucine-rich repeat kinase 2 phosphorylates brain tubulin-β isoforms 169

and modulates microtubule stability - a point of convergence in parkinsonian neurodegeneration? Journal of Neurochemistry 2009;110(5):1514-22. 108. Gloeckner CJ, Schumacher A, Boldt K, Ueffing M. The Parkinson disease- associated protein kinase LRRK2 exhibits MAPKKK activity and phosphorylates MKK3/6 and MKK4/7, in vitro. Journal of Neurochemistry 2009;109(4):959-68. 109. Lv X, Liu J, Shi Q, Tan Q, Wu D, Skinner JJ, et al. In vitro expression and analysis of the 826 human G protein-coupled receptors. Protein & Cell 2016;7(5):325-37. 110. Harmar AJ, Hills RA, Rosser EM, Jones M, Buneman OP, Dunbar DR, et al. IUPHAR-DB: the IUPHAR database of G protein-coupled receptors and ion channels. Nucleic Acids Research 2009;37(Database Issue):d680-d5. 111. Zhao J, Deng Y, Jiang Z, Qing H. G protein-coupled receptors (GPCRs) in Alzheimer’s disease: a focus on BACE1 related GPCRs. Frontiers in Aging Neuroscience 2016;8(1):58-72. 112. Cvicek V, Goddard WA, III, Abrol R. Structure-based sequence alignment of the transmembrane domains of all human GPCRs: phylogenetic, structural and functional implications. Public Library of Science Computational Biology 2016;12(3):e1004805-e35. 113. Birnbaumer L, Abramowitz J, Brown AM. Receptor-effector coupling by G proteins. Biochimica et Biophysica Acta 1990;1031(2):163-224. 114. Fang Y, Kenakin T, Liu C. Orphan GPCRs as emerging drug targets. Frontiers in Pharmacology 2015;6(295):1-2. 115. Choi HY, Saha SK, Kim K, Kim S, Yang G-M, Kim B, et al. G protein-coupled receptors in stem cell maintenance and somatic reprogramming to pluripotent or cancer stem cells. Biochemistry and Molecular Biology Reports 2015;48(2):68- 80. 116. Salazar NC, Chen J, Rockman HA. Cardiac GPCRs: GPCR signaling in healthy and failing hearts. Biochimica et Biophysica Acta 2007;1768(4):1006-18. 117. Capote LA, Mendez Perez R, Lymperopoulos A. GPCR signaling and cardiac function. The European Journal of Pharmacology 2015;763, Part B:143-8. 118. Fernandez-Patron C, Filep JG. GPCRs in cardiovascular pathologies. Drug Discovery Today. Disease Mechanisms 2012;9(3):e75-e8. 119. Sun L, Ye RD. Role of G protein-coupled receptors in inflammation. Acta Pharmacologica Sinica 2012;33(3):342-50. 120. Cash JL, Norling LV, Perretti M. Resolution of inflammation: targeting GPCRs that interact with lipids and peptides. Drug Discovery Today 2014;19(8):1186-92. 170

121. Packiriswamy N, Parameswaran N. G-protein coupled receptor kinases in inflammation and disease. Genes and Immunity 2015;16(6):367-77. 122. Harrison C. G protein-coupled receptors: a double attack on pain. Nature Reviews Drug Discovery 2013;12(9):665. 123. Stone LS, Molliver DC. In search of analgesia: emerging poles of GPCRs in pain. Molecular Interventions 2009;9(5):234-51. 124. Ahmad S, Dray A. Novel G protein-coupled receptors as pain targets. Current Opinion in Investigational Drugs 2004;5(1):67-70. 125. Bar-Shavit R, Maoz M, Kancharla A, Nag JK, Agranovich D, Grisaru-Granovsky S, et al. G protein-coupled receptors in cancer. International Journal of Molecular Sciences 2016;17(8):1320-35. 126. Liu Y, An S, Ward R, Yang Y, Guo XX, Li W, et al. G protein-coupled receptors as promising cancer targets. Cancer Letters 2016;376(2):226-39. 127. Dorsam RT, Gutkind JS. G protein-coupled receptors and cancer. Nature Reviews Cancer 2007;7(2):79-94. 128. Lappano R, Maggiolini M. G protein-coupled receptors: novel targets for drug discovery in cancer. Nature Reviews Drug Discovery 2011;10(1):47-60. 129. Reimann F, Gribble FM. G protein-coupled receptors as new therapeutic targets for type 2 diabetes. Diabetologia 2016;59:229-33. 130. Oh DY, Olefsky JM. G protein-coupled receptors as targets for anti-diabetic therapeutics. Nature Reviews Drug Discovery 2016;15(3):161-72. 131. Madiraju SRM, Poitout V. GPCRs and insulin secretion: 119 and counting. Endocrinology 2007;148(6):2598-600.

132. Cristalli G, Lambertucci C, Marucci G, Volpini R, Dal Ben D. A2A adenosine receptor and its modulators: overview on a druggable GPCR and on structure- activity relationship analysis and binding requirements of agonists and antagonists. Current Pharmaceutical Design 2008;14(15):1525-52. 133. Moreno JL, Holloway T, González-Maeso J. G protein-coupled receptor heterocomplexes in neuropsychiatric disorders. Progress in Molecular Biology and Translational Science 2013;117:187-205. 134. Proia RL, Hla T. Emerging biology of sphingosine-1-phosphate: its role in pathogenesis and therapy. Journal of Clinical Investigation 2015;125(4):1379-87. 135. Makide K, Uwamizu A, Shinjo Y, Ishiguro J, Okutani M, Inoue A, et al. Novel lysophosphoplipid receptors: their structure and function. Journal of Lipid Research 2014;55(10):1986-95. 171

136. Choi JW, Chun J. Lysophospholipids and their receptors in the central nervous system. Biochimica et Biophysica Acta 2013;1831(1):20-32. 137. Insel PA, Tang CM, Hahntow I, Michel MC. Impact of GPCRs in clinical medicine: monogenic diseases, genetic variants and drug targets. Biochimica et Biophysica Acta 2007;1768(4):994-1005. 138. Julius D, Nathans J. Signaling by sensory receptors. Cold Spring Harbor Perspectives in Biology 2012;4(1):a005991-a6005. 139. Rodrigues-Ferreira S, Nahmias C. G-protein coupled receptors of the renin- angiotensin system: new targets against breast cancer? Frontiers in Pharmacology 2015;6(24):1-7. 140. Bennett LD, Fox JM, Signoret N. Mechanisms regulating chemokine receptor activity. Immunology 2011;134(3):246-56. 141. Stone MJ, Hayward JA, Huang C, E. Huma Z, Sanchez J. Mechanisms of regulation of the chemokine-receptor network. International Journal of Molecular Sciences 2017;18(2):342-74. 142. Hazell GGJ, Hindmarch CC, Pope GR, Roper JA, Lightman SL, Murphy D, et al. G protein-coupled receptors in the hypothalamic paraventricular and supraoptic nuclei – serpentine gateways to neuroendocrine homeostasis. Frontiers in Neuroendocrinology 2012;33(1):45-66. 143. Michino M, Beuming T, Donthamsetti P, Newman AH, Javitch JA, Shi L. What can crystal structures of aminergic receptors tell us about designing subtype- selective ligands? Pharmacological Reviews 2015;67(1):198-213. 144. Cornil CA, Ball GF. Interplay among catecholamine systems: dopamine binds to

α2-adrenergic receptors in birds and mammals. Journal of Comparative Neurology 2008;511(5):610-27. 145. Miller LJ, Sexton PM, Dong M, Harikumar KG. The class B G-protein-coupled GLP-1 receptor: an important target for the treatment of type-2 diabetes mellitus. International Journal of Obesity Supplements 2014;4(Suppl 1):S9-S13. 146. Song G, Yang D, Wang Y, de Graaf C, Zhou Q, Jiang S, et al. Human GLP-1 receptor transmembrane domain structure in complex with allosteric modulators. Nature 2017;546(7657):312-5.

147. Howlett AC, Blume LC, Dalton GD. CB1 cannabinoid receptors and their associated proteins. Current Medicinal Chemistry 2010;17(14):1382-93. 148. Sánchez-Blázquez P, Rodríguez-Muñoz M, Garzón J. The cannabinoid receptor 1 associates with NMDA receptors to produce glutamatergic hypofunction: 172

implications in psychosis and schizophrenia. Frontiers in Pharmacology 2013;4(169):1-10. 149. Hua T, Vemuri K, Nikas SP, Laprairie RB, Wu Y, Qu L, et al. Crystal structures of agonist-bound human cannabinoid receptor CB1. Nature 2017;547(7664):468-71. 150. Hua T, Vemuri K, Pu M, Qu L, Han GW, Wu Y, et al. Crystal Structure of the

Human Cannabinoid Receptor CB1. Cell 2016;167(3):750-62. 151. Fredriksson R, Lagerström MC, Lundin L-G, Schiöth HB. The G-Protein- Coupled Receptors in the Human Genome Form Five Main Families. Phylogenetic Analysis, Paralogon Groups, and Fingerprints. Molecular Pharmacology 2003;63(6):1256-72. 152. Christopoulos A. Allosteric binding sites on cell-surface receptors: novel targets for drug discovery. Nature Reviews Drug Discovery 2002;1(3):198-210. 153. Gether U. Uncovering Molecular Mechanisms Involved in Activation of G Protein-Coupled Receptors. Endocrine Reviews 2000;21(1):90-113. 154. Cherezov V, Abola E, Stevens RC. Toward drug design: recent progress in the structure determination of GPCRs, a membrane protein family with high potential as pharmaceutical targets. Methods in Molecular Biology 2010;654:141-68. 155. Warne T, Edwards Patricia C, Leslie Andrew G, Tate Christopher G. Crystal

Structures of a Stabilized β1-Adrenoceptor Bound to the Biased Agonists Bucindolol and Carvedilol. Structure 2012;20-540(5-7):841-9. 156. Kobilka BK, Deupi X. Conformational complexity of G-protein-coupled receptors. Trends in Pharmacological Sciences 2007;28(8):397-406. 157. Brink CB, Harvey BH, Bodenstein J, Venter DP, Oliver DW. Recent advances in drug action and therapeutics: relevance of novel concepts in G-protein-coupled receptor and signal transduction pharmacology. British Journal of Clinical Pharmacology 2004;57(4):373-87. 158. Kobilka BK. G protein coupled receptor structure and activation. Biochimica et Biophysica Acta - Biomembranes 2007;1768(4):794-807. 159. Mystek P, Tworzydło M, Dziedzicka-Wasylewska M, Polit A. New insights into

the model of dopamine D1 receptor and G-proteins interactions. Biochimica et Biophysica Acta - Molecular Cell Research 2015;1853(3):594-603. 160. Tilley DG. G protein-dependent and G protein-independent signaling pathways and their impact on cardiac function. Circulation Research 2011;109(2):217-30. 161. Giulietti M, Vivenzio V, Piva F, Principato G, Bellantuono C, Nardi B. How much do we know about the coupling of G-proteins to serotonin receptors? Molecular 173

Brain 2014;7:49. 162. Billington CK, Penn RB. Signaling and regulation of G protein-coupled receptors in airway smooth muscle. Respiratory Research 2003;4(1):2-24. 163. Woulfe DS. REVIEW ARTICLES: Platelet G protein-coupled receptors in hemostasis and thrombosis. Journal of Thrombosis and Haemostasis 2005;3(10):2193-200.

164. Nguyen TD, Meichle S, Kim US, Wong T, Moody MW. P2Y11, a purinergic receptor acting via cAMP, mediates secretion by pancreatic duct epithelial cells. American Journal of Physiology - Gastrointestinal and Liver Physiology 2001;280(5):G795-G804. 165. Lohse MJ, Hoffmann C. Arrestin interactions with G protein-coupled receptors. Handbook of Experimental Pharmacology 2014;219:15-56. 166. Miller WE, Houtz DA, Nelson CD, Kolattukudy PE, Lefkowitz RJ. G-protein- coupled receptor (GPCR) kinase phosphorylation and β-arrestin recruitment regulate the constitutive signaling activity of the human cytomegalovirus US28 GPCR. Journal of Biological Chemistry 2003;278(24):21663-71. 167. Kang DS, Tian X, Benovic JL. Role of β-arrestins and arrestin domain-containing proteins in G protein-coupled receptor trafficking. Current Opinion in Cell Biology 2014;27:63-71. 168. Goodman OB, Krupnick JG, Santini F, Gurevich VV, Penn RB, Gagnon AW, et al.

β-Arrestin acts as a clathrin adaptor in endocytosis of the β2-adrenergic receptor. Nature 1996;383(6599):447-50. 169. Reiter E, Ahn S, Shukla AK, Lefkowitz RJ. Molecular mechanism of β-arrestin- biased agonism at seven-transmembrane receptors. Annual Review of Pharmacology and Toxicology 2012;52:179-97. 170. Luttrell LM, Ferguson SS, Daaka Y, Miller WE, Maudsley S, Della Rocca GJ, et

al. β-arrestin-dependent formation of β2 adrenergic receptor-Src protein kinase complexes. Science 1999;283(5402):655-61. 171. Whalen EJ, Rajagopal S, Lefkowitz RJ. Therapeutic potential of β-arrestin- and G protein-biased agonists. Trends in Molecular Medicine 2011;17(3):126-39. 172. Violin JD, Lefkowitz RJ. β-arrestin-biased ligands at seven-transmembrane receptors. Trends in Pharmacological Sciences 2007;28(8):416-22. 173. Hodavance SY, Gareri C, Torok RD, Rockman HA. G protein-coupled receptor biased agonism. Journal of Cardiovascular Pharmacology 2016;67(3):193-202. 174. Roy I, Getschman AE, Volkman BF, Dwinell MB. Exploiting agonist biased 174

signaling of chemokines to target cancer. Molecular Carcinogenesis 2017;56(3):804-13. 175. Geppetti P, Veldhuis Nicholas A, Lieu T, Bunnett Nigel W. G protein-coupled receptors: dynamic machines for signaling pain and itch. Neuron 2015;88(4):635- 49. 176. Shukla AK, Singh G, Ghosh E. Emerging structural insights into biased GPCR signaling. Trends in Biochemical Sciences 2014;39(12):594-602. 177. Rankovic Z, Brust TF, Bohn LM. Biased agonism: An emerging paradigm in GPCR drug discovery. Bioorganic & Medicinal Chemistry Letters 2016;26(2):241-50. 178. Urban JD, Clarke WP, von Zastrow M, Nichols DE, Kobilka B, Weinstein H, et al. Functional selectivity and classical concepts of quantitative pharmacology. Journal of Pharmacology and Experimental Therapeutics 2007;320(1):1-13. 179. Berg KA, Maayani S, Goldfarb J, Scaramellini C, Leff P, Clarke WP. Effector pathway-dependent relative efficacy at serotonin type 2A and 2C receptors: evidence for agonist-directed trafficking of receptor stimulus. Molecular Pharmacology 1998;54(1):94-104. 180. Felder CC, Kanterman RY, Ma AL, Axelrod J. Serotonin stimulates phospholipase

A2 and the release of arachidonic acid in hippocampal neurons by a type 2 serotonin receptor that is independent of inositolphospholipid hydrolysis. Proceedings of the National Academy of Sciences 1990;87(6):2187-91.

181. Lopez-Gimenez JF, Gonzalez-Maeso J. Hallucinogens and serotonin 5-HT2A receptor-mediated signaling pathways. Current Topics in Behavioral Neurosciences 2018;36:45-73. 182. Halberstadt AL, Geyer MA. Effect of hallucinogens on unconditioned behavior. Current Topics in Behavioral Neurosciences 2018;36:159-99. 183. Nichols DE, Nichols CD. Serotonin Receptors. Chemical Reviews 2008;108(5):1614-41. 184. Moreno JL, Holloway T, Rayannavar V, Sealfon SC, González-Maeso J. Chronic

treatment with LY341495 decreases 5-HT2A receptor binding and hallucinogenic effects of LSD in mice. Neuroscience Letters 2013;536:69-73. 185. Gonzalez-Maeso J, Weisstaub NV, Zhou M, Chan P, Ivic L, Ang R, et al.

Hallucinogens recruit specific cortical 5-HT2A receptor-mediated signaling pathways to affect behavior. Neuron 2007;53(3):439-52. 186. Geyer MA, Vollenweider FX. Serotonin research: contributions to understanding 175

psychoses. Trends in Pharmacological Sciences 2008;29(9):445-53.

187. Berg KA, Maayani S, Goldfarb J, Clarke WP. Pleiotropic behavior of 5-HT2A and

5-HT2C receptor agonists. Annals of the New York Academy of Sciences 1998;861:104-10. 188. Kurrasch-Orbaugh DM, Watts VJ, Barker EL, Nichols DE. Serotonin 5-

hydroxytryptamine2A receptor-coupled phospholipase C and phospholipase A2 signaling pathways have different receptor reserves. Journal of Pharmacology and Experimental Therapeutics 2003;304(1):229-37.

189. Stout BD, Clarke WP, Berg KA. Rapid desensitization of the serotonin2C receptor system: effector pathway and agonist dependence. Journal of Pharmacology and Experimental Therapeutics 2002;302(3):957-62. 190. Ames RS, Sarau HM, Chambers JK, Willette RN, Aiyar NV, Romanic AM, et al. Human urotensin-II is a potent vasoconstrictor and agonist for the orphan receptor GPR14. Nature 1999;401(6750):282-6. 191. Flohr S, Kurz M, Kostenis E, Brkovich A, Fournier A, Klabunde T. Identification of nonpeptidic urotensin II receptor antagonists by virtual screening based on a pharmacophore model derived from structure−activity relationships and nuclear magnetic resonance studies on urotensin II. Journal of Medicinal Chemistry 2002;45(9):1799-805. 192. Graminski GF, Jayawickreme CK, Potenza MN, Lerner MR. Pigment dispersion in frog melanophores can be induced by a phorbol ester or stimulation of a recombinant receptor that activates phospholipase C. Journal of Biological Chemistry 1993;268(8):5957-64. 193. Daniolos A, Lerner AB, Lerner MR. Action of light on frog pigment cells in culture. Pigment Cell Research 1990;3(1):38-43. 194. Bhattacharyya N, Hu X, Chen CZ, Mathews Griner LA, Zheng W, Inglese J, et al. A high throughput screening assay system for the identification of small molecule inhibitors of gsp. Public Library of Science One 2014;9(3):e90766-e74. 195. Thomsen W, Frazer J, Unett D. Functional assays for screening GPCR targets. Current Opinion in Biotechnology 2005;16(6):655-65. 196. Adams JW, Unett DJ, Chen X, Gaidarov I. Functional activity of etrasimod and a diverse panel of sphingosine-1-phosphate receptor (S1PR1) modulators at S1P4 receptors. The Federation of American Societies for Experimental Biology Journal 2017;31(1):993.12. 197. Ghosh RN, DeBiasio R, Hudson CC, Ramer ER, Cowan CL, Oakley RH. 176

Quantitative cell-based high-content screening for vasopressin receptor agonists using Transfluor® technology. Journal of Biomolecular Screening 2005;10(5):476-84. 198. Hudson CC, Oakley RH, Cruickshank RD, Rhem SM, Loomis CR. Automation and validation of the Transflour technology: a universal screening assay for G protein-coupled receptors. In: International Symposium on Biomedical Optics, 2002; San Jose, CA, United States: Proceedings of the SPIE; 2002. p. 548-55. 199. Namkung Y, Le Gouill C, Lukashova V, Kobayashi H, Hogue M, Khoury E, et al. Monitoring G protein-coupled receptor and β-arrestin trafficking in live cells using enhanced bystander BRET. Nature Communications 2016;7(1):12178-90. 200. Donthamsetti P, Quejada JR, Javitch JA, Gurevich VV, Lambert NA. Using bioluminescent resonance energy transfer (BRET) to characterize agonist-induced arrestin recruitment to modified and unmodified G protein-coupled receptors (GPCRs). Current Protocols in Pharmacology 2015;70(1):2.14.1-2.14.14. 201. Mengjie L, Beili W. Structural studies of G protein‐coupled receptors. International Union of Biochemistry and Molecular Biology Life 2016;68(11):894-903. 202. Katritch V, Cherezov V, Stevens RC. Structure-function of the G protein-coupled receptor superfamily. Annual Review of Pharmacology and Toxicology 2013;53:531-56. 203. Milic D, Veprintsev DB. Large-scale production and protein engineering of G protein-coupled receptors for structural studies. Frontiers in Pharmacology 2015;6(66):1-24. 204. Henderson R, Unwin PNT. Three-dimensional model of purple membrane obtained by electron microscopy. Nature 1975;257(5521):28-32. 205. Deisenhofer J, Epp O, Miki K, Huber R, Michel H. X-ray structure analysis of a membrane protein complex. Electron density map at 3 A resolution and a model of the chromophores of the photosynthetic reaction center from Rhodopseudomonas viridis. Journal of Molecular Biology 1984;180(2):385-98. 206. Palczewski K, Kumasaka T, Hori T, Behnke CA, Motoshima H, Fox BA, et al. Crystal structure of rhodopsin: a G protein-coupled receptor. Science 2000;289(5480):739-45. 207. Cherezov V, Rosenbaum DM, Hanson MA, Rasmussen SG, Thian FS, Kobilka

TS, et al. High-resolution crystal structure of an engineered human β2-adrenergic G protein-coupled receptor. Science 2007;318(5854):1258-65. 177

208. Rasmussen SG, Choi HJ, Rosenbaum DM, Kobilka TS, Thian FS, Edwards PC,

et al. Crystal structure of the human β2 adrenergic G-protein-coupled receptor. Nature 2007;450(7168):383-7. 209. Brogi S, Tafi A, Désaubry L, Nebigil CG. Discovery of GPCR ligands for probing signal transduction pathways. Frontiers in Pharmacology 2014;5(1):255-68. 210. Zhang D, Zhao Q, Wu B. Structural studies of G Protein-coupled receptors. Molecules and Cells 2015;38(10):836-42. 211. White S. Membrane proteins of known 3D structure. 2017 [11/08/2017]; Available from: http://blanco.biomol.uci.edu/mpstruc/#id_5UEN_21_3 212. Rose PW, Prlić A, Altunkaya A, Bi C, Bradley AR, Christie CH, et al. Yearly Growth of Total Structures. Protein data bank; 2017 [11/08/2017]; Available from: http://www.rcsb.org/pdb/statistics/contentGrowthChart.do?content=total 213. Heydenreich FM, Vuckovic Z, Matkovic M, Veprintsev DB. Stabilization of G protein-coupled receptors by point mutations. Frontiers in Pharmacology 2015;6(82):1-15. 214. Chun E, Thompson AA, Liu W, Roth CB, Griffith MT, Katritch V, et al. Fusion partner toolchest for the stabilization and crystallization of G protein-coupled receptors. Structure 2012;20(6):967-76. 215. Ghosh E, Kumari P, Jaiman D, Shukla AK. Methodological advances: the unsung heroes of the GPCR structural revolution. Nature Reviews Molecular Cell Biology 2015;16(2):69-81.

216. Ashok Y, Nanekar RT, Jaakola VP. Crystallogenesis of adenosine A2A receptor-T4 lysozyme fusion protein: a practical route for the structure. Methods in Enzymology 2013;520:175-98. 217. Landau EM, Rosenbusch JP. Lipidic cubic phases: A novel concept for the crystallization of membrane proteins. Proceedings of the National Academy of Sciences 1996;93(25):14532-5. 218. Ruiz-Gomez A, Molnar C, Holguin H, Mayor F, Jr., de Celis JF. The cell biology of Smo signalling and its relationships with GPCRs. Biochimica et Biophysica Acta 2007;1768(4):901-12. 219. Jastrzebska B. GPCR – G protein complexes – the fundamental signaling assembly. Amino Acids 2013;45(6):1302-14. 220. de March CA, Yu Y, Ni MJ, Adipietro KA, Matsunami H, Ma M, et al. Conserved residues control activation of mammalian G protein-coupled odorant receptors. Journal of the American Chemical Society 2015;137(26):8611-6. 178

221. Wheatley M, Wootten D, Conner MT, Simms J, Kendrick R, Logan RT, et al. Lifting the lid on GPCRs: the role of extracellular loops. British Journal of Pharmacology 2012;165(6):1688-703. 222. Katritch V, Cherezov V, Stevens RC. Diversity and modularity of G protein- coupled receptor structures. Trends in Pharmacological Sciences 2012;33(1):17- 27. 223. Hanson MA, Stevens RC. Discovery of new GPCR biology: one receptor structure at a Time. Structure 2009;17(1):8-14. 224. Park JH, Scheerer P, Hofmann KP, Choe H-W, Ernst OP. Crystal structure of the ligand-free G-protein-coupled receptor opsin. Nature 2008;454(7201):183-7. 225. Scheerer P, Park JH, Hildebrand PW, Kim YJ, Krausz N, Choe H-W, et al. Crystal structure of opsin in its G-protein-interacting conformation. Nature 2008;455(7212):497-502. 226. Miller-Gallacher JL, Nehmé R, Warne T, Edwards PC, Schertler GFX, Leslie

AGW, et al. The 2.1 Å resolution structure of cyanopindolol-bound β1- adrenoceptor identifies an intramembrane Na+ ion that stabilises the ligand-free receptor. Public Library of Science One 2014;9(3):e92727-e35.

227. Huang J, Chen S, Zhang JJ, Huang X-Y. Crystal structure of oligomeric β1- adrenergic G protein-coupled receptors in ligand-free basal state. Nature Structural & Molecular Biology 2013;20(4):419-25. 228. Jaakola V-P, Griffith MT, Hanson MA, Cherezov V, Chien EYT, Lane JR, et al.

The 2.6 angstrom crystal structure of a human A2A adenosine receptor bound to an antagonist. Science 2008;322(5905):1211-7. 229. Dore AS, Robertson N, Errey JC, Ng I, Hollenstein K, Tehan B, et al. Structure of

the adenosine A2A receptor in complex with ZM241385 and the xanthines XAC and caffeine. Structure 2011;19(9):1283-93. 230. Lebon G, Warne T, Edwards PC, Bennett K, Langmead CJ, Leslie AGW, et al.

Agonist-bound adenosine A2A receptor structures reveal common features of GPCR activation. Nature 2011;474(7352):521-5. 231. Xu F, Wu H, Katritch V, Han GW, Jacobson KA, Gao ZG, et al. Structure of an

agonist-bound human A2A adenosine receptor. Science 2011;332(6027):322-7. 232. Rasmussen SG, DeVree BT, Zou Y, Kruse AC, Chung KY, Kobilka TS, et al.

Crystal structure of the β2 adrenergic receptor-Gs protein complex. Nature 2011;477(7366):549-55. 233. Chung KY, Rasmussen SG, Liu T, Li S, DeVree BT, Chae PS, et al. 179

Conformational changes in the G protein Gs induced by the β2 adrenergic receptor. Nature 2011;477(7366):611-5. 234. Liang YL, Khoshouei M, Radjainia M, Zhang Y, Glukhova A, Tarrasch J, et al. Phase-plate cryo-EM structure of a class B GPCR-G-protein complex. Nature 2017;546(7656):118-23. 235. Rasmussen SGF, DeVree BT, Zou Y, Kruse AC, Chung KY, Kobilka TS, et al.

Crystal structure of the β2 adrenergic receptor-Gs protein complex. Nature 2011;477(7366):549-55. 236. Jastrzebska B, Ringler P, Lodowski DT, Moiseenkova-Bell V, Golczak M, Muller SA, et al. Rhodopsin-transducin heteropentamer: three-dimensional structure and biochemical characterization. Journal of Structural Biology 2011;176(3):387-94. 237. Zhang Y, Sun B, Feng D, Hu H, Chu M, Qu Q, et al. Cryo-EM structure of the activated GLP-1 receptor in complex with a G protein. Nature 2017;546(7657):248-53. 238. Carpenter B, Nehmé R, Warne T, Leslie AGW, Tate CG. Structure of the

adenosine A2A receptor bound to an engineered G protein. Nature 2016;536(7614):104-7. 239. Baidya M, Dwivedi H, Shukla AK. Frozen in action: cryo-EM structure of a GPCR–G-protein complex. Nature Structural & Molecular Biology 2017;24(1):500-2. 240. Tate CG, García-Nafría J, Lee Y, Bai X, Carpenter B. Cryo-EM structure of the

adenosine A2A receptor coupled to an engineered heterotrimeric G protein. eLife 2018;7(1):e35946-e64. 241. Liang Y-L, Khoshouei M, Glukhova A, Furness SGB, Zhao P, Clydesdale L, et al. Phase-plate cryo-EM structure of a biased agonist-bound human GLP-1 receptor– Gs complex. Nature 2018;555(7694):121-5. 242. Westfield GH, Rasmussen SG, Su M, Dutta S, DeVree BT, Chung KY, et al.

Structural flexibility of the Gαs α-helical domain in the β2-adrenoceptor Gs complex. Proceedings of the National Academy of Sciences 2011;108(38):16086- 91. 243. Shukla AK, Westfield GH, Xiao K, Reis RI, Huang LY, Tripathi-Shukla P, et al. Visualization of arrestin recruitment by a G-protein-coupled receptor. Nature 2014;512(7513):218-22. 244. Toufektsian MC, Yang Z, Prasad KM, Overbergh L, Ramos SI, Mathieu C, et al.

Stimulation of A2A-adenosine receptors after myocardial infarction suppresses 180

inflammatory activation and attenuates contractile dysfunction in the remote left ventricle. American Journal of Physiology - Heart and Circulatory Physiology 2006;290(4):H1410-H8.

245. Lasley RD, Jahania MS, Mentzer RM. Beneficial effects of adenosine A2a agonist CGS-21680 in infarcted and stunned porcine myocardium. American Journal of Physiology - Heart and Circulatory Physiology 2001;280(4):H1660-H6.

246. Okusa MD, Linden J, Macdonald T, Huang L. Selective A2A adenosine receptor activation reduces ischemia-reperfusion injury in rat kidney. American Journal of Physiology 1999;277(3):H404-H12. 247. Day Y-J, Huang L, McDuffie MJ, Rosin DL, Ye H, Chen J-F, et al. Renal

protection from ischemia mediated by A2A adenosine receptors on bone marrow– derived cells. Journal of Clinical Investigation 2003;112(6):883-91. 248. Reutershan J, Cagnina RE, Chang D, Linden J, Ley K. Therapeutic anti-

inflammatory effects of myeloid cell adenosine receptor A2a stimulation in lipopolysaccharide-induced lung injury. Journal of Immunology 2007;179(2):1254-63. 249. He W, Mazumder A, Wilder T, Cronstein BN. Adenosine regulates bone

metabolism via A1, A2A, and A2B receptors in bone marrow cells from normal humans and patients with multiple myeloma. The Federation of American Societies for Experimental Biology Journal 2013;27(9):3446-54.

250. Palmer TM, Jacobson KA, Stiles GL. Immunological identification of A2 adenosine receptors by two antipeptide antibody preparations. Molecular Pharmacology 1992;42(3):391-7. 251. Day Y-J, Marshall MA, Huang L, McDuffie MJ, Okusa MD, Linden J. Protection

from ischemic liver injury by activation of A2A adenosine receptors during reperfusion: inhibition of chemokine induction. American Journal of Physiology - Gastrointestinal and Liver Physiology 2004;286(2):G285-G93.

252. de Lera Ruiz M, Lim Y-H, Zheng J. Adenosine A2A receptor as a drug discovery target. Journal of Medicinal Chemistry 2014;57(9):3623-50. 253. Blackburn MR, Vance CO, Morschl E, Wilson CN. Adenosine receptors and inflammation. In: Wilson C, Mustafa S, editors. Adenosine Receptors in Health and Disease. 2009/07/30 ed. Berlin, Heidelberg: Springer; 2009. vol 193 p. 215- 69. 254. Haskó G, Pacher P, Vizi ES, Illes P. Adenosine receptor signaling in the brain immune system. Trends in Pharmacological Sciences 2005;26(10):511-6. 181

255. Li Y, Oskouian RJ, Day YJ, Rieger JM, Liu L, Kern JA, et al. Mouse spinal cord

compression injury is reduced by either activation of the adenosine A2A receptor

on bone marrow–derived cells or deletion of the A2A receptor on non-bone marrow–derived cells. Neuroscience 2006;141(4):2029-39. 256. Fredholm BB, IJzerman A, Jacobson KA, Klotz KN, Linden J. International Union of Pharmacology. XXV. Nomenclature and classification of adenosine receptors. Pharmacological Reviews 2001;53(4):527-52. 257. Verzijl D, Ijzerman AP. Functional selectivity of adenosine receptor ligands. Purinergic Signalling 2011;7(2):171-92.

258. Fredholm BB, Cunha RA, Svenningsson P. Pharmacology of adenosine A2A receptors and therapeutic applications. Current Topics in Medicinal Chemistry 2003;3(4):413-26. 259. Ohta A, Sitkovsky M. Role of G-protein-coupled adenosine receptors in downregulation of inflammation and protection from tissue damage. Nature 2001;414(6866):916-20. 260. Ingwersen J, Wingerath B, Graf J, Lepka K, Hofrichter M, Schröter F, et al. Dual

roles of the adenosine A2a receptor in autoimmune neuroinflammation. Journal of Neuroinflammation 2016;13(48):1-11.

261. Zarek PE, Huang C-T, Lutz ER, Kowalski J, Horton MR, Linden J, et al. A2A receptor signaling promotes peripheral tolerance by inducing T-cell anergy and the generation of adaptive regulatory T cells. Blood 2008;111(1):251-9. 262. Flögel U, Burghoff S, van Lent PLEM, Temme S, Galbarz L, Ding Z, et al.

Selective activation of adenosine A2A receptors on immune cells by a CD73- dependent prodrug suppresses joint inflammation in experimental rheumatoid arthritis. Science Translational Medicine 2012;4(146):146ra08. 263. Dai S-S, Zhou Y-G, Li W, An J-H, Li P, Yang N, et al. Local glutamate level

dictates adenosine A2A receptor regulation of neuroinflammation and traumatic brain injury. Journal of Neuroscience 2010;30(16):5802-10.

264. Mills JH, Kim D-G, Krenz A, Chen J-F, Bynoe MS. A2A adenosine receptor signaling in lymphocytes and the central nervous system regulates inflammation during experimental autoimmune encephalomyelitis. Journal of Immunology 2012;188(11):5713-22. 265. Yao S-Q, Li Z-Z, Huang Q-Y, Li F, Wang Z-W, Augusto E, et al. Genetic

inactivation of the adenosine A2A receptor exacerbates brain damage in mice with experimental autoimmune encephalomyelitis. Journal of Neurochemistry 182

2012;123(1):100-12. 266. Rebola N, Simões AP, Canas PM, Tomé AR, Andrade GM, Barry CE, et al.

Adenosine A2A receptors control neuroinflammation and consequent hippocampal neuronal dysfunction. Journal of Neurochemistry 2011;117(1):100-11. 267. Haskó G, Linden J, Cronstein B, Pacher P. Adenosine receptors: therapeutic aspects for inflammatory and immune diseases. Nature Reviews Drug discovery 2008;7(9):759-70.

268. Peirce SM, Skalak TC, Rieger JM, Macdonald TL, Linden J. Selective A2A adenosine receptor activation reduces skin pressure ulcer formation and inflammation. American Journal of Physiology - Heart and Circulatory Physiology 2001;281(1):H67-H74. 269. Hack SP, Christie MJ. Adaptations in adenosine signaling in drug dependence: therapeutic implications. Critical Reviews in Neurobiology 2003;15(3):235-74. 270. Ferré S, Diamond I, Goldberg SR, Yao L, Hourani SMO, Huang ZL, et al.

Adenosine A2A receptors in ventral striatum, hypothalamus and nociceptive circuitry. Implications for drug addiction, sleep and pain. Progress in Neurobiology 2007;83(5):332-47.

271. Schiffmann SN, Fisone G, Moresco R, Cunha RA, Ferré S. Adenosine A2A receptors and basal ganglia physiology. Progress in Neurobiology 2007;83(5):277-92.

272. Brown RM, Short JL. Adenosine A2A receptors and their role in drug addiction. Journal of Pharmacy and Pharmacology 2008;60(11):1409-30. 273. Li W, Dai S, An J, Xiong R, Li P, Chen X, et al. Genetic inactivation of adenosine

A2A receptors attenuates acute traumatic brain injury in the mouse cortical impact model. Experimental Neurology 2009;215(1):69-76. 274. Paterniti I, Melani A, Cipriani S, Corti F, Mello T, Mazzon E, et al. Selective

adenosine A2A receptor agonists and antagonists protect against spinal cord injury through peripheral and central effects. Journal of Neuroinflammation 2011;8(1):31-44. 275. Melani A, Cipriani S, Vannucchi MG, Nosi D, Donati C, Bruni P, et al. Selective

adenosine A2a receptor antagonism reduces JNK activation in oligodendrocytes after cerebral ischaemia. Brain 2009;132(6):1480-95.

276. Chen J-F, Huang Z, Ma J, Zhu J, Moratalla R, Standaert D, et al. A2A adenosine receptor deficiency attenuates brain injury induced by transient focal ischemia in mice. Journal of Neuroscience 1999;19(21):9192-200. 183

277. Yu L, Huang Z, Mariani J, Wang Y, Moskowitz M, Chen J-F. Selective

inactivation or reconstitution of adenosine A2A receptors in bone marrow cells reveals their significant contribution to the development of ischemic brain injury. Nature Medicine 2004;10(10):1081-7. 278. Melani A, Pugliese AM, Pedata F. Adenosine receptors in cerebral ischemia. International Review of Neurobiology 2014;119:309-48.

279. Stockwell J, Jakova E, Cayabyab FS. Adenosine A1 and A2A receptors in the brain: current research and their Role in neurodegeneration. Molecules 2017;22(4):676- 93. 280. Mayne M, Fotheringham J, Yan H-J, Power C, Del Bigio MR, Peeling J, et al.

Adenosine A2A receptor activation reduces proinflammatory events and decreases cell death following intracerebral hemorrhage. Annals of Neurology 2001;49(6):727-35. 281. Ådén U, Halldner L, Lagercrantz H, Dalmau I, Ledent C, Fredholm BB.

Aggravated brain damage after hypoxic ischemia in immature adenosine A2A knockout mice. Stroke 2003;34(3):739-44. 282. Jones PA, Smith RA, Stone TW. Protection against kainate-induced excitotoxicity

by adenosine A2A receptor agonists and antagonists. Neuroscience 1998;85(1):229-37. 283. Linden J. Molecular approach to adenosine receptors: receptor-mediated mechanisms of tissue protection. Annual Review of Pharmacology and Toxicology 2001;41(1):775-87. 284. Borea PA, Gessi S, Merighi S, Varani K. Adenosine as a multi-signalling guardian angel in human diseases: when, where and how does it exert its protective effects? Trends in Pharmacological Sciences 2016;37(6):419-34. 285. Ledent C, Vaugeois JM, Schiffmann SN, Pedrazzini T, El Yacoubi M, Vanderhaeghen JJ, et al. Aggressiveness, hypoalgesia and high blood pressure in

mice lacking the adenosine A2a receptor. Nature 1997;388(6643):674-8. 286. Fredholm BB, Chen JF, Masino SA, Vaugeois JM. Actions of adenosine at its receptors in the CNS: insights from knockouts and drugs. Annual Review of Pharmacology and Toxicology 2005;45:385-412. 287. Muller CE, Jacobson KA. Recent developments in adenosine receptor ligands and their potential as novel drugs. Biochimica et Biophysica Acta 2011;1808(5):1290- 308. 288. Li F-J, Ji H-F, Shen L. A meta-analysis of tea drinking and risk of Parkinson's 184

disease. The Scientific World Journal 2012;2012(923464):1-6. 289. Calon F, Dridi M, Hornykiewicz O, Bedard PJ, Rajput AH, Di Paolo T. Increased

adenosine A2A receptors in the brain of Parkinson's disease patients with dyskinesias. Brain 2004;127(5):1075-84. 290. Sokol LL, Young MJ, Espay AJ, Postuma RB. Cautionary optimism: caffeine and Parkinson’s disease risk. Journal of Clinical Movement Disorders 2016;3(1):7-9. 291. Eriksson A-K, Löfving S, Callaghan RC, Allebeck P. Alcohol use disorders and risk of Parkinson’s disease: findings from a Swedish national cohort study 1972– 2008. BioMed Central Neurology 2013;13(1):190-5. 292. Nam HW, Hinton DJ, Kang NY, Kim T, Lee MR, Oliveros A, et al. Adenosine transporter ENT1 regulates the acquisition of goal-directed behavior and ethanol

drinking through A2A receptor in the dorsomedial striatum. Journal of Neuroscience 2013;33(10):4329-38. 2+ 293. Higley MJ, Sabatini BL. Competitive regulation of synaptic Ca influx by D2

dopamine and A2A adenosine receptors. Nature Neuroscience 2010;13(8):958-66. 294. Fuxe K, Ferre S, Canals M, Torvinen M, Terasmaa A, Marcellino D, et al.

Adenosine A2A and dopamine D2 heteromeric receptor complexes and their function. Journal of Molecular Neuroscience 2005;26(2-3):209-20. 295. Hisahara S, Shimohama S. Dopamine receptors and Parkinson's disease. International Journal of Medicinal Chemistry 2011;2011(1):1-16. 296. Congreve M, Andrews SP, Doré AS, Hollenstein K, Hurrell E, Langmead CJ, et

al. Discovery of 1,2,4-triazine derivatives as adenosine A2A antagonists using structure based drug design. Journal of Medicinal Chemistry 2012;55(5):1898- 903.

297. Kanda T, Tashiro T, Kuwana Y, Jenner P. Adenosine A2A receptors modify motor function in MPTP-treated common marmosets. Neuroreport 1998;9(12):2857-60.

298. Aoyama S, Kase H, Borrelli E. Rescue of locomotor impairment in dopamine D2

receptor-deficient mice by an adenosine A2A receptor antagonist. Journal of Neuroscience 2000;20(15):5848-52. 299. Jenner P, Mori A, Hauser R, Morelli M, Fredholm BB, Chen JF. Adenosine,

adenosine A2A antagonists, and Parkinson's disease. Parkinsonism & Related Disorders 2009;15(6):406-13.

300. Cieślak M, Komoszyński M, Wojtczak A. Adenosine A2A receptors in Parkinson’s disease treatment. Purinergic Signalling 2008;4(4):305-12.

301. Salamone JD. Preladenant, a novel adenosine A2A receptor antagonist for the 185

potential treatment of parkinsonism and other disorders. IDrugs 2010;13(10):723- 31. 302. Cash R, Ruberg M, Raisman R, Agid Y. Adrenergic receptors in Parkinson's disease. Brain Research 1984;322(2):269-75. 303. Broese M, Riemann D, Hein L, Nissen C. α-Adrenergic receptor function, arousal and sleep: mechanisms and therapeutic implications. Pharmacopsychiatry 2012;45(6):209-16.

304. Henry B, Fox SH, Peggs D, Crossman AR, Brotchie JM. The α2-adrenergic receptor antagonist idazoxan reduces dyskinesia and enhances anti-parkinsonian actions of L-dopa in the MPTP-lesioned primate model of Parkinson's disease. Movement Disorders 1999;14(5):744-53. 305. Ruuskanen JO, Xhaard H, Marjamäki A, Salaneck E, Salminen T, Yan Y-L, et al.

Identification of duplicated fourth α2-adrenergic receptor subtype by cloning and mapping of five receptor genes in zebrafish. Molecular Biology and Evolution 2004;21(1):14-28.

306. Philipp M, Brede M, Hein L. Physiological significance of α2-adrenergic receptor subtype diversity: one receptor is not enough. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 2002;283(2):R287-R95.

307. Tan CM, Limbird LE. The α2-adrenergic receptors. In: Perez DM, editor. The Adrenergic Receptors: In the 21st Century. Totowa, NJ: Humana Press; 2006. p. 241-65.

308. Pao CS, Benovic JL. Structure/function analysis of α2A-adrenergic receptor interaction with G protein-coupled receptor kinase 2. Journal of Biological Chemistry 2005;280(12):11052-1108. 309. Lahdesmaki J, Sallinen J, MacDonald E, Kobilka BK, Fagerholm V, Scheinin M.

Behavioral and neurochemical characterization of α2A-adrenergic receptor knockout mice. Neuroscience 2002;113(2):289-99. 310. Altman JD, Trendelenburg AU, MacMillan L, Bernstein D, Limbird L, Starke K,

et al. Abnormal regulation of the sympathetic nervous system in α2A-adrenergic receptor knockout mice. Molecular Pharmacology 1999;56(1):154-61.

311. Bucheler MM, Hadamek K, Hein L. Two α2-adrenergic receptor subtypes, α2A and

α2C, inhibit transmitter release in the brain of gene-targeted mice. Neuroscience 2002;109(4):819-26. 312. Ihalainen JA, Tanila H. In vivo regulation of dopamine and noradrenaline release

by α2A-adrenoceptors in the mouse prefrontal cortex. The European Journal of 186

Neuroscience 2002;15(11):1789-94.

313. Schaak S, Cussac D, Cayla C, Devedjian J, Guyot R, Paris H, et al. α2 adrenoceptors regulate proliferation of human intestinal epithelial cells. Gut 2000;47(2):242-50. 314. Valet P, Senard JM, Devedjian JC, Planat V, Salomon R, Voisin T, et al.

Characterization and distribution of α2-adrenergic receptors in the human intestinal mucosa. Journal of Clinical Investigation 1993;91(5):2049-57.

315. Surprenant A, Horstman D, Akbarali H, Limbird L. A point mutation of the α2- adrenoceptor that blocks coupling to potassium but not calcium currents. Science 1992;257(5072):977-80.

316. Zhu QM, Lesnick JD, Jasper JR, MacLennan SJ, Dillon MP, Eglen RM, et al. α2A-

adrenoceptors, not I1-imidazoline receptors, mediate the hypotensive effects of rilmenidine and moxonidine in conscious mice. In vivo and in vitro studies. Annals of the New York Academy of Sciences 1999;881:287-9.

317. Schramm NL, McDonald MP, Limbird LE. The α2A-adrenergic receptor plays a protective role in mouse behavioral models of depression and anxiety. Journal of Neuroscience 2001;21(13):4875-82. 318. Arnsten AFT. The use of α-2A adrenergic agonists for the treatment of attention- deficit/hyperactivity disorder. Expert Review of Neurotherapeutics 2010;10(10):1595-605.

319. Meyer H, Palchaudhuri M, Scheinin M, Flugge G. Regulation of α2A-adrenoceptor expression by chronic stress in neurons of the brain stem. Brain Research 2000;880(1-2):147-58.

320. Stone LS, MacMillan LB, Kitto KF, Limbird LE, Wilcox GL. The α2A adrenergic

receptor subtype mediates spinal analgesia evoked by α2 agonists and is necessary for spinal adrenergic-opioid synergy. Journal of Neuroscience 1997;17(18):7157- 65. 321. Lakhlani PP, MacMillan LB, Guo TZ, McCool BA, Lovinger DM, Maze M, et al.

Substitution of a mutant α2A-adrenergic receptor via "hit and run" gene targeting reveals the role of this subtype in sedative, analgesic, and anesthetic-sparing responses in vivo. Proceedings of the National Academy of Sciences 1997;94(18):9950-555.

322. Avery RA, Franowicz JS, Studholme C, van Dyck CH, Arnsten AF. The α2A- adrenoceptor agonist, guanfacine, increases regional cerebral blood flow in dorsolateral prefrontal cortex of monkeys performing a spatial working memory 187

task. Neuropsychopharmacology 2000;23(3):240-9. 323. Franowicz JS, Kessler LE, Borja CM, Kobilka BK, Limbird LE, Arnsten AF.

Mutation of the α2A-adrenoceptor impairs working memory performance and annuls cognitive enhancement by guanfacine. Journal of Neuroscience 2002;22(19):8771-7. 324. Tan CM, Wilson MH, MacMillan LB, Kobilka BK, Limbird LE. Heterozygous

α2A-adrenergic receptor mice unveil unique therapeutic benefits of partial agonists. Proceedings of the National Academy of Sciences 2002;99(19):12471-6. 325. Ferrante A, Martire A, Armida M, Chiodi V, Pézzola A, Potenza RL, et al.

Influence of CGS 21680, a selective adenosine A2A receptor agonist, on NMDA receptor function and expression in the brain of Huntington's disease mice. Brain Research 2010;1323:184-91. 326. Vincenzi F, Corciulo C, Targa M, Merighi S, Gessi S, Casetta I, et al. Multiple

sclerosis lymphocytes upregulate A2A adenosine receptors that are antiinflammatory when stimulated. European Journal of Immunology 2013;43(8):2206-16.

327. Chen Y, Peng Y, Che P, Gannon M, Liu Y, Li L, et al. α2A adrenergic receptor promotes amyloidogenesis through disrupting APP-SorLA interaction. Proceedings of the National Academy of Sciences 2014;111(48):17296-301. 328. Singer SJ, Nicolson GL. The fluid mosaic model of the structure of cell membranes. Science 1972;175(4023):720-31. 329. Rawyler A, van der Schaft PH, Roelofsen B, Op den Kamp JA. Phospholipid localization in the plasma membrane of Friend erythroleukemic cells and mouse erythrocytes. Biochemistry 1985;24(7):1777-83. 330. de Kroon AIPM, Rijken PJ, De Smet CH. Checks and balances in membrane phospholipid class and acyl chain homeostasis, the yeast perspective. Progress in Lipid Research 2013;52(4):374-94. 331. Yamaji-Hasegawa A, Tsujimoto M. Asymmetric distribution of phospholipids in biomembranes. Biological & Pharmaceutical Bulletin 2006;29(8):1547-53. 332. Gurtovenko AA, Vattulainen I. Membrane potential and electrostatics of phospholipid bilayers with asymmetric transmembrane distribution of anionic lipids. Journal of Physical Chemistry B 2008;112(15):4629-34. 333. Krause MR, Regen SL. The structural role of cholesterol in cell membranes: from condensed bilayers to lipid rafts. Accounts of Chemical Research 2014;47(12):3512-21. 188

334. Seddon AM, Curnow P, Booth PJ. Membrane proteins, lipids and detergents: not just a soap opera. Biochimica et Biophysica Acta 2004;1666(1):105-17. 335. Grisshammer R. Purification of recombinant G-protein-coupled receptors. Methods in Enzymology 2009;463:631-45. 336. Bayburt TH, Carlson JW, Sligar SG. Reconstitution and imaging of a membrane protein in a nanometer-size phospholipid bilayer. Journal of Structural Biology 1998;123(1):37-44. 337. Bayburt TH, Carlson JW, Sligar SG. Single molecule height measurements on a membrane protein in nanometer-scale phospholipid bilayer disks. Langmuir 2000;16(14):5993-7. 338. Bayburt TH, Grinkova YV, Sligar SG. Self-assembly of discoidal phospholipid bilayer nanoparticles with membrane scaffold proteins. Nano Letters 2002;2(8):853-6. 339. Bayburt TH, Sligar SG. Single-molecule height measurements on microsomal cytochrome P450 in nanometer-scale phospholipid bilayer disks. Proceedings of the National Academy of Sciences 2002;99(10):6725-30. 340. Banerjee S, Huber T, Sakmar TP. Rapid incorporation of functional rhodopsin into nanoscale apolipoprotein bound bilayer (NABB) particles. Journal of Molecular Biology 2008;377(4):1067-81. 341. Serebryany E, Zhu GA, Yan ECY. Artificial membrane-like environments for in vitro studies of purified G-protein coupled receptors. Biochimica et Biophysica Acta - Biomembranes 2012;1818(2):225-33. 342. Chen B, Ren X, Neville T, Jerome WG, Hoyt DW, Sparks D, et al. Apolipoprotein AI tertiary structures determine stability and phospholipid-binding activity of discoidal high-density lipoprotein particles of different sizes. Protein Science 2009;18(5):921-35. 343. Wang M, Briggs MR. HDL: the metabolism, function, and therapeutic importance. Chemical Reviews 2004;104(1):119-37. 344. Martinez LO, Jacquet S, Tercé F, Collet X, Perret B, Barbaras R. New insight on the molecular mechanisms of high-density lipoprotein cellular interactions. Cellular and Molecular Life Sciences 2004;61(18):2343-60. 345. Meagher EA. Addressing cardiovascular risk beyond low-density lipoprotein cholesterol: the high-density lipoprotein cholesterol story. Current Cardiology Reports 2004;6(6):457-63. 346. McLachlan AD. Repeated helical pattern in apolipoprotein-A-I. Nature 189

1977;267(5610):465-6. 347. Thomas MJ, Bhat S, Sorci-Thomas MG. Three-dimensional models of HDL apoA-I: implications for its assembly and function. Journal of Lipid Research 2008;49(9):1875-83. 348. Marcel YL, Kiss RS. Structure-function relationships of apolipoprotein A-I: a flexible protein with dynamic lipid associations. Current Opinion in Lipidology 2003;14(2):151-7. 349. Borhani DW, Rogers DP, Engler JA, Brouillette CG. Crystal structure of truncated human apolipoprotein A-I suggests a lipid-bound conformation. Proceedings of the National Academy of Sciences 1997;94(23):12291-6. 350. Mei X, Atkinson D. Crystal structure of C-terminal truncated apolipoprotein A-I reveals the assembly of high density lipoprotein (HDL) by dimerization. Journal of Biological Chemistry 2011;286(44):38570-82. 351. Bibow S, Polyhach Y, Eichmann C, Chi CN, Kowal J, Albiez S, et al. Solution structure of discoidal high-density lipoprotein particles with a shortened apolipoprotein A-I. Nature Structural & Molecular Biology 2017;24(2):187-93. 352. Ajees AA, Anantharamaiah GM, Mishra VK, Hussain MM, Murthy HM. Crystal structure of human apolipoprotein A-I: insights into its protective effect against cardiovascular diseases. Proceedings of the National Academy of Sciences 2006;103(7):2126-31. 353. Segrest JP, Jones MK, Klon AE, Sheldahl CJ, Hellinger M, De Loof H, et al. A detailed molecular belt model for apolipoprotein A-I in discoidal high density lipoprotein. Journal of Biological Chemistry 1999;274(45):31755-8. 354. Koppaka V, Silvestro L, Engler JA, Brouillette CG, Axelsen PH. The structure of human lipoprotein A-I. Evidence for the "belt" model. Journal of Biological Chemistry 1999;274(21):14541-4. 355. Li Y, Kijac AZ, Sligar SG, Rienstra CM. Structural analysis of nanoscale self- assembled discoidal lipid bilayers by solid-state NMR spectroscopy. Biophysical Journal 2006;91(10):3819-28. 356. Davidson WS, Hilliard GM. The spatial organization of apolipoprotein A-I on the edge of discoidal high density lipoprotein particles: a mass specrometry study. Journal of Biological Chemistry 2003;278(29):27199-207. 357. Silva RA, Hilliard GM, Li L, Segrest JP, Davidson WS. A mass spectrometric determination of the conformation of dimeric apolipoprotein A-I in discoidal high density lipoproteins. Biochemistry 2005;44(24):8600-7. 190

358. Martin DDO, Budamagunta MS, Ryan RO, Voss JC, Oda MN. Apolipoprotein A- I assumes a “looped belt” conformation on reconstituted high density lipoprotein. Journal of Biological Chemistry 2006;281(29):20418-26. 359. Shih AY, Denisov IG, Phillips JC, Sligar SG, Schulten K. Molecular dynamics simulations of discoidal bilayers assembled from truncated human lipoproteins. Biophysical Journal 2005;88(1):548-56. 360. Mei X, Atkinson D. Lipid-free apolipoprotein A-I structure: insights into HDL formation and atherosclerosis development. Archives of Medical Research 2015;46(5):351-60. 361. Borch J, Hamann T. The nanodisc: a novel tool for membrane protein studies. Biological Chemistry 2009;390(8):805-14. 362. Gao T, Petrlova J, He W, Huser T, Kudlick W, Voss J, et al. Characterization of de novo synthesized GPCRs supported in nanolipoprotein discs. Public Library of Science One 2012;7(9):e44911-e8. 363. Gurevich VV, Gurevich EV. GPCR monomers and oligomers: it takes all kinds. Trends in Neurosciences 2008;31(2):74-81. 364. Henrich E, Peetz O, Hein C, Laguerre A, Hoffmann B, Hoffmann J, et al. Analyzing native membrane protein assembly in nanodiscs by combined non- covalent mass spectrometry and synthetic biology. eLife 2017;6(1):e20954-e72. 365. Nath A, Atkins WM, Sligar SG. Applications of phospholipid bilayer nanodiscs in the study of membranes and membrane proteins. Biochemistry 2007;46(8):2059-69. 366. Ranaghan MJ, Schwall CT, Alder NN, Birge RR. Green proteorhodopsin reconstituted into nanoscale phospholipid bilayers (nanodiscs) as photoactive monomers. Journal of the American Chemical Society 2011;133(45):18318-27. 367. Ritchie TK, Grinkova YV, Bayburt TH, Denisov IG, Zolnerciks JK, Atkins WM, et al. Reconstitution of membrane proteins in phospholipid bilayer nanodiscs. Methods in Enzymology 2009;464:211-31. 368. Bayburt TH, Sligar SG. Membrane protein assembly into nanodiscs. The Federation of European Biochemical Societies Letters 2010;584(9):1721-7. 369. Bayburt TH, Leitz AJ, Xie G, Oprian DD, Sligar SG. Transducin activation by nanoscale lipid bilayers containing one and two rhodopsins. Journal of Biological Chemistry 2007;282(20):14875-81. 370. El Moustaine D, Granier S, Doumazane E, Scholler P, Rahmeh R, Bron P, et al. Distinct roles of metabotropic glutamate receptor dimerization in agonist 191

activation and G-protein coupling. Proceedings of the National Academy of Sciences 2012;109(40):16342-7. 371. Bayburt TH, Sligar SG. Self-assembly of single integral membrane proteins into soluble nanoscale phospholipid bilayers. Protein Science 2003;12(11):2476-81. 372. Grinkova YV, Denisov IG, Sligar SG. Engineering extended membrane scaffold proteins for self-assembly of soluble nanoscale lipid bilayers. Protein Engineering, Design & Selection 2010;23(11):843-8. 373. Denisov IG, Grinkova YV, Lazarides AA, Sligar SG. Directed Self-Assembly of Monodisperse Phospholipid Bilayer Nanodiscs with Controlled Size. Journal of the American Chemical Society 2004;126(11):3477-87. 374. Kynde SA, Skar-Gislinge N, Pedersen MC, Midtgaard SR, Simonsen JB, Schweins R, et al. Small-angle scattering gives direct structural information about a membrane protein inside a lipid environment. Acta Crystallographica Section D: Biological Crystallography 2014;70(2):371-83. 375. Skar-Gislinge N, Kynde SAR, Denisov IG, Ye X, Lenov I, Sligar SG, et al. Small- angle scattering determination of the shape and localization of human cytochrome P450 embedded in a phospholipid nanodisc environment. Acta Crystallographica Section D: Biological Crystallography 2015;71(Pt 12):2412-21. 376. Denisov IG, McLean MA, Shaw AW, Grinkova YV, Sligar SG. Thermotropic phase transition in soluble nanoscale lipid bilayers. Journal of Physical Chemistry B 2005;109(32):15580-8. 377. Siuda I, Tieleman DP. Molecular models of nanodiscs. Journal of Chemical Theory and Computation 2015;11(10):4923-32. 378. Marty MT, Zhang H, Cui W, Blankenship RE, Gross ML, Sligar SG. Native mass spectrometry characterization of intact nanodisc lipoprotein complexes. Analytical Chemistry 2012;84(21):8957-60. 379. Berne BJ, Pecora R. Dynamic light scattering: with applications to chemistry, biology, and physics. Dover Publications; 2000. 380. Leitz AJ, Bayburt TH, Barnakov AN, Springer BA, Sligar SG. Functional

reconstitution of β2-adrenergic receptors utilizing self-assembling nanodisc technology. Biotechniques 2006;40(5):601-6. 381. Tsukamoto H, Szundi I, Lewis JW, Farrens DL, Kliger DS. Rhodopsin in nanodiscs has native membrane-like photointermediates. Biochemistry 2011;50(22):5086-91. 382. Van Eps N, Caro LN, Morizumi T, Kusnetzow AK, Szczepek M, Hofmann KP, et 192

al. Conformational equilibria of light-activated rhodopsin in nanodiscs. Proceedings of the National Academy of Sciences 2017;114(16):e3268-e75. 383. Calebiro D, Rieken F, Wagner J, Sungkaworn T, Zabel U, Borzi A, et al. Single- molecule analysis of fluorescently labeled G-protein–coupled receptors reveals complexes with distinct dynamics and organization. Proceedings of the National Academy of Sciences 2013;110(2):743-8. 384. Ferré S, Casadó V, Devi LA, Filizola M, Jockers R, Lohse MJ, et al. G protein– coupled receptor oligomerization revisited: functional and pharmacological perspectives. Pharmacological Reviews 2014;66(2):413-34. 385. Sommer ME, Hofmann KP, Heck M. Distinct loops in arrestin differentially regulate ligand binding within the GPCR opsin. Nature Communications 2012;3(995):1-10. 386. Fung JJ, Deupi X, Pardo L, Yao XJ, Velez‐Ruiz GA, DeVree BT, et al. Ligand‐

regulated oligomerization of β2-adrenoceptors in a model lipid bilayer. The European Molecular Biology Organizatoin Journal 2009;28(21):3315-28. 387. White JF, Grodnitzky J, Louis JM, Trinh LB, Shiloach J, Gutierrez J, et al. Dimerization of the class A G protein-coupled neurotensin receptor NTS1 alters G protein interaction. Proceedings of the National Academy of Sciences 2007;104(29):12199-204. 388. D’Antona AM, Xie G, Sligar SG, Oprian DD. Assembly of an activated rhodopsin/transducin complex in nanoscale lipid bilayers. Biochemistry 2014;53(1):127-34. 389. Whorton MR, Jastrzebska B, Park PS-H, Fotiadis D, Engel A, Palczewski K, et al. Efficient coupling of transducin to monomeric rhodopsin in a phospholipid bilayer. Journal of Biological Chemistry 2008;283(7):4387-94. 390. Bayburt TH, Vishnivetskiy SA, McLean MA, Morizumi T, Huang CC, Tesmer JJ, et al. Monomeric rhodopsin is sufficient for normal rhodopsin kinase (GRK1) phosphorylation and arrestin-1 binding. Journal of Biological Chemistry 2011;286(2):1420-8. 391. Tsukamoto H, Sinha A, DeWitt M, Farrens DL. Monomeric rhodopsin is the minimal functional unit required for arrestin binding. Journal of Molecular Biology 2010;399(3):501-11. 392. McIntosh TJ, Simon SA. Bilayers as protein solvents: role of bilayer structure and elastic properties. Journal of General Physiology 2007;130(2):225-7.

193

393. Dijkman PM, Watts A. Lipid modulation of early G protein-coupled receptor signalling events. Biochimica et Biophysica Acta - Biomembranes 2015;1848(11, Part A):2889-97. 394. Andersen OS, Koeppe RE. Bilayer thickness and membrane protein function: an energetic perspective. Annual Review of Biophysics and Biomolecular Structure 2007;36:107-30. 395. Cross TA, Murray DT, Watts A. Helical membrane protein conformations and their environment. European Biophysics Journal 2013;42(10):731-55. 396. Lee AG. How lipids affect the activities of integral membrane proteins. Biochimica et Biophysica Acta 2004;1666(1-2):62-87. 397. Kawai T, Caaveiro JM, Abe R, Katagiri T, Tsumoto K. Catalytic activity of MsbA reconstituted in nanodisc particles is modulated by remote interactions with the bilayer. The Federation of European Biochemical Societies Letters 2011;585(22):3533-7. 398. Nasvik Ojemyr L, von Ballmoos C, Gennis RB, Sligar SG, Brzezinski P. Reconstitution of respiratory oxidases in membrane nanodiscs for investigation of proton-coupled electron transfer. The Federation of European Biochemical Societies Letters 2012;586(5):640-5. 399. Grinkova YV, Denisov IG, Sligar SG. Functional reconstitution of monomeric CYP3A4 with multiple cytochrome P450 reductase molecules in Nanodiscs. Biochemical and Biophysical Research Communications 2010;398(2):194-8. 400. Luthra A, Gregory M, Grinkova YV, Denisov IG, Sligar SG. Nanodiscs in the studies of membrane-bound cytochrome P450 enzymes. Methods in Molecular Biology 2013;987:115-27. 401. Das A, Sligar SG. Modulation of the cytochrome P450 reductase redox potential by the phospholipid bilayer. Biochemistry 2009;48(51):12104-12. 402. Schwall CT, Greenwood VL, Alder NN. The stability and activity of respiratory complex II is cardiolipin-dependent. Biochimica et Biophysica Acta 2012;1817(9):1588-96. 403. Andersen KK, Oliveira CL, Larsen KL, Poulsen FM, Callisen TH, Westh P, et al. The role of decorated SDS micelles in sub-CMC protein denaturation and association. Journal of Molecular Biology 2009;391(1):207-26. 404. Shenkarev ZO, Lyukmanova EN, Butenko IO, Petrovskaya LE, Paramonov AS, Shulepko MA, et al. Lipid–protein nanodiscs promote in vitro folding of transmembrane domains of multi-helical and multimeric membrane proteins. 194

Biochimica et Biophysica Acta - Biomembranes 2013;1828(2):776-84. 405. Zocher M, Roos C, Wegmann S, Bosshart PD, Dotsch V, Bernhard F, et al. Single- molecule force spectroscopy from nanodiscs: an assay to quantify folding, stability, and interactions of native membrane proteins. American Chemical Society Nano 2012;6(1):961-71. 406. Nasr ML, Singh SK. Radioligand binding to nanodisc-reconstituted membrane transporters assessed by the scintillation proximity assay. Biochemistry 2014;53(1):4-6. 407. Xia L, de Vries H, Ijzerman AP, Heitman LH. Scintillation proximity assay (SPA) as a new approach to determine a ligand’s kinetic profile. A case in point for the

adenosine A1 receptor. Purinergic Signalling 2016;12(1):115-26. 408. Mills A, Le HT, Coulton JW, Duong F. FhuA interactions in a detergent-free nanodisc environment. Biochimica et Biophysica Acta 2014;1838(1):364-71. 409. Inagaki S, Ghirlando R, Grisshammer R. Biophysical characterization of membrane proteins in nanodiscs. Methods 2013;59(3):287-300. 410. Dorr JM, Koorengevel MC, Schafer M, Prokofyev AV, Scheidelaar S, van der Cruijsen EA, et al. Detergent-free isolation, characterization, and functional reconstitution of a tetrameric K+ channel: the power of native nanodiscs. Proceedings of the National Academy of Sciences 2014;111(52):18607-12. 411. Denisov IG, Sligar SG. Nanodiscs in membrane biochemistry and biophysics. Chemical Reviews 2017;117(6):4669-713. 412. Henrich E, Ma Y, Engels I, Münch D, Otten C, Schneider T, et al. Lipid requirements for the enzymatic activity of MraY translocases and in vitro reconstitution of the lipid II synthesis pathway. Journal of Biological Chemistry 2016;291(5):2535-46. 413. Shi L, Howan K, Shen Q-T, Wang YJ, Rothman JE, Pincet F. Preparation and characterization of SNARE-containing nanodiscs and direct study of cargo release through fusion pores. Nature Protocols 2013;8(5):935-48. 414. Ma J, Lu Y, Wu D, Peng Y, Loa-Kum-Cheung W, Peng C, et al. Ligand

identification of the adenosine A2A receptor in self-assembled nanodiscs by affinity mass spectrometry. Analytical Methods 2017;9(40):5851-8. 415. Inagaki S, Ghirlando R, White JF, Gvozdenovic-Jeremic J, Northup JK, Grisshammer R. Modulation of the interaction between neurotensin receptor

NTS1 and Gq potein by lipid. Journal of Molecular Biology 2012;417(1–2):95- 111. 195

416. van den Brink-van der Laan E, Killian JA, de Kruijff B. Nonbilayer lipids affect peripheral and integral membrane proteins via changes in the lateral pressure profile. Biochimica et Biophysica Acta 2004;1666(1):275-88. 417. Frolov VA, Shnyrova AV, Zimmerberg J. Lipid polymorphisms and membrane shape. Cold Spring Harbor Perspectives in Biology 2011;3(11):a004747-a60. 418. Opekarová M, Tanner W. Specific lipid requirements of membrane proteins—a putative bottleneck in heterologous expression. Biochimica et Biophysica Acta - Biomembranes 2003;1610(1):11-22. 419. Wadsater M, Maric S, Simonsen JB, Mortensen K, Cardenas M. The effect of using binary mixtures of zwitterionic and charged lipids on nanodisc formation and stability. Soft Matter 2013;9(7):2329-37. 420. Schuler MA, Denisov IG, Sligar SG. Nanodiscs as a new tool to examine lipid- protein interactions. Methods in Molecular Biology 2013;974:415-33. 421. Rues RB, Dotsch V, Bernhard F. Co-translational formation and pharmacological

characterization of β1-adrenergic receptor/nanodisc complexes with different lipid environments. Biochimica et Biophysica Acta 2016;1858(6):1306-16. 422. Roos C, Zocher M, Müller D, Münch D, Schneider T, Sahl H-G, et al. Characterization of co-translationally formed nanodisc complexes with small multidrug transporters, proteorhodopsin and with the E. coli MraY translocase. Biochimica et Biophysica Acta - Biomembranes 2012;1818(12):3098-106. 423. Duggal R, Liu Y, Gregory MC, Denisov IG, Kincaid JR, Sligar SG. Evidence that

cytochrome b5 acts as a redox donor in CYP17A1 mediated androgen synthesis. Biochemical and Biophysical Research Communications 2016;477(2):202-8. 424. Gregory M, Mak PJ, Sligar SG, Kincaid JR. Differential hydrogen bonding in human CYP17 dictates hydroxylation versus lyase chemistry. Angewandte Chemie (International Edition in English) 2013;52(20):5342-5. 425. Gregory MC, Denisov IG, Grinkova YV, Khatri Y, Sligar SG. Kinetic solvent isotope effect in human P450 CYP17A1-mediated androgen formation: evidence for a reactive peroxoanion intermediate. Journal of the American Chemical Society 2013;135(44):16245-7. 426. Grinkova YV, Denisov IG, McLean MA, Sligar SG. Oxidase uncoupling in heme monooxygenases: human cytochrome P450 CYP3A4 in Nanodiscs. Biochemical and Biophysical Research Communications 2013;430(4):1223-7. 427. Mak PJ, Denisov IG, Grinkova YV, Sligar SG, Kincaid JR. Defining CYP3A4 structural responses to substrate binding. Raman spectroscopic studies of a 196

nanodisc-incorporated mammalian cytochrome P450. Journal of the American Chemical Society 2011;133(5):1357-66. 428. Mak PJ, Gregory MC, Denisov IG, Sligar SG, Kincaid JR. Unveiling the crucial intermediates in androgen production. Proceedings of the National Academy of Sciences 2015;112(52):15856-61. 429. Mak PJ, Gregory MC, Sligar SG, Kincaid JR. Resonance raman spectroscopy reveals That substrate structure selectively impacts the heme-bound diatomic ligands of CYP17. Biochemistry 2014;53(1):90-100. 430. Mak PJ, Luthra A, Sligar SG, Kincaid JR. Resonance raman spectroscopy of the oxygenated intermediates of human CYP19A1 implicates a compound I intermediate in the final lyase step. Journal of the American Chemical Society 2014;136(13):4825-8. 431. Bender G, Schexnaydre EE, Murphy RC, Uhlson C, Newcomer ME. Membrane- dependent activities of human 15-LOX-2 and its Murine counterpart: implications for murine models of atherosclerosis. Journal of Biological Chemistry 2016;291(37):19413-24. 432. Kim C, Ye F, Hu X, Ginsberg MH. Talin activates integrins by altering the topology of the β transmembrane domain. Journal of Cell Biology 2012;197(5):605-11. 433. Meling DD, Zelasko S, Kambalyal A, Roy J, Das A. Functional role of the conserved i-helix residue I346 in CYP5A1-Nanodiscs. Biophysical Chemistry 2015;200-201:34-40. 434. Wang B, Zhao A, Xie Q, Olinares PD, Chait BT, Novick RP, et al. Functional Plasticity of the AgrC Receptor Histidine Kinase Required for Staphylococcal Virulence. Cell Chemical Biology 2017;24(1):76-86. 435. Francois-Martin C, Pincet F. Actual fusion efficiency in the lipid mixing assay - Comparison between nanodiscs and liposomes. Scientific Reports 2017;7(43860):1-8. 436. Wang X, Mu Z, Li Y, Bi Y, Wang Y. Smaller nanodiscs are suitable for studying protein lipid interactions by solution NMR. The Protein Journal 2015;34(3):205- 11. 437. Xu H, Hill JJ, Michelsen K, Yamane H, Kurzeja RJ, Tam T, et al. Characterization of the direct interaction between KcsA-Kv1.3 and its inhibitors. Biochimica et Biophysica Acta 2015;1848(10):1974-80. 438. Tark S-H, Das A, Sligar S, Dravid VP. Nanomechanical detection of cholera toxin 197

using microcantilevers functionalized with ganglioside nanodiscs. Nanotechnology 2010;21(43):435502.1-435502.7. 439. Ye X, McLean MA, Sligar SG. Phosphatidylinositol 4,5-bisphosphate modulates the phospholipid bilayers and activates its autoinhibited form. Biochemistry 2016;55(36):5038-48. 440. Xu XP, Kim E, Swift M, Smith JW, Volkmann N, Hanein D. Three-dimensional

structures of full-length, membrane-embedded human αIIbβ3 integrin complexes. Biophysical Journal 2016;110(4):798-809. 441. Shaw AW, Pureza VS, Sligar SG, Morrissey JH. The local phospholipid environment modulates the activation of blood clotting. Journal of Biological Chemistry 2007;282(9):6556-63. 442. Boettcher JM, Davis-Harrison RL, Clay MC, Nieuwkoop AJ, Ohkubo YZ, Tajkhorshid E, et al. Atomic view of calcium-induced clustering of phosphatidylserine in mixed lipid bilayers. Biochemistry 2011;50(12):2264-73. 443. Her C, Filoti DI, McLean MA, Sligar SG, Alexander Ross JB, Steele H, et al. The charge properties of phospholipid nanodiscs. Biophysical Journal 2016;111(5):989-98. 444. Nasr ML, Baptista D, Strauss M, Sun Z-YJ, Grigoriu S, Huser S, et al. Covalently circularized nanodiscs for studying membrane proteins and viral entry. Nature Methods 2017;14(1):49-52. 445. Mörs K, Roos C, Scholz F, Wachtveitl J, Dötsch V, Bernhard F, et al. Modified lipid and protein dynamics in nanodiscs. Biochimica et Biophysica Acta - Biomembranes 2013;1828(4):1222-9. 446. Ye X, McLean MA, Sligar SG. Conformational equilibrium of talin is regulated by anionic lipids. Biochimica et Biophysica Acta 2016;1858(8):1833-40. 447. Zhang P, Ye F, Bastidas AC, Kornev AP, Wu J, Ginsberg MH, et al. An isoform specific myristylation switch targets type II PKA holoenzymes to Membranes. Structure 2015;23(9):1563-72. 448. Yao Y, Nisan D, Fujimoto LM, Antignani A, Barnes A, Tjandra N, et al. Characterization of the membrane-inserted C-terminus of cytoprotective BCL-XL. Protein Expression and Purification 2016;122:56-63. 449. Bajaj R, Bruce KE, Davidson AL, Rued BE, Stauffacher CV, Winkler ME. Biochemical characterization of essential cell division proteins FtsX and FtsE that mediate peptidoglycan hydrolysis by PcsB in Streptococcus pneumoniae. MicrobiologyOpen 2016;5(5):738-52. 198

450. Gao Y, Cao E, Julius D, Cheng Y. TRPV1 structures in nanodiscs reveal mechanisms of ligand and lipid action. Nature 2016;534(7607):347–51. 451. Cao E, Cordero-Morales Julio F, Liu B, Qin F, Julius D. TRPV1 channels are intrinsically heat sensitive and negatively regulated by phosphoinositide lipids. Neuron 2013;77(4):667-79. 452. Rice AJ, Alvarez FJD, Davidson AL, Pinkett HW. Effects of lipid environment on the conformational changes of an ABC importer. Channels 2014;8(4):327-33. 453. Boldog T, Grimme S, Li M, Sligar SG, Hazelbauer GL. Nanodiscs separate chemoreceptor oligomeric states and reveal their signaling properties. Proceedings of the National Academy of Sciences 2006;103(31):11509-14. 454. Eggensperger S, Fisette O, Parcej D, Schäfer LV, Tampé R. An annular lipid belt is essential for allosteric coupling and viral inhibition of the antigen translocation complex TAP (Transporter Asssociated with Antigen Processing). Journal of Biological Chemistry 2014;289(48):33098-108. 455. Finkenwirth F, Sippach M, Landmesser H, Kirsch F, Ogienko A, Grunzel M, et al. ATP-dependent conformational changes trigger substrate capture and release by an ECF-type biotin transporter. Journal of Biological Chemistry 2015;290(27):16929-42. 456. Hörnschemeyer P, Liss V, Heermann R, Jung K, Hunke S. Interaction analysis of a two-component system using nanodiscs. Public Library of Science One 2016;11(2):e0149187-e204. 457. Dominik PK, Borowska MT, Dalmas O, Kim SS, Perozo E, Keenan RJ, et al. Conformational chaperones for structural studies of membrane proteins using antibody phage display with nanodiscs. Structure 2016;24(2):300-9. 458. Mi LZ, Grey MJ, Nishida N, Walz T, Lu C, Springer TA. Functional and structural stability of the epidermal growth factor receptor in detergent micelles and phospholipid nanodiscs. Biochemistry 2008;47(39):10314-23. 459. Kuszak AJ, Pitchiaya S, Anand JP, Mosberg HI, Walter NG, Sunahara RK. Purification and functional reconstitution of monomeric mu-opioid receptors:

allosteric modulation of agonist binding by Gi2. Journal of Biological Chemistry 2009;284(39):26732-41. 460. Denisov IG, Sligar SG. Cytochromes P450 in Nanodiscs. Biochimica et Biophysica Acta - Proteins and Proteomics 2011;1814(1):223-9. 461. Bayburt TH, Grinkova YV, Sligar SG. Assembly of single bacteriorhodopsin trimers in bilayer nanodiscs. Archives of Biochemistry and Biophysics 199

2006;450(2):215-22. 462. Hagn F, Etzkorn M, Raschle T, Wagner G. Optimized phospholipid bilayer nanodiscs facilitate high-resolution structure determination of membrane proteins. Journal of the American Chemical Society 2013;135(5):1919-25. 463. Johnson PJ, Halpin A, Morizumi T, Brown LS, Prokhorenko VI, Ernst OP, et al. The photocycle and ultrafast vibrational dynamics of bacteriorhodopsin in lipid nanodiscs. Physical Chemistry Chemical Physics 2014;16(39):21310-20. 464. Ishida H, Garcia-Herrero A, Vogel HJ. The periplasmic domain of Escherichia coli outer membrane protein A can undergo a localized temperature dependent structural transition. Biochimica et Biophysica Acta 2014;1838(12):3014-24. 465. Shi L, Shen QT, Kiel A, Wang J, Wang HW, Melia TJ, et al. SNARE proteins: one to fuse and three to keep the nascent fusion pore open. Science 2012;335(6074):1355-9. 466. Whorton MR, Bokoch MP, Rasmussen SG, Huang B, Zare RN, Kobilka B, et al. A monomeric G protein-coupled receptor isolated in a high-density lipoprotein particle efficiently activates its G protein. Proceedings of the National Academy of Sciences 2007;104(18):7682-7. 467. Vestergaard M, Kraft JF, Vosegaard T, Thøgersen L, Schiøtt B. Bicelles and other membrane mimics: comparison of structure, properties, and dynamics from MD simulations. Journal of Physical Chemistry B 2015;119(52):15831-43. 468. Mitra N, Liu Y, Liu J, Serebryany E, Mooney V, DeVree BT, et al. Calcium- dependent ligand binding and G-protein signaling of family B GPCR parathyroid hormone 1 receptor purified in nanodiscs. American Chemical Society Chemical Biology 2013;8(3):617-25. 469. Goldsmith BR, Mitala JJ, Josue J, Castro A, Lerner MB, Bayburt TH, et al. Biomimetic chemical sensors using nanoelectronic readout of olfactory receptor proteins. American Chemical Society Nano 2011;5(7):5408-16. 470. Proverbio D, Roos C, Beyermann M, Orban E, Dotsch V, Bernhard F. Functional properties of cell-free expressed human endothelin A and endothelin B receptors in artificial membrane environments. Biochimica et Biophysica Acta 2013;1828(9):2182-92. 471. Yoshiura C, Kofuku Y, Ueda T, Mase Y, Yokogawa M, Osawa M, et al. NMR analyses of the interaction between CCR5 and its ligand using functional reconstitution of CCR5 in lipid bilayers. Journal of the American Chemical Society 2010;132(19):6768-77. 200

472. Marty MT, Wilcox KC, Klein WL, Sligar SG. Nanodisc-solubilized membrane protein library reflects the membrane proteome. Analytical and Bioanalytical Chemistry 2013;405(12):4009-16. 473. Long AR, O'Brien CC, Malhotra K, Schwall CT, Albert AD, Watts A, et al. A detergent-free strategy for the reconstitution of active enzyme complexes from native biological membranes into nanoscale discs. BioMed Central Biotechnology 2013;13(1):41-50. 474. Roy J, Pondenis H, Fan TM, Das A. Direct Capture of Functional Proteins from Mammalian Plasma Membranes into Nanodiscs. Biochemistry 2015;54(41):6299-302. 475. Etzkorn M, Raschle T, Hagn F, Gelev V, Rice AJ, Walz T, et al. Cell-free expressed bacteriorhodopsin in different soluble membrane mimetics: biophysical properties and NMR accessibility. Structure 2013;21(3):394-401. 476. Rajesh S, Knowles T, Overduin M. Production of membrane proteins without cells or detergents. New Biotechnology 2011;28(3):250-4. 477. Vinothkumar KR. Membrane protein structures without crystals, by single particle electron cryomicroscopy. Current Opinion in Structural Biology 2015;33:103-14. 478. Zhou Q, Huang X, Sun S, Li X, Wang HW, Sui SF. Cryo-EM structure of SNAP- SNARE assembly in 20S particle. Cell Research 2015;25(5):551-60. 479. Kudryashev M, Castano-Diez D, Deluz C, Hassaine G, Grasso L, Graf-Meyer A,

et al. The structure of the mouse serotonin 5-HT3 receptor in lipid vesicles. Structure 2016;24(1):165-70. 480. Trahey M, Li MJ, Kwon H, Woodahl EL, McClary WD, Atkins WM. Applications of lipid nanodiscs for the study of membrane proteins by surface plasmon resonance. Current Protocols in Protein Science 2015;81(1):29.13.1-29.13.6. 481. Borch J, Torta F, Sligar SG, Roepstorff P. Nanodiscs for immobilization of lipid bilayers and membrane receptors: kinetic analysis of cholera toxin binding to a glycolipid receptor. Analytical Chemistry 2008;80(16):6245-52. 482. Viegas A, Viennet T, Etzkorn M. The power, pitfalls and potential of the nanodisc system for NMR-based studies. Biological Chemistry 2016;397(12):1335-54. 483. Puthenveetil R, Nguyen K, Vinogradova O. Nanodiscs and solution NMR: preparation, application and challenges. Nanotechnology Reviews 2017;6(1):111- 26. 484. Kijac AZ, Li Y, Sligar SG, Rienstra CM. Magic-angle spinning solid-state NMR spectroscopy of nanodisc-embedded human CYP3A4. Biochemistry 201

2007;46(48):13696-703. 485. Okude J, Ueda T, Kofuku Y, Sato M, Nobuyama N, Kondo K, et al. Identification of a conformational equilibrium that determines the efficacy and functional selectivity of the μ-opioid receptor. Angewandte Chemie (International Edition in English) 2015;54(52):15771-6. 486. Park SH, Berkamp S, Cook GA, Chan MK, Viadiu H, Opella SJ. Nanodiscs vs. macrodiscs for NMR of membrane proteins. Biochemistry 2011;50(42):8983-5. 487. Nikolaev M, Round E, Gushchin I, Polovinkin V, Balandin T, Kuzmichev P, et al. Integral membrane proteins can be crystallized directly from nanodiscs. Crystal Growth & Design 2017;17(3):945-8. 488. Daury L, Taveau JC, Salvador D, Glavier M, Lambert O. Reconstitution of membrane proteins into nanodiscs for single-particle electron microscopy. Methods in Molecular Biology 2017;1635:317-27. 489. Efremov RG, Gatsogiannis C, Raunser S. Lipid nanodiscs as a tool for high- resolution structure determination of membrane proteins by single-particle Cryo- EM. Methods in Enzymology 2017;594:1-30. 490. Rouck JE, Krapf JE, Roy J, Huff HC, Das A. Recent advances in nanodisc technology for membrane protein studies (2012-2017). The Federation of European Biochemical Societies Letters 2017;591(14):2057-88. 491. Khoshouei M, Radjainia M, Baumeister W, Danev R. Cryo-EM structure of haemoglobin at 3.2 Å determined with the Volta phase plate. Nature Communications 2017;8(1):16099-104. 492. Merk A, Bartesaghi A, Banerjee S, Falconieri V, Rao P, Davis MI, et al. Breaking Cryo-EM resolution barriers to facilitate drug discovery. Cell 2016;165(7):1698- 707. 493. Henderson R. The potential and limitations of neutrons, electrons and X-rays for atomic resolution microscopy of unstained biological molecules. Quarterly Reviews of Biophysics 1995;28(2):171-93. 494. Thompson RF, Walker M, Siebert CA, Muench SP, Ranson NA. An introduction to sample preparation and imaging by cryo-electron microscopy for structural biology. Methods 2016;100:3-15. 495. Kim J, Wu S, Tomasiak TM, Mergel C, Winter MB, Stiller SB, et al. Subnanometre-resolution electron cryomicroscopy structure of a heterodimeric ABC exporter. Nature 2015;517(7534):396-400. 496. Malhotra K, Alder NN. Reconstitution of mitochondrial membrane proteins into 202

nanodiscs by cell-Free expression. Methods in Molecular Biology 2017;1567:155-78. 497. Lyukmanova EN, Shenkarev ZO, Khabibullina NF, Kopeina GS, Shulepko MA, Paramonov AS, et al. Lipid–protein nanodiscs for cell-free production of integral membrane proteins in a soluble and folded state: comparison with detergent micelles, bicelles and liposomes. Biochimica et Biophysica Acta - Biomembranes 2012;1818(3):349-58. 498. Junge F, Haberstock S, Roos C, Stefer S, Proverbio D, Dotsch V, et al. Advances in cell-free protein synthesis for the functional and structural analysis of membrane proteins. New Biotechnology 2011;28(3):262-71. 499. Cappuccio JA, Blanchette CD, Sulchek TA, Arroyo ES, Kralj JM, Hinz AK, et al. Cell-free co-expression of functional membrane proteins and apolipoprotein, forming soluble nanolipoprotein particles. Molecular & Cellular Proteomics 2008;7(11):2246-53. 500. Yang J-P, Cirico T, Katzen F, Peterson TC, Kudlicki W. Cell-free synthesis of a functional G protein-coupled receptor complexed with nanometer scale bilayer discs. BioMed Central Biotechnology 2011;11(1):1-8. 501. Bocquet N, Kohler J, Hug MN, Kusznir EA, Rufer AC, Dawson RJ, et al. Real-

time monitoring of binding events on a thermostabilized human A2A receptor embedded in a lipid bilayer by surface plasmon resonance. Biochimica et Biophysica Acta 2015;1848(5):1224-33. 502. Lemieux RU, Spohr U. How Emil Fischer was led to the lock and key concept for enzyme specificity. Advances in Carbohydrate Chemistry and Biochemistry 1994;50:1-20. 503. Hofstadler SA, Sannes-Lowery KA. Applications of ESI-MS in drug discovery: interrogation of noncovalent complexes. Nature Reviews Drug Discovery 2006;5(7):585-95. 504. Severson WE, McDowell M, Ananthan S, Chung D-H, Rasmussen L, Sosa MI, et al. High-throughput screening of a 100,000 compound library for inhibitors of influenza A virus (H3N2). Journal of Biomolecular Screening 2008;13(9):879-87. 505. Szymański P, Markowicz M, Mikiciuk-Olasik E. Adaptation of high-throughput screening in drug discovery—toxicological screening tests. International Journal of Molecular Sciences 2012;13(1):427-52. 506. Tang X, Wang Y, Li D, Luo J, Liu M. Orphan G protein-coupled receptors (GPCRs): biological functions and potential drug targets. Acta Pharmacologica 203

Sinica 2012;33(3):363-71. 507. Dove A. Drug screening- beyond the bottleneck. Nature Biotechnology 1999;17(9):859-63. 508. Major J. Challenges and opportunities in high throughput screening: implications for new technologies. Journal of Biomolecular Screening 1998;3(1):13-7. 509. Turek-Etienne TC, Small EC, Soh SC, Xin TA, Gaitonde PV, Barrabee EB, et al. Evaluation of fluorescent compound interference in 4 fluorescence polarization assays: 2 kinases, 1 protease, and 1 phosphatase. Journal of Biomolecular Screening 2003;8(2):176-84. 510. Ma H, Deacon S, Horiuchi K. The challenge of selecting protein kinase assays for lead discovery optimization. Expert Opinion on Drug Discovery 2008;3(6):607- 21. 511. McKoy AF, Chen J, Schupbach T, Hecht MH. Structure activity relationships for a series of compounds that inhibit aggregation of the Alzheimer’s peptide, Aβ42: SAR of Aβ42 inhibitors. Chemical Biology & Drug Design 2014;84(5):505-12. 512. Wawer M, Bajorath J. Extraction of structure-activity relationship information from high-throughput screening data. Current Medicinal Chemistry 2009;16(31):4049-57. 513. Cunningham AR, Carrasquer CA, Qamar S, Maguire JM, Cunningham SL, Trent JO. Global structure–activity relationship model for nonmutagenic carcinogens using virtual ligand-protein interactions as model descriptors. Carcinogenesis 2012;33(10):1940-5. 514. Sterling T, Irwin JJ. ZINC 15 – ligand discovery for everyone. Journal of Chemical Information and Modeling 2015;55(11):2324-37. 515. Meng X-Y, Zhang H-X, Mezei M, Cui M. Molecular Docking: A powerful approach for structure-based drug discovery. Current Computer-Aided Drug Design 2011;7(2):146-57. 516. Chemi G, Gemma S, Campiani G, Brogi S, Butini S, Brindisi M. Computational tool for fast in silico evaluation of hERG K+ channel affinity. Frontiers in Chemistry 2017;5(7):1-9. 517. Wang L, Deng Y, Wu Y, Kim B, LeBard DN, Wandschneider D, et al. Accurate modeling of scaffold hopping transformations in drug discovery. Journal of Chemical Theory and Computation 2017;13(1):42-54. 518. Kroemer RT. Structure-based drug design: docking and scoring. Current Protein & Peptide Science 2007;8(4):312-28. 204

519. Pang YP, Xu K, Yazal JE, Prendergas FG. Successful molecular dynamics simulation of the zinc-bound farnesyltransferase using the cationic dummy atom approach. Protein Science 2000;9(10):1857-65. 520. Khandelwal A, Lukacova V, Comez D, Kroll DM, Raha S, Balaz S. A combination of Docking, QM/MM methods, and MD simulation for binding affinity estimation of metalloprotein ligands. Journal of Medicinal Chemistry 2005;48(17):5437-47. 521. Shah S, Friedman SH. An ESI-MS method for characterization of native and modified oligonucleotides used for RNA interference and other biological applications. Nature Protocols 2008;3(3):351-6. 522. Domon B, Aebersold R. Mass spectrometry and protein analysis. Science 2006;312(5771):212-7. 523. Chen J, Yin S, Wu Y, Ouyang J. Development of a native nanoelectrospray mass spectrometry method for determination of the drug-to-antibody ratio of antibody– drug conjugates. Analytical Chemistry 2013;85(3):1699-704. 524. Mann M, Hendrickson RC, Pandey A. Analysis of proteins and proteomes by mass spectrometry. Annual Review of Biochemistry 2001;70(1):437-73. 525. Loo JA. Studying noncovalent protein complexes by electrospray ionization mass spectrometry. Mass Spectrometry Reviews 1997;16(1):1-23. 526. Vivat Hannah V, Atmanene C, Zeyer D, Van Dorsselaer A, Sanglier-Cianferani S. Native MS: an 'ESI' way to support structure- and fragment-based drug discovery. Future Medicinal Chemistry 2010;2(1):35-50. 527. Beu SC, Senko MW, Quinn JP, Wampler FM, McLafferty FW. Fourier-transform electrospray instrumentation for tandem high-resolution mass spectrometry of large molecules. Journal of the American Society for Mass Spectrometry 1993;4(7):557-65. 528. Winger BE, Hofstadler SA, Bruce JE, Udseth HR, Smith RD. High-resolution accurate mass measurements of biomolecules using a new electrospray ionization ion cyclotron resonance mass spectrometer. Journal of the American Society for Mass Spectrometry 1993;4(7):566-77. 529. McLafferty FW. High-resolution tandem FT mass spectrometry above 10 kDa. Accounts of Chemical Research 1994;27(11):379-86. 530. Feng R, Konishi Y. Analysis of antibodies and other large glycoproteins in the mass range of 150,000-200,000 Da by electrospray ionization mass spectrometry. Analytical Chemistry 1992;64(18):2090-5. 531. Chen X, Li L, Chen S, Xu Y, Xia Q, Guo Y, et al. Identification of inhibitors of 205

the antibiotic-resistance target New Delhi metallo-β-lactamase 1 by both nanoelectrospray ionization mass spectrometry and ultrafiltration Liquid chromatography/mass spectrometry approaches. Analytical Chemistry 2013;85(16):7957-65. 532. Chen X, Qin S, Chen S, Li J, Li L, Wang Z, et al. A Ligand-observed mass spectrometry approach integrated into the fragment based Lead discovery pipeline. Scientific Reports 2015;5(1):8361-8. 533. Gelpi E. Interfaces for coupled liquid-phase separation/mass spectrometry techniques. An update on recent developments. Journal of Mass Spectrometry 2002;37(3):241-53. 534. Przybylski M, Glocker MO. Electrospray mass spectrometry of biomacromolecular complexes with noncovalent interactions—new analytical perspectives for supramolecular chemistry and molecular recognition processes. Angewandte Chemie (International Edition in English) 1996;35(8):806-26. 535. Fenn JB, Mann M, Meng CK, Wong SF, Whitehouse CM. Electrospray ionization–principles and practice. Mass Spectrometry Reviews 1990;9(1):37-70. 536. Fenn JB, Mann M, Meng CK, Wong SF, Whitehouse CM. Electrospray ionization for mass spectrometry of large biomolecules. Science 1989;246(4926):64-71. 537. Ho CS, Lam CWK, Chan MHM, Cheung RCK, Law LK, Lit LCW, et al. Electrospray ionisation mass spectrometry: principles and clinical applications. The Clinical Biochemist Reviews 2003;24(1):3-12. 538. Bruins AP. Mechanistic aspects of electrospray ionization. Journal of Chromatography A 1998;794(1):345-57. 539. Rosu F, De Pauw E, Gabelica V. Electrospray mass spectrometry to study drug- nucleic acids interactions. Biochimie 2008;90(7):1074-87. 540. de Hoffmann E, Stroobant V. Mass spectrometry: principles and applications. 3rd edition ed. The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England: John Wiley & Sons Ltd; 2007. 541. Zeleny J. Instability of electrified liquid surfaces. Physical Review 1917;10(1):1- 6. 542. Taflin DC, Ward TL, Davis EJ. Electrified droplet fission and the Rayleigh limit. Langmuir 1989;5(2):376-84. 543. Rostom AA, Robinson CV. Detection of the intact GroEL chaperonin assembly by mass spectrometry. Journal of the American Chemical Society 1999;121(19):4718-9. 206

544. Boys BL, Konermann L. Folding and assembly of hemoglobin monitored by electrospray mass spectrometry using an on-line dialysis system. Journal of the American Society for Mass Spectrometry 2007;18(1):8-16. 545. Kuprowski MC, Boys BL, Konermann L. Analysis of protein mixtures by electrospray mass spectrometry: effects of conformation and desolvation behavior on the signal intensities of hemoglobin subunits. Journal of the American Society for Mass Spectrometry 2007;18(7):1279-85. 546. Hunter CL, Mauk AG, Douglas DJ. Dissociation of heme from myoglobin and

cytochrome b5: comparison of behavior in solution and the gas phase. Biochemistry 1997;36(5):1018-25. 547. Kraunsoe JA, Aplin RT, Green B, Lowe G. An investigation of the binding of protein proteinase inhibitors to trypsin by electrospray ionization mass spectrometry. The Federation of European Biochemical Societies Letters 1996;396(1):108-12. 548. Kapur A, Beck JL, Brown SE, Dixon NE, Sheil MM. Use of electrospray ionization mass spectrometry to study binding interactions between a replication terminator protein and DNA. Protein Science 2002;11(1):147-57. 549. Akashi S, Osawa R, Nishimura Y. Evaluation of protein-DNA binding affinity by electrospray ionization mass spectrometry. Journal of the American Society for Mass Spectrometry 2005;16(1):116-25. 550. Kamadurai HB, Subramaniam S, Jones RB, Green-Church KB, Foster MP. Protein folding coupled to DNA binding in the catalytic domain of bacteriophage λ integrase detected by mass spectrometry. Protein Science 2003;12(3):620-6. 551. Gale DC, Goodlett DR, Light-Wahl KJ, Smith RD. Observation of duplex DNA- drug noncovalent complexes by electrospray ionization mass spectrometry. Journal of the American Chemical Society 1994;116(13):6027-8. 552. Wickham G, Iannitti P, Boschenok J, Sheil MM. Electrospray ionization mass spectrometry of covalent ligand-oligonucleotide adducts: evidence for specific duplex ion formation. Rapid Communications in Mass Spectrometry 1995;9(3):S197-S203. 553. Wu Q, Gao J, Joseph-McCarthy D, Sigal GB, Bruce JE, Whitesides GM, et al. Carbonic anhydrase-inhibitor binding: from solution to the gas phase. Journal of the American Chemical Society 1997;119(5):1157-8. 554. Xu N, Pasa-Tolic L, Smith RD, Ni S, Thrall BD. Electrospray ionization-mass spectrometry study of the interaction of cisplatin-adducted oligonucleotides with 207

human XPA minimal binding domain protein. Analytical Biochemistry 1999;272(1):26-33. 555. Whitehouse CM, Dreyer RN, Yamashita M, Fenn JB. Electrospray interface for liquid chromatographs and mass spectrometers. Analytical Chemistry 1985;57(3):675-9. 556. Zhu M, Gao Y, Wu Z, Zhao Y. A rapid HPLC/ESI-MS/MS method for quantitative analysis of isovalerylshikonin in rat plasma. Biomedical Chromatography 2010;24(4):413-9. 557. Breuker K. The study of protein–ligand interactions by mass spectrometry—a personal view. International Journal of Mass Spectrometry 2004;239(1):33-41. 558. Smith RD, Bruce JE, Wu Q, Lei QP. New mass spectrometric methods for the study of noncovalent associations of biopolymers. Chemical Society Reviews 1997;26(3):191-202. 559. Smith RD, Light-Wahl KJ. The observation of non-covalent interactions in solution by electrospray ionization mass spectrometry: Promise, pitfalls and prognosis. Biological Mass Spectrometry 1993;22(9):493-501. 560. Ganem B, Li YT, Henion JD. Detection of noncovalent receptor-ligand complexes by mass spectrometry. Journal of the American Chemical Society 1991;113(16):6294-6. 561. Potier N, Rogniaux H, Chevreux G, Van Dorsselaer A. Ligand–metal ion binding to proteins: investigation by ESI mass spectrometry. In: Burlingame AL, editor. Methods in Enzymology: Academic Press; 2005. vol 402 p. 361-89. 562. Cheng X, Chen R, Bruce JE, Schwartz BL, Anderson GA, Hofstadler SA, et al. Using electrospray ionization FTICR mass spectrometry to study competitive binding of inhibitors to carbonic anhydrase. Journal of the American Chemical Society 1995;117(34):8859-60. 563. Loo JA, Peifeng H, McConnell P, Tom Mueller W, Sawyer TK, Thanabal V. A study of Src SH2 domain protein—phosphopeptide binding interactions by electrospray ionization mass spectrometry. Journal of the American Society for Mass Spectrometry 1997;8(3):234-43. 564. Hao C, March RE, Croley TR, Smith JC, Rafferty SP. Electrospray ionization tandem mass spectrometric study of salt cluster ions. Part 1— Investigations of alkali metal chloride and sodium salt cluster ions. Journal of Mass Spectrometry 2001;36(1):79-96. 565. Zhou S, Hamburger M. Formation of sodium cluster ions in electrospray mass 208

dpectrometry. Rapid Communications in Mass Spectrometry 1996;10(7):797-800. 566. Flick TG, Cassou CA, Chang TM, Williams ER. Solution additives that desalt protein ions in native mass spectrometry. Analytical Chemistry 2012;84(17):7511-7. 567. Susa AC, Xia Z, Williams ER. Small emitter tips for native mass spectrometry of proteins and protein complexes from nonvolatile buffers that mimic the intracellular environment. Analytical Chemistry 2017;89(5):3116-22. 568. Susa AC, Xia Z, Williams ER. Native mass spectrometry from common buffers with salts that mimic the extracellular environment. Angewandte Chemie (International Edition in English) 2017;56(27):7912-5. 569. Andren PE, Emmett MR, Caprioli RM. Micro-Electrospray: Zeptomole/attomole per microliter sensitivity for peptides. Journal of the American Society for Mass Spectrometry 1994;5(9):867-9. 570. Kriger MS, Cook KD, Ramsey RS. Durable gold-coated fused silica capillaries for use in electrospray mass spectrometry. Analytical Chemistry 1995;67(2):385- 9. 571. Marshall AG, Hendrickson CL, Jackson GS. Fourier transform ion cyclotron resonance mass spectrometry: a primer. Mass Spectrometry Reviews 1998;17(1):1-35. 572. Gross JH, Roepstorff P. Mass Spectrometry: A textbook. 2nd ed: Springer-Verlag Berlin Heidelberg; 2011. 573. Paša Tolić L, Harms AC, Anderson GA, Smith RD, Willie A, Jorns MS. Electrospray ionization–mass spectrometry characterization of heterotetrameric sarcosine oxidase. Journal of the American Society for Mass Spectrometry 1998;9(5):510-5. 574. Benesch JLP. Collisional activation of protein complexes: picking up the pieces. Journal of the American Society for Mass Spectrometry 2009;20(3):341-8. 575. Mitchell Wells J, McLuckey SA. Collision‐induced dissociation (CID) of peptides and proteins. Methods in Enzymology 2005;402:148-85. 576. Valaskovic GA, Kelleher NL, McLafferty FW. Attomole protein characterization by capillary electrophoresis-mass spectrometry. Science 1996;273(5279):1199- 202. 577. Jensen PK, Pasa-Tolic L, Anderson GA, Horner JA, Lipton MS, Bruce JE, et al. Probing proteomes using capillary isoelectric focusing-electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Analytical 209

Chemistry 1999;71(11):2076-84. 578. Zhang L-K, Rempel D, Pramanik BN, Gross ML. Accurate mass measurements by Fourier transform mass spectrometry. Mass Spectrometry Reviews 2005;24(2):286-309. 579. Marshall AG. Milestones in fourier transform ion cyclotron resonance mass spectrometry technique development. International Journal of Mass Spectrometry 2000;200(1):331-56. 580. Siuzdak G. Mass analyzers and ion detectors. In: Siuzdak G, editor. Mass Spectrometry for Biotechnology. San Diego: Academic Press; 1996. p. 32-55. 581. Barbirz S, Jakob U, Glocker MO. Mass spectrometry unravels disulfide bond formation as the mechanism that activates a molecular chaperone. Journal of Biological Chemistry 2000;275(25):18759-66. 582. Lacorte S, Fernandez‐Alba AR. Time of flight mass spectrometry applied to the liquid chromatographic analysis of pesticides in water and food. Mass Spectrometry Reviews 2006;25(6):866-80. 583. Parker CH, Morgan CR, Rand KD, Engen JR, Jorgenson JW, Stafford DW. A conformational investigation of propeptide binding to the integral membrane protein γ-glutamyl carboxylase using nanodisc hydrogen exchange mass spectrometry. Biochemistry 2014;53(9):1511-20. 584. Hebling CM, Morgan CR, Stafford DW, Jorgenson JW, Rand KD, Engen JR. Conformational analysis of membrane proteins in phospholipid bilayer nanodiscs by hydrogen exchange Mass spectrometry. Analytical Chemistry 2010;82(13):5415-9. 585. Sokolovski M, Cveticanin J, Hayoun D, Korobko I, Sharon M, Horovitz A. Measuring inter-protein pairwise interaction energies from a single native mass spectrum by double-mutant cycle analysis. Nature Communications 2017;8(212):1-7. 586. Marty MT, Zhang H, Cui W, Gross ML, Sligar SG. Interpretation and deconvolution of nanodisc native mass spectra. Journal of the American Society for Mass Spectrometry 2014;25(2):269-77. 587. Hoi KK, Robinson CV, Marty MT. Unraveling the composition and behavior of heterogeneous lipid nanodiscs by mass spectrometry. Analytical Chemistry 2016;88(12):6199-204. 588. Han L, Kitova EN, Li J, Nikjah S, Lin H, Pluvinage B, et al. Protein–glycolipid interactions studied in vitro using ESI-MS and nanodiscs: insights into the 210

mechanisms and energetics of binding. Analytical Chemistry 2015;87(9):4888-96. 589. Mata IF, Wedemeyer WJ, Farrer MJ, Taylor JP, Gallo KA. LRRK2 in Parkinson's disease: protein domains and functional insights. Trends in Neurosciences 2006;29(5):286-93. 590. Lesage S, Brice A. Parkinson's disease: from monogenic forms to genetic susceptibility factors. Human Molecular Genetics 2009;18(R1):R48-R59. 591. Lesage S, Condroyer C, Lannuzel A, Lohmann E, Troiano A, Tison F, et al. Molecular analyses of the LRRK2 gene in European and North African autosomal dominant Parkinson's disease. Journal of Medical Genetics 2009;46(7):458-64. 592. Guruprasad K, Reddy BV, Pandit MW. Correlation between stability of a protein and its dipeptide composition: a novel approach for predicting in vivo stability of a protein from its primary sequence. Protein Engineering 1990;4(2):155-61. 593. Monera OD, Sereda TJ, Zhou NE, Kay CM, Hodges RS. Relationship of sidechain hydrophobicity and α-helical propensity on the stability of the single-stranded amphipathic α-helix. Journal of Peptide Science 1995;1(5):319-29. 594. Kyte J, Doolittle RF. A simple method for displaying the hydropathic character of a protein. Journal of Molecular Biology 1982;157(1):105-32. 595. Chang E, Ge B, Lee M, So M, Wang W. Investigation of the ligation efficiency of NdeI digested fragments. Journal of Experimental Microbiology and Immunology 2005;7:68-72. 596. Desjardins P, Conklin D. NanoDrop microvolume quantitation of nucleic acids. Journal of Visualized Experiments 2010;22(45):e2565-e8. 597. Structural Genomics C, Architecture et Fonction des Macromolécules B, Berkeley Structural Genomics C, China Structural Genomics C, Integrated Center for S, Function I, et al. Protein production and purification. Nature Methods 2008;5(2):135-46. 598. Vancraenenbroeck R, Lobbestael E, Weeks SD, Strelkov SV, Baekelandt V, Taymans JM, et al. Expression, purification and preliminary biochemical and structural characterization of the leucine rich repeat namesake domain of leucine rich repeat kinase 2. Biochimica et Biophysica Acta 2012;1824(3):450-60. 599. Greggio E, Taymans J-M, Zhen EY, Ryder J, Vancraenenbroeck R, Beilina A, et al. The Parkinson’s disease kinase LRRK2 autophosphorylates its GTPase domain at multiple sites. Biochemical and Biophysical Research Communications 2009;389(3):449-54. 600. Liu W, Chun E, Thompson AA, Chubukov P, Xu F, Katritch V, et al. Structural 211

basis for allosteric regulation of GPCRs by sodium ions. Science 2012;337(6091):232-6. 601. Alexandrov AI, Mileni M, Chien EYT, Hanson MA, Stevens RC. Microscale fluorescent thermal stability assay for membrane proteins. Structure 2008;16(3):351-9. 602. Jones KR, Choi U, Gao J-L, Thompson RD, Rodman LE, Malech HL, et al. A novel method for screening adenosine receptor specific agonists for use in adenosine drug development. Scientific Reports 2017;7(44816):1-10. 603. El-Tayeb A, Michael S, Abdelrahman A, Behrenswerth A, Gollos S, Nieber K, et

al. Development of polar adenosine A2A receptor agonists for inflammatory bowel

disease: synergism with A2B antagonists. American Chemical Society Medicinal Chemistry Letters 2011;2(12):890-5. 604. Huang P-C, Hsiao Y-T, Kao S-Y, Chen C-F, Chen Y-C, Chiang C-W, et al.

Adenosine A2A receptor up-regulates retinal wave frequency via starburst amacrine cells in the developing rat retina. Public Library of Science One 2014;9(4):e95090-e102.

605. O'Malley MA, Naranjo AN, Lazarova T, Robinson AS. Analysis of adenosine A2a receptor stability: effects of ligands and disulfide bonds. Biochemistry 2010;49(43):9181-9. 606. Jamshad M, Charlton J, Lin YP, Routledge SJ, Bawa Z, Knowles TJ, et al. G- protein coupled receptor solubilization and purification for biophysical analysis and functional studies, in the total absence of detergent. Bioscience Reports 2015;35(2):e00188-e97. 607. Isberg V, de Graaf C, Bortolato A, Cherezov V, Katritch V, Marshall FH, et al. Generic GPCR residue numbers - aligning topology maps minding the gaps. Trends in Pharmacological Sciences 2015;36(1):22-31. 608. Isberg V, Mordalski S, Munk C, Rataj K, Harpsøe K, Hauser AS, et al. GPCRdb: an information system for G protein-coupled receptors. Nucleic Acids Research 2016;44(Database issue):D356-D64. 609. Mi L-Z, Grey MJ, Nishida N, Walz T, Lu C, Springer TA. Functional and structural stability of the epidermal growth factor receptor in detergent micelles and phospholipid nanodiscs. Biochemistry 2008;47(39):10314-23. 610. Boldog T, Li M, Hazelbauer GL. Using nanodiscs to create water-soluble transmembrane chemoreceptors inserted in lipid bilayers. Methods in Enzymology 2007;423:317-35. 212

611. Li G, Kim J, Huang Z, St. Clair JR, Brown DA, London E. Efficient replacement of plasma membrane outer leaflet phospholipids and sphingolipids in cells with exogenous lipids. Proceedings of the National Academy of Sciences 2016;113(49):14025-30. 612. Bevers EM, Williamson PL. Getting to the outer leaflet: physiology of phosphatidylserine exposure at the plasma membrane. Physiological Reviews 2016;96(2):605-45. 613. Bryksin AV, Matsumura I. Overlap extension PCR cloning: a simple and reliable way to create recombinant plasmids. BioTechniques 2010;48(6):463-5. 614. Yen H-Y, Hopper JTS, Liko I, Allison TM, Zhu Y, Wang D, et al. Ligand binding to a G protein–coupled receptor captured in a mass spectrometer. Science Advances 2017;3(6):e1701016-e22. 615. Campuzano IDG, Li H, Bagal D, Lippens JL, Svitel J, Kurzeja RJM, et al. Native MS analysis of bacteriorhodopsin and an empty nanodisc by orthogonal acceleration Time-of-Flight, Orbitrap and ion cyclotron resonance. Analytical Chemistry 2016;88(24):12427-36. 616. Gault J, Donlan JAC, Liko I, Hopper JTS, Gupta K, Housden NG, et al. High- resolution mass spectrometry of small molecules bound to membrane proteins. Nature Methods 2016;13(4):333-6. 617. Hernandez H, Robinson CV. Determining the stoichiometry and interactions of macromolecular assemblies from mass spectrometry. Nature Protocols 2007;2(3):715-26. 618. Poulsen SA, Quinn RJ. Adenosine receptors: new opportunities for future drugs. Bioorganic & Medicinal Chemistry 1998;6(6):619-41. 619. Lagerstrom MC, Schioth HB. Structural diversity of G protein-coupled receptors and significance for drug discovery. Nature Reviews Drug Discovery 2008;7(4):339-57. 620. Venkatakrishnan AJ, Deupi X, Lebon G, Tate CG, Schertler GF, Babu MM. Molecular signatures of G-protein-coupled receptors. Nature 2013;494(7436):185-94. 621. Santos R, Ursu O, Gaulton A, Bento AP, Donadi RS, Bologa CG, et al. A comprehensive map of molecular drug targets. Nature Reviews Drug Discovery 2017;16(1):19-34. 622. Vu H, Pham NB, Quinn RJ. Direct screening of natural product extracts using mass spectrometry. Journal of Biomolecular Screening 2008;13(4):265-75. 213

623. Vu H, Roullier C, Campitelli M, Trenholme KR, Gardiner DL, Andrews KT, et al. Plasmodium Gametocyte Inhibition Identified from a Natural-Product-Based Fragment Library. American Chemical Society Chemical Biology 2013;8(12):2654-9. 624. Maresca A, Temperini C, Vu H, Pham NB, Poulsen SA, Scozzafava A, et al. Non- zinc mediated inhibition of carbonic anhydrases: coumarins are a new class of suicide inhibitors. Journal of the American Chemical Society 2009;131(8):3057- 62. 625. Vetting MW, Al-Obaidi N, Zhao S, San Francisco B, Kim J, Wichelecki DJ, et al. Experimental strategies for functional annotation and metabolism discovery: targeted screening of solute binding proteins and unbiased panning of metabolomes. Biochemistry 2015;54(3):909-31. 626. Pedro L, Quinn RJ. Native Mass Spectrometry in Fragment-Based Drug Discovery. Molecules 2016;21(8). 627. Jonker N, Kool J, Irth H, Niessen WM. Recent developments in protein-ligand affinity mass spectrometry. Analytical and Bioanalytical Chemistry 2011;399(8):2669-81. 628. Imaduwage KP, Go EP, Zhu Z, Desaire H. HAMS: high-affinity mass spectrometry screening. A high-throughput screening method for identifying the tightest-binding lead compounds for target proteins with no false positive identifications. Journal of the American Society for Mass Spectrometry 2016;27(11):1870-7. 629. Cologna CT, Gilles N, Echterbille J, Degueldre M, Servent D, de Pauw E, et al. Mass-spectrometry-based method for screening of new peptide ligands for G- protein-coupled receptors. Analytical and Bioanalytical Chemistry 2015;407(18):5299-307. 630. Locatelli-Hoops S, Yeliseev AA, Gawrisch K, Gorshkova I. Surface plasmon resonance applied to G protein-coupled receptors. Biomedical Spectroscopy and Imaging 2013;2(3):155-81. 631. Covy JP, Giasson BI. Identification of compounds that inhibit the kinase activity of leucine-rich repeat kinase 2. Biochemical and Biophysical Research Communications 2009;378(3):473-7. 632. Martiny-Baron G, Kazanietz MG, Mischak H, Blumberg PM, Kochs G, Hug H, et al. Selective inhibition of protein kinase C isozymes by the indolocarbazole Go 6976. Journal of Biological Chemistry 1993;268(13):9194-7. 214

633. Kramer T, Lo Monte F, Goring S, Okala Amombo GM, Schmidt B. Small molecule kinase inhibitors for LRRK2 and their application to Parkinson's disease models. American Chemical Society Chemical Neuroscience 2012;3(3):151-60. 634. Reichling LJ, Riddle SM. Leucine-rich repeat kinase 2 mutants I2020T and G2019S exhibit altered kinase inhibitor sensitivity. Biochemical and Biophysical Research Communications 2009;384(2):255-8. 635. Roux PP, Dorval G, Boudreau M, Angers-Loustau A, Morris SJ, Makkerh J, et al. K252a and CEP1347 are neuroprotective compounds that inhibit mixed-lineage kinase-3 and induce activation of Akt and ERK. Journal of Biological Chemistry 2002;277(51):49473-80. 636. Kase H, Iwahashi K, Nakanishi S, Matsuda Y, Yamada K, Takahashi M, et al. K- 252 compounds, novel and potent inhibitors of protein kinase C and cyclic nucleotide-dependent protein kinases. Biochemical and Biophysical Research Communications 1987;142(2):436-40. 637. Hidaka H, Kobayashi R. Pharmacology of protein kinase inhibitors. Annual Review of Pharmacology and Toxicology 1992;32(1):377-97. 638. Porter J. Small molecule c-Met kinase inhibitors: a review of recent patents. Expert Opinion on Therapeutic Patents 2010;20(2):159-77. 639. Maroney AC, Finn JP, Connors TJ, Durkin JT, Angeles T, Gessner G, et al. Cep- 1347 (KT7515), a semisynthetic inhibitor of the mixed lineage kinase family. Journal of Biological Chemistry 2001;276(27):25302-8. 640. Han SY. c-Jun N-terminal kinase signaling inhibitors under development. Toxicological Research 2008;24:93-100. 641. Ola A, Kerkelä R, Tokola H, Pikkarainen S, Skoumal R, Vuolteenaho O, et al. The mixed-lineage kinase 1–3 signalling pathway regulates stress response in cardiac myocytes via GATA-4 and AP-1 transcription factors. British Journal of Pharmacology 2010;159(3):717-25. 642. Sen S, Webber PJ, West AB. Dependence of leucine-rich repeat kinase 2 (LRRK2) kinase activity on dimerization. Journal of Biological Chemistry 2009;284(52):36346-56. 643. Smith IM, Hoshi N. ATP competitive protein kinase C inhibitors demonstrate distinct state-dependent inhibition. Public Library of Science One 2011;6(10):e26338-e46. 644. Chen H-M, Wang L, D’Mello SR. A chemical compound commonly used to inhibit PKR, {8-(imidazol-4-ylmethylene)-6H-azolidino[5,4-g]benzothiazol-7- 215

one}, protects neurons by inhibiting cyclin-dependent kinase. The European Journal of Neuroscience 2008;28(10):2003-16. 645. Nichols RJ, Dzamko N, Hutti JE, Cantley LC, Deak M, Moran J, et al. Substrate specificity and inhibitors of LRRK2, a protein kinase mutated in Parkinson's disease. The Biochemical Journal 2009;424(1):47-60. 646. Ramsden N, Perrin J, Ren Z, Lee BD, Zinn N, Dawson VL, et al. Chemoproteomics-based design of potent LRRK2-selective lead compounds that attenuate Parkinson's disease-related toxicity in human neurons. American Chemical Society Chemical Biology 2011;6(10):1021-8. 647. Lee BD, Shin J-H, VanKampen J, Petrucelli L, West AB, Ko HS, et al. Inhibitors of leucine-rich repeat kinase-2 protect against models of Parkinson's disease. Nature Medicine 2010;16(9):998-1000. 648. Perabo FG, Frossler C, Landwehrs G, Schmidt DH, von Rucker A, Wirger A, et al. Indirubin-3'-monoxime, a CDK inhibitor induces growth inhibition and apoptosis-independent up-regulation of survivin in transitional cell cancer. Anticancer Research 2006;26(3a):2129-35. 649. Dieter P, Fitzke E. RO 31-8220 and RO 31-7549 show improved selectivity for protein kinase C over staurosporine in macrophages. Biochemical and Biophysical Research Communications 1991;181(1):396-401. 650. Davis PD, Hill CH, Keech E, Lawton G, Nixon JS, Sedgwick AD, et al. Potent selective inhibitors of protein kinase C. The Federation of European Biochemical Societies Letters 1989;259(1):61-3. 651. Beltman J, McCormick F, Cook SJ. The selective protein kinase C inhibitor, Ro- 31-8220, inhibits mitogen-activated protein kinase phosphatase-1 (MKP-1) expression, induces c-Jun expression, and activates Jun N-terminal kinase. Journal of Biological Chemistry 1996;271(43):27018-24. 652. Yun H, Heo HY, Kim HH, DooKim N, Seol W. Identification of chemicals to inhibit the kinase activity of leucine-rich repeat kinase 2 (LRRK2), a Parkinson's disease-associated protein. Bioorganic & Medicinal Chemistry Letters 2011;21(10):2953-7. 653. Deng X, Dzamko N, Prescott A, Davies P, Liu Q, Yang Q, et al. Characterization of a selective inhibitor of the Parkinson's disease kinase LRRK2. Nature Chemical Biology 2011;7(4):203-5. 654. Deng X, Yang Q, Kwiatkowski N, Sim T, McDermott U, Settleman JE, et al. Discovery of a benzo[e]pyrimido-[5,4-b][1,4]diazepin-6(11H)-one as a potent and 216

selective inhibitor of big MAP kinase 1. American Chemical Society Medicinal Chemistry Letters 2011;2(3):195-200. 655. Kavanagh ME, Doddareddy MR, Kassiou M. The development of CNS-active LRRK2 inhibitors using property-directed optimisation. Bioorganic & Medicinal Chemistry Letters 2013;23(13):3690-6. 656. Davies SP, Reddy H, Caivano M, Cohen P. Specificity and mechanism of action of some commonly used protein kinase inhibitors. The Biochemical Journal 2000;351(Pt 1):95-105. 657. Liu M, Dobson B, Glicksman MA, Yue Z, Stein RL. Kinetic mechanistic studies of wild-type leucine-rich repeat kinase 2: characterization of the kinase and GTPase activities. Biochemistry 2010;49(9):2008-17. 658. Anekpankul T, Goto M, Sasaki M, Pavasant P, Shotipruk A. Extraction of anticancer damnacanthal from roots of Morinda citrifolia by subcritical water. Separation and Purification Technology 2007;55(3):343-9. 659. Lee BD, Shin J-H, VanKampen J, Petrucelli L, West AB, Ko HS, et al. Inhibitors of leucine rich repeat kinase 2 (LRRK2) protect against LRRK2-models of Parkinson’s disease. Nature Medicine 2010;16(9):998-1000. 660. Faltynek CR, Schroeder J, Mauvais P, Miller D, Wang S, Murphy D, et al. Damnacanthal is a highly potent, selective inhibitor of p56lck tyrosine kinase activity. Biochemistry 1995;34(38):12404-10. 661. Eynott PR, Xu L, Bennett BL, Noble A, Leung S-Y, Nath P, et al. Effect of an inhibitor of Jun N-terminal protein kinase, SP600125, in single allergen challenge in sensitized rats. Immunology 2004;112(3):446-53. 662. Bennett BL, Sasaki DT, Murray BW, O'Leary EC, Sakata ST, Xu W, et al. SP600125, an anthrapyrazolone inhibitor of Jun N-terminal kinase. Proceedings of the National Academy of Sciences 2001;98(24):13681-6. 663. Liu M, Poulose S, Schuman E, Zaitsev AD, Dobson B, Auerbach K, et al. Development of a mechanism-based high-throughput screen assay for leucine- rich repeat kinase 2 - discovery of LRRK2 inhibitors. Analytical Biochemistry 2010;404(2):186-92. 664. Choi HG, Zhang J, Deng X, Hatcher JM, Patricelli MP, Zhao Z, et al. Brain penetrant LRRK2 inhibitor. American Chemical Society Medicinal Chemistry Letters 2012;3(8):658-62. 665. Liu Z, Galemmo RA, Fraser KB, Moehle MS, Sen S, Volpicelli-Daley LA, et al. Unique functional and structural properties of the LRRK2 protein ATP-binding 217

pocket. Journal of Biological Chemistry 2014;289(47):32937-51. 666. Estrada AA, Liu X, Baker-Glenn C, Beresford A, Burdick DJ, Chambers M, et al. Discovery of highly potent, selective, and brain-penetrable leucine-rich repeat kinase 2 (LRRK2) small molecule inhibitors. Journal of Medicinal Chemistry 2012;55(22):9416-33. 667. Fuji RN, Flagella M, Baca M, MA SB, Brodbeck J, Chan BK, et al. Effect of selective LRRK2 kinase inhibition on nonhuman primate lung. Science Translational Medicine 2015;7(273):273ra15. 668. Zhang J, Deng X, Choi HG, Alessi DR, Gray NS. Characterization of TAE684 as a potent LRRK2 kinase inhibitor. Bioorganic & Medicinal Chemistry Letters 2012;22(5):1864-9. 669. Schonherr C, Ruuth K, Yamazaki Y, Eriksson T, Christensen J, Palmer RH, et al. Activating ALK mutations found in neuroblastoma are inhibited by Crizotinib and NVP-TAE684. The Biochemical Journal 2011;440(3):405-13. 670. Bossi RT, Saccardo MB, Ardini E, Menichincheri M, Rusconi L, Magnaghi P, et al. Crystal structures of anaplastic lymphoma kinase in complex with ATP competitive inhibitors. Biochemistry 2010;49(32):6813-25. 671. Reith AD, Bamborough P, Jandu K, Andreotti D, Mensah L, Dossang P, et al. GSK2578215A; a potent and highly selective 2-arylmethyloxy-5-substitutent-N- arylbenzamide LRRK2 kinase inhibitor. Bioorganic & Medicinal Chemistry Letters 2012;22(17):5625-9. 672. Chan BK, Estrada AA, Chen H, Atherall J, Baker-Glenn C, Beresford A, et al. Discovery of a highly selective, brain-penetrant aminopyrazole LRRK2 inhibitor. American Chemical Society Medicinal Chemistry Letters 2013;4(1):85-90. 673. Garofalo AW, Adler M, Aubele DL, Bowers S, Franzini M, Goldbach E, et al. Novel cinnoline-based inhibitors of LRRK2 kinase activity. Bioorganic & Medicinal Chemistry Letters 2013;23(1):71-4. 674. Franzini M, Ye XM, Adler M, Aubele DL, Garofalo AW, Gauby S, et al. Triazolopyridazine LRRK2 kinase inhibitors. Bioorganic & Medicinal Chemistry Letters 2013;23(7):1967-73. 675. Hoflack J, Blom P, inventors; Ipsen Pharma S.A.S.Oncodesign S.A., assignee. Patent: Macrocyclic LRRK2 kinase inhibitors. 2013.

218

Appendix

SI 1: A phylogenetic tree of TKL group. The LRRK2 is a close relative of the receptor- interacting protein kinases (RIPKs) and the mixed lineage kinases (MLK) families. The scale on the bottom represents genetic distances in substitutions per nucleotide (generated by http://www.kinase.com).

219

SI 2: Sequences of LRRK2 subdomains of MAPKKK kinase domain, Roc domain, and

COR domain.

Alignment: E:\LRRK2\Confirmation\LRRK2.gb

....|....| ....|....| ....|....| ....|....| ....|....| 10 20 30 40 50 kinase caagctccag agtttctcct aggtgatggc agttttggat cagtttaccg Roc aaccgaatga aacttatgat tgtgggaaat actgggagtg gtaaaaccac COR ttcaagatcc gagatcagct tgttgttgga cagctgattc cagactgcta

....|....| ....|....| ....|....| ....|....| ....|....| 60 70 80 90 100 kinase agcagcctat gaaggagaag aagtggctgt gaagattttt aataaacata Roc cttattgcag caattaatga aaaccaagaa atcagatctt ggaatgcaaa COR tgtagaactt gaaaaaatca ttttatcgga gcgtaaaaat gtgccaattg

....|....| ....|....| ....|....| ....|....| ....|....| 110 120 130 140 150 kinase catcactcag gctgttaaga caagagcttg tggtgctttg ccacctccac Roc gtgccacagt tggcatagat gtgaaagact ggcctatcca aataagagac COR aatttcccgt aattgaccgg aaacgattat tacaactagt gagagaaaat

....|....| ....|....| ....|....| ....|....| ....|....| 160 170 180 190 200 kinase caccccagtt tgatatcttt gctggcagct gggattcgtc cccggatgtt Roc aaaagaaaga gagatctcgt cctaaatgtg tgggattttg caggtcgtga COR cagctgcagt tagatgaaaa tgagcttcct cacgcagttc actttctaaa

....|....| ....|....| ....|....| ....|....| ....|....| 210 220 230 240 250 kinase ggtgatggag ttagcctcca agggttcctt ggatcgcctg cttcagcagg Roc ggaattctat agtactcatc cccattttat gacgcagcga gcattgtacc COR tgaatcagga gtccttcttc attttcaaga cccagcactg cagttaagtg

....|....| ....|....| ....|....| ....|....| ....|....| 260 270 280 290 300 kinase acaaagccag cctcactaga accctacagc acaggattgc actccacgta Roc ttgctgtcta tgacctcagc aagggacagg ctgaagttga tgccatgaag COR acttgtactt tgtggaaccc aagtggcttt gtaaaatcat ggcacagatt

....|....| ....|....| ....|....| ....|....| ....|....| 310 320 330 340 350 kinase gctgatggtt tgagatacct ccactcagcc atgattatat accgagacct Roc ccttggctct tcaatataaa ggctcgcgct tcttcttccc ctgtgattct COR ttgacagtga aagtggaagg ttgtccaaaa caccctaagg gaattatttc

....|....| ....|....| ....|....| ....|....| ....|....| 360 370 380 390 400 kinase gaaaccccac aatgtgctgc ttttcacact gtatcccaat gctgccatca Roc cgttggcaca catttggatg tttctgatga gaagcaacgc aaagcctgca COR gcgtagagat gtggaaaaat ttctttcaaa gaaaaggaaa tttccaaaga

220

....|....| ....|....| ....|....| ....|....| ....|....| 410 420 430 440 450 kinase ttgcaaagat tgctgactac ggcattgctc agtactgctg tagaatgggg Roc tgagtaaaat caccaaggaa ctcctgaata agcgagggtt ccctgccata COR actacatgac acagtatttt aagctcctag aaaaattcca gattgctttg

....|....| ....|....| ....|....| ....|....| ....|....| 460 470 480 490 500 kinase ataaaaacat cagagggcac accagggttt cgtgcacctg aagttgccag Roc cgagattacc actttgtgaa tgccaccgag gaatctgatg ctttggcaaa COR ccaataggag aagaatattt gctggttcca agcagtttgt ctgaccacag

....|....| ....|....| ....|....| ....|....| ....|....| 510 520 530 540 550 kinase aggaaatgtc atttataacc aacaggctga tgtttattca tttggtttac Roc acttcggaaa accatcataa acgagagcct taat...... COR gcctgtgata gagcttcccc attgtgagaa ctctgaaatt atcatccgac

....|....| ....|....| ....|....| ....|....| ....|....| 560 570 580 590 600 kinase tactctatga cattttgaca actggaggta gaatagtaga gggtttgaag Roc ...... COR tatatgaaat gccttatttt ccaatgggat tttggtcaag attaatcaat

....|....| ....|....| ....|....| ....|....| ....|....| 610 620 630 640 650 kinase tttccaaatg agtttgatga attagaaata caaggaaaat tacctgatcc Roc ...... COR cgattacttg agatttcacc ttacatgctt tcagggagag aacgagcact

....|....| ....|....| ....|....| ....|....| ....|....| 660 670 680 690 700 kinase agttaaagaa tatggttgtg ccccatggcc tatggttgag aaattaatta Roc ...... COR tcgcccaaac agaatgtatt ggcgacaagg catttactta aattggtctc

....|....| ....|....| ....|....| ....|....| ....|....| 710 720 730 740 750 kinase aacagtgttt gaaagaaaat cctcaagaaa ggcctacttc tgcccaggtc Roc ...... COR ctgaagctta ttgtctggta ggatctgaag tcttagacaa tcatccagag

....|....| ....|....| ....|....| ....|....| ....|....| 760 770 780 790 800 kinase tttgacattt tgaattcagc tgaattagtc ...... Roc ...... COR agtttcttaa aaattacagt tccttcttgt agaaaaggct gtattctttt

....|....| ....|....| ....|....| ....|....| ....|....| 810 820 830 840 850 kinase ...... Roc ...... COR gggccaagtt gtggaccaca ttgattctct catggaagaa tggtttcctg

221

....|....| ....|....| ....|....| ....|....| ....|....| 860 870 880 890 900 kinase ...... Roc ...... COR ggttgctgga gattgatatt tgtggtgaag gagaaactct gttgaagaaa

....|....| ....|....| ....|....| ....|....| ....|....| 910 920 930 940 950 kinase ...... Roc ...... COR tgggcattat atagttttaa tgatggcgaa gaacatcaaa aaatcttact

....|....| ....|....| ....|....| ....|....| ....|....| 960 970 980 990 1000 kinase ...... Roc ...... COR tgatgacttg atgaagaaag cagaggaagg agatctctta gtaaatccag

....|....| ....|....| ....|....| ....|....| ....|....| 1010 1020 1030 1040 1050 kinase ...... Roc ...... COR atcaaccaag gctcaccatt ccaatatctc agattgcccc tgacttgatt

....|....| ....|....| ....|....| ....|....| ....|....| 1060 1070 1080 1090 1100 kinase ...... Roc ...... COR ttggctgacc tgcctagaaa tattatgttg aataatgatg agttggaatt

....|....| ....|....| ....|....| ....|....| ....|....| 1110 1120 1130 1140 1150 kinase ...... Roc ...... COR tgaacaagct ccagagtttc tcctaggtga tggcagtttt ggatcagttt

....|....| ....|....| ....|....| ....|....| ....|....| 1160 1170 1180 1190 1200 kinase ...... Roc ...... COR accgagcagc ctatgaagga gaagaagtgg ctgtgaagat ttttaataaa

....|....| ....|....| ....|....| ....|....| ....|....| 1210 1220 1230 1240 1250 kinase ...... Roc ...... COR catacatcac tcaggctgtt aagacaagag cttgtggtgc tttgccacct

....|....| ....|....| ....|....| ....|....| ....|....| 1260 1270 1280 1290 1300 kinase ...... Roc ...... COR ccaccacccc agtttgatat ctttgctggc agctgggatt cgtccccgga

222

....|....| ....|....| ....|....| ....|....| ....|....| 1310 1320 1330 1340 1350 kinase ...... Roc ...... COR tgttggtgat ggagttagcc tccaagggtt ccttggatcg cctgcttcag

....|....| ....|....| ....|....| ....|....| ....|....| 1360 1370 1380 1390 1400 kinase ...... Roc ...... COR caggacaaag ccagcctcac tagaacccta cagcacagga ttgcactcca

....|....| ....|....| ....|....| ....|....| ....|....| 1410 1420 1430 1440 1450 kinase ...... Roc ...... COR cgtagctgat ggtttgagat acctccactc agccatgatt atataccgag

....|....| ....|....| ....|....| ....|....| ....|....| 1460 1470 1480 1490 1500 kinase ...... Roc ...... COR acctgaaacc ccacaatgtg ctgcttttca cactgtatcc caatgctgcc

....|....| ....|....| ....|....| ....|....| ....|....| 1510 1520 1530 1540 1550 kinase ...... Roc ...... COR atcattgcaa agattgctga ctacggcatt gctcagtact gctgtagaat

....|....| ....|....| ....|....| ....|....| ....|....| 1560 1570 1580 1590 1600 kinase ...... Roc ...... COR ggggataaaa acatcagagg gcacaccagg gtttcgtgca cctgaagttg

....|....| ....|....| ....|....| ....|....| ....|....| 1610 1620 1630 1640 1650 kinase ...... Roc ...... COR ccagaggaaa tgtcatttat aaccaacagg ctgatgttta ttcatttggt

....|....| ....|....| ....|....| ....|....| ....|....| 1660 1670 1680 1690 1700 kinase ...... Roc ...... COR ttactactct atgacatttt gacaactgga ggtagaatag tagagggttt

....|....| ....|....| ....|....| ....|....| ....|....| 1710 1720 1730 1740 1750 kinase ...... Roc ...... COR gaagtttcca aatgagtttg atgaattaga aatacaagga aaattacctg

223

....|....| ....|....| ....|....| ....|....| ....|....| 1760 1770 1780 1790 1800 kinase ...... Roc ...... COR atccagttaa agaatatggt tgtgccccat ggcctatggt tgaaaaatta

....|....| ....|....| ....|....| ....|....| ....|....| 1810 1820 1830 1840 1850 kinase ...... Roc ...... COR attaaacagt gtttgaaaga aaatcctcaa gaaaggccta cttctgccca

....|....| ....|....| ....|....| ....|....| ....|....| 1860 1870 1880 1890 1900 kinase ...... Roc ...... COR ggtctttgac attttgaatt cagctgaatt agtctgtctg acgagacgca

....|....| ....|....| ....|....| ....|....| ....|....| 1910 1920 1930 1940 1950 kinase ...... Roc ...... COR ttttattacc taaaaacgta attgttgaat gcatggttgc tacacatcac

....|....| ....|....| ....|....| ....|....| ....|....| 1960 1970 1980 1990 2000 kinase ...... Roc ...... COR aacagcagga atgcaagcat ttggctgggc tgtgggcaca ccgacagagg

....|....| ....|....| ....|....| ....|....| ....|....| 2010 2020 2030 2040 2050 kinase ...... Roc ...... COR acagctctca tttcttgact taaatactga aggatacact tctgaggaag

....|....| ....|....| ....|....| ....|....| ....|....| 2060 2070 2080 2090 2100 kinase ...... Roc ...... COR ttgctgatag tagaatattg tgcttagcct tggtgcatct tcctgttgaa

....|....| ....|....| ....|....| ....|....| ....|....| 2110 2120 2130 2140 2150 kinase ...... Roc ...... COR aaggaaagct ggattgtgtc tgggacacag tctggtactc tcctggtcat

....|....| ....|....| ....|....| ....|....| ....|....| 2160 2170 2180 2190 2200 kinase ...... Roc ...... COR caataccgaa gatgggaaaa agagacatac cctagaaaag atgactgatt

224

....|....| ....|....| ....|....| ....|....| ....|....| 2210 2220 2230 2240 2250 kinase ...... Roc ...... COR ctgtcacttg tttgtattgc aattcctttt ccaagcaaag caaacaaaaa

....|....| ....|....| ....|....| ....|....| ....|....| 2260 2270 2280 2290 2300 kinase ...... Roc ...... COR aattttcttt tggttggaac cgctgatggc aagttagcaa tttttgaaga

....|....| ....|....| ....|....| ....|....| ....|....| 2310 2320 2330 2340 2350 kinase ...... Roc ...... COR taagactgtt aagcttaaag gagctgctcc tttgaagata ctaaatatag

....|....| ....|....| ....|....| ....|....| ....|....| 2360 2370 2380 2390 2400 kinase ...... Roc ...... COR gaaatgtcag tactccattg atgtgtttga gtgaatccac aaattcaacg

....|....| ....|....| ....|....| ....|....| ....|....| 2410 2420 2430 2440 2450 kinase ...... Roc ...... COR gaaagaaatg taatgtgggg aggatgtggc acaaagattt tctccttttc

....|....| ....|....| ....|....| ....|....| ....|....| 2460 2470 2480 2490 2500 kinase ...... Roc ...... COR taatgatttc accattcaga aactcattga gacaagaaca agccaactgt

....|....| ....|....| ....|....| ....|....| ....|....| 2510 2520 2530 2540 2550 kinase ...... Roc ...... COR tttcttatgc agctttcagt gattccaaca tcataacagt ggtggtagac

....|....| ....|....| ....|....| ....|....| ....|....| 2560 2570 2580 2590 2600 kinase ...... Roc ...... COR actgctctct atattgctaa gcaaaatagc cctgttgtgg aagtgtggga

....|....| ....|....| ....|....| ....|....| ....|....| 2610 2620 2630 2640 2650 kinase ...... Roc ...... COR taagaaaact gaaaaactct gtggactaat agactgcgtg cactttttaa

225

....|....| ....|....| ....|....| ....|....| ....|....| 2660 2670 2680 2690 2700 kinase ...... Roc ...... COR gggaggtaac ggtaaaagaa aacaaggaat caaaacacaa aatgtcttat

....|....| ....|....| ....|....| ....|....| ....|....| 2710 2720 2730 2740 2750 kinase ...... Roc ...... COR tctgggagag tgaaaaccct ctgccttcag aagaacactg ctctttggat

....|....| ....|....| ....|....| ....|....| ....|....| 2760 2770 2780 2790 2800 kinase ...... Roc ...... COR aggaactgga ggaggccata ttttactcct ggatctttca actcgtcgac

....|....| ....|....| ....|....| ....|....| ....|....| 2810 2820 2830 2840 2850 kinase ...... Roc ...... COR ttatacgtgt aatttacaac ttttgtaatt cggtcagagt catgatgaca

....|....| ....|....| ....|....| ....|....| ....|....| 2860 2870 2880 2890 2900 kinase ...... Roc ...... COR gcacagctag gaagccttaa aaatgtcatg ctggtattgg gctacaaccg

....|....| ....|....| ....|....| ....|....| ....|....| 2910 2920 2930 2940 2950 kinase ...... Roc ...... COR gaaaaatact gaaggtacac aaaagcagaa agagatacaa tcttgcttga

....|....| ....|....| ....|....| ....|....| ....|....| 2960 2970 2980 2990 3000 kinase ...... Roc ...... COR ccgtttggga catcaatctt ccacatgaag tgcaaaattt agaaaaacac

....|....| ....|....| ....|....| ....|....| ....|....| 3010 3020 3030 3040 3050 kinase ...... Roc ...... COR attgaagtga gaaaagaatt agctgaaaaa atgagacgaa catctgttga

. kinase . Roc . COR g

226

SI 3: Cloned plasmids with sequences of different MSP variants and A2AR constructs.

Alignment: E:\Results\hMSP\msp variants.gb

....|....| ....|....| ....|....| ....|....| ....|....| 10 20 30 40 50 MSP1D1 CGTCCGGCGT AGAGGATCGA GATCTCGATC CCGCGAAATT AATACGACTC MSP1 ATGGGTCATC ATCATCATCA TCACATTGAG GGACGTCTGA AGCTGTTGGA MSP1D2 ATGGGTCATC ATCATCATCA TCATCACGAT TATGATATTC CTACTACTGA MSP2N2 ATGGGTCATC ATCATCATCA TCATCACGAT TATGATATTC CTACTACTGA MSP2N3 ATGGGTCATC ATCATCATCA TCATCACGAT TATGATATTC CTACTACTGA MSP1E1D1 TAATACGACT CACTATAGGG GAATTGTGAG CGGATAACAA TTCCCCTCTA MSP1E3D1 ATGGGTCATC ATCATCATCA TCATCACGAT TATGATATTC CTACTACTGA MSP1E2D1 CGTCCGGCGT AGAGGATCGA GATCTCGATC CCGCGAAATT AATACGACTC A2AR ATGCCCATCA TGGGCTCCTC GGTGTACATC ACGGTGGAGC TGGCCATTGC A2AR-GFP ATGAAGACGA TCATCGCCCT GAGCTACATC TTCTGCCTGG TGTTCGCCGA A2AR-GS li ATGCCCATCA TGGGCTCCTC GGTGTACATC ACGGTGGAGC TGGCCATTGC

....|....| ....|....| ....|....| ....|....| ....|....| 60 70 80 90 100 MSP1D1 ACTATAGGGG AATTGTGAGC GGATAACAAT TCCCCTCTAG AAATAATTTT MSP1 CAATTGGGAC TCTGTTACGT CTACCTTCAG TAAACTTCGC GAACAACTGG MSP1D2 GAATTTGTAT TTTCAGGGTC CCGTGACGCA GGAATTCTGG GACAACCTGG MSP2N2 GAATTTGTAT TTTCAGGGTT CTACCTTCAG TAAACTTCGC GAACAACTGG MSP2N3 GAATTTGTAT TTTCAGGGTT CTACCTTCAG TAAACTTCGC GAACAACTGG MSP1E1D1 GAAATAATTT TGTTTAACTT TAAGAAGGAG ATATACCATG GGTCATCATC MSP1E3D1 GAATTTGTAT TTTCAGGGTT CTACCTTCAG TAAACTTCGC GAACAACTGG MSP1E2D1 ACTATAGGGG AATTGTGAGC GGATAACAAT TCCCCTCTAG AAATAATTTT A2AR TGTGCTGGCC ATCCTGGGCA ATGTGCTGGT GTGCTGGGCC GTGTGGCTCA A2AR-GFP CTACAAGGAC GATGATGACA AGGGCGCGCC ACCCATCATG GGCTCCTCGG A2AR-GS li TGTGCTGGCC ATCCTGGGCA ATGTGCTGGT GTGCTGGGCC GTGTGGCTCA

....|....| ....|....| ....|....| ....|....| ....|....| 110 120 130 140 150 MSP1D1 GTTTAACTTT AAGAAGGAGA TATACCATGG GTCATCATCA TCATCATCAT MSP1 GCCCCGTGAC GCAGGAATTC TGGGACAACC TGGAAAAAGA AACCGAGGGA MSP1D2 AAAAAGAAAC CGAGGGACTG CGTCAGGAAA TGTCCAAAGA TTTAGAAGAG MSP2N2 GCCCCGTGAC GCAGGAATTC TGGGACAACC TGGAAAAAGA AACCGAGGGA MSP2N3 GCCCCGTGAC GCAGGAATTC TGGGACAACC TGGAAAAAGA AACCGAGGGA MSP1E1D1 ATCATCATCA TCACGATTAT GATATTCCTA CTACTGAGAA TTTGTATTTT MSP1E3D1 GCCCCGTGAC GCAGGAATTC TGGGACAACC TGGAAAAAGA AACCGAGGGA MSP1E2D1 GTTTAACTTT AAGAAGGAGA TATACCATGG GTCATCATCA TCATCATCAT A2AR ACAGCAACCT GCAGAACGTC ACCAACTACT TTGTGGTGTC ACTGGCGGCG A2AR-GFP TGTACATCAC GGTGGAGCTG GCCATTGCTG TGCTGGCCAT CCTGGGCAAT A2AR-GS li ACAGCAACCT GCAGAACGTC ACCAACTACT TTGTGGTGTC ACTGGCGGCG

....|....| ....|....| ....|....| ....|....| ....|....| 160 170 180 190 200 MSP1D1 CACGATTATG ATATTCCTAC TACTGAGAAT TTGTATTTTC AGGGTTCTAC MSP1 CTGCGTCAGG AAATGTCCAA AGATTTAGAA GAGGTGAAGG CCAAGGTTCA MSP1D2 GTGAAGGCCA AGGTTCAGCC ATATCTCGAT GACTTTCAGA AAAAATGGCA MSP2N2 CTGCGTCAGG AAATGTCCAA AGATTTAGAA GAGGTGAAGG CCAAGGTTCA MSP2N3 CTGCGTCAGG AAATGTCCAA AGATTTAGAA GAGGTGAAGG CCAAGGTTCA MSP1E1D1 CAGGGTTCTA CCTTCAGTAA ACTTCGCGAA CAACTGGGCC CCGTGACGCA MSP1E3D1 CTGCGTCAGG AAATGTCCAA AGATTTAGAA GAGGTGAAGG CCAAGGTTCA MSP1E2D1 CACGATTATG ATATTCCTAC TACTGAGAAT TTGTATTTTC AGGGTTCTAC A2AR GCCGACATCG CAGTGGGTGT GCTCGCCATC CCCTTTGCCA TCACCATCAG A2AR-GFP GTGCTGGTGT GCTGGGCCGT GTGGCTCAAC AGCAACCTGC AGAACGTCAC A2AR-GS li GCCGACATCG CAGTGGGTGT GCTCGCCATC CCCTTTGCCA TCACCATCAG

227

....|....| ....|....| ....|....| ....|....| ....|....| 210 220 230 240 250 MSP1D1 CTTCAGTAAA CTTCGCGAAC AACTGGGCCC CGTGACGCAG GAATTCTGGG MSP1 GCCATATCTC GATGACTTTC AGAAAAAATG GCAGGAAGAG ATGGAATTAT MSP1D2 GGAAGAGATG GAATTATATC GTCAAAAGGT GGAACCGCTG CGTGCGGAAC MSP2N2 GCCATATCTC GATGACTTTC AGAAAAAATG GCAGGAAGAG ATGGAATTAT MSP2N3 GCCATATCTC GATGACTTTC AGAAAAAATG GCAGGAAGAG ATGGAATTAT MSP1E1D1 GGAATTCTGG GACAACCTGG AAAAAGAAAC CGAGGGACTG CGTCAGGAAA MSP1E3D1 GCCATATCTC GATGACTTTC AGAAAAAATG GCAGGAAGAG ATGGAATTAT MSP1E2D1 CTTCAGTAAA CTTCGCGAAC AACTGGGCCC CGTGACGCAG GAATTCTGGG A2AR CACCGGGTTC TGCGCTGCCT GCCACGGCTG CCTCTTCATT GCCTGCTTCG A2AR-GFP CAACTACTTT GTGGTGTCAC TGGCGGCGGC CGACATCGCA GTGGGTGTGC A2AR-GS li CACCGGGTTC TGCGCTGCCT GCCACGGCTG CCTCTTCATT GCCTGCTTCG

....|....| ....|....| ....|....| ....|....| ....|....| 260 270 280 290 300 MSP1D1 ACAACCTGGA AAAAGAAACC GAGGGACTGC GTCAGGAAAT GTCCAAAGAT MSP1 ATCGTCAAAA GGTGGAACCG CTGCGTGCGG AACTGCAAGA GGGGGCACGC MSP1D2 TGCAAGAGGG GGCACGCCAA AAACTCCATG AGCTCCAAGA GAAGCTCAGC MSP2N2 ATCGTCAAAA GGTGGAACCG CTGCGTGCGG AACTGCAAGA GGGGGCACGC MSP2N3 ATCGTCAAAA GGTGGAACCG CTGCGTGCGG AACTGCAAGA GGGGGCACGC MSP1E1D1 TGTCCAAAGA TTTAGAAGAG GTGAAGGCCA AGGTTCAGCC ATATCTCGAT MSP1E3D1 ATCGTCAAAA GGTGGAACCG CTGCGTGCGG AACTGCAAGA GGGGGCACGC MSP1E2D1 ACAACCTGGA AAAAGAAACC GAGGGACTGC GTCAGGAAAT GTCCAAAGAT A2AR TCCTGGTCCT CACGCAGAGC TCCATCTTCA GTCTCCTGGC CATCGCCATT A2AR-GFP TCGCCATCCC CTTTGCCATC ACCATCAGCA CCGGGTTCTG CGCTGCCTGC A2AR-GS li TCCTGGTCCT CACGCAGAGC TCCATCTTCA GTCTCCTGGC CATCGCCATT

....|....| ....|....| ....|....| ....|....| ....|....| 310 320 330 340 350 MSP1D1 TTAGAAGAGG TGAAGGCCAA GGTTCAGCCA TATCTCGATG ACTTTCAGAA MSP1 CAAAAACTCC ATGAGCTCCA AGAGAAGCTC AGCCCATTAG GCGAAGAAAT MSP1D2 CCATTAGGCG AAGAAATGCG CGATCGCGCC CGTGCACATG TTGATGCACT MSP2N2 CAAAAACTCC ATGAGCTCCA AGAGAAGCTC AGCCCATTAG GCGAAGAAAT MSP2N3 CAAAAACTCC ATGAGCTCCA AGAGAAGCTC AGCCCATTAG GCGAAGAAAT MSP1E1D1 GACTTTCAGA AAAAATGGCA GGAAGAGATG GAATTATATC GTCAAAAGGT MSP1E3D1 CAAAAACTCC ATGAGCTCCA AGAGAAGCTC AGCCCATTAG GCGAAGAAAT MSP1E2D1 TTAGAAGAGG TGAAGGCCAA GGTTCAGCCA TATCTCGATG ACTTTCAGAA A2AR GACCGCTACA TTGCCATCCG CATCCCGCTC CGGTACAATG GCTTGGTGAC A2AR-GFP CACGGCTGCC TCTTCATTGC CTGCTTCGTC CTGGTCCTCA CGCAGAGCTC A2AR-GS li GACCGCTACA TTGCCATCCG CATCCCGCTC CGGTACAATG GCTTGGTGAC

....|....| ....|....| ....|....| ....|....| ....|....| 360 370 380 390 400 MSP1D1 AAAATGGCAG GAAGAGATGG AATTATATCG TCAAAAGGTG GAACCGCTGC MSP1 GCGCGATCGC GCCCGTGCAC ATGTTGATGC ACTCCGGACT CATTTGGCGC MSP1D2 CCGGACTCAT TTGGCGCCGT ATTCGGATGA ACTTCGCCAG CGTTTGGCCG MSP2N2 GCGCGATCGC GCCCGTGCAC ATGTTGATGC ACTCCGGACT CATTTGGCGC MSP2N3 GCGCGATCGC GCCCGTGCAC ATGTTGATGC ACTCCGGACT CATTTGGCGC MSP1E1D1 GGAACCATAT CTCGATGACT TTCAGAAAAA ATGGCAGGAA GAGATGGAAT MSP1E3D1 GCGCGATCGC GCCCGTGCAC ATGTTGATGC ACTCCGGACT CATTTGGCGC MSP1E2D1 AAAATGGCAG GAAGAGATGG AATTATATCG TCAAAAGGTG GAACCGCTGC A2AR CGGCACGAGG GCTAAGGGCA TCATTGCCAT CTGCTGGGTG CTGTCGTTTG A2AR-GFP CATCTTCAGT CTCCTGGCCA TCGCCATTGA CCGCTACATT GCCATCCGCA A2AR-GS li CGGCACGAGG GCTAAGGGCA TCATTGCCAT CTGCTGGGTG CTGTCGTTTG

228

....|....| ....|....| ....|....| ....|....| ....|....| 410 420 430 440 450 MSP1D1 GTGCGGAACT GCAAGAGGGG GCACGCCAAA AACTCCATGA GCTCCAAGAG MSP1 CGTATTCGGA TGAACTTCGC CAGCGTTTGG CCGCACGTCT CGAGGCGCTG MSP1D2 CACGTCTCGA GGCGCTGAAA GAAAACGGGG GTGCCCGCTT GGCTGAGTAC MSP2N2 CGTATTCGGA TGAACTTCGC CAGCGTTTGG CCGCACGTCT CGAGGCGCTG MSP2N3 CGTATTCGGA TGAACTTCGC CAGCGTTTGG CCGCACGTCT CGAGGCGCTG MSP1E1D1 TATATCGTCA AAAGGTGGAA CCGCTGCGTG CGGAACTGCA AGAGGGGGCA MSP1E3D1 CATATCTCGA TGACTTTCAG AAAAAATGGC AGGAAGAGAT GGAATTATAT MSP1E2D1 GTGCGGAACT GCAAGAGGGG GCACGCCAAA AACTCCATGA GCTCCAAGAG A2AR CCATCGGCCT GACTCCCATG CTAGGTTGGA ACAACTGCGG TCAGCCAAAG A2AR-GFP TCCCGCTCCG GTACAATGGC TTGGTGACCG GCACGAGGGC TAAGGGCATC A2AR-GS li CCATCGGCCT GACTCCCATG CTAGGTTGGA ACAACTGCGG TCAGCCAAAG

....|....| ....|....| ....|....| ....|....| ....|....| 460 470 480 490 500 MSP1D1 AAGCTCAGCC CATTAGGCGA AGAAATGCGC GATCGCGCCC GTGCACATGT MSP1 AAAGAAAACG GGGGTGCCCG CTTGGCTGAG TACCACGCGA AAGCGACAGA MSP1D2 CACGCGAAAG CGACAGAACA CCTGAGCACC TTGAGCGAAA AAGCGAAACC MSP2N2 AAAGAAAACG GGGGTGCCCG CTTGGCTGAG TACCACGCGA AAGCGACAGA MSP2N3 AAAGAAAACG GGGGTGCCCG CTTGGCTGAG TACCACGCGA AAGCGACAGA MSP1E1D1 CGCCAAAAAC TCCATGAGCT CCAAGAGAAG CTCAGCCCAT TAGGCGAAGA MSP1E3D1 CGTCAAAAGG TGGAACCGCT GCGTGCGGAA CTGCAAGAGG GGGCACGCCA MSP1E2D1 AAGCTCAGCC CATATCTCGA TGACTTTCAG AAAAAATGGC AGGAAGAGAT A2AR GAGGGCAAGA ACCACTCCCA GGGCTGCGGG GAGGGCCAAG TGGCCTGTCT A2AR-GFP ATTGCCATCT GCTGGGTGCT GTCGTTTGCC ATCGGCCTGA CTCCCATGCT A2AR-GS li GAGGGCAAGA ACCACTCCCA GGGCTGCGGG GAGGGCCAAG TGGCCTGTCT

....|....| ....|....| ....|....| ....|....| ....|....| 510 520 530 540 550 MSP1D1 TGATGCACTC CGGACTCATT TGGCGCCGTA TTCGGATGAA CTTCGCCAGC MSP1 ACACCTGAGC ACCTTGAGCG AAAAAGCGAA ACCGGCGCTG GAAGATCTAC MSP1D2 GGCGCTGGAA GATCTACGCC AGGGCTTATT GCCTGTTCTT GAGAGCTTTA MSP2N2 ACACCTGAGC ACCTTGAGCG AAAAAGCGAA ACCGGCGCTG GAAGATCTAC MSP2N3 ACACCTGAGC ACCTTGAGCG AAAAAGCGAA ACCGGCGCTG GAAGATCTAC MSP1E1D1 AATGCGCGAT CGCGCCCGTG CACATGTTGA TGCACTCCGG ACTCATTTGG MSP1E3D1 AAAACTCCAT GAGCTCCAAG AGAAGCTCAG CCCATTAGGC GAAGAAATGC MSP1E2D1 GGAATTATAT CGTCAAAAGG TGGAACCGCT GCGTGCGGAA CTGCAAGAGG A2AR CTTTGAGGAT GTGGTCCCCA TGAACTACAT GGTGTACTTC AACTTCTTTG A2AR-GFP AGGTTGGAAC AACTGCGGTC AGCCAAAGGA GGGCAAGAAC CACTCCCAGG A2AR-GS li CTTTGAGGAT GTGGTCCCCA TGAACTACAT GGTGTACTTC AACTTCTTTG

....|....| ....|....| ....|....| ....|....| ....|....| 560 570 580 590 600 MSP1D1 GTTTGGCCGC ACGTCTCGAG GCGCTGAAAG AAAACGGGGG TGCCCGCTTG MSP1 GCCAGGGCTT ATTGCCTGTT CTTGAGAGCT TTAAAGTCAG TTTTCTGTCA MSP1D2 AAGTCAGTTT TCTGTCAGCT CTGGAAGAAT ATACTAAAAA GCTGAATACC MSP2N2 GCCAGGGCTT ATTGCCTGTT CTTGAGAGCT TTAAAGTCAG TTTTCTGTCA MSP2N3 GCCAGGGCTT ATTGCCTGTT CTTGAGAGCT TTAAAGTCAG TTTTCTGTCA MSP1E1D1 CGCCGTATTC GGATGAACTT CGCCAGCGTT TGGCCGCACG TCTCGAGGCG MSP1E3D1 GCGATCGCGC CCGTGCACAT GTTGATGCAC TCCGGACTCA TTTGGCGCCG MSP1E2D1 GGGCACGCCA AAAACTCCAT GAGCTCCAAG AGAAGCTCAG CCCATTAGGC A2AR CCTGTGTGCT GGTGCCCCTG CTGCTCATGC TGGGTGTCTA TTTGCGGATC A2AR-GFP GCTGCGGGGA GGGCCAAGTG GCCTGTCTCT TTGAGGATGT GGTCCCCATG A2AR-GS li CCTGTGTGCT GGTGCCCCTG CTGCTCATGC TGGGTGTCTA TTTGCGGATC

229

....|....| ....|....| ....|....| ....|....| ....|....| 610 620 630 640 650 MSP1D1 GCTGAGTACC ACGCGAAAGC GACAGAACAC CTGAGCACCT TGAGCGAAAA MSP1 GCTCTGGAAG AATATACTAA AAAGCTGAAT ACCCAG...... MSP1D2 CAG...... MSP2N2 GCTCTGGAAG AATATACTAA AAAGCTGAAT ACCCAGGGTA CCCCCGTGAC MSP2N3 GCTCTGGAAG AATATACTAA AAAGCTGAAT ACCCAGGGTA CCCGCGAACA MSP1E1D1 CTGAAAGAAA ACGGGGGTGC CCGCTTGGCT GAGTACCACG CGAAAGCGAC MSP1E3D1 TATTCGGATG AACTTCGCCA GCGTTTGGCC GCACGTCTCG AGGCGCTGAA MSP1E2D1 GAAGAAATGC GCGATCGCGC CCGTGCACAT GTTGATGCAC TCCGGACTCA A2AR TTCCTGGCGG CGCGACGACA GCTGGCTGAT CTGGAAGACA ATTGGGAAAC A2AR-GFP AACTACATGG TGTACTTCAA CTTCTTTGCC TGTGTGCTGG TGCCCCTGCT A2AR-GS li TTCCTGGCGG CGCGACGACA GCTGGCTGAT CTGGAAGACA ATTGGGAAAC

....|....| ....|....| ....|....| ....|....| ....|....| 660 670 680 690 700 MSP1D1 AGCGAAACCG GCGCTGGAAG ATCTACGCCA GGGCTTATTG CCTGTTCTTG MSP1 ...... MSP1D2 ...... MSP2N2 GCAGGAATTC TGGGACAACC TGGAAAAAGA AACCGAGGGA CTGCGTCAGG MSP2N3 ACTGGGCCCC GTGACGCAGG AATTCTGGGA CAACCTGGAA AAAGAAACCG MSP1E1D1 AGAACACCTG AGCACCTTGA GCGAAAAAGC GAAACCGGCG CTGGAAGATC MSP1E3D1 AGAAAACGGG GGTGCCCGCT TGGCTGAGTA CCACGCGAAA GCGACAGAAC MSP1E2D1 TTTGGCGCCG TATTCGGATG AACTTCGCCA GCGTTTGGCC GCACGTCTCG A2AR TCTGAACGAC AATCTCAAGG TGATCGAGAA GGCTGACAAT GCTGCACAAG A2AR-GFP GCTCATGCTG GGTGTCTATT TGCGGATCTT CCTGGCGGCG CGACGACAGC A2AR-GS li TCTGAACGAC AATCTCAAGG TGATCGAGAA GGCTGACAAT GCTGCACAAG

....|....| ....|....| ....|....| ....|....| ....|....| 710 720 730 740 750 MSP1D1 AGAGCTTTAA AGTCAGTTTT CTGTCAGCTC TGGAAGAATA TACTAAAAAG MSP1 ...... MSP1D2 ...... MSP2N2 AAATGTCCAA AGATTTAGAA GAGGTGAAGG CCAAGGTTCA GCCATATCTC MSP2N3 AGGGACTGCG TCAGGAAATG TCCAAAGATT TAGAAGAGGT GAAGGCCAAG MSP1E1D1 TACGCCAGGG CTTATTGCCT GTTCTTGAGA GCTTTAAAGT CAGTTTTCTG MSP1E3D1 ACCTGAGCAC CTTGAGCGAA AAAGCGAAAC CGGCGCTGGA AGATCTACGC MSP1E2D1 AGGCGCTGAA AGAAAACGGG GGTGCCCGCT TGGCTGAGTA CCACGCGAAA A2AR TCAAAGACGC TCTGACCAAG ATGAGGGCAG CAGCCCTGGA CGCTCAGAAG A2AR-GFP TGGCTGATCT GGAAGACAAT TGGGAAACTC TGAACGACAA TCTCAAGGTG A2AR-GS li TCAAAGACGC TCTGACCAAG ATGAGGGCAG CAGCCCTGGA CGCTCAGAAG

....|....| ....|....| ....|....| ....|....| ....|....| 760 770 780 790 800 MSP1D1 CTGAATACCC AGTAAGCTTG CGGCCGCACT CGAGCACCAC CACCACCACC MSP1 ...... MSP1D2 ...... MSP2N2 GATGACTTTC AGAAAAAATG GCAGGAAGAG ATGGAATTAT ATCGTCAAAA MSP2N3 GTTCAGCCAT ATCTCGATGA CTTTCAGAAA AAATGGCAGG AAGAGATGGA MSP1E1D1 TCAGCTCTGG AAGAATATAC TAAAAAGCTG AATACCCAGT AAGCTTGCGG MSP1E3D1 CAGGGCTTAT TGCCTGTTCT TGAGAGCTTT AAAGTCAGTT TTCTGTCAGC MSP1E2D1 GCGACAGAAC ACCTGAGCAC CTTGAGCGAA AAAGCGAAAC CGGCGCTGGA A2AR GCCACTCCAC CTAAGCTCGA GGACAAGAGC CCAGATAGCC CTGAAATGAA A2AR-GFP ATCGAGAAGG CTGACAATGC TGCACAAGTC AAAGACGCTC TGACCAAGAT A2AR-GS li GCCACTCCAC CTAAGCTCGA GGACAAGAGC CCAGATAGCC CTGAAATGAA

230

....|....| ....|....| ....|....| ....|....| ....|....| 810 820 830 840 850 MSP1D1 ACTGAGATCC GGCTGCTAAC AAAGCCCGAA AGGAAGCTGA GTTGGCTGCT MSP1 ...... MSP1D2 ...... MSP2N2 GGTGGAACCG CTGCGTGCGG AACTGCAAGA GGGGGCACGC CAAAAACTCC MSP2N3 ATTATATCGT CAAAAGGTGG AACCGCTGCG TGCGGAACTG CAAGAGGGGG MSP1E1D1 CCGCACTCGA GCACCACCAC CACCACCACT GAGATCCGGC TGCTAACAAA MSP1E3D1 TCTGGAAGAA TATACTAAAA AGCTGAATAC CCAG...... MSP1E2D1 AGATCTACGC CAGGGCTTAT TGCCTGTTCT TGAGAGCTTT AAAGTCAGTT A2AR AGACTTTCGG CATGGATTCG ACATTCTGGT GGGACAGATT GATGATGCAC A2AR-GFP GAGGGCAGCA GCCCTGGACG CTCAGAAGGC CACTCCACCT AAGCTCGAGG A2AR-GS li AGACTTTCGG CATGGATTCG ACATTCTGGT GGGACAGATT GATGATGCAC

....|....| ....|....| ....|....| ....|....| ....|....| 860 870 880 890 900 MSP1D1 GCCACCGCTG A...... MSP1 ...... MSP1D2 ...... MSP2N2 ATGAGCTCCA AGAGAAGCTC AGCCCATTAG GCGAAGAAAT GCGCGATCGC MSP2N3 CACGCCAAAA ACTCCATGAG CTCCAAGAGA AGCTCAGCCC ATTAGGCGAA MSP1E1D1 GCCCGAAAGG AAGCTGAGTT GGCTGC...... MSP1E3D1 ...... MSP1E2D1 TTCTGTCAGC TCTGGAAGAA TATACTAAAA AGCTGAATAC CCAGTAAGCT A2AR TCAAGCTGGC CAATGAAGGG AAAGTCAAGG AAGCACAAGC AGCCGCTGAG A2AR-GFP ACAAGAGCCC AGATAGCCCT GAAATGAAAG ACTTTCGGCA TGGATTCGAC A2AR-GS li TCAAGCTGGC CAATGAAGGG AAAGTCAAGG AAGCACAAGC AGCCGCTGAG

....|....| ....|....| ....|....| ....|....| ....|....| 910 920 930 940 950 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 GCCCGTGCAC ATGTTGATGC ACTCCGGACT CATTTGGCGC CGTATTCGGA MSP2N3 GAAATGCGCG ATCGCGCCCG TGCACATGTT GATGCACTCC GGACTCATTT MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 TGCGGCCGCA CTCGAGCACC ACCACCACCA CCACTGAGAT CCGGCTGCTA A2AR CAGCTGAAGA CCACCCGGAA TGCATACATT CAGAAGTACC TGTCCACACT A2AR-GFP ATTCTGGTGG GACAGATTGA TGATGCACTC AAGCTGGCCA ATGAAGGGAA A2AR-GS li CAGCTGAAGA CCACCCGGAA TGCATACATT CAGAAGTACC TGTCCACACT

....|....| ....|....| ....|....| ....|....| ....|....| 960 970 980 990 1000 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 TGAACTTCGC CAGCGTTTGG CCGCACGTCT CGAGGCGCTG AAAGAAAACG MSP2N3 GGCGCCGTAT TCGGATGAAC TTCGCCAGCG TTTGGCCGCA CGTCTCGAGG MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ACAAAGCCCG AAAGGAAGCT GAGTTGGCTG CTGCCACCGC TGA...... A2AR GCAGAAGGAG GTCCATGCTG CCAAGTCACT GGCCATCATT GTGGGGCTCT A2AR-GFP AGTCAAGGAA GCACAAGCAG CCGCTGAGCA GCTGAAGACC ACCCGGAATG A2AR-GS li GCAGAAGGAG GTCCATGCTG CCAAGTCACT GGCCATCATT GTGGGGCTCT

231

....|....| ....|....| ....|....| ....|....| ....|....| 1010 1020 1030 1040 1050 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 GGGGTGCCCG CTTGGCTGAG TACCACGCGA AAGCGACAGA ACACCTGAGC MSP2N3 CGCTGAAAGA AAACGGGGGT GCCCGCTTGG CTGAGTACCA CGCGAAAGCG MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR TTGCCCTCTG CTGGCTGCCC CTACACATCA TCAACTGCTT CACTTTCTTC A2AR-GFP CATACATTCA GAAGTACCTG GAGCGCGCTC GGTCCACACT GCAGAAGGAG A2AR-GS li TTGCCCTCTG CTGGCTGCCC CTACACATCA TCAACTGCTT CACTTTCTTC

....|....| ....|....| ....|....| ....|....| ....|....| 1060 1070 1080 1090 1100 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ACCTTGAGCG AAAAAGCGAA ACCGGCGCTG GAAGATCTAC GCCAGGGCTT MSP2N3 ACAGAACACC TGAGCACCTT GAGCGAAAAA GCGAAACCGG CGCTGGAAGA MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR TGCCCCGACT GCAGCCACGC CCCTCTCTGG CTCATGTACC TGGCCATCGT A2AR-GFP GTCCATGCTG CCAAGTCACT GGCCATCATT GTGGGGCTCT TTGCCCTCTG A2AR-GS li TGCCCCGACT GCAGCCACGC CCCTCTCTGG CTCATGTACC TGGCCATCGT

....|....| ....|....| ....|....| ....|....| ....|....| 1110 1120 1130 1140 1150 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ATTGCCTGTT CTTGAGAGCT TTAAAGTCAG TTTTCTGTCA GCTCTGGAAG MSP2N3 TCTACGCCAG GGCTTATTGC CTGTTCTTGA GAGCTTTAAA GTCAGTTTTC MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR CCTCTCCCAC ACCAATTCGG TTGTGAATCC CTTCATCTAC GCCTACCGTA A2AR-GFP CTGGCTGCCC CTACACATCA TCAACTGCTT CACTTTCTTC TGCCCCGACT A2AR-GS li CCTCTCCCAC ACCAATTCGG TTGTGAATCC CTTCATCTAC GCCTACCGTA

....|....| ....|....| ....|....| ....|....| ....|....| 1160 1170 1180 1190 1200 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 AATATACTAA AAAGCTGAAT ACCCAGTAA...... MSP2N3 TGTCAGCTCT GGAAGAATAT ACTAAAAAGC TGAATACCCA G...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR TCCGCGAGTT CCGCCAGACC TTCCGCAAGA TCATTCGCAG CCACGTCCTG A2AR-GFP GCAGCCACGC CCCTCTCTGG CTCATGTACC TGGCCATCGT CCTCTCCCAC A2AR-GS li TCCGCGAGTT CCGCCAGACC TTCCGCAAGA TCATTCGCAG CCACGTCCTG

232

....|....| ....|....| ....|....| ....|....| ....|....| 1210 1220 1230 1240 1250 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR AGGCAGCAAG AACCTCACCA TCACCATCAC CATCACCATC ACTGA..... A2AR-GFP ACCAATTCGG TTGTGAATCC CTTCATCTAC GCCTACCGTA TCCGCGAGTT A2AR-GS li AGGCAGCAAG AACCTGGAGG CTCAGGTGGC GGATCGGGGC ACCATCACCA

....|....| ....|....| ....|....| ....|....| ....|....| 1260 1270 1280 1290 1300 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CCGCCAGACC TTCCGCAAGA TCATTCGCAG CCACGTCCTG AGGCAGCAAG A2AR-GS li TCACCATCAC CATCACTGA......

....|....| ....|....| ....|....| ....|....| ....|....| 1310 1320 1330 1340 1350 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP AACCTTTCAA GGCAGAATTC GAGAACCTGT ACTTCCAGGG TGGCTCCGGT A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1360 1370 1380 1390 1400 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP AGCGGCTCTG GTTCCATGTC AAAGGGCGAG GAATTGTTCA CTGGCGTGGT A2AR-GS li ......

233

....|....| ....|....| ....|....| ....|....| ....|....| 1410 1420 1430 1440 1450 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP GCCTATCTTG GTCGAGTTGG ACGGCGATGT CAACGGTCAC AAGTTCAGCG A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1460 1470 1480 1490 1500 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP TGTCGGGTGA GGGCGAAGGA GACGCTACTT ACGGCAAGCT GACATTGAAG A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1510 1520 1530 1540 1550 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP TTCATCTGCA CCACTGGCAA GCTGCCCGTT CCTTGGCCAA CACTGGTGAC A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1560 1570 1580 1590 1600 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP AACCCTCACC TACGGAGTCC AGTGTTTCTC AAGGTACCCT GACCACATGA A2AR-GS li ......

234

....|....| ....|....| ....|....| ....|....| ....|....| 1610 1620 1630 1640 1650 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP AGCAACACGA TTTCTTCAAG TCGGCTATGC CAGAGGGCTA CGTGCAGGAA A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1660 1670 1680 1690 1700 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CGTACCATCT TCTTCAAGGA CGATGGAAAC TACAAGACTC GCGCCGAGGT A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1710 1720 1730 1740 1750 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CAAGTTCGAA GGAGACACAC TCGTTAACAG GATCGAGTTG AAGGGTATCG A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1760 1770 1780 1790 1800 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP ACTTCAAGGA AGATGGAAAC ATCTTGGGTC ACAAGCTGGA GTACAACTAC A2AR-GS li ......

235

....|....| ....|....| ....|....| ....|....| ....|....| 1810 1820 1830 1840 1850 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP AACTCTCACA ACGTGTACAT CATGGCTGAC AAGCAGAAGA ACGGTATCAA A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1860 1870 1880 1890 1900 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP GGTCAACTTC AAGATCAGAC ACAACATCGA AGACGGCTCC GTTCAATTGG A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1910 1920 1930 1940 1950 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CCGATCACTA CCAGCAAAAC ACCCCAATCG GTGACGGACC AGTCCTGCTC A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 1960 1970 1980 1990 2000 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CCCGATAACC ACTACCTGTC CACTCAATCT AAGCTCAGCA AGGACCCAAA A2AR-GS li ......

236

....|....| ....|....| ....|....| ....|....| ....|....| 2010 2020 2030 2040 2050 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CGAGAAGCGT GATCACATGG TGTTGCTGGA GTTCGTCACC GCTGCCGGCA A2AR-GS li ......

....|....| ....|....| ....|....| ....|....| ....|....| 2060 2070 2080 2090 2100 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP TCACTCACGG AATGGACGAA CTCTACAAGC ACCACCATCA CCATCACCAT A2AR-GS li ......

....|....| .. 2110 MSP1D1 ...... MSP1 ...... MSP1D2 ...... MSP2N2 ...... MSP2N3 ...... MSP1E1D1 ...... MSP1E3D1 ...... MSP1E2D1 ...... A2AR ...... A2AR-GFP CACCATCACT GA A2AR-GS li ......

237

SI 4: Some reported LRRK2 kinase inhibitors. Compounds were grouped according by

their common structure 57,631-651 77,104,631,633,645-647,651-675.

Compounds Structure

Nonspecific Gö6976 (IC50 250nM), Synthesized

Nonspecific K-252A (IC50 25nM ), metabolite of Nocardiopsis sp. bacterium

Nonspecific CEP-1347 (IC50 23-51nM), derivative of natural product K252A

Nonspecific K-252B (IC50 50nM), derivative of K-252A and also metabolite of Nocardiopsis sp. bacterium

Nonspecific staurosporine/St (IC50 1-10nM), extract from Streptomyces staurosporeus bacteria

JAK3 Inhibitor VI (IC50 22nM)

238

PKR Inhibitor (IC50 13nM)

Raf1 kinase Inhibitor I (IC50 500nM)

Nonspecific Su-11248/Sunitinib (IC50 79nM), synthesized

Indirubin-3’-monoxime (I3MO) (IC50 1.75μM), derivative of natural product indirubin

Nonspecific GW5074 (3-(3, 5-dibromo-4-hydroxybenzylidine-

5-iodo-1,3-dihydro-indol-2-one)) (a Raf-1 inhibitor) (IC50 0.88μM)

Ro-31-8220 (IC50 1.64μM), derivatives from staurosporine and K252A

GF109203X or Bisindolylmaleimide I (BIM I), highly selective, cell-permeable, and reversible protein kinase C

(PKC) inhibitor (IC50 5.27μM)

5-Iodotubercidin (IC50 3.21μM)

(E)-4-methyl-2-(2-(pyridin-4-

ylmethylene)hydrazinyl)quinoline (IC50 4.1μM)

239

Specific LRRK2-IN-1 (ATP-competitive) (IC50: wt, 13nM; mutant, 6nM) (benzo[e]pyrimido-[5,4-b][1,4]diazepin-6(11H)- one)

Nonspecific H-1152 (IC50 244nM)

Nonspecific Y-27632 (IC50 2.4μM)

Raf inhibitor IV (IC50 5.43μM)

Sorafenib (IC50 14.56μM)

LDN-73794 (IC50 3.5μM)

Damnacanthal (IC50 19.14μM), extracted from Morinda citrifolia

SP600125 (IC50 3.71μM)

LDN-22684 (IC50 5.4μM)

240

HG-10-102-01 (2-anilino-4-methylamino-5-chloropyrimidine)

GNE-7915

Potent and selective CZC-25146

CZC-54252

NVP-TAE684

ponatinib

imatinib

Bosutinib

S7 (la-S7)

241

la-sunitinib

GSK3-XIII

GNE-0877 (0877)

GSK 2578215A

LRRK2-IN-2 (HG-9-106-03)

SR94444

Compound 18

Compound 2 (4-aminoquinoline-3-carboxamide derivative)

Compound 11 (4-aminoquinoline-3-carboxamide derivative)

242

SI 5. ESI-FT-ICR MS parameters optimisation against 10 µM bCA standard protein (m/z

1200 – 3000, referred to Figure 5.1). Figure A was a standard ESI-FT-ICR MS method acquired for protein molecules. Figure B to Figure E were the optimization of different instrumental parameters.

243

244

245

246

247