Glutamate-Gated Chloride Channel Receptors and Mechanisms of Drug Resistance in Pathogenic Species

Glutamate-gated chloride channel receptors and mechanisms of drug resistance in pathogenic species

Mohammed Atif B. Pharmacy, M. Pharmacy (Pharmacology)

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2019 Queensland Brain Institute

Dedicated to my beloved parents & my demised brother who I miss everyday

Thesis Abstract

Pentameric ligand-gated ion channels (pLGICs) are important therapeutic targets for a wide range of neurological disorders that include cognitive impairment, stroke, psychiatric conditions and peripheral pain. They are also targets for treating parasite infections and controlling pest species in agriculture, veterinary practice and human health. Here we focus on one family of the pLGICs i.e., the glutamate-gated chloride channel receptors (GluClRs) which are expressed at inhibitory synapses of invertebrates. Ivermectin (IVM) is one of the main drugs used to control pest species and parasites, and it works by activating GluClRs in nematode and arthropod muscle and nerves. IVM resistance is becoming a major problem in many invertebrate pathogens, necessitating the development of novel anti-parasitic drugs. This project started with the simple aim of determining the sensitivity to glutamate and IVM of GluClRs from two different pest species: the parasitic nematode Haemonchus contortus (HcoGluClRs) and the mosquito malaria vector Anopheles gambiae (AgGluClRs).

In chapter 3, we found that the β homomeric GluClRs of H.contortus were insensitive

to IVM (EC50> 10 µM), whereas α homomeric HcoGluClRs were highly sensitive (EC50 = 20 nM).

Heteromeric αβ HcoGluClRs exhibited an intermediate sensitivity to IVM (EC50 = 135 nM). By contrast, the EC50 values for glutamate at α homomeric and αβ heteromeric receptors were not distinguishable; falling between 20-30 µM. GluClR-mediated IPSCs were significantly different for α relative to αβ receptors, with αβ GluClRs exhibiting a faster decay rate of 17 ms in contrast to α GluClRs which had a decay rate of 40ms. 5 nM IVM slowed the decay rates of IPSCs mediated by both isoforms with α GluClRs reaching a mean decay rate of 100 ms and αβ GluClRs decaying at a mean rate of 72ms. Evaluating the biophysical properties of single heteromeric αβ and homomeric β GluClRs revealed that different stoichiometric GluClR isoforms exhibited different open state amplitudes. The α GluClR had the biggest current (1.8 pA) followed by the αβ (1:1) GluClR (two different amplitudes – 1.2 pA, 90% and 0.7 pA, 10%) and the smallest being β homomer (0.4 pA). The β and αβ GluClRs had mean active durations of 5.5 ms and 43 ms, respectively, in response to 2 µM glutamate. This duration increased to 126 ms and 876 ms, respectively, when 2 µM glutamate was co-applied with 5 nM IVM. Collectively, the data from both the IPSCs and the single channel measurements show a correlation between shorter active periods; faster IPSC decay rates and decreased IVM sensitivity. iii

For better drug optimisation, a quantitative understanding of activation and modulation mechanisms of the GluClR is necessary. For this purpose, in chapter 4 we evaluated the biophysical properties of the α homomeric HcoGluClR and a mutation to a position that confers IVM resistance in several pest species (TM3-36’) in presence of glutamate and IVM. We showed that the G36’A mutation decreased the duration of active periods and increased receptor desensitisation. As the results showed a common effect for both glutamate- and IVM-gated currents, we infer that the intrinsic properties of the receptor were affected by the G36’A mutation and not the IVM binding mechanism as previously proposed. In Chapter 5, we looked to characterize the effects of glutamate, IVM, lindane, fipronil and picrotoxin on 2 splice isoforms of GluClRs expressed by A. gambiae. The two splice isoforms AgGluCl-a1 and AgGluCl-b when evaluated as homomeric GluClRs for effects of glutamate and IVM disclosed something interesting. We found that the AgGluCl-b was sensitive to IVM, contrary to what was reported by (Meyers et al. 2015). Lindane and fipronil both block glycine receptors (Islam and Lynch 2012), which, like the GluClR, is an inhibitory member of pLGIC family. In addition, many insecticides target GluClRs, this prompted us to investigate how lindane and fipronil affected the two GluClR isoforms. Both compounds were effective at inhibiting glutamate-gated currents mediated by both isoforms. Glutamate-gated currents completely recovered in a time dependent manner for AgGluClR-a1. By contrast, only a partial recovery of currents was observed for AgGluClR-b, suggesting a greater affinity of lindane and fipronil at AgGluClR-b. This differential affinity implies different binding mechanisms between the two isoforms. Picrotoxin, a classical pore blocker, equally blocked glutamate currents for both the isoforms. In chapter 6, we screened libraries of natural marine extracts against the hippocampal-dominant α5 GABAAR isoform (α5β3γ2L) using an automated yellow fluorescent protein-based screening assay to identify fractions with selective activity. Hits were confirmed using electrophysiology. Although the hits were potent, further screening of the compounds was not proceeded due to low compound availability. In conclusion, we demonstrated that reduced IVM sensitivity is due to shorter active periods and faster decay rates. This conclusion was supported by the findings from chapter 3 that showed that heteromeric population of αβ HcoGluClRs were less sensitive to the potentiating effects of IVM compared to α HcoGluClRs. The briefer single receptor periods

iv seen in β-containing GluClRs correspond closely with, faster IPSC decay rates. All the results point towards β HcoGluClRs being responsible for reduced IVM sensitivity.

Declaration by author

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

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

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

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis and have sought permission from co- authors for any jointly authored works included in the thesis.

Mohammed Atif

Publications during candidature

Peer-reviewed journal publications

Atif M, Estrada-Mondragon A, Nguyen B, Lynch JW, Keramidas A (2017) Effects of glutamate and ivermectin on single glutamate-gated chloride channels of the parasitic nematode H. contortus. (Chapter 4) PLoSPathog 13(10): e1006663. https://doi.org/10.1371/journal.ppat.1006663.

Submitted manuscripts included in this thesis

Atif M, Smith JJ, Estrada-Mondragon A, Xiao X, Capon RJ, Lynch JW and Keramidas A (2018) GluClR-mediated inhibitory postsynaptic currents reveal targets for ivermectin and potential mechanisms of IVM resistance (Chapter 3) (In press – PLoS pathogens).

Presentations

1. ASB 2017; Annual Meeting of the Australian society for Biophysics, Sydney, Australia.

Atif M, Lynch JW and Keramidas A. Modulating ivermectin sensitivity at glutamate- gated chloride channels of H.contortus and A.gambiae (Poster)

2. SfN 2018; Society for Neuroscience, San Diego, USA.

Atif M, Smith JJ, Lynch JW and Keramidas A. Modulating ivermectin sensitivity at glutamate-gated chloride channels of H. contortus and study of inhibitory postsynaptic currents mediated by α and αβ HcoGluCl receptors (Poster).

3. ANS 2018; 38th Annual scientific Meeting of the Australasian Neuroscience Society, Brisbane, Australia

vii

Publications included in this thesis Atif M, Estrada-Mondragon A, Nguyen B, Lynch JW, Keramidas A (2017) Effects of glutamate and ivermectin on single glutamate-gated chloride channels of the parasitic nematode H. contortus. (Incorporated as Chapter 4) PLoSPathog 13(10): e1006663. https://doi.org/10.1371/journal.ppat.1006663.

Contributor Statement of contribution

Author Mohammed Atif (Candidate) Electrophysiological recordings and data analysis (40%).

Figure preparations (20%).

Subsequent draft editing (20%)

Author Argel Estrada-Mondragon Molecular study and Figure 11 (100%).

Subsequent draft editing (20%)

Author Bindi Nguyen Electrophysiological recordings (10%)

Subsequent draft editing (20%)

Author Joseph W. Lynch Writing and editing first draft (50%)

Subsequent draft editing (20%)

Author Angelo Keramidas Electrophysiological recordings and data analysis (10%).

Figure preparation (80%)

Writing and editing first draft (50%)

Subsequent draft editing (20%)

viii

Submitted manuscripts included in this thesis Atif M, Smith JJ, Estrada-Mondragon A, Xiao X, Salim AA, Capon RJ, Lynch JW and Keramidas A (2018) GluClR-mediated inhibitory postsynaptic currents reveal targets for ivermectin and potential mechanisms of IVM resistance (Incorporated as Chapter 3) (In press – PLOS Pathogens).

Contributor Statement of contribution

Author Mohammed Atif (Candidate) Electrophysiological recordings and data analysis (80%).

Figure preparation (50%).

Microscopy image experiment & analysis (50%).

Subsequent draft editing (11%)

Author Jennifer Smith Electrophysiological recordings (10%).

Subsequent draft editing (11%)

Author Argel Estrada-Mondragon Microscopy experiments (50%).

Subsequent draft editing (11%)

Author Xiao X Synthesis of IVM analogues (100%)

Subsequent draft editing (11%)

Author Angela Salim Pooling data on synthesis of IVM analogues (100%)

Subsequent draft editing (11%)

Author Robert J Capon Lab head that synthesised all the IVM analogues.

Subsequent draft editing (11%)

Author Joseph W. Lynch Writing and editing first draft (50%)

Subsequent draft editing (11%)

Author Angelo Keramidas Electrophysiological recordings and data analysis (10%).

Microscopy image analysis (50%)

Figure preparation (50%)

Writing and editing first draft (50%)

Subsequent draft editing (11%)

Other publications during candidature

No other publications

Contributions by other to the thesis

Our collaborators, Professor Robert J Capon, Dr. Pritesh Prasad, Shamsunnahar Khushi and Dr. Krishantha Wardamune Gedara (Institute of Molecular Biosciences, The University of Queensland), kindly provided the libraries of natural marine extracts listed in Chapter 6. Dr. Pritesh Prasad initially isolated the pure compounds, which was then taken over by Shamsunnahar Khushi. Dr Krishantha Wardamune Gedara prepared and isolated quinazoline alkaloids. Dr. Mogens Nielsen from Gabather, Sweden, kindly provided us with the GT-002 drug. Other contributors were listed in the above ‘Publications included in the thesis & submitted manuscripts in this thesis’ section.

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

None.

Research Involving Human or Animal subjects

All the X. laevis handling and surgical procedures were conducted under conditions approved by the University of Queensland Animal Ethics Committee (approval number: QBI/AIBN/087/16/NHMRC/ARC), certificate attached in the appendices.

All the animal handling and surgical procedures for rats/rat embryos were conducted under conditions approved by the University of Queensland Animal Ethics Committee (approval number: QBI/142/16/NHMRC/ARC), certificate attached in the appendices.

Acknowledgements

My PhD journey started with an email to Professor Joseph (Joe) Lynch. Since that day until today Joe has been the best supervisor/mentor/guide/friend that I could have hoped for my PhD. His easy-going attitude and friendly nature always prompted me have a chat with him, be it about my PhD or things like cricket and PlayStation. He has always been a supportive figure in the lab and helped me get through the initial struggling phase of drug discovery. You have been and will always be a role model that I look up to.

I also owe a great deal to my co-supervisor Dr. Angelo Keramidas, who has taught me how to perform one of the most laborious techniques in the lab and helped me to get out of tight spots. I would miss the weird conversations that we used to have while doing experiments.

I am extremely grateful to Justine Haddrill, our senior lab manager. Justine has just not been a lab manager but a mother-like figure in the lab, organising all the weekend trips, helping us with difficult molecular techniques and keeping everyone sane in the lab with her ever-calm nature. Wherever life takes me, I will miss you in the lab dearly.

I would specially like to thank some of the previous lab members, firstly Dr. Ming Soh, who was extremely generous in helping me during my initial stages of PhD and trying to teach me all the basics of drug discovery, TEVC and VCF. Dr. Robiul Islam, for constantly guiding me and teaching me things about patching. Dr. Argel Estrada-Mondragon, you my friend have been the best support a PhD could ever find from a postdoc, you’re calm, supportive and realistic attitude always made me see the brighter picture.

I would also like to thank all the members of the Lynch lab, both past (Ming, Robi, Argel, Yang, Sahil, Sharifun, Kooi, Bindi, and Thao) and present (Nela, Parnya, Chen, Edmund, Maleeha, Dejan) not only for the help and support, but also for contributing to a friendly lab environment. It saddens me that I will no longer be able to enjoy the weekends away, lunch get-together or karaoke sessions with all of you.

Special thanks to Professor Robert Capon, Dr. Pritesh Prasad, Shamsunnahar Khushi and Dr. Krishantha Wardamune Gedara for providing me with the compound libraries for screening. I would also like to extend my gratitude to QBI for infrastructure support, and to UQI, Lynch group and QBI travel award for the financial support. xi

Dearest Mom and Dad, I cannot thank you enough for the support, the unconditional love and belief that you had me in me, I dedicate this thesis to you hoping that it will make you proud. I know that I cannot in any way make up to the things that you have endured in bringing me up but I hope that this thesis would be a small token of my appreciation to you. To my dearest brother Mohammed Asif, writing this sentence was probably the hardest part in my thesis, I love you, and I miss you, and I know that you would have been the proudest one. To my sister Talat Afreen, thank you for being a constant support and helping me throughout my PhD journey and the two bundle of joys that you brought along Abu Sufyaan and Adiba, naughty as they might be, they are best thing that happened to our family. Finally yet importantly, my deepest appreciation to my wife Dr. Asmath Jabeen, I know that it has been tough being away for the past year but know that without your help, support and love I would not be able to achieve this mammoth task today. Thank you for all the priceless memories and hope to spend the rest of the days with you.

xii

Financial support

This research was supported by The University of Queensland International scholarship and University of Queensland living allowance scholarship. Travel support to conferences was provided by Queensland Brain Institute travel award and Lynch lab.

Keywords

Glutamate, IVM, HcoGluClR, AgGluClR, lindane, fipronil, PTX and α5 GABAARs

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 060601 Animal Physiology – Biophysics, 50%

ANZSRC code: 060110, Receptor and Membrane Biology, 30%

ANZSRC code: 110903, Central Nervous System, 20%

Fields of Research (FoR) Classification

FoR code: 0606, Physiology, 50%

FoR code: 0601, Biochemistry and Cell Biology, 30%

FoR code: 1109, Neuroscience, 20%

xiii

Table of Contents

Title page ...... i

Thesis Abstract ...... iii

Declaration by author...... vi

Publication during candidature ...... vii

Publications included in this thesis ...... viii

Submitted manuscripts included in this thesis ...... ix

Contributions by others to the thesis ...... x

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

Research involving animal or human subjects ...... x

Acknowledgements ...... xi

Financial support ...... xiii

Keywords ...... xiii

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

Fields of Research (FoR) Classification… ...... xiii

Table of contents ...... xiv

List of Figures and Tables ...... xviii

Abbreviations ...... xxv 1. Chapter 1: Introduction 1.1 Pentameric ligand-gated Ion channels (pLGICs)………………………………………………….2 1.2 General structure and features of pLGICs…………………………………………………………..3 1.2.1 Ligand binding domain……………………………………………………………………………3 1.2.2 Transmembrane domain………………………………………………………………………..5 1.3 Glutamate-gated chloride channel receptors (GluClRs) 1.3.1 Physiological role of GluClR genes………………………………………………………….7 1.3.2 Haemonchus contortus………………………………………………………………………….9 a) Introduction…………………………………………………………………………………….9 xiv

b) Life cycle of H.contortus………………………………………………………………….10 c) Pathogenicity, prevention and treatment……………………………………….10 d) Different type of GluClR isoforms from H. contortus (HcoGluClRs) and their distribution pattern………………………………………………………………..11 1.3.3 Anopheles gambiae a) Introduction……………………………………………………………………………………13 b) Life cycle of A. gambiae………………………………………………………………….14 c) Malaria – its medical significance and management……………………....14 d) Different GluClR splice isoforms from A. gambiae (AgGluClRs)……….14 e) Expression profile of GluClRs in mature A.gambiae………………………..15 1.3.4 Agonists of GluClRs 1.3.4.1 Glutamate and ibotenate……………………………………………………………15 1.3.4.2 Ivermectin (IVM)………………………………………………………………………..16 a) Discovery, synthesis and structure………………………………………………16 b) Use of IVM in animal and human health……………………………………..17 c) GluClRs as targets for IVM…………………………………………………………..17 d) Mode of action of IVM and its binding sites on GluClRs………………18 e) Resistance to anthelmintics………………………………………………………..20 f) New targets and novel applications…………………………………………….20 1.3.5 Antagonists of GluClRs……………………………………………………………………………21

1.4 γ−amino butyric acid type A receptors (GABAARs)…………………………………………22

1.4.1 Physiological role of GABAARs……………………………………………………………22

1.4.2 GABAAR subunit composition and binding sites.

a) Different subunit composition of GABAARs……………………………….23

b) The GABAAR binding sites: GABA and Benzodiazepines (BZD)……24

1.4.3 Overview of GABAAR pharmacology…………………………………………………..24

1.4.4 Therapeutic significance of modulators specifically targeting α5

containing GABAARs……………………………………………………………………………26 1.5 Aims of PhD thesis………………………………………………………………………………………… 29 1.6 References……………………………………………………………………………………………………..31

2 Chapter 2: General Methods 2.1 Tissue culture and transfection of HEK 293 cells ……………………………………………………41 2.2 Patch clamp electrophysiology – whole cell, single channels and rapid application.41 2.3 Isolation of oocytes from Xenopus frogs and expression of DNA……………………………42 2.4 Two-electrode voltage clamp (TEVC) electrophysiology ………………………………………..43 2.5 Heteroculture preparation and IPSCs recording…………………………………………………....43 2.6 Fluorescence recovery after photobleaching (FRAP)……………………………………………..43 2.7 Marine extracts library, extraction and isolation…………………………………………………...43 2.8 High-throughput yellow fluorescent protein (YFP)- based anion influx assay (ToxFinder) …………………………………………………….…………………………………………………….44 2.9 Automated electrophysiology ……………………………….……………………………………………..45 2.10 Data analysis……………………………………………………………………………………………………….46 3 Chapter 3: GluClR-mediated inhibitory postsynaptic currents reveal targets for IVM and potential mechanisms of IVM resistance (In press – PLOS Pathogens) 3.1 Introduction…………………………………………………………………………………………………………55 3.2 Results…………………………………………………………………………………………………………………59 3.3 Discussion and conclusion……………………………………………………………………………………70 4 Chapter 4: Effects of glutamate and IVM on single GluClRs of parasitic nematode H.contortus (Published in PLOS Pathogens) 4.1 Introduction……………………………………………………………………………………………………….115 4.2 Results……………………………………………………………………………………………………………….118 4.3 Discussion and conclusion………………………………………………………………………………….131 5 Chapter 5: Pharmacological characterization of lindane, fipronil and picrotoxin on two GluClR isoforms of A.gambiae. 5.1 Introduction………………………………………………………………………………………………………163 5.2 Results………………………………………………………………………………………………………………164 5.3 Discussion and conclusion…………………………………………………………………………………171 6 Chapter 6: Other projects as part of my thesis 6.1 Introduction………………………………………………………………………………………………………177 6.2 Results………………………………………………………………………………………………………………178 6.3 Discussion and conclusion…………………………………………………………………………………186

xvi

7 Chapter 7: General discussion 7.1 The β subunit: a potential cause of culprit IVM resistance in H. contortus…………191 7.2 G36’A effects the intrinsic properties of the receptor and not IVM binding……….192 7.3 Management strategies to delay resistance……………………………………………………….194 7.4 Differential effects of lindane, fipronil and PTX at AgGluClR-a1 and AgGluClR-b…195 7.5 Other projects as part of my thesis…………………………………………………………………….197

Appendices: Supplementary Information Section 1: Effect of Glutamate and IVM at two β homomeric mutants of H. contortus……204 Section 2: Discovering new drugs for autism, cognitive impairment and stroke………………208 Animal Ethics approval certificate for X. laevis………………………………………………………………..220 Animal Ethics approval certificate for rats and their embryos…………………………………………223

xvii

List of figures and tables

Figures

Chapter 1: Introduction

Figure 1.1 General structure of pLGICs 4

Figure 1.2 Schematic distribution of different GluClR subunits in H. contortus. 13

Figure 1.3 Chemical structure of neurotransmitter glutamate and different binding 15 interactions in the binding pocket.

Figure 1.4 Chemical structure for IVM 18

Figure 1.5 Mechanism of allosteric activation and modulation by IVM 19

Figure 1.6 IVM-binding sites and atomic interactions. 19

Figure 1.7 Chemical structures for different GluClR antagonists. 22

Figure 1.8 Different binding sites on GABAARs 25

Figure 1.9 Structures of GABAARs agonists and antagonists 26

Chapter 3: GluClR-mediated inhibitory postsynaptic currents (IPSCs) reveal targets for IVM and potential mechanisms of IVM resistance.

Figure 1 Activity of glutamate and IVM at homo- and heteromeric GluClRs 78

Figure 2 Activity of IVM at homo- and heteromeric GluClRs 79

Figure 3 IVM analogue potency 80

Figure 4 IPSCs mediated by α and αβ GluClRs 81

xviii

Figure 5 Effect of IVM at IPSCs mediated at by α and αβ GluClRs 82

Figure 6 Single channel properties of β and αβ GluClRs 83

Figure 7 IVM-bdpy fluorescent probe 84

Figure 8 Membrane partitioning and lateral diffusion of IVM-bdpy 85

Figure S1 Synthetic scheme for IVM analogues, bdpy and IVM-bdpy 91

1 Figure S2 H NMR (DMSO-d6, 600 MHz) spectra of IVM 97

13 Figure S3 C NMR (DMSO-d6, 150 MHz) spectra of IVM 98

1 Figure S4 H NMR (CDCl3, 600 MHz) spectra of IVM-1 98

13 Figure S5 C NMR (CDCl3, 150 MHz) spectra of IVM-1 99

Figure S6 HRMS data for IVM-1 100

1 Figure S7 H NMR (CDCl3, 600 MHz) spectra of IVM-2 101

13 Figure S8 C NMR (CDCl3, 150 MHz) spectra of IVM-2 102

Figure S9 HRMS data for IVM-2 103

1 Figure S10 H NMR (CDCl3, 600 MHz) spectra of IVM-3 104

13 Figure S11 C NMR (CDCl3, 150 MHz) spectra of IVM-3 104

Figure S12 HRMS data for IVM-3 105

1 Figure S13 H NMR (MeOH-d4, 600 MHz) spectra of IVM-bdpy 106

13 Figure S14 C NMR (MeOH-d4, 150 MHz) spectra of IVM-bdpy 107

Figure S15 HRMS data for IVM-bdpy 108

xix

Figure A1 Effect of glutamate on WT and two β HcoGluClR mutants using TEVC 204

Figure A2 Effect of IVM on I-18’L 205

Chapter 4: Effects of glutamate and IVM on single GluClRs of parasitic nematode H.contortus

Figure 1 Single channel conductance and kinetic properties of GluClRs in response to 143 glutamate activation.

Figure 2 Kinetic properties of GluClRs at low glutamate concentrations. 144

Figure 3 Concentration-dependence of glutamate effects on GluClRs. 145

Figure 4 Ensemble glutamate-induced activation properties of GluClRs. 146

Figure 5 Comparison of the effect of 1 mM glutamate on wild-type and G36’A mutant 147 GluClRs.

Figure 6 Concentration-dependence of glutamate effects on G36’A mutant GluClRs. 148

Figure 7 Estimation of desensitization rate in patches expressing a known number of 149 wild-type or G36’A mutant channels.

Figure 8 Outside-out macropatch recordings of currents mediated by the indicated 150 receptors.

Figure 9 Direct activation by 5 nM IVM. 151

Figure 10 Current potentiation in the presence of 2 µM glutamate and 5 nM IVM. 152

Figure 11 Structural representations of TM3 domains in shut and IVM bound 153 configurations.

Chapter 5: Pharmacological characterization of lindane, fipronil and picrotoxin on two GluClR isoforms of A.gambiae.

Figure 5.1 Glutamate and IVM dose-response relationship on two GluClR splice variants 165 from A. gambiae.

Figure 5.2 Inhibitory dose-response relationship for lindane, fipronil and PTX at 168 AgGluClR-a1 and AgGluClR-b.

Figure 5.3 Time dependence recovery of lindane and fipronil induced inhibition at 170 AgGluClR-a1 and AgGluClR-b.

Figure 5.4 Ability of lindane, fipronil and PTX to block currents mediated by IVM. 172

Chapter 6: Other projects as part of my thesis

Figure 6.1 Flow chart showing the partition and trituration of ethanol crude extracts. 178

Figure 6.2 Summary of marine extract screening. 179

Figure 6.3 Concentration response of Aurantoside A against α5β3γ2L GABAARs. 181

Figure 6.4 Concentration response of Fumiquinazoline J against α5β3γ2L GABAARs. 182

Figure 6.5 Properties of spontaneous IPSCs recorded from heterosynapses 184

incorporating α5β1γ2L, α5β2γ2L and α5β3γ2L GABAARs.

Figure 6.6 Representative recordings of IPSCs from HEK293 cells expressing each 185 isoform.

Figure A3 GABAAR concentration responses on ToxFinder and using whole cell patch 208 clamp electrophysiology.

Figure A4 A sample dose response trace for GABA at α5β3γ2L GABAARs using the 211 patchliner.

xxi

Figure A5 Traces from the fractions belonging to CMB-01871 group. 212

Figure A6 Traces from the fractions belonging to CMB-03405 group. 213

Figure A7 Traces from the fractions belonging to CMB-03208 and CMB-02130 group. 214

Figure A8 Bar graph showing the different sub-fractions from CMB-02130-1-3 and 215

CMB-03405-1-3 and their effects on α5β3γ2L GABAARs when normalised to 1 µM GABA.

Figure A9 Traces from the sub-fractions belonging to fraction CMB-02130-1-3. 216

Figure A10 Traces from the sub-fractions belonging to fraction CMB-03405-1-3. 217

Figure A11 Traces from the sub-fractions belonging to fraction CMB-03405-1-3. 218

List of tables

Chapter 1: Introduction

Table 1.1 The GluClR genes of C. elegans and H. contortus and the subunits they 8 encode.

Table 1.2 Different locations of avermectin receptor subunits in parasitic nematode H. 13 contortus.

Chapter 3: GluClR-mediated inhibitory postsynaptic currents (IPSCs) reveal targets for IVM and potential mechanisms of IVM resistance.

Table 1 Glutamate and IVM concentration-response parameters. 60

Table A1 EC50 values of glutamate and IVM on two mutants from β homomeric 206 HcoGluClRs

Chapter 4: Effects of glutamate and IVM on single GluClRs of parasitic nematode H.contortus.

xxii

Table 1 Macropatch deactivation and desensitisation. 125

Table 2 IVM-dependent active durations and open probability (PO). 129

Table S1 Dwell time components and fractions for wild-type and G36’A mutant 159 GluClRs for glutamate.

Table S2 Glutamate-dependent active durations and open probability (PO). 160

Table S3 Dwell time components and fractions for wild-type and G36’A mutant 161 GluClRs for IVM.

Chapter 5: Pharmacological characterization of lindane, fipronil and picrotoxin on two GluClR isoforms of A.gambiae.

Table 5.1 Different nomenclature used for AgGluClR splice variants in paper and the 164 nucleotide database.

Table 5.2 Different EC50 values and Hill co-efficient for glutamate and IVM at two splice 165 isoforms of AgGluClRs.

Table 5.3 Different IC50 values and an approximate percentage inhibition by lindane, 167 fipronil and PTX at AgGluClR-a1 and AgGluClR-b.

Table 5.4 Time dependence recovery from lindane and fipronil induced inhibition using 169

EC50 glutamate at AgGluClR-a1 and AgGluClR-b.

Table 5.5 Different IC50 values and an approximate percentage of IVM currents 171 inhibited by lindane, fipronil and PTX at AgGluClR-a1 and AgGluClR-b.

Chapter 6: Other projects as part of my thesis.

Table 6.1 Percentage modulation by the active marine extracts, partitions and 181 triturates from the 14 fractions belonging to four crude extracts screened on ToxFinder.

xxiii

Table 6.2 Effect of 10 nM and 1 µM GT-002 on IPSC parameters for three different 185

GABAAR isoforms.

Table A5 GABA EC50 values from ToxFinder. 208

Table A6 GABA EC50 values from whole cell patch clamp recordings. 208

Table A7 Full list of crude extracts with more than 50% activity on the ToxFinder. The 209 subsequent activities of fractions also included.

Table A8 Different fractions tested on the Patchliner. 210

Table A9 Further partitioning products from the fractions CMB-02130-1-3 and 03405- 215 1-3.

xxiv

Abbreviations

5-HT3 5-hydroxytryptamine type 3

A.gambiae Anopheles gambiae

AFC Amphid neurons with finger like ciliated endings

AgGluClRs Anopheles gambiae glutamate gated chloride channel receptors.

AIY Amphid interneurons

ASD Autism spectrum disorder.

AWC amphid wing “C” cell

BES N,N-Bis(2-hydroxyethyl)-2-aminoethanesulfonic acid

BZD Benzodiazepine

C Cuticle

C.elegans Caenorhabditis elegans

CD4 Cluster differentiation 4

CNS Central nervous system

Cryo-EM Cryo-electron microscopy

D. melanogaster Drosophila melanogaster

Dc Dorsal nerve cord

DCM Dichloromethane

DMEM Dulbecco’s Modified Eagle Medium

DmGluClR D. melanogaster GluClR

DMSO Dimethyl sulfoxide

Dslc Dorsal sublateral cord

EC50 Half maximal effective concentration

ELIC Erwinia chrysanthemi

GABA γ-Aminobutyric acid

xxv

GABAAR γ-Aminobutyric acid type A receptor

GLIC Gleobacter violaceus

GluClRs Glutamate gated chloride channel receptors

GlyRs Glycine Receptors

H. contortus Haemonchus contortus

HcoGluClRs Haemonchus contortus glutamate gated chloride channel receptors

HEK 293 Human embryonic kidney type 293

HPLC High performance liquid chromatography

IC50 Half maximal inhibitory concentration

IMB Institute of Molecular Biosciences

IRS Indoor residual spraying

IVM Ivermectin

KCC2 K-Cl co-transporters

Lc Lateral cord

MeOH Methanol

MLs Macrocyclic lactones

MNC Motor neuron commissures mRNA Messenger ribonucleic acid nAChRs Nicotinic acetylcholine receptors n-BuOH n-butanol nH Hill coefficient

NMR Nuclear magnetic resonance

O. volvulus Onchocerca volvulus

P. xylostella Plutella xylostella

PBS Phosphate buffered saline

xxvi

PDL Poly-D-lysine

PenStrep Penicillin streptomycin

PGE2 Prostaglandin E2

Ph Pharynx pLGIC Pentameric ligand-gated ion channel

PO Intra-activation open probability.

PTX Picrotoxin

SCL Stable cell line

T. urticae Tetranychus urticae

TEVC Two electrode voltage clamp

TMD Transmembrane domain

Vc Ventral nerve cord

WT Wild-type

X. laevis Xenopus laevis

YFP Yellow fluorescent protein

α5 GABAARs α5 containing γ-Aminobutyric acid type A receptors

δ Desensitisation rate constant.

δ/ω Equilibrium constant.

ω Resensitisation rate constant

xxvii

Introduction

1. Introduction 1.1 Pentameric ligand-gated ion channels (pLGICs) Intercellular communication mediated by pLGICs is a result of conversion from binding of a neurotransmitter that is released from the presynaptic terminal into ion flux in the postsynaptic membrane (Lemoine et al. 2012). The members of this superfamily include

nicotinic acetylcholine receptors (nAChRs), the serotonin type 3 receptors (5-HT3Rs), the γ- aminobutyric acid type A receptors (GABAARs), the glutamate gated chloride channel receptors (GluClRs) and the glycine receptors (GlyRs). When a neurotransmitter binds to the above-mentioned receptors, it leads to activation of channels causing fast excitatory or inhibitory neurotransmission in the peripheral and the central nervous systems (CNS) (Lynch 2004). Having a broad distribution in the CNS, the pLGICs are involved in a number of biological processes. From the above-mentioned superfamily, some of them are cation selective which are involved in depolarisation of neurons through conducting Na+ and K+ such

as nAChRs and 5HT3Rs. On the other hand, the anion selective receptors such as GluClRs,

- GABAARs and GlyRs are involved in hyperpolarisation, mainly conducting Cl ions (Hille 2001). Drugs involved in treating disorders such as epilepsy, anxiety and use in anaesthetics involve GABAARs as their targets (Sigel and Steinmann 2012). On the contrary, a review by (Webb and Lynch 2007) has proposed GlyR subtypes as potential targets for treating chronic inflammatory pain and hyperekplexia. GluClRs on the other hand, are only found in invertebrates (Wolstenholme 2012) and are targets for treating parasitic infections caused by invertebrate pest species such as Haemonchus contortus (Chiejina et al. 2015). In addition to being used widely in agriculture, aquaculture and veterinary practice, drugs that target invertebrate GluClRs are also used to treat human conditions such as river blindness (Sneader 2005), lymphatic filariasis (Mehlhorn 2008) and scabies (Panahi et al. 2015). Anthelminthics such as ivermectin (IVM) are one class of drugs that target the GluClR where they bind in the transmembrane domain of the receptor and keep the channel open for a long duration resulting in a Cl- influx. This causes hyperpolarization of muscle and nerve cells resulting in paralysis and death of the organism (Cully et al. 1994; Arena et al. 1991; Hibbs and Gouaux 2011). However, IVM resistance is becoming a major problem in pest species (Wolstenholme et al. 2004), especially in the parasitic nematode H. contortus (McCavera et al. 2009). One focus of my PhD is to understand

one of the possible mechanisms by which the H. contortus develops resistance to anthelmenthics. 1.2 General structure and features of pLGICs As the name suggests, pLGICs consist of five subunits, which are arranged around a central ion-conducting pore. Using techniques such as X-ray crystallography, cryo electron microscopy (cryo-EM) and nuclear magnetic resonance (NMR) spectroscopy, various structures from the pLGIC family have been resolved. Channels from the prokaryotic bacteria Erwinia chrysanthemi (ELIC) and Gleobacter violaceus (GLIC) were the first complete structures to be resolved in their open and closed conformations using these techniques (Hilf and Dutzler 2008, 2009; Bocquet et al. 2009; Nury et al. 2010). (Basak et al. 2018) resolved

the crystal structure of 5-HT3AR in its resting confirmation using cryo-EM. Also (Hibbs and Gouaux 2011) crystallized the GluClRs from C. elegans in its apo and IVM bound state. The high-resolution crystal structure of the extracellular domain of nAChR α1 bound to α- bungarotoxin was solved using these techniques (Dellisanti et al. 2007). The crystal structure

of human GlyR α3 bound to antagonist strychnine, human GlyR α3 bound to IVM and a novel

class of analgesic potentiator bound to human GlyR α3 were solved using crystallographic techniques (Huang et al. 2015; Huang, Chen, et al. 2017; Huang, Shaffer, et al. 2017). All these crystallized structures share a similar structure. Using X-ray crystallography (Miller and

Aricescu 2014) recently solved the structure of human GABAAR. Using cryo-EM, (Phulera et

al. 2018) presented a new benzodiazepine sensitive α1β1γ2S triheteropentameric GABAAR in complex with GABA. Crystallographic techniques helped in revealing a new endogenous

neurosteroid-binding site on GABAAR chimera (Laverty et al. 2017). The data from these resolved structures indicate that the large N-terminal ligand-binding domain of pLGIC receptors are rich in β-sheets. Four α-helical regions form the transmembrane domains (TM1- TM4). A large intracellular loop connects the TM3 and TM4. Each subunit contributes the TM2 domain to form the channel pore (Figure 1.1). 1.2.1 Ligand binding domain The ligand-binding domain is made up of an α helix near its N-terminal which is followed by a number of β-strands which organise themselves as three strand outer and seven strand inner sheets that form a sandwich. The subunit interfaces form the ligand-binding pockets. Three loops (A, B and C) from a single subunit situated on the clockwise side of the interface forms the (+) side of the pocket whereas the three β-sheets (loop D, E and F) from 3

Figure 1.1: General structure of pLGICs. (A) Crystal structurer of C. elegans of α GluClR showing the channel pore and pentameric arrangement of the channel viewed from the top. (B) Parallel view of two individual subunits, red molecule showing the glutamate binding in the extracellular domain and IVM in orange binding in the TM domain. (C) A single GluClR subunit from two different angles. Loops of particular relevance to agonist binding and allosteric gating linkage are also shown. Image generated from PyMOL, Delano scientific LLC (PDB: 3RIF) (Hibbs and Gouaux 2011).

the adjacent subunit make up the complementary (-) side (Brejc et al. 2001; Lynch 2004) (Figure 1.1). A series of conformational rearrangements take place upon binding of a ligand in the pocket between the two subunits thereby causing the channel to open. Protruding towards the transmembrane region of the channel is the conserved cys-loop and the loop that connects the β sheets 2 and 3 (loop 2). Ligand-binding information is transmitted from the ligand-binding pocket to the channel gate via loop 2, which leads to the opening of the pore for review see (Lynch 2004; Webb and Lynch 2007). The ligand-binding pocket is rich in aromatic residues, which have been shown by previous studies to contribute to the ligand binding through cation-π interactions with the ligand (Pless et al. 2008). 1.2.2 Transmembrane (TM) Domains Made of four α-helical structures the TM domains in the pLGICs are embedded in the lipid membrane making a passageway through the apolar lipid bilayer for the aqueous ion permeation pathway. Connected to each other through three flexible linkers, the TM domains consists of an intracellular TM1-TM2 and TM3-TM4 linkers and an extracellular TM2-TM3 linker (Hibbs and Gouaux 2011; Miller and Aricescu 2014; Miyazawa et al. 2003). The water filled pore is formed by the TM2 domains. Surrounding the TM2 domain are the other three domains TM1, TM3 and TM4, working as a barrier in separating the TM2 from the hydrophobic environment of the lipid bilayer (Lynch 2004; Miller and Aricescu 2014). TM1 plays a role as a linker between the binding site and the channel activation gate due to its direct contact with the ligand-binding domain. A number of studies have determined the role of pre-TM1 residues in channel gating (Keramidas et al. 2006; Mercado and Czajkowski 2006). Spitzmaul et. al, showed that mutations of TM1 residues lead to an increased current desensitization and altered channel gating (Spitzmaul et al. 2004). TM2 helices are slightly kinked towards the axis of the channel pore, forming an impenetrable hydrophobic girdle due to side-by-side hydrophobic interactions of the hefty 9’ leucine side chain as suggested by the nAChR crystal structure (Miyazawa et al. 2003). The general importance of 9’ residues in the TM2 domain in forming the channel gate were

supported by the recent crystal structures of the β3 GABAAR and α GluClR (Hibbs and Gouaux 2011; Miller and Aricescu 2014). Many pharmacological and physiological probes bind to the TM2 domain, making a drug-binding site of interest.

The TM2 and TM3 domains are connected by a short flexible linker. The crucial role of this linker in coupling the agonist-binding site to the channel gate has been investigated by various studies (Pless et al. 2007; Bera et al. 2002). The TM3 domain is involved the lining of a water-filled pocket distinct from the channel pore (Williams and Akabas 1999, 2001, 2002). The role of TM3 residues as determinants for the binding of molecules such as alcohol, IVM and anaesthetics has been demonstrated by several studies (Lynch 2004). Mutational studies of the TM3 residues and the recent GluClR crystal structure have yielded a rational picture of the IVM binding site (Lynagh and Lynch 2012; Hibbs and Gouaux 2011). The role of the TM3 mutation, G36’A in changing the intrinsic properties of the receptor has recently been studied as part of the current project (Atif et al. 2017). There is also a large intracellular linker connecting the TM3 with the TM4 domain called the TM3-TM4 loop. Wild-type (WT) like activity was seen when this loop was replaced with a heptapeptide in chimeric receptors of GABAARs and 5HT3R, suggesting that this loop does not have much of a role in receptor assembly and function (Jansen et al. 2008). Nevertheless, this loop plays a role in phosphorylation and modulation of receptor function.

For example, PGE2 – mediated phosphorylation of the S346 residue in the TM3-TM4 loop of

α3 GlyR results in current inhibition leading to chronic inflammatory pain sensitization (Harvey et al. 2004). In addition, phosphorylation of the same residue in the TM3-TM4 intracellular loop produces a conformational change at the extracellular ligand-binding domain (Han, Talwar, Wang, et al. 2013). Making close contact along its entire length with the TM3 are the TM4 and TM1 domains (Hilf and Dutzler 2008). A difference in the glycine efficacy is observed at the homomeric α1 and α3 GlyRs due to non-conserved residues in the TM4 domain (Chen et al.

2009). The TM3 and TM4 domain in α1 and α3 GlyRs were compared for structural interactions, in order to get a clear picture of how the non-conserved residues from TM4 domain gave rise to a difference in receptor function. The results indicated a differential

orientation of TM4 around α1 and α3 GlyRs (Han, Talwar, and Lynch 2013). 1.3 Glutamate-gated chloride channel receptors (GluClRs)

First identified in invertebrate phyla, such as insects and crustaceans (Cull-Candy 1976; Gration et al. 1979; Marder and Paupardin-Tritsch 1978), GluClRs are closely related to

mammalian GlyRs (Wolstenholme 2012). They are found at neuronal and neuromuscular inhibitory synapses that selectively permeate anions (Wolstenholme 2012). Platyhelminthes, nematodes (e.g., H. contortus) and arthropods (e.g., A. gambiae) also express GluClRs (Wolstenholme 2012) however there is no evidence for their expression in vertebrates. GluClRs can assemble as homomers or heteromers (Cully et al. 1994). GluClRs are involved in a number of functions in these organisms that can range from feeding to mediating sensory inputs and locomotion. IVM, a drug belonging to the class of macrocyclic lactones, selectively targets GluClRs making them an important therapeutic target. The following section will discuss the different roles of GluClRs in these organisms.

1.3.1 Physiological role of GluClRs

In the invertebrate nervous system the GluClRs account for a number of different roles including, a) governing and regulation of movement, b) control of feeding and 3) intercession of sensory inputs (Wolstenholme 2012). The GluClRs are also involved in reproduction and maturation of nematodes, however these functions have not been studied in detail (Wolstenholme 2012). The main reason for choosing nematodes, as a model to study the function of GluClRs is the abundant expression of these channels in motor neurons, sensory neurons and interneurons and their importance as parasitic pest species (Horoszok et al. 2001; Cook et al. 2006). The below section highlights the role of GluClRs in nematodes as well as arthropods.

a) Locomotion

Role of GluClRs in nematode locomotion

There is an increased frequency in direction changes of C. elegans in response to odour and temperature; this increase is linked to the mutations in genes of GluClR such as avr-15, RNAi of avr-15 and glc-1, however an opposite effect is observed after knocking down the expression of glc-3 (Glendinning et al. 2011). Treatment of C. elegans worms with macrocyclic lactones inhibited locomotion and pharyngeal pumping (Wolstenholme and Rogers 2005; Kotze et al. 2012).

Role of GluClRs in arthropod locomotion The expression of GluClRs is abundant in insect neurons including the dorsal unpaired median neurons of locusts and cockroaches (Janssen et al. 2007; Janssen et al. 2010; Washio 7

2002). As these neurons innervate the leg and flight muscles, it is likely that GluClRs have an effect on the flight and movement control in these organisms.

b) Feeding

Role of GluClRs in foraging behaviour of nematodes

Mutations in avr-14 and other GluClR genes like avr-15, glc-1, and glc-3, which are widely expressed in C. elegans motor neurons and interneurons, affect the feeding patterns of the worm (Dent et al. 2000; Horoszok et al. 2001; Cook et al. 2006). It is clearly understood that maintenance of hydrostatic pressure and feeding of the worms in nematodes involves pharyngeal pumping, which is controlled by GluClRs. GluClRs expressed from the avr-15 gene mediate rapid relaxation of pharyngeal muscles which is facilitated by M3 glutamatergic motor neurons (Avery 1993; Raizen and Avery 1994) through a Cl- dependent hyperpolarization (Pemberton et al. 2001).

Role of GluClRs in feeding pattern of arthropods.

Human head lice hatched from eggs treated with IVM exhibited reduced feeding suggesting a role of GluClRs in regulation of feeding (Strycharz et al. 2011).

c) Sensory inputs

Apart from being expressed in the muscle cells of pharynx, avr-15 gene is also diversely expressed in the nervous system (Dent et al. 1997; Pemberton et al. 2001) along with glc-2 (Laughton et al. 1997). Gustatory plasticity is affected by mutations in avr-15 (Hukema et al. 2008) which mediates mechanosensory inhibition of pharyngeal pumping along with avr-14 (Keane and Avery 2003). Responses to odour (Chalasani et al. 2007) and temperature (Ohnishi et al. 2011) are regulated by amphid wing “C” cell (AWC) neurons and amphid neurons with finger like ciliated endings (AFD neurons) respectively which receive inputs from amphid interneurons (AIY interneurons) that expresses glc-3.

Table 1.1: The GluClR genes of C. elegans and H. contortus and the subunits they encode

Gene (C. elegans) Gene (H. contortus) Subunits encoded Properties

avr-14 avr-14 GluCl α3A Alternatively spliced to form 2 subunits. avr-14B forms glutamate and IVM-

GluCl α3B gated channels (Jagannathan et al. 1999; Laughton et al. 1997; Dent et al. 2000; McCavera et al. 2009).

avr-15 avr-15# GluCl α2A Alternatively spliced to form 2 subunits. Forms glutamate and IVM sensitive GluCl α2B channels (Dent et al. 1997; Vassilatis et al. 1997) # (Doyle et al. 2019)

glc-1 - GluCl α1 When expressed alone forms IVM sensitive channels, forms glutamate channels when co-expressed with GLC- 2 (Cully et al. 1994).

glc-2 glc-2 GluCl β GLC-2 from C. elegans forms glutamate channels that are sensitive to IVM when co-expressed with GlC-1 (Cully et al. 1994).

glc-3 glc-3 GluCl α4 Both glutamate and IVM channels are formed with C. elegans GLC-3. Hco-glc- 3 has been recently identified (Horoszok et al. 2001).

glc-4 glc-4 No data No functional data

- glc-5 GLC-5a Forms IVM and GluClRs (Forrester et al. 2004; Beech et al. 2010). GLC-5b*

- glc-6 GLC-6 Forms IVM sensitive GluClRs (Glendinning et al. 2011)

*=has not been localised yet

1.3.2 H.contortus a) Introduction

This parasite of ruminant animals is also known as barber’s pole worm due to the characteristic stripes that are present on its body. This parasite is mostly active during summer in warm humid climates where it infects the gastrointestinal tracts of sheep, cows and goats resulting in death due to anaemia and oedema (Burke 2005). Farmers around the world experience huge economic losses due the infection by H. contortus i.e., haemonchosis. Drugs belonging to the group of macrocyclic lactones are used to treat these and other worm infections, however resistance is becoming a major problem (Chiejina et al. 2015). b) Life cycle of H. contortus

The adult female H. contortus worm lays 5,000-15,000 eggs every day, which is then excreted in the faeces (Emery et al. 2016). By feeding on the bacteria present in the moist dung, the eggs continue to develop into L1 (rhabditiform) and L2 juvenile stages. Under optimal conditions of 24-29oC, it takes about four to six days for the eggs to reach L1 stage (CRC 2017). The eggs then develop into L3 infective larvae by shedding the cuticle of the L2 rhabditiform (Rahman and Collins 1990; El-Ashram and Suo 2017; CRC 2017). It is during grazing that sheep, goats and other ruminants are infected with L3 infective larvae. Passing through the three stomachs, they reach the abomasum followed by shedding the cuticle and burrowing into the internal layer of abomasum and develop into L4 in about 48hrs. These then molt and develop into L5 adult form where they live, mate and feed on the blood of the animal (El-Ashram and Suo 2017; CRC 2017). c) Pathogenicity, prevention and treatment

In order to prevent infection in the ruminants, prophylactic anthelmintic treatment is necessary however, given the rapid rise of anthelmintic resistance a reduction on reliance on chemical treatment is warranted. The only observation of acute infection is the sudden death of the animal. Apart from this, there are other symptoms such as anaemia, oedema, pallor and lethargy (Villarroel 2013). A common phenomenon called “bottle jaw” (in sheep and goats, also in cattle) may be observed wherein there is accumulation of fluid in submandibular tissue (Villarroel 2013).

Reducing the egg count and thereby pasture contamination can be done by vaccines that are commercially available, such as Barbevax or Wirevax available in Australia and South Africa respectively (Anderson 2013). An increased demand for cheap organic meat and dairy

products devoid of drug residues has motivated research on management strategies that are independent of anthelmintics. This is achieved by culling the animal that is most susceptible or by using parasite-resistant breeds (Villarroel 2013; Besier et al. 2012). In order to resist parasitic infections, selective breeding of sheep and goats seems to be a sustainable solution for small ruminant population. Over the past decade, a number of studies have shown different breeds that are resistant to parasitic infection (Alba-Hurtado and Munoz-Guzman 2013). A relatively fewer number of eggs were shed by St. Croix lambs and they harboured less than 1% of worms in their abomasum when compared with age-matched Dorset lambs both in vivo and in vitro (Gamble and Zajac 1992). South African Dorper breed were more prone to H. contortus infection than the Red Maasai sheep during natural exposure to parasites in Kenya (Mugambi et al. 1997).

d) Different types of GluClR isoforms from H.contortus and their distribution pattern

The different isoforms and the way they are distributed in the organism would help us to understand how and where the drugs targeting these genes would act. In H. contortus, to date three genes encoding four GluClR subunits have been identified. Two subunits of HcoGluClR α3A and –α3B are encoded by Hcgbr-2 which are orthologous to avr-14, both in gene structure and sequence (Portillo et al. 2003; Jagannathan et al. 1999). A β subunit gene that is orthologous to its C. elegans counterpart glc-2 was also identified i.e., HcoGluClR β (Delany et al. 1998). HcoGluClR α is third gene encoded by an α subunit which is not orthologous to any of the C. elegans GluClR genes and binds IVM when expressed in mammalian cultured cells (Forrester et al. 1999; Forrester et al. 2002; Cheeseman et al. 2001).

• HcoGluClR α and HcoGluClR β

HcoGluClR α subunits were detected on H. contortus motor neuron commissures (MNCs), this was done using immunofluorescence (Table 1.2). Apart from that, a brighter HcoGluClR α immunostaining was observed at the point where motor neurons made contact with nerve cords (Table 1.2) (Delany et al. 1998; Jagannathan et al. 1999). Adult H. contortus worms when

treated with anti-HcoGluClR α and anti-GABAAR antibodies showed that HcoGluClR α subunit is expressed on inhibitory motor neurons (Portillo et al. 2003). Avermectin sensitive receptors contain both subunits of, HcoGluClR α and HcoGluClR β, as suggested by their similar expression pattern in the motor nervous system that innervate the body-wall muscles of H.

contortus (Delany et al. 1998). The point where commissures made contact with the nerve cords showed brighter staining when labelled with anti-HcoGluClR α. The motor neuron commissures in the anterior and middle parts of H. contortus were completely co-localised with HcoGluClR α and –β antibodies (Table 1.2), confirming the presence of these two avermectin receptor subunits at this site.

MNCs of H. contortus were also stained with rat HcoGluClR β antibodies suggesting their presence in this cell type (Table 1.2). Sub-lateral and lateral nerve cords were stained anti- HcoGluClR β rat antibody in the midanterior region of the worm, showing new sites of expression for HcoGluClR β (Table 1.2).

• HcoGluClR α3A and HcoGluClR α3B

Experiments with immunofluorescence techniques suggested the expression of HcoGluClR α3A splice variants along different parts of the worm. Lateral neurons in the worm head were stained with an anti-HcoGluClR α3A antibody (Table 1.2). Cell bodies with a pair of nerve processes coming out of them could be easily identified. These nerve processes correspond to the bipolar sensory neurons from amphids. Based on the staining of cell bodies that were about 300 µm from the anterior tip of the worm, it is suggested that the amphidial neurons are located possibly on the lateral ganglion, posterior of the nerve ring. HcoGluClR α3 expression in the head was confirmed by (Portillo et al. 2003) (Table 1.2).

About 500 µm from the anterior tip, anti-HcoGluClR α3B antibodies were found to stain with the structures present in the head of the worm. MNCs in the anterior and middle regions of worms were stained with both HcoGluClR α3A and HcoGluClR α3B antibodies. However, expression of HcoGluClR α3B is much higher in the MNCs than that of HcoGluClR α3A (Table 1.2). Interestingly, HcoGluClR α3B expresses in nerve cords whereas HcoGluClR α3A does not (Portillo et al. 2003).

Even though there have been a number of studies done to describe the expression of GluClR subunits in nematodes and arthropods, the physiological combination of these subunits either in vivo or in vitro is not known. Part of the work in my thesis will work to address this question.

Table 1.2: Different locations of avermectin receptor subunits in parasitic nematode H. contortus. Taken from (Portillo et al. 2003).

Subunits Location of avermectin receptor subunits in H.contortus

GluClR α Motor neuron commissures (MNCs)

GluClR β MNCs, lateral and sublateral nerve cords.

GluClR α3A Amphidial neurons, MNCs, nerve cord and nerve rings

GluClR α3B Pharyngeal neurons, MNCs, sublateral and ventral nerve cords, and nerve rings

Figure 1.2: The above figure depicts the schematic distribution of different GluClR subunits in H. contortus. (A) A side view of the complete worm. (B) A side view of the complete dorsal-ventral extent of the motor nervous system, without the cuticle showing the expression of individual subunits and a combination of different subunits. (C) A top view of pharynx exhibiting the possible expression of GluClR α3 subunits in the pharyngeal motor neurons (M1, M2l and M2r) and in amphid neurons (A). Anterior end of the worm is towards the right in all the panels and in panel A and B; the dorsal side is towards the top. c = Cuticle, dc = dorsal nerve cord, dslc = dorsal sublateral cord, lc = lateral cord, mnc = motor neuron commissures, ph = pharynx, vc = ventral nerve cord, vslc = ventral sublateral cord. Figure from (Portillo et al. 2003).

1.3.3 A. gambiae a) Introduction There were around 198 million cases of malaria reported worldwide in the year of 2013, from 584,000 resulted in death according to the 2014 WHO world malaria report. Africa

accounted for approximately 90% of these deaths where A. gambiae is the primary vector of malaria (WHO 2014). Spraying of pyrethroid-based insecticides forms one of the current malaria control programs. Pyrethroid resistance however is becoming widespread among the A.gambiae population across Africa (Ranson et al. 2011; Trape et al. 2011). b) Life cycle of A. gambiae

A. gambiae like any other species of mosquitoes has four life stages i.e., egg, larva, pupa and adult. Apart from feeding on the nectar from plants, female mosquitoes also bloodfeed on vertebrates which provides the nutrients for her eggs (Foster and Walker 2009). These mosquitoes prefer to bloodfeed from humans and are therefore called as anthrophilic (White 1974) thereby making them efficient carriers of human malaria parasite (CDC. 2010). Larvae hatch from the eggs that have been laid on water surfaces in different aquatic habitats (Gillies and Meillon 1968). After the fourth instar, the larvae pupate (Foster and Walker 2009) and this stage encompasses a non-feeding stage during which the pupa develops into the adult mosquito (Gillies and Meillon 1968).

c) Malaria – its medical significance and management

In order to control the mosquito population and limit malaria transmission, a number of management techniques are in place. These include educating the community about malaria and the way it spreads and different modifications that can be done to prevent the entry of the mosquitoes in the house, such as using nets on the windows, bed nets, spatial repellents and use of insecticides as indoor residual spraying (IRS) (Takken and Knols 2009). However, continuous use of insecticides can cause the mosquitoes to develop resistance (Brogdon and McAllister 1998; Mawejje et al. 2013). A. gambiae have already been found to develop resistance to insecticides in the regions of eastern Uganda (Mawejje et al. 2013), thereby leading to the decreased efficacy of the bed nets that were treated with pyrethroid.

e) Different GluClR splice isoforms from A.gambiae

Four splice isoforms were discovered from the cloning of AgGluClR gene from the mRNA of A. gambiae (Meyers et al. 2015). These four genes were labelled as AgGluClR a1, a2, b and c. A protein structure that corresponds to the Cl- permeable members of the pLGICs was disclosed upon the analysis of AgGluClR gene (Meyers et al. 2015). A protein with conserved residues that is associated the glutamate and IVM binding sites was revealed after

aligning the gene product from the recently crystallised GluClR α from C. elegans (CeGluClR) with the GluClR from Drosophila melanogaster (DmGluClR) (Hibbs and Gouaux 2011).

f) Expression profile of GluClR genes in mature A. gambiae

Studies using immunolabelling found that AgGluClR genes were expressed all over the body of the mosquito, it was found at Johnston’s organ, antennal segments, thoracic ganglia, supraesophageal ganglion and the optic lobe (Meyers et al. 2015). Antennal segments that might be associated with antennal sensilla were found to have small puncta of AgGluClR staining. They were found in each layer of the optic lobe and at the photoreceptors. Involved in the audition and flight coordination, the Johnston’s organ which is a mechanosensory organ was found to have specific AgGluClR staining. The staining was also prominent at the cell bodies along supraesophageal ganglion as well as on the neuronal processes. As the three thoracic ganglia showed AgGluClR staining, it suggested that AgGluClRs are involved in locomotion and flight of the mosquito, as thoracic ganglia comprises of motor neurons that govern the flight and the leg muscles (Meyers et al. 2015).

1.3.4 Agonists of GluClRs 1.3.4.1 Glutamate and ibotenate

Glutamate is a neurotransmitter that activates GluClRs by binding in the orthosteric ligand-binding pocket present in the extracellular domain where it lodges itself between two subunits (Hibbs and Gouaux 2011).

Ibotenate also called, as ibotenic acid is a psychoactive compound from the plant Amanita muscaria and acts as a non-selective agonist for GluClRs because of its structural similarity with the neurotransmitter glutamate (Becker et al. 1999; Isacson et al. 1984).

Figure 1.3: (A) Chemical structure of the neurotransmitter glutamate. (B) Glutamate binding site looking parallel to the membrane with loop C removed for clarity. Blown up image showing different interactions of glutamate with the residues in the binding pocket with hydrogen bonds shown in dashed lines and a cation-π interaction Y200 from (Hibbs and Gouaux 2011).

1.3.4.2 IVM

Belonging to a class of broad-spectrum pesticides, IVM binds and potentiates GluClR activity. The impact of IVM on human and veterinary health has been extraordinary, from its discovery on a Japanese golf course to winning a Nobel Prize.

a) Discovery and synthesis

In the year 1970, on the southeast coast of Honshu, Japan, a microbiologist Satoshi Omura collected a soil sample from the woods near a golf course in Kawana (Van Voorhis et al. 2015). From this soil sample, the then unknown species of a Gram-positive bacteria, Streptomyces (named as sample NRRL 8165-a) was isolated and cultured by S. Omura. Heligomosoides polygyrus infection in mice was significantly reduced when tested against NRRL 8165 cultures and when these cultures were purified to get the active compound, it was revealed to be from a family of macrocyclic lactones. In order to reflect the worm free averminous conditions produced by these naturally occurring compounds, they were named avermectins and subsequently the bacteria as, Streptomyces avermitilis (Burg et al. 1979; Campbell 1981).

Avermectins that exists naturally are a mixture of four different compounds i.e., avermectin A1, A2, B1 and B2 (Campbell 1981; Campbell et al. 1983). These four compounds had different biological properties as they differed from each other subtly in their structural properties. Although, all the four compounds showed some efficacy against nematodes, the most significant effect was seen with avermectins belonging to the B series (Blair and Campbell 1979). Moreover, a difference in efficacy was seen when there was a change in the mode of drug administration. That is, avermectin B1 was more effective when administered orally whereas avermectin B2 was more active when the dose was administered parenterally (Campbell et al. 1983). Commercial anthelmintic development then focused on the B series and changes in the chemical structure due to the fact that B series was more significant than the other series. Based on this, using the naturally occurring avermectin B1, IVM was synthesised which is comprised of 80%̴ 22, 23-dihydro-avermectin B1a and 20%̴ 22, 23 dihydro-avermectin B1b. IVM has significant activity against parasitic nematodes when administered both orally and parenterally (Campbell 1981; Campbell et al. 1983).

16 b) Use of IVM in animal and human health

Although IVM is licensed to control lungworm, ectoparasites in cows and in aquaculture industry, its market has remained particularly strong for the control of gastrointestinal roundworms such as H. contortus. Anthelmintics like IVM, at present are the most commonly sought after treatment in the UK sheep industry (Burgess et al. 2012) and cattle industry in the US (McArthur and Reinemeyer 2014). Apart from that, the control of equine roundworms also involves IVM (Relf et al. 2012; Stratford et al. 2014), and macrocyclic lactones forms a large market for the treatment and control of parasitic nematodes and ectoparasites in domestic pets. Control of gastrointestinal roundworms and canine heartworms involves the use of IVM. Transmission of malaria from the adult mosquito can be prevented by targeting the developing larvae with IVM (Laing et al. 2017).

Even though the market for IVM in livestock and animal market was substantial from the start, there was very little thought that went into developing IVM for human health. Dr Roy Vagelos, CEO of Merck & Co., was moved by the efficacy of IVM against onchocerciasis and lymphatic filariasis and decided to donate as much IVM (licensed as Mectizan) that was needed, for anyone that needed it and for long as they needed it (Molyneux and Ward 2015). A single oral dose of IVM (150 mg/kg) given annually suppresses the production of microfilaria thereby preventing the disease progression however does not kill the adult Onchocerca volvulus (Taylor and Greene 1989). The same is the case in lymphatic filariasis by Wuchereria bancrofti, where IVM does not kill the adult organism but its microfilaria (Diallo et al. 1987; Endeshaw et al. 2015; Richards et al. 2005). Complete clearance of Wuchereria bancrofti microfilariae was observed when treated with a single dose of IVM in combination with albendazole and diethylcarbamazine (DEC) (Thomsen et al. 2016). c) GluClRs as targets for IVM

The testing for the potentiating effects of IVM using electrophysiological investigations began as early as first reports of anthelmintic activity that suggested pLGICs as their target. IVM increases the Cl- conductance in lobster neurons that have ligand-gated sodium and Cl- channels (Fritz et al. 1979) and IVM diminished the responses of motor neurons to excitatory input in the nematode preparations (Kass et al. 1984). Two GluClR subunits from C. elegans were isolated, where one of it i.e., α, was sensitive to IVM whereas the other subunit, β, was

17 sensitive to glutamate, when expressed as homomeric GluClRs. The subunits also assembled as heteromers and were active against both glutamate and IVM (Cully et al. 1994). Sharing up to 40% of the amino acid sequence, the GluClRs from helminths clearly fit into pLGIC superfamily with the closest human relative being GlyRs (Vassilatis et al. 1997; Dent 2006). IVM has been shown to have variable activity at Histamine gated chloride channel receptors (HisClRs) and proton chloride channel receptors (pHClRs) which share up to 45% sequence identity with GlyRs (Zheng et al. 2002; Schnizler et al. 2005; Mounsey et al. 2007). d) Mode of action of IVM and its binding sites at GluClR

Even though the efficacy of IVM against the infections from parasitic nematodes is clearly understood, the way it exerts is action at GluClRs is unclear because we do not know whether it is the diffusion properties or its partitioning parameters or both that exerts the action. The binding pocket for IVM is formed by the subunit interface of two transmembrane domains proximal to the extracellular region of the membrane bilayer (Figure 1.5 & 1.6). Like a foot wedged in the door, the IVM molecule wedges itself between the a helix from TM3 that makes up the principal or (+) subunit and the complimentary or (-) subunit that is formed from the a helix of TM1 of a different subunit (Hibbs and Gouaux 2011). IVM makes important contacts with the TM2-TM3 loop and the TM2 (+) pore lining a helix, by inserting itself deeply into the subunit interface (Figure 1.6) (Hibbs and Gouaux 2011).

Figure 1.4: Chemical structure of IVM. IVM is a mixture of two homologues 5-O-dimethyl-

22, 23-dihydroavermectin B1a and B1b in a ratio of 80:20.

Figure 1.5: A cartoon representation showing the mechanism of allosteric activation and modulation by IVM. (A) Single subunit interface, parallel to membrane. (B) View from the top. IVM (shown as green block) binds at the subunit interface of TM3 and TM1 α-helices spreading it apart allowing for ion conductance. Adapted from (Hibbs and Gouaux 2011).

Figure 1.6: IVM-binding sites and atomic interactions. (A) & (B) Different orientations of a GluClR subunit interface showing binding sites for IVM (shown as a yellow molecule) (Hibbs and Gouaux 2011).

e) Resistance to anthelmintics

Mass drug administration of IVM has been used extensively in animals for prevention and treatment against parasitic nematodes, but unfortunately, they have often developed resistance due to the implementation of this strategy. This resistance towards anthelmintics is now a grave issue in controlling gastrointestinal roundworms in cattle, horses and sheep (Kaplan and Vidyashankar 2012). Concerns about IVM resistance evolving in human parasites have been raised in Ghana and Cameroon where there have been reports of reduced embryostatic effects of IVM on O. volvulus (Osei-Atweneboana et al. 2011; Osei- Atweneboana et al. 2007; Nana-Djeunga et al. 2014).

A number of genes involved in IVM resistance have been identified in C. elegans (Laing et al. 2017). High level of IVM resistance was conferred in C. elegans after a simultaneous mutation of three GluClR genes i.e., glc-1, avr-14 and avr-15 (Dent et al. 2000). A four amino acid deletion in glc-1 gene from field isolates of C. elegans resulted in resistance towards abamectin and IVM in multiple populations (Ghosh et al. 2012). More recently, a number of candidate genes like glc-1, avr-14 and avr-15 have been investigated in the parasitic nematode H. contortus to determine their link in IVM resistance, however, an association between IVM resistance and these candidate genes could not be proved (Gilleard 2013). Wild populations of Plutella xylostella (A30’V) (Wang et al. 2017; Wang et al. 2016) and Tetranychus urticae (G36’D and G36’E) (Dermauw et al. 2012; Kwon et al. 2010; Mermans et al. 2017) have been shown to be resistant IVM due to missense point mutations to GluClR subunits in their genome. Apart from this, many studies have focused on the role of drug uptake and their metabolism/excretion (Kotze et al. 2014) where a number of polymorphisms and/or changes in the expression pattern of the candidate gene were correlated with IVM resistance. However, to date studies have not been able to define the major mechanism behind IVM resistance.

f) New targets and novel applications

Having a wide safety margin, IVM affects the susceptible nematode at nanomolar concentrations but has a broad range of effects across different organisms at higher concentrations. There is an increased Cl- conductance in mammalian neuronal cells in response to a high IVM application, and based on this, patients with spinal cord injuries have been treated for severe muscle spasticity using a high dose of IVM (up to 1.6 mg/kg) (Costa 20

and Diazgranados 1994). In vitro experiments involving human leukaemia cells have shown that IVM induces an intracellular Cl- flux that selectively kills leukaemia cells (Sharmeen et al. 2010). A deregulation of WNT-TCF (WNT-T cell factor) signalling pathway has been observed in many diseases, including cancers of breast, ovary, colon, prostate and skin. In vivo experiments showed that IVM is effective against lung carcinoma xenografts and WNT-TCF- dependant human colon cancer, suggesting that IVM specifically blocks the WNT-TCF response (Melotti et al. 2014). A recent study by has shown that IVM possesses anti- inflammatory properties against T cell-mediated skin disease (Ventre et al. 2017). By inhibiting a viral helicase, IVM stops the replication process of a number of flaviviruses (Mastrangelo et al. 2012). IVM activity has been reported against various species of Mycobacterium, including the causative agent of tuberculosis Mycobacterium tuberculosis (Lim et al. 2013).

1.3.5 Antagonists Insecticides such as lindane, fipronil and a pore blocker picrotoxin which are non-

competitive antagonists have been shown to inhibit responses mediated by GABAARs, GlyRs and GluClRs (Chen et al. 2006; Islam and Lynch 2012; Hirata et al. 2008). a) Lindane

An organochlorine chemical and an isomer of hexachlorocyclohexane, it is also called gamma-hexachlorocyclohexane. Used both as a pharmaceutical and an insecticide for scabies and lice treatment (Goldust et al. 2013; Panahi et al. 2015; Wooltorton 2003), lindane binds

to the Drosophila melanogaster RDL GABAARs and acts a neurotoxin by interfering with the GABA neurotransmission (Cole et al. 1995) and also inhibits the GlyRs (Islam and Lynch 2012). It affects kidneys, liver and the nervous system in humans (Gerberding 2005).

b) Fipronil

A broad-spectrum insecticide, fipronil comes from the phenylpyrazole family. It blocks

Drosophila melanogaster RDL GABAARs (Hosie et al. 1995) and GluClRs in the insect nervous system leading to hyperexcitation of the insects (Hirata et al. 2008). Fipronil has also been show to inhibit currents mediated by GlyRs (Islam and Lynch 2012). Its effectiveness towards insects is believed to be due to the GluClRs that are only present in invertebrates and not in the mammals (Raymond-Delpech et al. 2005) apart from its action on the insect GABA

21 receptors. Due to this, a large number of pharmaceuticals for controlling fleas incorporate fipronil as its active ingredient (Maddison et al. 2008). c) Picrotoxin

Also called as cocculin, it is a mixture of two related compounds (picrotin and picrotoxin) and is found naturally in fruits of the plant Anamirta occulus. Picrotoxin conveys its action as a stimulant and a convulsant by interacting with neurotransmitter GABA thereby causing respiratory paralysis and seizures by influencing the CNS at high doses (Overington et al. 2006). Picrotoxin inhibits α homomeric GlyRs but shows a reduced sensitivity at the heteromeric αβ GlyRs (Pribilla et al. 1992; Handford et al. 1996; Islam and Lynch 2012).

Figure 1.7: Chemical structures for different GluClR antagonists

1.4 GABAARs

1.4.1 Physiological role of GABAARs

GABAARs belong to the pLGIC family, which mediates fast synaptic

- - neurotransmission in the central nervous system (CNS). They permeate HCO3 and Cl ions. When activated in adult neurons, where the intracellular Cl- concentration is low, they induce

- a Cl influx thereby hyperpolarising the cell. GABAARs can also cause neuronal depolarization in some regions of CNS and in early embryonic development due to an increased resting intracellular Cl- concentration. K-Cl co-transporters maintain the low intracellular Cl- concentration in adults (Rivera et al. 2005). The low efficacy of GABAAR-mediated inhibition in nigral dopaminergic neurons could possibly be due to the level of KCC2 expression, which is low in adult dopaminergic neurons in the substansia nigra pars compacta (Gulacsi et al.

2003). Furthermore, the local Cl- gradients are affected due to varied expression levels of KCC2 in subdomains of neurons (Szabadics et al. 2006).

1.4.2 GABAAR subunit composition and binding sites a) Different subunit composition of GABAARs

Being a member of the pLGIC family, functional GABAARs are comprised of five subunits symmetrically arranged around a central ion-conducting pore. So far, 19 subunits

(α1–6, β1–3, γ1–3, δ, ɛ, θ, and π) of GABAARs have been identified in mammals with the most common composition being 2α, 2β and 1γ subunit (Sieghart 2006; Hedblom and Kirkness 1997; Wisden et al. 1992). The neuroanatomical distribution of individual subunits reveal distinct but overlapping patterns based on in situ hybridization and immunohistochemical research. The α1 containing GABAARs are the most abundant isoforms of all the GABAARs present in the brain and accounts for 70%-90% of the total GABAARs and is involved in sedation (Sieghart and Sperk 2002; Pirker et al. 2000). About 50-70% of GABAARs contain β2 and γ2 subunits (Pirker et al. 2000). Distributed widely throughout the brain are α1, β1-3 and γ2 subunits, apart from that the other subunits have a limited distribution but may be locally enriched in certain brain regions. For example, 31% of GABAARs containing α5 subunits and

13% of GABAARs containing α4 subunits are present in hippocampus (Sieghart and Sperk 2002).

Because cerebellar granule cells are abundant and strongly express α6 subunits, this

subunit accounts for about 56% of GABAAR subunits in the cerebellum (Sieghart and Sperk 2002). Although they are highly expressed in the central and medial amygdaloid nuclei, the

general distribution of γ1 subunit is extremely sparse in the brain (Sieghart and Sperk 2002). Although reported as being expressed in the brain, ρ subunits preferentially express in the retina whereas the distribution pattern of ε, π and θ subunits is still not clear (Moragues et al. 2000; Moragues et al. 2002; Hedblom and Kirkness 1997; Sinkkonen et al. 2000). The common

stoichiometric formation of GABAARs is two α subunits, two β subunits and one γ subunit, where the γ can be replaced with δ, ε, θ or π subunits. The most common stoichiometry of

GABAARs in the brain is α1β2γ2 (Sieghart and Sperk 2002; Rudolph and Mohler 2006; Ernst et

al. 2003). The most common combination containing the α5 GABAARs is α5β3γ2 and is implicated in autism (McCauley et al. 2004), cognitive impairment (Vithlani et al. 2011) and

stroke (Clarkson et al. 2010). Only about 100 combinations of GABAARs are thought to exist in the brain out of approximately 10,000 possible stoichiometries (Ernst et al. 2003). b) The GABAAR binding sites: GABA and BZDs.

The Ernst group made the first structural models of GABAAR binding pockets for GABA and BZDs. This was done by alignment and computation studies based on the crystal structure of AChBP (Ernst et al. 2003). The structure that was obtained was imprecise as the sequence alignment between GABAAR and AChBP was uncertain.

The subunit interface between β+ (loops A, B and C) and α- (loops D, E and F) forms the GABA binding pocket. The above mentioned loops incorporate the aromatic side chains that provide sites for the interaction with the ligands via cation-π interactions (Brejc et al. 2001; Dougherty 2008).

γ2 knockdown experiments confirmed that γ subunit is necessary for the high affinity-binding BZD site (Gunther et al. 1995). The interface between α+ and γ- subunits forms the high affinity BZD binding pocket. Larger BZDs molecules can fit in this binding pocket as this BZD binding site is much deeper and more open that the GABA binding pocket, as suggested by modelling studies (Cromer et al. 2002; Ernst et al. 2003).

1.4.3 Overview of GABAAR pharmacology

When the neurotransmitter GABA binds to the GABAAR, it results in the influx of Cl- through the pore via an inwards electrochemical gradient. The GABA-mediated hyperpolarization is antagonised by a plant alkaloid, bicuculline, which competes with GABA for its orthosteric site (Beitz and Larson 1985). There are several other allosteric sites present on the receptor in addition to the orthosteric site (Figure 1.8). For example, BZDs and barbiturates are called positive modulators or potentiators, as they potentiate the effect of GABA by binding to the allosteric sites (Johnston 2005). For the past 50 years BZDs have been used for a different range of therapeutic purposes such as sedatives in epilepsy, anxiolytic, hypnotics and muscle relaxation (Sigel and Buhr 1997). Drugs such as barbiturates, BZDs,

neurosteroids, alcohols and anaesthetics have different sites of action on GABAARs such as on individual subunits or in clefts between two subunits. Barbiturates, BZDs and neurosteroids are used to treat conditions such as anxiety, sedation, insomnia, epilepsy and muscle spasms etc. Negative modulators or inverse agonists are the agents that bind to the allosteric site and 24 reduce the GABA induced response. Whereas there are some agents referred to as neutralising modulators due to their neutralising effect, where they inhibit the effects of both the positive and negative modulators, one such example is a drug that binds to the BZD site i.e., flumazenil (Johnston 2005). Modulators targeting the allosteric sites on GABAARs have been widely used as therapeutics; this is because GABAARs play consequential role in several neurological disorders such as autism spectrum disorder, cognitive impairment, stroke, epilepsy, Alzheimer’s disease, anxiety, insomnia and Down syndrome. However, currently available medications target different GABAAR subtypes non-selectively producing untoward effects such as sedation, memory impairment and tolerance. In order to minimise these undesirable side effects, increasing number of studies are trying to develop and discover modulators that target specific GABAAR subtypes selectively.

- Figure 1.8: Different binding sites on GABAARs. The Cl ion channel as seen from parallel to the membrane has the orthosteric GABA and other allosteric binding positions. The β subunit in the front is removed for clarity.

Figure 1.9: Structures of GABAAR agonists and antagonists. The endogenous agonist GABA binds to the orthosteric site. The positive allosteric modulators, diazepam and pentobarbital, bind at the allosteric sites and the antagonists, bicuculline and PTX, bind in the orthosteric site and pore, respectively.

1.4.4 Therapeutic significance of modulators specifically targeting α5 containing GABAARs

Drugs selectively targeting α5 containing GABAARs (α5 GABAARs) are of specific

interest. α5 GABAARs have a relatively restricted distribution, being primarily expressed in the hippocampus, a region of the brain associated with learning and memory (Chambers et al.

2003). Although the α5 receptors account for less than 5% of the total GABAAR population in the brain (McKernan and Whiting 1996), in the hippocampus they represent 20% of all

GABAARs (Sieghart 1995), thereby implicating this GABAAR subtype in learning and memory (Sur et al. 1998). A number of disorders including stroke, cognitive enhancement, dementia-

related conditions and schizophrenia have been linked with α5 GABAARs as therapeutic

targets. Both positive and negative modulators of α5 GABAARs have the potential to be

clinically useful, unlike other GABAAR subtypes (Soh and Lynch 2015).

Autism Genetic studies conducted by (McCauley et al. 2004) reported the linkage of one

autism subgroup to chromosome 15q11-13 in the α5 GABAAR coding regions. While the α5

GABAAR has been postulated to be a candidate gene for autism spectrum disorder, no convincing evidence for a potential link of chromosome 15 to this disorder is available (Delong

2007). Altered expression levels of GABAARs in the brains of subjects with autism suggested a GABA/glutamate system dysregulation (Esmaeili and Ghaedi 2010). A significant decrease in

α1, α2, α3 and β3 in parietal cortex was seen (D'Hulst et al. 2006). These results revealed that

GABAAR expression was reduced in the brain regions that have previously been associated with the pathogenesis of autism, suggesting a widespread GABAergic dysfunction in the brain of subjects with autism (Rice et al. 1996). In α5 GABAAR knockout mice, there were identified gamma frequency oscillations in the electroencephalogram, evoked by increased network drive, which were significantly greater than in wild type mice (Towers et al. 2004). Similar increased gamma activity is found in individuals with autism identifying the presence or absence of an illusory Kanizsa shape (an illusory shape shown to a patient suffering with autism to measure the gamma abnormalities) (Brown et al. 2005). Abnormal gamma activity has been hypothesized to underlie cognitive deficits in autism (Brock et al. 2002). Modulation of associative learning (acquisition and retention) involves hippocampal α5-containing

GABAARs which supports the involvement of α5 GABAARs in cognitive dysfunction and memory impairment associated with ASD (Cellot and Cherubini 2014). Cognitive Impairment

The α5 subunits have been implicated in learning and memory following the identification of a single point mutation (α5 H105R) that facilitates a particular form of

associative learning in mice (Crestani et al. 2002). This impairs the interaction of α5 receptor subtype with diazepam, thereby abolishing its memory-impairing effects (Vithlani et al. 2011).

α5 containing GABAARs have been suggested to play a potential role in hippocampal- dependent learning and memory (Mohler and Rudolph 2017; Rudolph and Mohler 2014). Another study conducted by (Collinson et al. 2002) demonstrated that decreased GABA

mediated synaptic inhibition in the hippocampus in α5 deficient mice is most likely the

underlying mechanism for the enhanced performance in learning and memory without having an effect on anxiety or sedation. A study done by (Atack 2010), upheld the findings by showing, triazolophthalazine, a negative modulator that decreases GABA responses particularly at α5 GABAARs, enhanced intellectual performance without causing anxiety or

convulsions in rodents. Cognitive performance was impaired by the use of α5 GABAARs positive-allosteric modulators indicating the involvement of this subunit in learning and memory and cognitive tasks (Xu et al. 2018; Mohamad and Has 2019). Thus, compounds created to be specific for α5 GABAARs can possibly be utilized as memory enhancing

medications without symptoms such as, sedation, which is caused by α1-containing GABAARs. This class of medications might be used to treat cognitive impairment, as is found in patients suffering from Alzheimer’s and other dementias. Stroke Stroke is one of the leading causes of disability, and is without any pharmacological therapy for promoting recovery (Clarkson et al. 2010). Current therapies that promote functional recovery after stroke are limited to physical rehabilitation (Dobkin 2008). After

stroke, tonic inhibition mediated by extra synaptic GABAAR is increased in the peri-infarct zone through an impairment of GABA transporter (GAT-3/GAT-4) function (Clarkson et al.

2010). In α5 knockout mice, motor recovery improved, suggesting that the use of α5-specific inverse agonists may be a novel approach to promote recovery after stroke (Clarkson et al. 2010).

The discrete localization of the α5 subtype mainly in the hippocampus and the excellent sensitivity of the inhibitory GABAergic system to pharmacological intervention

makes the α5-containing GABAAR an attractive novel target for treatment of disorders associated with cognitive defects.

Aims The main aim of my PhD project was to determine whether the α HcoGluClR subunit in combination with β HcoGluClR affects IVM sensitivity. A subsidiary aim was to determine the possible mechanisms behind IVM resistance in H. contortus. As IVM, resistance is becoming a major problem and causing huge capital losses to farmers around the world, one of the ways to address this problem is to identify one of the mechanisms of how H. contortus develops resistance and thereby develop techniques that can address this issue. Previous studies on C. elegans have shown that changes in the subunit composition can have differential effects on IVM sensitivity. We hypothesised that one of the GluClR subunits could be the reason that H.contortus is developing resistance to IVM. Thorough studies have been carried out on many pLGIC family members including nAChRs (Auerbach 2014; Purohit et al. 2013; Mukhtasimova et al. 2016). GlyRs (Burzomato et al. 2004), 5-HT3Rs (Corradi et al. 2009), and GABAARs (Keramidas and Harrison 2010; Dixon et al. 2014; Lema and Auerbach 2006), to determine how agonists bind to and activate the receptor. However, a detailed study on the biophysical properties of GluClRs has not been performed. The second part of this project was to investigate the biophysical properties of α homomeric GluClRs from H. contortus. Here we tried to determine how glutamate and IVM produce different activation properties and to delve into the way in which the mutation, TM3-

G36’A reduces IVM sensitivity to a level that is similar to vertebrates GlyRs and GABAARs (Lynagh and Lynch 2010; Estrada-Mondragon and Lynch 2015). Two GluClR splice isoforms from A. gambiae were shown to have different sensitivities towards IVM (Meyers et al. 2015). As these splice variants only have a marginal difference in their amino acid sequence, in the third aim we wanted to see whether there would be differential effects towards different insecticides such as lindane, fipronil, PTX and also towards glutamate and IVM. The last aim of my PhD was to discover and characterise new molecules that

specifically targets α5 GABAARs. The eventual aim was to identify a novel lead compound as a potential treatment for neurological disorders including autism, age related cognitive impairment and stroke.

Aim 1: To determine one of the possible mechanisms of IVM resistance in H. contortus Aim 2: To determine the biophysical properties of α HcoGluClR and a mutation in its TM3 domain. Aim 3: To pharmacologically characterize the effects of a few insecticides as well as a classical pore blocker picrotoxin on two splice variants of GluClRs from A. gambiae.

Aim 4: To discover new lead compounds that selectively target α5 GABAARs.

References

Alba-Hurtado, F., and M. A. Munoz-Guzman. 2013. 'Immune responses associated with resistance to haemonchosis in sheep', Biomed Res Int, 2013: 162158. Anderson, Samuel. 2013. 'New England small Ruminant Survey Results'. Arena, J. P., et al. 1991. 'Avermectin-sensitive chloride currents induced by Caenorhabditis elegans RNA in Xenopus oocytes', Mol Pharmacol, 40: 368-74. Atack, J. R. 2010. 'Preclinical and clinical pharmacology of the GABAA receptor alpha5 subtype- selective inverse agonist alpha5IA', Pharmacol Ther, 125: 11-26. Atif, M., et al. 2017. 'Effects of glutamate and ivermectin on single glutamate-gated chloride channels of the parasitic nematode H. contortus', PLoS Pathog, 13: e1006663. Auerbach, A. 2014. 'Agonist activation of a nicotinic acetylcholine receptor', Neuropharmacology. Avery, L. 1993. 'Motor neuron M3 controls pharyngeal muscle relaxation timing in Caenorhabditis elegans', J Exp Biol, 175: 283-97. Basak, S., et al. 2018. 'Cryo-EM structure of 5-HT3A receptor in its resting conformation', Nat Commun, 9: 514. Becker, A., et al. 1999. 'Social behaviour in rats lesioned with ibotenic acid in the hippocampus: quantitative and qualitative analysis', Psychopharmacology (Berl), 144: 333-8. Beech, R., et al. 2010. 'Association of ion-channel genotype and macrocyclic lactone sensitivity traits in Haemonchus contortus', Mol Biochem Parasitol, 171: 74-80. Beitz, A. J., and A. A. Larson. 1985. 'Inhibition of intrathecally administered picrotoxin- and bicuculline- induced convulsions in mice by pipecolic acid or GABA', Eur J Pharmacol, 114: 181-7. Bera, A. K., et al. 2002. 'GABA(A) receptor M2-M3 loop secondary structure and changes in accessibility during channel gating', J Biol Chem, 277: 43002-10. Besier, B, et al. 2012. "Towards a commercial vaccine against Haemonchus contortus-a field trial in Western Australia." In Proc. Australian Sheep Vet Conf, 14-18. Blair, L. S., and W. C. Campbell. 1979. 'Efficacy of avermectin B1a against microfilariae of Dirofilaria immitis', Am J Vet Res, 40: 1031-2. Bocquet, N., et al. 2009. 'X-ray structure of a pentameric ligand-gated ion channel in an apparently open conformation', Nature, 457: 111-4. Brejc, K., et al. 2001. 'Crystal structure of an ACh-binding protein reveals the ligand-binding domain of nicotinic receptors', Nature, 411: 269-76. Brock, J., et al. 2002. 'The temporal binding deficit hypothesis of autism', Dev Psychopathol, 14: 209- 24. Brogdon, W. G., and J. C. McAllister. 1998. 'Insecticide resistance and vector control', Emerg Infect Dis, 4: 605-13. Brown, C., et al. 2005. 'Gamma abnormalities during perception of illusory figures in autism', Cortex, 41: 364-76. Burg, R. W., et al. 1979. 'Avermectins, new family of potent anthelmintic agents: producing organism and fermentation', Antimicrob Agents Chemother, 15: 361-7. Burgess, C. G., et al. 2012. 'A survey of the trichostrongylid nematode species present on UK sheep farms and associated anthelmintic control practices', Vet Parasitol, 189: 299-307. Burke, Joan. 2005. 'Management of Barber pole Worm in Sheep and Goats in the Southern U.S. '. Burzomato, V., et al. 2004. 'Single-channel behavior of heteromeric alpha1beta glycine receptors: an attempt to detect a conformational change before the channel opens', J Neurosci, 24: 10924- 40. Campbell, W. C. 1981. 'An introduction to the avermectins', N Z Vet J, 29: 174-8. Campbell, W. C., et al. 1983. 'Ivermectin: a potent new antiparasitic agent', Science, 221: 823-8. Cellot, Giada, and Enrico Cherubini. 2014. 'GABAergic Signaling as Therapeutic Target for Autism Spectrum Disorders', Frontiers in Pediatrics, 2. Chalasani, S. H., et al. 2007. 'Dissecting a circuit for olfactory behaviour in Caenorhabditis elegans', Nature, 450: 63-70. 31

Chambers, M. S., et al. 2003. 'Identification of a novel, selective GABA(A) alpha5 receptor inverse agonist which enhances cognition', J Med Chem, 46: 2227-40. Cheeseman, C. L., et al. 2001. 'High-affinity ivermectin binding to recombinant subunits of the Haemonchus contortus glutamate-gated chloride channel', Mol Biochem Parasitol, 114: 161- 8. Chen, L., et al. 2006. 'Structural model for gamma-aminobutyric acid receptor noncompetitive antagonist binding: widely diverse structures fit the same site', Proc Natl Acad Sci U S A, 103: 5185-90. Chen, X., et al. 2009. 'The M4 transmembrane segment contributes to agonist efficacy differences between alpha1 and alpha3 glycine receptors', Mol Membr Biol, 26: 321-32. Chiejina, S. N., et al. 2015. 'Haemonchotolerance in West African Dwarf goats: contribution to sustainable, anthelmintics-free helminth control in traditionally managed Nigerian dwarf goats', Parasite, 22: 7. Clarkson, A. N., et al. 2010. 'Reducing excessive GABA-mediated tonic inhibition promotes functional recovery after stroke', Nature, 468: 305-9. Cole, L. M., et al. 1995. 'Drosophila GABA-gated chloride channel: modified [3H]EBOB binding site associated with Ala-->Ser or Gly mutants of Rdl subunit', Life Sci, 56: 757-65. Collinson, N., et al. 2002. 'Enhanced learning and memory and altered GABAergic synaptic transmission in mice lacking the alpha 5 subunit of the GABAA receptor', J Neurosci, 22: 5572- 80. Cook, A., et al. 2006. 'Caenorhabditis elegans ivermectin receptors regulate locomotor behaviour and are functional orthologues of Haemonchus contortus receptors', Mol Biochem Parasitol, 147: 118-25. Corradi, J., et al. 2009. 'Single-channel kinetic analysis for activation and desensitization of homomeric 5-HT(3)A receptors', Biophys J, 97: 1335-45. Costa, J. L., and J. A. Diazgranados. 1994. 'Ivermectin for spasticity in spinal-cord injury', Lancet, 343: 739. CRC, Sheep. 2017. 'Roundworm life cycle'. http://www.wormboss.com.au/worms/roundworms/roundworm-life-cycle.php. Crestani, F., et al. 2002. 'Trace fear conditioning involves hippocampal alpha5 GABA(A) receptors', Proc Natl Acad Sci U S A, 99: 8980-5. Cromer, B. A., et al. 2002. 'Anxiety over GABA(A) receptor structure relieved by AChBP', Trends Biochem Sci, 27: 280-7. Cull-Candy, S. G. 1976. 'Two types of extrajunctional L-glutamate receptors in locust muscle fibres', J Physiol, 255: 449-64. Cully, D. F., et al. 1994. 'Cloning of an avermectin-sensitive glutamate-gated chloride channel from Caenorhabditis elegans', Nature, 371: 707-11. D'Hulst, C., et al. 2006. 'Decreased expression of the GABAA receptor in fragile X syndrome', Brain Res, 1121: 238-45. Delany, N. S., et al. 1998. 'Cloning and localisation of an avermectin receptor-related subunit from Haemonchus contortus', Mol Biochem Parasitol, 97: 177-87. Dellisanti, C. D., et al. 2007. 'Crystal structure of the extracellular domain of nAChR alpha1 bound to alpha-bungarotoxin at 1.94 A resolution', Nat Neurosci, 10: 953-62. Delong, R. 2007. 'GABA(A) receptor alpha5 subunit as a candidate gene for autism and bipolar disorder: a proposed endophenotype with parent-of-origin and gain-of-function features,with or without oculocutaneous albinism', Autism, 11: 135-47. Dent, J. A. 2006. 'Evidence for a diverse Cys-loop ligand-gated ion channel superfamily in early bilateria', J Mol Evol, 62: 523-35. Dent, J. A., et al. 1997. 'avr-15 encodes a chloride channel subunit that mediates inhibitory glutamatergic neurotransmission and ivermectin sensitivity in Caenorhabditis elegans', EMBO J, 16: 5867-79.

Dent, J. A., et al. 2000. 'The genetics of ivermectin resistance in Caenorhabditis elegans', Proc Natl Acad Sci U S A, 97: 2674-9. Dermauw, W., et al. 2012. 'The cys-loop ligand-gated ion channel gene family of Tetranychus urticae: implications for acaricide toxicology and a novel mutation associated with abamectin resistance', Insect Biochem Mol Biol, 42: 455-65. Diallo, S., et al. 1987. 'Dose-ranging study of ivermectin in treatment of filariasis due to Wuchereria bancrofti', Lancet, 1: 1030. Dixon, C., et al. 2014. 'GABAA receptor alpha and gamma subunits shape synaptic currents via different mechanisms', J Biol Chem, 289: 5399-411. Dobkin, B. H. 2008. 'Training and exercise to drive poststroke recovery', Nat Clin Pract Neurol, 4: 76- 85. Dougherty, D. A. 2008. 'Cys-loop neuroreceptors: structure to the rescue?', Chem Rev, 108: 1642-53. Doyle, S. R., et al. 2019. 'Population genomic and evolutionary modelling analyses reveal a single major QTL for ivermectin drug resistance in the pathogenic nematode, Haemonchus contortus', BMC Genomics, 20: 218. El-Ashram, S., and X. Suo. 2017. 'Erratum: Exploring the microbial community (microflora) associated with ovine Haemonchus contortus (macroflora) field strains', Sci Rep, 7: 12403. Emery, D. L., et al. 2016. 'Haemonchus contortus: the then and now, and where to from here?', Int J Parasitol, 46: 755-69. Endeshaw, T., et al. 2015. 'Presence of Wuchereria bancrofti microfilaremia despite 7 years of annual ivermectin monotherapy mass drug administration for onchocerciasis control: a study in north-west Ethiopia', Pathog Glob Health, 109: 344-51. Ernst, M., et al. 2003. 'Comparative modeling of GABA(A) receptors: limits, insights, future developments', Neuroscience, 119: 933-43. Esmaeili, A, and K Ghaedi. 2010. 'GABAA receptors as novel drug targets for treatment of mental disorders'. Estrada-Mondragon, A., and J. W. Lynch. 2015. 'Functional characterization of ivermectin binding sites in alpha1beta2gamma2L GABA(A) receptors', Front Mol Neurosci, 8: 55. Forrester, S. G., et al. 2004. 'Agonist enhacement of macrocyclic lactone activity at a glutamate-gated chloride channel subunit from Haemonchus contortus', Biochem Pharmacol, 67: 1019-24. Forrester, S. G., et al. 1999. 'Cloning, sequencing, and developmental expression levels of a novel glutamate-gated chloride channel homologue in the parasitic nematode Haemonchus contortus', Biochem Biophys Res Commun, 254: 529-34. Forrester, S. G., et al. 2002. 'A glutamate-gated chloride channel subunit from Haemonchus contortus: expression in a mammalian cell line, ligand binding, and modulation of anthelmintic binding by glutamate', Biochem Pharmacol, 63: 1061-8. Foster, W.A., and E.D. Walker. 2009. Medical and Veterinary Entomology (Academic Press). Fritz, L. C., et al. 1979. 'Avermectin B1a irreversibly blocks postsynaptic potentials at the lobster neuromuscular junction by reducing muscle membrane resistance', Proc Natl Acad Sci U S A, 76: 2062-6. Gamble, H. R., and A. M. Zajac. 1992. 'Resistance of St. Croix lambs to Haemonchus contortus in experimentally and naturally acquired infections', Vet Parasitol, 41: 211-25. Gerberding, Julie Louise. 2005. 'Toxicological profile of alpha, beta, gamma and delta- hexachlorocyclohexane'. Ghosh, R., et al. 2012. 'Natural variation in a chloride channel subunit confers avermectin resistance in C. elegans', Science, 335: 574-8. Gilleard, J. S. 2013. 'Haemonchus contortus as a paradigm and model to study anthelmintic drug resistance', Parasitology, 140: 1506-22. Gillies, M.T., and B. de Meillon. 1968. 'The Anophelinae of Africa South of the Sahara (Ethiopian Zoogeographical Region): Second Edition, by M. T. Gillies, School of Biological Sciences, University of Sussex, Brighton, England, and Botha De Meillon, Project Leader, South East Asia Mosquito Project, Smithsonian Institution, Washington, D.C. Publication of the South African

Institute for Medical Research, No. 54. 343 pages, illustrated. The South African Institute for Medical Research, P.O. Box 1038, Johannesburg, South Africa. 1968. R 15.00', The American Journal of Tropical Medicine and Hygiene, 19: 736-36. Glendinning, S. K., et al. 2011. 'Glutamate-gated chloride channels of Haemonchus contortus restore drug sensitivity to ivermectin resistant Caenorhabditis elegans', PLoS One, 6: e22390. Goldust, M., et al. 2013. 'Ivermectin vs. lindane in the treatment of scabies', Ann Parasitol, 59: 37-41. Gration, K. A., et al. 1979. 'Three types of L-glutamate receptor on junctional membrane of locust muscle fibres', Brain Res, 171: 360-4. Gulacsi, A., et al. 2003. 'Cell type-specific differences in chloride-regulatory mechanisms and GABA(A) receptor-mediated inhibition in rat substantia nigra', J Neurosci, 23: 8237-46. Gunther, U., et al. 1995. 'Benzodiazepine-insensitive mice generated by targeted disruption of the gamma 2 subunit gene of gamma-aminobutyric acid type A receptors', Proc Natl Acad Sci U S A, 92: 7749-53. Han, L., et al. 2013. 'The relative orientation of the TM3 and TM4 domains varies between alpha1 and alpha3 glycine receptors', ACS Chem Neurosci, 4: 248-54. Han, L., et al. 2013. 'Phosphorylation of alpha3 glycine receptors induces a conformational change in the glycine-binding site', ACS Chem Neurosci, 4: 1361-70. Handford, C. A., et al. 1996. 'The human glycine receptor beta subunit: primary structure, functional characterisation and chromosomal localisation of the human and murine genes', Brain Res Mol Brain Res, 35: 211-9. Harvey, R. J., et al. 2004. 'GlyR alpha3: an essential target for spinal PGE2-mediated inflammatory pain sensitization', Science, 304: 884-7. Hedblom, E., and E. F. Kirkness. 1997. 'A novel class of GABAA receptor subunit in tissues of the reproductive system', J Biol Chem, 272: 15346-50. Hibbs, R. E., and E. Gouaux. 2011. 'Principles of activation and permeation in an anion-selective Cys- loop receptor', Nature, 474: 54-60. Hilf, R. J., and R. Dutzler. 2008. 'X-ray structure of a prokaryotic pentameric ligand-gated ion channel', Nature, 452: 375-9. Hilf, R. J., and R. Dutzler. 2009. 'Structure of a potentially open state of a proton-activated pentameric ligand-gated ion channel', Nature, 457: 115-8. Hille, Bertil. 2001. Ion channels of excitable membranes (Sinauer: Sunderland, Mass.). Hirata, K., et al. 2008. 'Role of a serine residue (S278) in the pore-facing region of the housefly L- glutamate-gated chloride channel in determining sensitivity to noncompetitive antagonists', Insect Mol Biol, 17: 341-50. Horoszok, L., et al. 2001. 'GLC-3: a novel fipronil and BIDN-sensitive, but picrotoxinin-insensitive, L- glutamate-gated chloride channel subunit from Caenorhabditis elegans', Br J Pharmacol, 132: 1247-54. Hosie, A. M., et al. 1995. 'Actions of the insecticide fipronil, on dieldrin-sensitive and- resistant GABA receptors of Drosophila melanogaster', Br J Pharmacol, 115: 909-12. Huang, X., et al. 2015. 'Crystal structure of human glycine receptor-alpha3 bound to antagonist strychnine', Nature, 526: 277-80. Huang, X., et al. 2017. 'Crystal Structures of Human GlyRalpha3 Bound to Ivermectin', Structure, 25: 945-50 e2. Huang, X., et al. 2017. 'Crystal structures of human glycine receptor alpha3 bound to a novel class of analgesic potentiators', Nat Struct Mol Biol, 24: 108-13. Hukema, R. K., et al. 2008. 'Gustatory plasticity in C. elegans involves integration of negative cues and NaCl taste mediated by serotonin, dopamine, and glutamate', Learn Mem, 15: 829-36. Isacson, O., et al. 1984. 'Functional neuronal replacement by grafted striatal neurones in the ibotenic acid-lesioned rat striatum', Nature, 311: 458-60. Islam, R., and J. W. Lynch. 2012. 'Mechanism of action of the insecticides, lindane and fipronil, on glycine receptor chloride channels', Br J Pharmacol, 165: 2707-20.

Jagannathan, S., et al. 1999. 'Ligand-gated chloride channel subunits encoded by the Haemonchus contortus and Ascaris suum orthologues of the Caenorhabditis elegans gbr-2 (avr-14) gene', Mol Biochem Parasitol, 103: 129-40. Jansen, M., et al. 2008. 'Modular design of Cys-loop ligand-gated ion channels: functional 5-HT3 and GABA rho1 receptors lacking the large cytoplasmic M3M4 loop', J Gen Physiol, 131: 137-46. Janssen, D., et al. 2007. 'Dorsal unpaired median neurons of locusta migratoria express ivermectin- and fipronil-sensitive glutamate-gated chloride channels', J Neurophysiol, 97: 2642-50. Janssen, D., et al. 2010. 'Cys-loop ligand-gated chloride channels in dorsal unpaired median neurons of Locusta migratoria', J Neurophysiol, 103: 2587-98. Johnston, G. A. 2005. 'GABA(A) receptor channel pharmacology', Curr Pharm Des, 11: 1867-85. Kaplan, R. M., and A. N. Vidyashankar. 2012. 'An inconvenient truth: global worming and anthelmintic resistance', Vet Parasitol, 186: 70-8. Kass, I. S., et al. 1984. 'The effects of avermectin and drugs related to acetylcholine and 4-aminobutyric acid on neurotransmission in Ascaris suum', Mol Biochem Parasitol, 13: 213-25. Keane, J., and L. Avery. 2003. 'Mechanosensory inputs influence Caenorhabditis elegans pharyngeal activity via ivermectin sensitivity genes', Genetics, 164: 153-62. Keramidas, A., and N. L. Harrison. 2010. 'The activation mechanism of alpha1beta2gamma2S and alpha3beta3gamma2S GABAA receptors', J Gen Physiol, 135: 59-75. Keramidas, A., et al. 2006. 'The pre-M1 segment of the alpha1 subunit is a transduction element in the activation of the GABAA receptor', J Physiol, 575: 11-22. Kotze, A. C., et al. 2012. 'A reappraisal of the relative sensitivity of nematode pharyngeal and somatic musculature to macrocyclic lactone drugs', Int J Parasitol Drugs Drug Resist, 2: 29-35. Kotze, A. C., et al. 2014. 'Recent advances in candidate-gene and whole-genome approaches to the discovery of anthelmintic resistance markers and the description of drug/receptor interactions', Int J Parasitol Drugs Drug Resist, 4: 164-84. Kwon, D. H., et al. 2010. 'A point mutation in a glutamate-gated chloride channel confers abamectin resistance in the two-spotted spider mite, Tetranychus urticae Koch', Insect Mol Biol, 19: 583- 91. Laing, R., et al. 2017. 'Ivermectin - Old Drug, New Tricks?', Trends Parasitol, 33: 463-72. Laughton, D. L., et al. 1997. 'Reporter gene constructs suggest that the Caenorhabditis elegans avermectin receptor beta-subunit is expressed solely in the pharynx', J Exp Biol, 200: 1509-14. Laverty, D., et al. 2017. 'Crystal structures of a GABAA-receptor chimera reveal new endogenous neurosteroid-binding sites', Nat Struct Mol Biol, 24: 977-85. Lema, G. M., and A. Auerbach. 2006. 'Modes and models of GABA(A) receptor gating', J Physiol, 572: 183-200. Lemoine, D., et al. 2012. 'Ligand-gated ion channels: new insights into neurological disorders and ligand recognition', Chem Rev, 112: 6285-318. Lim, L. E., et al. 2013. 'Anthelmintic avermectins kill Mycobacterium tuberculosis, including multidrug- resistant clinical strains', Antimicrob Agents Chemother, 57: 1040-6. Lynagh, T., and J. W. Lynch. 2010. 'A glycine residue essential for high ivermectin sensitivity in Cys- loop ion channel receptors', Int J Parasitol, 40: 1477-81. Lynagh, T., and J. W. Lynch. 2012. 'Molecular mechanisms of Cys-loop ion channel receptor modulation by ivermectin', Front Mol Neurosci, 5: 60. Lynch, J. W. 2004. 'Molecular structure and function of the glycine receptor chloride channel', Physiol Rev, 84: 1051-95. Maddison, Jill E., et al. 2008. Small animal clinical pharmacology (Saunders/Elsevier: Edinburgh ; New York). Marder, E., and D. Paupardin-Tritsch. 1978. 'The pharmacological properties of some crustacean neuronal acetylcholine, gamma-aminobutyric acid, and L-glutamate responses', J Physiol, 280: 213-36.

Mastrangelo, Eloise, et al. 2012. 'Ivermectin is a potent inhibitor of flavivirus replication specifically targeting NS3 helicase activity: new prospects for an old drug', Journal of Antimicrobial Chemotherapy, 67: 1884-94. Mawejje, H. D., et al. 2013. 'Insecticide resistance monitoring of field-collected Anopheles gambiae s.l. populations from Jinja, eastern Uganda, identifies high levels of pyrethroid resistance', Med Vet Entomol, 27: 276-83. McArthur, M. J., and C. R. Reinemeyer. 2014. 'Herding the U.S. cattle industry toward a paradigm shift in parasite control', Vet Parasitol, 204: 34-43. McCauley, J. L., et al. 2004. 'A linkage disequilibrium map of the 1-Mb 15q12 GABA(A) receptor subunit cluster and association to autism', Am J Med Genet B Neuropsychiatr Genet, 131B: 51-9. McCavera, S., et al. 2009. 'An ivermectin-sensitive glutamate-gated chloride channel from the parasitic nematode Haemonchus contortus', Mol Pharmacol, 75: 1347-55. McKernan, R. M., and P. J. Whiting. 1996. 'Which GABAA-receptor subtypes really occur in the brain?', Trends Neurosci, 19: 139-43. Mehlhorn, Heinz. 2008. Encyclopedia of parasitology (Springer: New York). Melotti, A., et al. 2014. 'The river blindness drug Ivermectin and related macrocyclic lactones inhibit WNT-TCF pathway responses in human cancer', EMBO Mol Med, 6: 1263-78. Mercado, J., and C. Czajkowski. 2006. 'Charged residues in the alpha1 and beta2 pre-M1 regions involved in GABAA receptor activation', J Neurosci, 26: 2031-40. Mermans, C., et al. 2017. 'A G326E substitution in the glutamate-gated chloride channel 3 (GluCl3) of the two-spotted spider mite Tetranychus urticae abolishes the agonistic activity of macrocyclic lactones', Pest Manag Sci, 73: 2413-18. Meyers, J. I., et al. 2015. 'Characterization of the target of ivermectin, the glutamate-gated chloride channel, from Anopheles gambiae', J Exp Biol, 218: 1478-86. Miller, P. S., and A. R. Aricescu. 2014. 'Crystal structure of a human GABAA receptor', Nature, 512: 270-5. Miyazawa, A., et al. 2003. 'Structure and gating mechanism of the acetylcholine receptor pore', Nature, 423: 949-55. Mohamad, Fatin H., and Ahmad Tarmizi Che Has. 2019. 'The α5-Containing GABAA Receptors—a Brief Summary', Journal of Molecular Neuroscience, 67: 343-51. Mohler, H., and U. Rudolph. 2017. 'Disinhibition, an emerging pharmacology of learning and memory', F1000Res, 6. Molyneux, D. H., and S. A. Ward. 2015. 'Reflections on the Nobel Prize for Medicine 2015--The Public Health Legacy and Impact of Avermectin and Artemisinin', Trends Parasitol, 31: 605-07. Moragues, N., et al. 2000. 'cDNA cloning and expression of a gamma-aminobutyric acid A receptor epsilon-subunit in rat brain', Eur J Neurosci, 12: 4318-30. Moragues, N., et al. 2002. 'Localisation of GABA(A) receptor epsilon-subunit in cholinergic and aminergic neurones and evidence for co-distribution with the theta-subunit in rat brain', Neuroscience, 111: 657-69. Mounsey, K. E., et al. 2007. 'Molecular characterisation of a pH-gated chloride channel from Sarcoptes scabiei', Invert Neurosci, 7: 149-56. Mugambi, J. M., et al. 1997. 'Resistance of four sheep breeds to natural and subsequent artificial Haemonchus contortus infection', Vet Parasitol, 69: 265-73. Mukhtasimova, N., et al. 2016. 'Improved resolution of single channel dwell times reveals mechanisms of binding, priming, and gating in muscle AChR', J Gen Physiol, 148: 43-63. Nana-Djeunga, H. C., et al. 2014. 'Reproductive status of Onchocerca volvulus after ivermectin treatment in an ivermectin-naive and a frequently treated population from Cameroon', PLoS Negl Trop Dis, 8: e2824. Nury, H., et al. 2010. 'Crystal structure of the extracellular domain of a bacterial ligand-gated ion channel', J Mol Biol, 395: 1114-27. Ohnishi, N., et al. 2011. 'Bidirectional regulation of thermotaxis by glutamate transmissions in Caenorhabditis elegans', EMBO J, 30: 1376-88.

Osei-Atweneboana, M. Y., et al. 2011. 'Phenotypic evidence of emerging ivermectin resistance in Onchocerca volvulus', PLoS Negl Trop Dis, 5: e998. Osei-Atweneboana, M. Y., et al. 2007. 'Prevalence and intensity of Onchocerca volvulus infection and efficacy of ivermectin in endemic communities in Ghana: a two-phase epidemiological study', Lancet, 369: 2021-29. Overington, J. P., et al. 2006. 'How many drug targets are there?', Nat Rev Drug Discov, 5: 993-6. Panahi, Y., et al. 2015. 'The efficacy of topical and oral ivermectin in the treatment of human scabies', Ann Parasitol, 61: 11-6. Pemberton, D. J., et al. 2001. 'Characterization of glutamate-gated chloride channels in the pharynx of wild-type and mutant Caenorhabditis elegans delineates the role of the subunit GluCl-alpha2 in the function of the native receptor', Mol Pharmacol, 59: 1037-43. Phulera, S., et al. 2018. 'Cryo-EM structure of the benzodiazepine-sensitive alpha1beta1gamma2S tri- heteromeric GABAA receptor in complex with GABA', Elife, 7. Pirker, S., et al. 2000. 'GABA(A) receptors: immunocytochemical distribution of 13 subunits in the adult rat brain', Neuroscience, 101: 815-50. Pless, S. A., et al. 2007. 'Conformational variability of the glycine receptor M2 domain in response to activation by different agonists', J Biol Chem, 282: 36057-67. Pless, S. A., et al. 2008. 'A cation-pi interaction in the binding site of the glycine receptor is mediated by a phenylalanine residue', J Neurosci, 28: 10937-42. Portillo, V., et al. 2003. 'Distribution of glutamate-gated chloride channel subunits in the parasitic nematode Haemonchus contortus', J Comp Neurol, 462: 213-22. Pribilla, I., et al. 1992. 'The atypical M2 segment of the beta subunit confers picrotoxinin resistance to inhibitory glycine receptor channels', EMBO J, 11: 4305-11. Purohit, P., et al. 2013. 'Functional anatomy of an allosteric protein', Nat Commun, 4: 2984. Rahman, W. A., and G. H. Collins. 1990. 'The establishment and development of Haemonchus contortus in goats', Vet Parasitol, 35: 189-93. Raizen, D. M., and L. Avery. 1994. 'Electrical activity and behavior in the pharynx of Caenorhabditis elegans', Neuron, 12: 483-95. Ranson, H., et al. 2011. 'Pyrethroid resistance in African anopheline mosquitoes: what are the implications for malaria control?', Trends Parasitol, 27: 91-8. Raymond-Delpech, V., et al. 2005. 'Ion channels: molecular targets of neuroactive insecticides', Invert Neurosci, 5: 119-33. Relf, V. E., et al. 2012. 'A questionnaire study on parasite control practices on UK breeding Thoroughbred studs', Equine Vet J, 44: 466-71. Rice, A., et al. 1996. 'Long-lasting reduction of inhibitory function and gamma-aminobutyric acid type A receptor subunit mRNA expression in a model of temporal lobe epilepsy', Proc Natl Acad Sci U S A, 93: 9665-9. Richards, F. O., Jr., et al. 2005. 'Mass ivermectin treatment for onchocerciasis: lack of evidence for collateral impact on transmission of Wuchereria bancrofti in areas of co-endemicity', Filaria J, 4: 6. Rivera, C., et al. 2005. 'Two developmental switches in GABAergic signalling: the K+-Cl- cotransporter KCC2 and carbonic anhydrase CAVII', J Physiol, 562: 27-36. Rudolph, U., and H. Mohler. 2006. 'GABA-based therapeutic approaches: GABAA receptor subtype functions', Curr Opin Pharmacol, 6: 18-23. Rudolph, U., and H. Mohler. 2014. 'GABAA receptor subtypes: Therapeutic potential in Down syndrome, affective disorders, schizophrenia, and autism', Annu Rev Pharmacol Toxicol, 54: 483-507. Schnizler, K., et al. 2005. 'A novel chloride channel in Drosophila melanogaster is inhibited by protons', J Biol Chem, 280: 16254-62. Sharmeen, S., et al. 2010. 'The antiparasitic agent ivermectin induces chloride-dependent membrane hyperpolarization and cell death in leukemia cells', Blood, 116: 3593-603.

Sieghart, W. 1995. 'Structure and pharmacology of gamma-aminobutyric acidA receptor subtypes', Pharmacol Rev, 47: 181-234. Sieghart, W. 2006. 'Structure, pharmacology, and function of GABAA receptor subtypes', Adv Pharmacol, 54: 231-63. Sieghart, W., and G. Sperk. 2002. 'Subunit composition, distribution and function of GABA(A) receptor subtypes', Curr Top Med Chem, 2: 795-816. Sigel, E., and A. Buhr. 1997. 'The benzodiazepine binding site of GABAA receptors', Trends Pharmacol Sci, 18: 425-9. Sigel, E., and M. E. Steinmann. 2012. 'Structure, function, and modulation of GABA(A) receptors', J Biol Chem, 287: 40224-31. Sinkkonen, S. T., et al. 2000. 'GABA(A) receptor epsilon and theta subunits display unusual structural variation between species and are enriched in the rat locus ceruleus', J Neurosci, 20: 3588-95. Sneader, W. 2005. 'Drug Discovery: A History', Drug Discovery: A History: 1-468. Soh, M. S., and J. W. Lynch. 2015. 'Selective Modulators of alpha5-Containing GABAA Receptors and their Therapeutic Significance', Curr Drug Targets, 16: 735-46. Spitzmaul, G., et al. 2004. 'Mechanistic contributions of residues in the M1 transmembrane domain of the nicotinic receptor to channel gating', Mol Membr Biol, 21: 39-50. Stratford, C. H., et al. 2014. 'A questionnaire study of equine gastrointestinal parasite control in Scotland', Equine Vet J, 46: 25-31. Strycharz, J. P., et al. 2011. 'Ivermectin acts as a posteclosion nymphicide by reducing blood feeding of human head lice (Anoplura: Pediculidae) that hatched from treated eggs', J Med Entomol, 48: 1174-82. Sur, C., et al. 1998. 'Rat and human hippocampal alpha5 subunit-containing gamma-aminobutyric AcidA receptors have alpha5 beta3 gamma2 pharmacological characteristics', Mol Pharmacol, 54: 928-33. Szabadics, J., et al. 2006. 'Excitatory effect of GABAergic axo-axonic cells in cortical microcircuits', Science, 311: 233-5. Takken, W., and B. G. Knols. 2009. 'Malaria vector control: current and future strategies', Trends Parasitol, 25: 101-4. Taylor, H. R., and B. M. Greene. 1989. 'The status of ivermectin in the treatment of human onchocerciasis', Am J Trop Med Hyg, 41: 460-6. Thomsen, E. K., et al. 2016. 'Efficacy, Safety, and Pharmacokinetics of Coadministered Diethylcarbamazine, Albendazole, and Ivermectin for Treatment of Bancroftian Filariasis', Clin Infect Dis, 62: 334-41. Towers, S. K., et al. 2004. 'Alpha 5 subunit-containing GABAA receptors affect the dynamic range of mouse hippocampal kainate-induced gamma frequency oscillations in vitro', J Physiol, 559: 721-8. Trape, J. F., et al. 2011. 'Malaria morbidity and pyrethroid resistance after the introduction of insecticide-treated bednets and artemisinin-based combination therapies: a longitudinal study', Lancet Infect Dis, 11: 925-32. Van Voorhis, W. C., et al. 2015. 'Profile of William C. Campbell, Satoshi Omura, and Youyou Tu, 2015 Nobel Laureates in Physiology or Medicine', Proc Natl Acad Sci U S A, 112: 15773-6. Vassilatis, D. K., et al. 1997. 'Genetic and biochemical evidence for a novel avermectin-sensitive chloride channel in Caenorhabditis elegans. Isolation and characterization', J Biol Chem, 272: 33167-74. Ventre, E., et al. 2017. 'Topical ivermectin improves allergic skin inflammation', Allergy, 72: 1212-21. Villarroel, Aurora. 2013. 'Internal parasites in sheep and goats'. Vithlani, M., et al. 2011. 'The dynamic modulation of GABA(A) receptor trafficking and its role in regulating the plasticity of inhibitory synapses', Physiol Rev, 91: 1009-22. Wang, X., et al. 2017. 'Mutations on M3 helix of Plutella xylostella glutamate-gated chloride channel confer unequal resistance to abamectin by two different mechanisms', Insect Biochem Mol Biol, 86: 50-57.

Wang, X., et al. 2016. 'A point mutation in the glutamate-gated chloride channel of Plutella xylostella is associated with resistance to abamectin', Insect Mol Biol, 25: 116-25. Washio, H. 2002. 'Glutamate receptors on the somata of dorsal unpaired median neurons in cockroach, Periplaneta americana, thoracic ganglia', Zoolog Sci, 19: 153-62. Webb, T. I., and J. W. Lynch. 2007. 'Molecular pharmacology of the glycine receptor chloride channel', Curr Pharm Des, 13: 2350-67. White, G. B. 1974. 'Anopheles gambiae complex and disease transmission in Africa', Trans R Soc Trop Med Hyg, 68: 278-301. WHO. 2014. 'World Malaria Report'. Williams, D. B., and M. H. Akabas. 1999. 'Gamma-aminobutyric acid increases the water accessibility of M3 membrane-spanning segment residues in gamma-aminobutyric acid type A receptors', Biophys J, 77: 2563-74. Williams, D. B., and M. H. Akabas. 2001. 'Evidence for distinct conformations of the two alpha 1 subunits in diazepam-bound GABA(A) receptors', Neuropharmacology, 41: 539-45. Williams, D. B., and M. H. Akabas. 2002. 'Structural evidence that propofol stabilizes different GABA(A) receptor states at potentiating and activating concentrations', J Neurosci, 22: 7417-24. Wisden, W., et al. 1992. 'The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon', J Neurosci, 12: 1040-62. Wolstenholme, A. J. 2012. 'Glutamate-gated chloride channels', J Biol Chem, 287: 40232-8. Wolstenholme, A. J., et al. 2004. 'Drug resistance in veterinary helminths', Trends Parasitol, 20: 469- 76. Wolstenholme, A. J., and A. T. Rogers. 2005. 'Glutamate-gated chloride channels and the mode of action of the avermectin/milbemycin anthelmintics', Parasitology, 131 Suppl: S85-95. Wooltorton, E. 2003. 'Concerns over lindane treatment for scabies and lice', CMAJ, 168: 1447-8. Xu, N. Z., et al. 2018. 'Negative allosteric modulation of alpha 5-containing GABAA receptors engenders antidepressant-like effects and selectively prevents age-associated hyperactivity in tau-depositing mice', Psychopharmacology (Berl), 235: 1151-61. Zheng, Y., et al. 2002. 'Identification of two novel Drosophila melanogaster histamine-gated chloride channel subunits expressed in the eye', J Biol Chem, 277: 2000-5.

General Methods

2.1 Tissue culture and transfection of HEK 293 cells

HEK AD293 cells (American Type Culture Collection, Manassas, VA, USA) were cultured in 15 mL DMEM (supplemented with 10% FCS and 1% PenStrep), until approximately 90% confluent. The medium was aspirated before the cells were washed with PBS and dissociated using 0.05% trypsin-EDTA. For single channels and macropatches, HEK AD293 cells were seeded onto poly-D-lysine coated glass coverslips and transfected with cDNAs encoding the GluClR subunit avr-14B (pcDNA 3.1+) of H. contortus using a calcium phosphate-DNA co- precipitate method previously described by (Chen and Okayama 1987). The cDNA encoding the CD4 surface antigen was also added to the transfection mixture and acted as a marker of transfected cells. Cells were used for experiments 2-3 days after transfection. The point mutation, TM3-G36’A, was incorporated into the subunit using the QuickChange site-directed mutagenesis method. Successful incorporation of mutation was confirmed by sequencing the mutated DNA.

2.2 Patch clamp electrophysiology – whole cell, single channels and rapid application.

All experiments were carried out at room temperature (21 – 24 oC). Single-channel and macropatch currents were recorded from outside-out excised patches at a clamped potential of −70 mV, unless indicated otherwise. The patches were continuously perfused via a gravity-fed double-barrelled glass tube. Out of one barrel flowed an extracellular bath

solution containing (in mM), 140 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 10 HEPES, and 10 D-glucose and titrated to pH 7.4. The adjacent barrel contained agonist dissolved in this extracellular solution. Glass electrodes were pulled from borosilicate glass (G150F-3; Warner Instruments), coated with a silicone elastomer (Sylgard-184; Dow Corning) and heat-polished to a final tip resistance of 4-15 MΩ when filled with an intracellular solution containing (in mM) 145 CsCl,

2 MgCl2, 2 CaCl2, 10 HEPES, and 5 EGTA, pH 7.4. Stock solutions of L-glutamate were also pH- adjusted to 7.4 with NaOH. Excised patches were directly perfused with extracellular solution by placing them in front of one barrel of the double-barrelled glass tube. Single channel currents were elicited by exposing the patch continuously to agonist containing solution, flowing through the adjacent barrel. Macropatches currents were elicited by lateral translation of the tube from the agonist free to agonist containing barrel using a piezo-electric stepper (Siskiyou). This achieved rapid solution exchange (<1 ms). Currents were recorded

using an Axopatch 200B amplifier (Molecular Devices), filtered at 5 kHz and digitized at 20 kHz using Clampex (pClamp 10 suite, Molecular Devices) via a Digidata 1440A digitizer.

2.3 Isolation of oocytes from Xenopus frogs and expression of DNA

Xenopus frogs were housed at the Australian Institute for Bioengineering and Nanotechnology University of Queensland, Australia. All animal handling procedure were conducted under conditions approved by the University of Queensland Animal Ethics Committee (approval number: QBI/AIBN/087/16/NHMRC/ARC) Adult female frogs were anaesthetized (in 1.3mg/ml MS-222) and oocytes were surgically removed via a small incision to the abdomen. Oocytes were then defolliculated with 1.5 mg/ml collagenase for 2 hours. Free oocytes were rinsed with calcium-free OR-2 solution (82.5 mM NaCl, 2 mM KCl, 1 mM

MgCl2.6H2O, 5 mM HEPES, pH 7.4) and mature oocytes stage V or VI were isolated for experiments. The DNAs coding for α and β HcoGluClRs subunits were nuclear injected into the oocytes at a ratio of 1:0, 0:1, 1:1, 1:50 (αβ), and also α and β AgGluClRs (1:0 and 0:1). Injected oocytes were incubated in ND96 storage solution (96 mM NaCl, 2 mM KCl, 1 mM

MgCl2.6H2O, 1.8 mM CaCl2, 5 mM HEPES, 50 µg/ml gentamicin, 2.5 mM sodium pyruvate, 0.5 mM theophylline, pH 7.4) at 16°C for 2-3 days before experiment.

2.4 Two-electrode voltage clamp (TEVC) using frog oocytes

During the recording, the oocyte was secured in a cell bath that was continually perfused with ND96 recording solution (ND96 storage solution without pyruvate, theophylline and gentamicin). Compounds diluted in ND96 recording solution were applied to the oocyte via bath perfusion. Two microelectrodes, the voltage and current electrodes, each containing 3 M KCl (0.2-2 MΩ resistance), clamped the oocytes, and the membrane voltage was kept constant at −40 mV. The change in glutamate and IVM-induced current was filtered at 200 Hz using an OC-725C Oocyte Clamp amplifier (Warner, Hamden, CT). The current analogs were then digitized by the Digidata 1440A interface and acquired at 2 kHz sampling rate using Clampex 10.2 software (Axon Instruments, Union City, CA). Current amplitudes were measured off-line using Clamp fit 10.2 (Axon Instruments, Union City, CA).

2.5 Heteroculture preparation and IPSC recording

Primary neuronal cultures were prepared from cortices of E18 rat embryos, which were triturated and plated at 100,000 cells per 18-mm poly-D-lysine-coated coverslip in DMEM with 10% fetal bovine serum. After 24 h, the entire medium was replaced with Neurobasal medium that included 2% B27 and 1% GlutaMAX supplements. After one week, half of this medium was replaced and the neurons were allowed to grow in vitro for 3-5 weeks before introducing the transfected HEK293 cells. Recordings of synaptic currents were done in whole-cell configuration (at −70 mV) using an Axopatch 200B amplifier (Molecular Devices), filtered at 5 kHz and digitized at 20 kHz using Clampex (pClamp 10 suite, Molecular Devices) via a Digidata 1440A digitizer.

2.6 Fluorescence recovery after photobleaching (FRAP)

Membrane partitioning and FRAP experiments were performed on an LSM 710 inverted 2- photon confocal microscope equipped with a 40X water immersion objective lens (1.2 NA/280 µM WD/0.208 µM/pixel and a GFP/Alexa 488 filter set. Untransfected HEK293 cells at a confluency of 50-75% were plated onto 35-mm glass bottom dishes one day prior to experiments. The IVM-bdpy fluorophore was dissolved in 100% DMSO and stored at −20 0C at a concentration of 24 mM.

A Mai Tai eHP 2-photon laser/760-1040 nm was used as an excitation source (power at 920 nm) for the FRAP experiments. The laser and bleaching power were set to 7% and 75% respectively. Membrane partitioning images were taken for 35 min at 0.5 Hz after replacing the cell medium (extracellular solution) with one containing the IVM-bdpy probe at a concentration of 500 nM. FRAP images were taken at a frequency of 0.5 Hz for 10 min after a 30 min equilibration period in 500 nM IVM-bdpy.

2.7 Marine extracts library, extraction and isolation A library of around 180 Australian and Antarctic marine invertebrate and algae samples were collected, extracted and stored in the laboratory of Professor Rob Capon (Institute of Molecular Biology, The University of Queensland, Australia) according to the protocol described by (Balansa et al. 2010). When preparing whole extract samples for ToxFinder screening, small aliquots of the stock n-BuOH or H2O partitions previously stored in deep 96- well plates were diluted with DMSO and transferred into new 384-well extract plates. These

plates were then stored at -20oC and thawed for use on the day of the experiment. A portion of the EtOH extracts that were hits from the ToxFinder experiments were further partitioned and the residue fractions were triturated according to the method outlined by (Balansa et al. 2010). These fractions were subsequently prepared in 384-well plates for a second ToxFinder screen in order to discriminate the fractions containing only the active compounds, which were then purified for both ToxFinder and electrophysiological experiments. Results from these screenings are reported in Chapter 6.

2.8 High-throughput yellow fluorescent protein (YFP)-based anion influx assay (ToxFinder)

To prepare for ToxFinder, the crude extracts and fractions (previously dissolved in

DMSO) were diluted with NaCl Ringer’s solution (140 mM NaCl, 5 mM KCl, 1 mM MgCl2.6H2O,

2 mM CaCl2, 10 mM glucose, 10 mM HEPES, pH 7.4) to a final volume of 30 µl. 10 µl of each diluted extract or fraction was next added to the cells in each well to a final volume of 25 µl (cell plate) and incubated for 5-10 minutes prior to the experiment. For individual compound testing, the stock compounds were diluted to 1 µM and 100 µM in a final volume of 25 µL NaCl Ringer’s solution. The medium in each well of the cell plate was then replaced with the diluted compounds and incubated for 5-10 minutes before experiment. 60 µL of GABA solution, dissolved in NaI Ringer’s solution (equimolar concentration of NaI replaced NaCl) was prepared correspondingly in a separate 384-well plate (drug plate).

ToxFinder experiments was conducted using an automated imaging system with a robot-controlled liquid handling system, as described by Kruger (Kruger et al. 2005). This high- throughput assay has the advantage of screening the effect of a drug on thousands of cells at a time. The cell plate was first placed on a motorised stage (Prior ProScan II/III, GT Vision, Hagerstown, MD, USA) of an Olympus IX51/Nikon Eclipse Ti inverted microscope before the cells were imaged with 10x objective, once before and then 8 seconds after the addition of drugs. A mercury/halide arc lamp was used to excite the YFP, producing fluorescence that was magnified through a diopter 8 mineral glass lens and imaged by a monochrome/CCD camera before being digitised onto a personal computer. The maximum image acquisition rate was 1.25 Hz and the images were formed with 696x520 pixels resolution. A 100-µl syringe of the LC PAL auto sampler (CTC Analytics, Zwingen, Switzerland) transferred the solution from drug to cell plate. All hardware control, image acquisition, storage, image analysis and data quantification were performed by LabView 8.2.1 software.

2.9 Automated electrophysiology The patch clamp technique permits single channel and whole cell electrophysiological recordings from small mammalian cells (Hamill et al. 1981). Electrophysiology recordings were performed using a high throughput automated planar patch clamp device developed by Nanion (Nanion Technologies GmbH, Munich, Germany). In addition to increasing throughput, this device permits complete exchange of the solution bathing the cells with 25 μL of new solution. This later feature was particularly important as the Capon lab provided us with only minute amounts of extracts and pure compounds. For Patchliner experiments, the

stable α1β2γ2L and α5β3γ2L GABAAR-HEK293 cells were suspended in the standard external solution containing (in mM): 140 NaCl, 4 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES/NaOH and 5 glucose (pH 7.4). The final cell suspension, the pre-prepared internal solution containing (in mM) 50 KCl, 10 NaCl, 60 KF, 2 MgCl2, 10 HEPES/KOH (pH 7.2), and 20 EGTA, the external solution and the compound-containing solutions were all placed at pre-defined positions on the patchliner workstation and were applied to cells via robotic pipette. The cell suspension was transferred to the cell hotel (i.e. a cell reservoir on the patchliner station) immediately before experiments. Intracellular solution, extracellular solution, seal enhancer, and compound (drug) solutions, were also placed in the allocated positions on the workstation. The 16- channel chip was placed on the chip wagon. Briefly, the protocol comprises the following steps – the robotic pipette first applies internal solution into the chip, and then extracellular solution, the chip then is docked, and the machine then forms an electrical contact between intracellular and extracellular solutions. Then the cell suspension (from the cell hotel) was applied to the external side of the chip. Once a cell has been caught in each channel (i.e., sucked onto the aperture), seal enhancer solution was applied into the chip. Finally, the ‘Patch Control HT’ protocol performed several seal formation and optimisation steps before it established the whole-cell recording configuration. For concentration-response experiments, cells were voltage clamped at –40 mV at room temperature (~24 °C) (Balansa et al. 2010). 100 mM GABA (Ajax Finechem, Seven Hills, NSW, Australia) stocks were prepared in water and kept at -20 ºC. Strychnine (Sigma-Aldrich, St Louis, MO, USA) and other compounds to be tested were dissolved in DMSO (Sigma-Aldrich) and kept at -20oC.

2.10 Data Analysis

A. Two-electrode voltage clamp (TEVC) and whole cell electrophysiology

Glutamate concentration-response was standardized as I% = I/Imax (%) x 100, where I is the peak current amplitude, and Imax is the maximal peak current produced by saturating glutamate in each oocyte. The responses were normalized as I% = I/IEC50 x 100, where IEC50 is the peak current produced by glutamate EC50. The responses of glutamate in the presence of

2+ diazepam and Zn were normalized as ΔI% = I/Icontrol x 100, where I is the peak current amplitude, and Icontrol is the peak current produced by glutamate control (EC20). Concentration-responses were fitted using the four-parameter non-linear regression, I% = bottom + (top-bottom)/ (1 + 10^ ((log (EC50 or IC50) – x)*nH)), where top and bottom refers to the maximal and basal responses respectively. EC50 and IC50 refer to the concentration that gives half the maximal effective or inhibitory response respectively, x refers to the log concentration of the agonist or inhibitor, and nH refers to the Hill coefficient (SigmaPlot 13.0, Systat software, Inc. SigmaPlot for Windows). Unless stated otherwise, the pooled data were contrasted using unpaired t-test where appropriate, and considered significant when p<0.05. All results were expressed as mean ± S.E.M from n≥3 oocyte recordings.

Group data were analysed in SigmaPlot 13.0 using one-way ANOVAs, where p < 0.05 was taken as the significance threshold and expressed as mean ± SEM.

B. Single channels and macropatches

Single channel recordings were analysed in QuB software. Single channel currents were idealised and separated into discrete activations by applying critical shut times of 50 ms in glutamate alone or 50-100 ms in glutamate plus IVM to separate and define single receptor active periods. Patches yielded between 30-250 individual active periods. Critical shut times were determined by generating an initial shut duration histogram to continuous data that included inactive periods corresponding to receptor desensitisation. The selected critical times eliminated periods of receptor desensitisation while retaining singe receptor active periods. Mean active period duration and intra-activation open probability was estimated from each patch. Group means were obtained by averaging across multiple patches for each recording condition.

References

Balansa, W., et al. 2010. 'Ircinialactams: subunit-selective glycine receptor modulators from Australian sponges of the family Irciniidae', Bioorg Med Chem, 18: 2912-9. Chen, C., and H. Okayama. 1987. 'High-efficiency transformation of mammalian cells by plasmid DNA', Mol Cell Biol, 7: 2745-52. Hamill, O. P., et al. 1981. 'Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches', Pflugers Arch, 391: 85-100. Kruger, Wade, et al. 2005. 'A yellow fluorescent protein-based assay for high-throughput screening of glycine and GABAA receptor chloride channels', Neuroscience Letters, 380: 340-45.

GluClR-mediated inhibitory postsynaptic currents reveal targets for IVM and potential

mechanisms of IVM resistance

Background

Although there have been a number of drugs targeting GluClRs in invertebrates over the period of time parasites such as nematodes and arthropods have managed to increase their resistance to all the drugs available on the market. Drug discovery strategy now involves either to find a molecule that can overcome this barrier of resistance or to find one of the possible mechanisms by which parasitic organisms such as H. contortus develop resistance to these drugs. I sought to investigate the pharmacological properties of HcoGluClRs by two-electrode voltage clamp

(TEVC). This is because we hypothesised that one of the HcoGluClR subunit is imparting resistance to the parasitic nematode. This study was undertaken in collaboration with Professor Robert

Capon’s group (Institute of Molecular Biosciences, The University of Queensland). The results of this study are detailed below, and the manuscript has been submitted to PLOS Pathogens.

Statement of contributions by authors

I designed and carried out 80% of all the electrophysiological recordings and data analysis. I also put all the figures together which were later edited by Dr. Angelo Keramidas and edited the first draft.

Dr. Jennifer Smith performed 10% of experiments where she performed single channel recordings and IPSCs recordings. Dr. Argel Estrada-Mondragon performed the membrane accumulation experiment and also helped in FRAP experiment. Dr. Xiao synthesised all the IVM analogues. Dr.

Angela Salim helped us to gather all the supplementary data for all the analogues due to Dr. Xiao’s absence. Professor Joseph Lynch and Dr. Angelo Keramidas advised on the experimental design and made the first draft. All authors edited the subsequent draft equally. Other acknowledgments are included in the manuscript. The table below lists the percentage contribution by the individual authors.

Contributor Statement of contribution

Author Mohammed Atif (Candidate) Electrophysiological recordings and data analysis (80%).

Figure preparation (50%).

Microscopy image experiment & analysis (50%).

Subsequent draft editing (11%)

Author Jennifer Smith Electrophysiological recordings (10%).

Subsequent draft editing (11%)

Author Argel Estrada-Mondragon Microscopy experiments (50%).

Subsequent draft editing (11%)

Author Xiao Synthesis of IVM analogues (100%)

Subsequent draft editing (11%)

Author Angela Salim Pooling data on synthesis of IVM analogues (100%)

Subsequent draft editing (11%)

Author Robert J Capon Lab head that synthesised all the IVM analogues.

Subsequent draft editing (11%)

Author Joseph W. Lynch Writing and editing first draft (50%)

Subsequent draft editing (11%)

Author Angelo Keramidas Electrophysiological recordings and data analysis (10%).

Microscopy image analysis (50%)

Figure preparation (50%)

Writing and editing first draft (50%)

Subsequent draft editing (11%)

GluClR-mediated inhibitory postsynaptic currents reveal targets for IVM and potential mechanisms of IVM resistance

Mohammed Atif1, Jennifer Smith 2, Argel Estrada-Mondragon 1, Xiao X2, Angela Salim 2, Robert J

Capon2, Joseph W. Lynch 1* and Angelo Keramidas1*.

1Queensland Brain Institute and 2Institute for Molecular Bioscience, The University of Queensland, Brisbane,

QLD 4072 Australia.

*Correspondence: Angelo Keramidas, Queensland Brain Institute, The University of Queensland, Brisbane,

QLD 4072 Australia E-mail: [email protected]; Joseph Lynch, Queensland Brain Institute, The

University of Queensland, Brisbane, QLD 4072 Australia E-mail: [email protected].

Abstract Glutamate-gated chloride channel receptors (GluClRs) mediate inhibitory neurotransmission at invertebrate synapses and are primary targets of parasites that impact drastically on agriculture and human health. Ivermectin (IVM) is a broad-spectrum pesticide that binds and potentiates GluClR activity. Resistance to IVM is a major economic and health concern, but the molecular and synaptic mechanisms of resistance are ill-defined. Here we focus on GluClRs of the agricultural endoparasite,

Haemonchus contortus. We demonstrate that IVM potentiates inhibitory input by inducing a tonic current that plateaus over 15 minutes and by enhancing post-synaptic current peak amplitude and decay times. We further demonstrate that IVM greatly enhances the active durations of single receptors. These effects are greatly attenuated when endogenous IVM-insensitive subunits are incorporated into GluClRs, suggesting a mechanism of IVM resistance that does not affect glutamate sensitivity. We discovered functional groups of IVM that contribute to tuning its potency at different isoforms and show that the dominant mode of access of IVM is via the cell membrane to the receptor.

Author Summary

Glutamate-gated chloride channel receptors (GluClRs) mediate chemoelectric inhibition in invertebrate animals, and are targets for broad-spectrum pesticides such as ivermectin. However, resistance to ivermectin threatens the effective control of invertebrates that cause a range of agricultural and human diseases. This study investigates different isoforms of GluClR expressed by the major agricultural endoparasite, Haemonchus contortus, on a synaptic and single receptor level.

We discovered that ivermectin enhances synaptic current amplitude and decay and lengthens single receptor activity. Furthermore, ivermectin is less efficacious at GluClRs that incorporate a naturally ivermectin-resistant subunit, suggesting a potential resistance mechanism. Finally, we identify two chemical interactions between the GluClR and ivermectin that determine its potency and show that ivermectin binds to GluClRs via cell membrane interactions.

Introduction

Glutamate-gated chloride channel receptors (GluClRs) are a major class of pentameric ligand

gated ion channels (pLGICs) (Cully et al. 1994) that mediate neuronal and muscular inhibition

(Marder and Paupardin-Tritsch 1978; Marder and Eisen 1984). They are expressed exclusively in invertebrates, making them ideal targets for the development of pesticides with minimal risk of unwanted activity at vertebrate pLGICs. GluClRs are of interest because they are an excellent model for understanding pLGIC structure at high resolution (Althoff et al. 2014;

Hibbs and Gouaux 2011), function and pharmacology (Lynagh and Lynch 2010; Callau-

Vazquez et al. 2018), and because many invertebrates that express them are commercially important pest species in agriculture (Paini et al. 2016; Fitzpatrick 2013; Tigchelaar et al.

2018), aquaculture (Horsberg 2012) and in veterinary and human health (Wolstenholme

2012).

Ivermectin (IVM) is a highly effective, broad-spectrum anthelminthic drug that targets

GluClRs. Its binding site is formed between the first (M1) and third (M3) transmembrane domains contributed by adjacent subunits (Hibbs and Gouaux 2011) (Fig. 1A, B). The drug is used extensively as an insecticide and anti-parasitic agent (Laing et al. 2017), where its

mechanisms of action include paralysis of feeding structures, such as the pharynx, in adult

organisms (Portillo et al. 2003; Wolstenholme and Rogers 2005; Geary et al. 1993) and impairment of larval viability (Wolstenholme et al. 2016; Laing et al. 2017). However, resistance to IVM and related macrocyclic lactones, such as abamectin, is an emerging concern that threatens food production (Oerke 2005; Fitzpatrick 2013) as well as human

(Osei-Atweneboana et al. 2011) and animal health (Kaplan and Vidyashankar 2012) on a global scale. Underscoring the imperative for a greater understanding of the molecular

targets of pesticides and mechanisms of resistance, studies predict that increasing seasonal

temperatures due to climate warming will significantly exacerbate agricultural loss that

results from an increase in populations of herbivorous pests (Tigchelaar et al. 2018; Bebber

et al. 2013) and veterinary parasites (Fitzpatrick 2013). The chemical synthesis of new IVM derivatives is a key strategy in overcoming IVM resistance as it has the potential to probe specific molecular groups of IVM (Michael et al. 2001) and their interactions with IVM sensitive and resistant forms of GluClR.

Wild, IVM-resistant pest isolates that express missense point mutations to GluClR

subunits have been identified in the herbivorous species, Plutella xylostella (Wang et al. 2017;

Wang et al. 2016) (A30’V) and Tetranychus urticae (Dermauw et al. 2012; Kwon et al. 2010;

Mermans et al. 2017) (G36’D, G36’E) and in a laboratory-induced mutation to a GluClR subunit of Drosophila melanogaster (P23’S) upon exposure to IVM (Kane et al. 2000) (Fig. 1B). Point mutations to GluClR subunits are also being reported in human head lice, Pediculus humanus, that include the A55’V, and S46P and H272R, the latter two mutations being located in the extracellular and intracellular domains of the receptor, respectively (Amanzougaghene et al.

2018).

In addition to mutations, the sensitivity of GluClRs to IVM is affected by subunit composition. Cully et al. 1994 first identified GluClR subunit isoforms in the nematode,

Caenorhabditis elegans and demonstrated that they could assemble as homo- and heteromeric receptors with different pharmacological properties (Cully et al. 1994). C. elegans expresses a β subunit that is insensitive to IVM when expressed as a homopentamer, and acts to reduce IVM sensitivity when co-assembled with the IVM-sensitive α subunit (Cully et al. 1994) (Fig. 1B). A naturally IVM-resistant variant of an α subunit isoform has also been

discovered in a strain of C. elegans. IVM insensitivity can be overcome by overexpressing IVM-

sensitive subunits in the resistant C. elegans strain (Ghosh et al. 2012). Moreover, IVM

effectively kills the larvae of the human pathogens, Onchocerca volvulus and Wuchereria

bancrofti, but is less effective against the adult stages of these organisms (Laing et al. 2017),

suggesting a developmental switch in GluClR subunit composition. Evidence for selection

pressure on GluClR isoform expression that confers IVM insensitivity comes from IVM-

resistant isolates of the bovine pathogens, Cooperia oncophora and Ostertagia ostertagi,

which demonstrate transcriptional downregulation of genes that encode IVM-sensitive

GluClR subunits (El-Abdellati et al. 2011).

H. contortus is one of the world’s most economically important agricultural parasites

that infects domestic ruminant animals, and field resistance to IVM is well-documented (Laing et al. 2016). It has also been reported to infect wild ruminants, including species that are critically endangered (Rose et al. 2014). This endoparasitic nematode has six GluClR genes encoding at least eight subunits (Williamson et al. 2011). Functional expression of wild-type

homomeric GluClRs comprising α (avr-14b) subunits have been reported previously, including

mutations to this subunit that greatly reduce IVM sensitivity (Lynagh and Lynch 2010; Atif et

al. 2017; McCavera et al. 2009). Although the β subunit of H. contortus was shown to have

overlapping distribution patterns with the α (avr-14b) and other α subunit isoforms (Portillo

et al. 2003), little is known about its function. Notably, it is not known if the β subunit forms

functional homomeric receptors or whether it can combine with other subunits to form

heteromers. It also remains to be determined which homomeric or heteromeric combinations

of α and β subunits cluster at postsynaptic sites to mediate inhibitory postsynaptic currents

(IPSCs). This deficiency represents an obstacle to understanding basic invertebrate

57 neurobiology and determining the effects of drugs on neuronal and muscular inhibitory input, including highly lipophilic drugs, such as IVM that exhibit quasi-irreversible effects on GluClRs

(Lynagh and Lynch 2012). Due to its lipophilic nature, IVM is believed to diffuse through the body cuticle of nematodes (Alvarez et al. 2007) and partition into cell membranes of target organisms where it reaches a high local concentration (Martin and Kusel 1992). We have recently shown that single receptor active periods of homomeric α (avr-14b) GluClRs (α

GluClRs) increase in duration over 1-2 minutes after exposure to IVM, suggesting that the drug equilibrates in the membrane over time to produce maximum receptor potentiation

(Atif et al. 2017).

In this study, our aims were to determine (1) whether the β subunit of H. contortus can assemble as both β homomeric and αβ heteromeric GluClRs (β GluClRs and αβ GluClRs),

(2) whether subunit composition determines IVM sensitivity, (3) the potency of synthetically modified analogues of IVM (4) whether β and αβ GluClRs mediate IPSCs with different properties, and finally, (5) whether the membrane partitioning and diffusion properties of

IVM correspond to the time-course of current potentiation.

Results

GluClRs comprising homo- and heteromeric combinations of α and β subunits of H.

contortus. To test for functional expression of β and αβ GluClRs, cDNA encoding the β subunit

was injected into oocytes either alone or with cDNA encoding the α subunit, at a ratio of (α:β)

1:1 or 1:50. Example glutamate-gated currents obtained from oocytes injected with the β

subunit alone and the α and β subunits at a ratio of 1:50 are shown in Fig. 1C. These data

clearly demonstrate that the β subunit can indeed form functional β GluClRs and, given the

differential glutamate sensitivity of co-injected oocytes, can assemble with the α subunit to

form αβ GluClRs. The maximal currents for the α, αβ and β GluClRs were (in µA) 3.7 ± 0.8, 2.0

± 0.4 (1:1), 2.2 ± 0.4 (1:50) and 1.9 ± 0.3, respectively and were not statistically different.

Complete glutamate concentration-response experiments were done in oocytes injected with the either cDNA encoding the α or β subunit and in oocytes co-injected with both cDNAs at ratios of 1:1 and 1:50 (Fig. 1D). These data show the β GluClRs are substantially less sensitive to glutamate with an EC50 of 394 µM compared to α homomers and the αβ heteromers (p <

0.001), all of which exhibited similar glutamate sensitivities, with EC50s of 28 µM (α

homomers), 40 µM (αβ, 1:1) and 44 µM (αβ, 1:50) (Table 1). Notably, the fitted

concentration-response plots for both heteromeric combinations of receptors did not exhibit

inflections typical of mixtures of distinct receptor stoichiometries with differential EC50s, as has been shown for GABA-gated pLGICs (Hartiadi et al. 2016). These data suggest that co-

injected oocytes express mostly homogeneous populations of heteromeric receptors, and

that a given injection ratio gives rise to a particular homogeneous stoichiometry. (Refer to appendices section 1 for additional experiments that were done to determine why β

HcoGluClRs were insensitive to IVM).

Table 1. Glutamate and IVM concentration-response parameters

Ligand GluClR EC50 (nM) Hill co-efficient n

α 28 ± 6 1.9 ± 0.3 10

αβ (1:1) 40 ± 6 1.7 ± 0.1 6 Glutamate αβ (1:50) 44 ± 3 1.7 ± 0.1 13

β 394 ± 57*** 1.5 ± 0.1 8

α 22 ± 3 1.9 ± 0.1 11

αβ (1:1) 86 ± 14**# 1.0 ± 0.1 6 IVM αβ (1:50) 141 ± 11*** 1.6 ± 0.1 14

β >10 µM − 6

One-way ANOVA **p < 0.01, ***p < 0.001 compared to α GluClRs, # p < 0.05 compared to

αβ (1:50) GluClRs. cDNA injection ratios in parentheses.

IVM concentration-response experiments were carried out to obtain additional evidence of populations of receptors that were a function of the oocyte injection ratio.

Example IVM-induced currents for α GluClRs and αβ GluClRs at both injection ratios are shown in Fig. 2A-C. The corresponding group concentration-response data for four combinations of GluClRs are shown in Fig. 2D. These plots indicate that α GluClRs are highly sensitive to IVM with an EC50 of 22 nM and a maximal response equivalent to 59% of that seen with saturating glutamate. Conversely, β GluClRs were effectively insensitive to IVM with an EC50 > 10 µm. The heteromeric combinations exhibited intermediate IVM sensitivities with

the 1:1 injected oocytes being more sensitive (EC50 of 86 nM, 26% of saturating glutamate

response) than the 1:50 injected oocytes (EC50 of 141 nM, 33% of saturating glutamate

response). The % of saturating glutamate response was not significantly different between

the two injection ratios. However, the EC50s for the two heteromeric combinations were

statistically different from the α GluClRs and from each other (Table 1).

Together these data demonstrate that β subunits of H. contortus and C. elegans (Cully et al. 1994) exhibit some functional similarities. They both form homomeric receptors that are insensitive to IVM and can combine with α subunits to reduce IVM sensitivity. However, unlike the β GluClRs of C. elegans, those of H. contortus are responsive to glutamate, albeit with a reduced sensitivity. The data also show that the β subunit can co-assemble with the α

subunit without reducing the sensitivity of the receptors to the neurotransmitter, glutamate.

Furthermore, IVM potency is reduced when oocytes are injected with an excess of β subunit

cDNA, implying that GluClRs with a greater content of β subunit are less sensitive to IVM. This

later inference also suggests that heteromeric αβ GluClRs of H. contortus can increase the

proportion of β subunit as a function of its expression level. This is in contrast to the reported

fixed stoichiometry of heteromeric GluClRs of C. elegans (Degani-Katzav et al. 2016).

IVM analogue potency at α and αβ GluClRs. In our next series of experiments, we wished to

see if modifying the IVM molecule might alter its potency differentially at α and αβ GluClRs.

Three analogues were screened (Fig. 3A) at α and αβ GluClRs (1:50) using a standard concentration of 30 nM (Fig. 3B). Changes to the IVM parent molecule were made on the basis of IVM and receptor or lipid membrane interactions as revealed by the crystallographic structure of the C. elegans α GluClR (Hibbs and Gouaux 2011).

The maximum current elicited by each IVM analogue was normalised to that induced by a saturating concentration of glutamate (5 mM, Fig. 3B) and its efficacy (Ratio of saturating

IVM current relative to saturating glutamate current) was compared to that of IVM and between α and αβ GluClRs (Fig. 3C). The disaccharide moiety is predicted to protrude into the upper leaflet of the cell membrane (Fig. 1A) and to form van der Waal (VDW) interactions with the loop that links the M2 and M3 domains of the α subunit of C. elegans (Hibbs and

Gouaux 2011) and H. contortus. Hydrolysis of IVM yielded the aglycone IVM-1 (synthesis details are given in Supporting Information, S1). IVM-1 retained differential efficacy between the two receptor isoforms but was significantly less efficacious than IVM only at α GluClRs (p

< 0.05). Replacing the disaccharide with a methoxymethyl ether group (MOMO) produced

IVM-2 (Supporting Information, S1). This derivative restored efficacy at α but not αβ GluClRs and abolished the differential sensitivity between the receptors. This result suggests that in addition to the predicted M2-M3 loop interactions, steric factors also contribute to the potentiating effects of IVM. IVM-2 (and IVM and IVM-1), with a β-configured 5-OH moiety, is predicted to form VDW interactions and hydrogen bonds with residues of the M2 and M1 domains (Hibbs and Gouaux 2011). The α-configured 5-OH epimer (IVM-3, Supporting

Information, S1), prepared from IVM-2, produced the most marked reduction in efficacy (p <

0.001 for both GluClRs) relative to IVM. Moreover, this change also ablated the differential efficacy between the two GluClR isoforms (Fig. 3C). This result is in accord with the predicted role of the 5–OH moiety, which forms a hydrogen bond with the polar residues within the pore-lining M2 domain (S15’ in C. elegans α (Hibbs and Gouaux 2011) or S16’ and Q15’in H. contortus α and β, respectively (Fig. 1B). A previous study also found this position was important, as subtle changes to 5-O or 5-NOH resulted in dramatic decreases in activity when tested in an H. contortus larval assay (Michael et al. 2001).

Overall, the maximal efficacies of α and αβ GluClRs were maintained for IVM and IVM-

1 whereas IVM-2 and IVM-3 exhibited similar maximal efficacies at both receptors. The

preservation of efficacy of IVM-1 and IVM-2 compared to IVM at αβ GluClRs suggests that the

β subunit likely contributes to IVM binding sites. The disaccharide and the 5-OH moiety of

IVM are salient determinants of the potentiating potency of IVM. The data reveal two IVM-

receptor interactions that accord well with predictions from structural studies (Hibbs and

Gouaux 2011) and suggest that with rational drug design it may be possible to develop IVM

analogues to selectively target specific GluClR isoforms.

IPSCs mediated by α and αβ (1:1) GluClRs. Evidence suggesting the existence of glutamate-

gated anion-selective post-synaptic receptors was first noted in arthropods over 40 years ago

via observing membrane potential changes (Marder and Paupardin-Tritsch 1978; Cull-Candy

1976; Marder and Eisen 1984). However, to our knowledge direct recordings of GluClR-

mediated IPSCs have not been made in any invertebrate species. We have recently developed

a heteroculture consisting of primary (mammalian) neurons and HEK293 cells transfected

with defined post-synaptic receptors of vertebrates, along with the trans-synaptic protein

neuroligin (Dixon et al. 2014; Chen et al. 2017; Zhang et al. 2015).

This heterosynapse preparation was used to examine if α and αβ GluClRs of H. contortus cluster at post-synaptic sites and respond to synaptically released glutamate. Cells transfected with the α (Fig. 4A) or α and β (1:1, Fig. 4B) subunits in co-culture with neurons

exhibited prominent currents typical of those expressed at vertebrate synapses, consisting of

a fast rise to peak followed by a slower, exponential decay back to baseline (Fig. 4C, D). These

IPSCs were only observed in transfected cells and were blocked by picrotoxin (Fig. 4E),

confirming that the receptors were anion-selective GluClRs (Hibbs and Gouaux 2011;

McCavera et al. 2009). These data clearly demonstrate that α and αβ GluClRs of H. contortus generate classic, picrotoxin-sensitive IPSCs. A clear, observable difference between IPSCs mediated by the two receptor isoforms was that αβ GluClRs decayed faster than those

mediated by α GluClRs (Fig. 4C, D). The mean rise-times, peak amplitudes and decay times

were measured and plotted as bar plots (Fig. 4F-H). The group averages confirmed that only

the decay times between the two receptor isoforms were significantly different, with those

of the α GluClRs decaying with a time constant of 40 ± 2 ms (n = 17), whereas the decay time

constant for the αβ GluClRs was 17 ± 2 ms (n = 8). The rise times were 4.6 ± 0.7 ms for the α

GluClRs and 3.2 ± 0.3 ms for the αβ GluClRs. The peak amplitudes were the most variable of the three parameters, being 60 ± 9 pA and 34 ± 8 pA for the α and αβ GluClRs, respectively, likely reflecting variable expression levels of receptors.

The effects of IVM at IPSCs of α and αβ (1:1) GluClRs. After validating the heterocultures as a reliable preparation for investigating IPSCs mediated by GluClRs, we made synaptic recordings while continuously applying IVM (5 nM) to cells similarly transfected with α and

αβ GluClRs. Two salient effects were observed in these recordings. Firstly, there was a steady downward deflection in baseline current over the course of the recording that was not different between cells expressing α (Fig. 5A) and αβ (1:1, Fig. 5B) GluClRs. These data were

subsequently pooled for analysis of the tonic current component. The tonic currents

plateaued at 17 ± 1 minutes (n = 13), had a fitted time constant of 483 ± 72 s and a mean

amplitude of 304 ± 54 pA. Changes to the kinetic properties of IPSCs were also observed upon

IVM application (Fig. 5C-G). These were analysed 15 minutes after the commencement of the

IVM application to ensure that the effects had equilibrated. For both receptor types, the

decay and rise times and peak amplitudes increased significantly in the presence of IVM

relative to IVM naïve cells. The decay time constant for α GluClRs slowed from 40 ± 2 ms to

100 ± 3 ms (n = 6) and for αβ GluClRs from 17 ± 2 ms to 72 ± 4 ms (n = 6, Fig. 5E). The increase

in activation times were from 4.6 ± 0.7 ms to 19 ± 1 ms for α GluClRs and from 3.0 ± 0.3 ms

to 10 ± 1 ms for the αβ heteromers (Fig. 5F). Again, the peak IPSC amplitudes were variable

but increased significantly in the presence of IVM from 60 ± 9 pA to 122 ± 39 pA and from 34

± 8 pA to 124 ± 18 pA for α and αβ GluClRs, respectively (Fig. 5G). A comparison of the magnitude changes of IPSC parameters was also made between the two receptor isoforms.

Consistently greater decay and rise times were observed for the α GluClRs (Fig. 5E, F), suggesting that IPSCs mediated by these receptors produce greater inhibitory input than αβ

GluClRs in the presence of IVM.

Single receptor properties of β and αβ (1:1) GluClRs. The decay times of IPSCs mediated by

αβ GluClRs were consistently and significantly faster than those of α GluClRs, in both the

absence and presence of IVM. We previously showed that the decay times of macropatch

currents mediated by α GluClRs incorporating an IVM-insensitive mutation (G36’A) were

faster than those mediated by wild-type α GluClRs, and that this was due to briefer single receptor active periods and enhanced receptor desensitisation (Atif et al. 2017).

To explore whether a similar mechanism applies to β-containing GluClRs, we recorded

single receptor currents in excised patches from cells transfected with either the α and β

subunits (1:1) or the β subunit alone. These experiments were recorded in 2 µM and 3 mM

glutamate as well as 2 µM glutamate plus 5 nM IVM at a clamped potential of −70 mV

(reversal potential = 4.0 mV, liquid junction potential = 4.7 mV and membrane potential =

78.7 mV) (Atif et al. 2017). Patches from cells transfected with both subunits revealed that

the majority (~90%) of single αβ receptors opened to an amplitude of 1.2 pA (Fig. 6A) with

the remainder opening to 0.7 pA (Fig. 6B). By contrast, patches from cells transfected with

the β subunit alone invariably opened to an amplitude of 0.4 pA (Fig. 6C). Amplitude histograms confirmed two amplitude levels for αβ GluClRs and a single amplitude for the β

GluClRs (Fig. 6D). Our data suggest that a 1:1 transfection ratio of α and β subunits results in

a predominant stoichiometry, which exhibits an amplitude of 1.2 pA and a less frequent

subunit combination that opens to an amplitude of 0.7 pA. We infer that the αβ GluClRs that

open to 0.7 pA likely contain a greater proportion of β subunits. This is consistent with β

GluClRs having the smallest amplitude and α GluClRs having the greatest amplitude (1.8 pA)

under similar recording conditions (Atif et al. 2017). The calculated conductance values for

the corresponding amplitudes for αβ GluClRs was 15.2 pS and 8.9 pS and for the β GluClRs

was 5.1 pS.

Single receptor active periods were analysed for mean duration and intra-activation

open probability (PO). For αβ GluClRs the activations from both stoichiometries were pooled

to obtain mean active periods of 43 ± 12 ms (n = 5) in 2 µM glutamate and 152 ± 27 ms (n =

5) in 3 mM glutamate. Consistent with the example recordings presented in Fig. 6A, B the

respective POs were 0.49 ± 0.04 and 0.87 ± 0.03 in 2 µM and 3 mM glutamate (Fig. 6E). Current

potentiation by IVM did not change the proportion of activations to 1.2 pA compared to 0.7

pA, but manifested as a marked increase in the mean duration of the active periods (pooled,

876 ± 178 ms, n = 6) (Fig. 6A, B, E). This represents an over 20-fold increase compared to 2

µM glutamate alone and 6-fold increase compared to 3 mM glutamate. The PO in 2 µM

glutamate plus 5 nM IVM was similar to that in 3 mM glutamate, being 0.83 ± 0.02 (Fig. 6E).

β GluClRs exhibited the briefest active periods (Fig. 6C). At 2 µM glutamate, β GluClRs opened for a mean duration of 5.5 ± 2.0 ms and had a PO of 0.56 ± 0.09 (n = 5) (Fig. 6F). The active

durations increased in 3 mM glutamate to 164 ± 40 ms, whereas the PO remained unchanged

(0.67 ± 0.14, n = 5) (Fig. 6F). In contrast to the supersaturating effects IVM has at α and αβ

GluClRs, it was ineffective at increasing the duration of active periods of β GluClRs beyond

that measured in 3 mM glutamate. The active periods in 2 µM glutamate plus 5 nM IVM were

126 ± 26 ms in duration and the PO was 0.65 ± 0.04 (n = 5) (Fig. 6F).

Our single receptor measurements show that by increasing the active durations and

PO of single GluClRs, IVM produces an increase in IPSC decay times and peak amplitude,

respectively. Our data also demonstrate a correlation between single receptor active

duration, IPSC decay times and IVM potentiation of IPSCs.

Partitioning and diffusion of IVM in cell membranes. Our IPSC and single channel data

suggest that the potentiating effect of IVM at GluClRs requires several minutes to stabilise

(Atif et al. 2017). This relatively long delay could reflect the slow binding of IVM directly from

the aqueous solution or slow IVM-induced conformational effects at GluClRs once bound

(Sykes et al. 2014; Degani-Katzav et al. 2017). Alternatively, it may reflect a slow accumulation

rate of IVM into the membrane, possibly coupled with a slow lateral diffusion rate within the

membrane that controls its access to receptor binding sites. To help distinguish between

these possibilities, we synthesised a fluorescent IVM-bodipy probe (IVM-bdpy, Fig. 7A,

Supporting Information, S1) to monitor the interaction of IVM with the cell membrane. This

probe consisted of IVM-1, which retains its activity at GluClRs (Fig. 3), attached to a bodipy

fluorophore. The resulting IVM-bdpy derivative exhibited spectral emission properties similar

to free bdpy (Fig. 7B) and was active at α GluClRs (Fig. 7C), albeit with a reduced potency (Fig.

7D).

Two types of experiments were conducted with IVM-bdpy. The first was aimed at monitoring the time-course of membrane accumulation. Untransfected HEK293 cells were incubated in extracellular solution containing 500 nM IVM-bdpy and time-lapse images were taken over a period of 30 min (Fig. 8A). The change in total membrane fluorescence averaged from 5 repeat experiments was then plotted against time and fitted to a standard exponential function (Fig. 8B). In a control experiment, the corresponding accumulation rate was also determined for bdpy. To obtain the net membrane partitioning time-course for IVM-1, the data for the bdpy alone was subtracted from the data obtained for IVM-bdpy (Fig. 8B). The time constants for the three plots were 6.5 min for the IVM-bdpy, 8.2 min for the bdpy fluorophore and 6.0 min for IVM-1. As the lipophilic properties of IVM and IVM-1 are similar

(logPs of 5.4 and 5.1, respectively), we infer this analysis quantitatively reflects the membrane partitioning properties of IVM. Notably, the resultant plot plateaued at about 18-20 minutes, which was similar to the estimated time to plateau for the IVM-activated current as presented in Fig. 5A and B. It should be noted, however, that the IVM-bdpy experiment was done using a 100-fold higher concentration of drug in un-transfected cells and represents passive membrane accumulation, whereas the current plateau involved IVM binding to and activating receptors. Nevertheless, these data suggest that membrane partitioning of IVM is the rate limiting factor controlling the activation rate of GluClRs.

The second experiment was aimed at determining the lateral membrane diffusion properties of IVM using the fluorescence recovery after photobleaching (FRAP) technique.

After a 30-min equilibration time, 5 cells were monitored for fluorescence recovery within a bleached circular patch of membrane of 3.7 µm diameter (Fig. 8C). Any bleaching and recovery of the surrounding area of the cell was measured, using a similar circular patch and used to correct the recovery rate of the membrane delineated by the bleached area. We also

noted that the recovery reached 98 ± 2% of control. The resultant plot, fitted to a standard

exponential, is shown in Fig. 8D. The time constant (τ) for recovery of 1.3 minutes was used to calculate the diffusion coefficient (D), using the radius (r) of the bleached patch, according to D = r2/4τ. The calculated value of D was 1.1 x 10−2 µm2s−1. Over the course of 1 s, the IVM-

bdpy probe would traverse 0.2 µm along the membrane surface (lateral diffusion rate of 0.2

µms−1). This value is comparable to that calculated using a similar probe that was used to

measure partitioning and lateral diffusion in muscle membranes of A. suum (Martin and Kusel

1992). We infer that this slow lateral diffusion rate coupled to membrane partitioning are the

major contributors to local increases of IVM around its target binding sites and the slow quasi-

irreversible actions of IVM at GluClRs.

Discussion

Resistance to IVM in pest species is a growing problem worldwide. Previous research efforts

have identified mutations and allelles of GluClRs associated with resistance to IVM, however,

there is a lack of information regarding the mechanisms that underlie resistance. In this study,

we investigated the synaptic and biophysical properties of different GluClR subtypes and

detailed a possible mode of resistance in the pest species H. contortus.

We found that β subunits of H. contortus can assemble both as homomers and as heteromeric combinations with the α subunit. Given that both subunits have common distribution patterns in motor neuron commissures and nerve cords (Portillo et al. 2003), it is possible that these two subunits combine to form heteromeric receptors, which are less sensitive to IVM but retain high sensitivity to neurotransmitter. Our data also support the idea that the subunit stoichiometry of heteromeric GluClRs is variable and that receptors with a greater proportion of β subunits are less sensitive to IVM. With a high EC50 (400 µM) for

glutamate, a low unitary conductance and brief active periods, homomeric β GluClRs are not

likely to be efficacious post-synaptic receptors. Moreover, given that our single channel recordings reveal little homomeric β GluClR activity in patches expressing both α and β

subunits, homomeric β GluClRs are likely to be scarce in cells that also express α subunits.

Using heterocultures of neurons and transfected HEK cells, we demonstrate that α

and αβ GluClRs mediate IPSCs in response to presynaptic glutamate release. HEK293-

neuronal co-cultures have been shown to recapitulate synaptic current profiles in excitatory

glutamatergic synapses expressing NMDA or AMPA receptors (Fu et al. 2003) suggesting that

glutamate release in heterosynapses is likely to be physiological. To our knowledge, no

published reports are available describing the properties of IPSCs mediated by native GluClRs

in invertebrate neurons or muscle cells. Our preparation therefore represents a technological

advance with the potential to improve our understanding of (1) the kinetics of IPSCs mediated

by defined GluClR isoforms, (2) the effects of drugs on IPSCs mediated by defined GluClR

isoforms, (3) the effect of posttranslational modifications (e.g., phosphorylation) and

resistance mutations on the formation and function of synapses, and (4) synaptogenesis and

synaptic clustering mechanisms. We observed a substantial increase in IPSC peak amplitude

and a general slowing of IPSC kinetics, particularly the decay times. IPSC experiments also

revealed a tonic inhibitory component, which is a well-studied mode of inhibition in

vertebrates that may also involve extrasynaptic pLGICs (Jia et al. 2008). In the presence of

IVM we demonstrate that whereas the tonic component of inhibition is similar in cells

expressing α and αβ GluClRs, the phasic IPSC component is potentiated to a smaller degree

in αβ GluClRs. This lower sensitivity is mostly due to the intrinsic activation properties of β-

containing receptors. Hence, the net inhibitory signal mediated by αβ GluClRs in the presence of IVM would be less than that of α GluClRs, thereby underlying a possible mechanism of IVM resistance.

Our single channel recordings showed that β-containing GluClRs had briefer active periods and a smaller conductance, both of which render receptors less sensitive to the potentiating effects of IVM when compared to α GluClRs (Atif et al. 2017).

Single receptor active durations also accord well with ensemble decay times as revealed in macropatches (Atif et al. 2017) and IPSCs (this study). GluClRs with faster decay times are potentiated to a lesser extent by IVM in absolute terms and thus produce less inhibitory input to target cells. The decay times for the α GluClRs were 67 ms (macropatch)

(Atif et al. 2017) and 40 ms (IPSCs). For GluClRs that exhibited reduced IVM sensitivity the

decay times were 11 ms (macropatch) for α(G36’A) (Atif et al. 2017) and 17 ms (IPSCs) for the

αβ GluClRs. It is also noteworthy that vertebrate pLGICs, such as glycine and GABA-gated receptors that are much less sensitive to IVM also exhibit relatively brief single receptor active periods and faster ensemble decay times (Dixon et al. 2014; Scott et al. 2015).

We hypothesised that a noteworthy component of the mode of action of IVM was its

interactions with cell membranes. A fluorescently tagged IVM (IVM-bdpy) was used to

measure the time-course of drug accumulation and lateral diffusion in membranes that was

compared to the onset of current potentiation. Single channel currents recorded from small,

~2 µM diameter patches equilibrated over 1-2 minutes (Atif et al. 2017). IVM-bdpy required

about 18 minutes to saturate whole cell membranes, which was similar to the time taken for

baseline currents to plateau (17 minutes) in our IPSC recordings. In addition to the

concentrating effects of membrane partitioning, a slow lateral diffusion rate (0.2 µms−1) of

IVM would also contribute to the apparent association rates of IVM to binding sites at GluClRs

(Sykes et al. 2014). The recovery after bleaching reached near control levels, but was not at

100 % for each cell, suggesting the presence of membrane compartments of relatively immobile lipid rafts (Lingwood and Simons 2010). Our experiments suggest that the dominant

pathway for IVM to reach its binding sites at GluClRs is by membrane partitioning and

diffusion rather than inducing slow conformational rearrangements to the receptors after

binding directly from the aqueous compartment.

Finally, we sought to identify a new IVM analogue with a differential effect on α and αβ

GluClRs as proof of principle that this pharmacophore may be useful in refining anthelminthic

treatments by targeting particular GluClR isoforms. We identified key receptor-drug

interactions that determine drug potency that include (1) the disaccharide group of IVM and

(2) the significance of the C-5 configuration of IVM, which is a key determinant of differential potency between GluClR isoforms.

In summary, we describe the functional properties of α homomeric and αβ heteromeric GluclRs using conventional whole cell recording, single channel analysis and heterosynaptic analysis. Our key findings support the general inference that the intrinsic activation properties of the inhibitory pLGICs is a critical determinant and predictor of the potentiating potency of IVM. Informed by our findings, drug design strategy may be directed towards increasing the active durations of single receptors and slowing decay times of IPSCs in synaptic isoforms of GluClRs.

Materials and Methods

Molecular biology. cDNAs encoding the α (avr-14b) (pcDNA 3.1+) or β (pUNIV) GluClR

subunits of H. contortus were nuclear injected into oocytes (NASCO, WI USA) or transfected

into HEK293AD cells (CellBank Australia) using a calcium phosphate-DNA co-precipitate

method. Synapse formation between neurons and HEK293 cells was promoted by co-

transfecting neuroligin 2A or 1B and the cDNA encoding CD4 surface antigen was also added

to the transfection mixture so as to identify transfected cells The. Oocytes and HEK cells were

used for experiments 2-3 days after the introduction of the cDNAs. All experiments were done

at room temperature (22 ± 1 0C).

Oocyte preparation and two-electrode voltage-clamp. Xenopus laevis oocytes were

harvested by surgical incision. The oocytes were then defolliculated with 1.5 mg/ml

collagenase for 2 hours. Free oocytes were rinsed with calcium-free OR-2 solution containing

(in mM) 82.5 NaCl, 2 KCl, 1 MgCl2, 5 HEPES, pH 7.4 and mature stage V or VI oocytes were

isolated for experiments. The DNAs coding for α and β subunits were nuclear injected into the oocytes using a Nanoliter 2000 microinjector (WPI Inc) at a ratio of (α:β), 1:0, 0:1, 1:1 and

1:50. The total amount of injected DNA in all combinations was 400 ng ml−1, including the

1:50 (α:β) combination (8 ng of DNA encoding the α and 392 ng of DNA encoding β subunits).

Injected oocytes were incubated in ND96 storage solution (in mM) 96 NaCl, 2 KCl, 1

MgCl2.6H2O, 1.8 CaCl2, 5 HEPES, 50 μg ml−1 gentamicin, 2.5 sodium pyruvate, 0.5 theophylline, pH 7.4 at 16°C for 2-3 days before experiment.

For the two-electrode voltage clamp recordings, each oocyte was secured in a cell bath

that was continually perfused with ND96 recording solution (ND96 storage solution without

pyruvate, theophylline and gentamicin). Glutamate, IVM and IVM analogues were diluted in

ND96 recording solution and were applied to the oocyte via bath perfusion. The two microelectrodes contained 3 M KCl and had resistances of 0.2-2 MΩ. Recordings were done using Clampex 10.2 software (Molecular Devices) at a clamped voltage of −40 mV. Currents were low-pass filtered at 200 Hz, sampled at 2 kHz using a Gene Clamp 500B amplifier and digitised by a Digidata 1440A interface. The IVM (IVM B1a) concentration-response experiments were standardised by using an application protocol that consisted of applying increasing concentrations of IVM for, respectively, 3 min, 3 min, 3 min, 2 min, 1 min, 1 min and 0.5 min. The IVM-induced currents were normalised to a saturating glutamate concentration of 5 mM (see example, Fig. 2).

Single channel recordings. Single-channel currents were recorded from outside-out excised patches at a clamped potential of −70 mV. The single channel conductance was calculated by dividing the mean current by the net driving force at a clamped potential of 70 mV, after accounting for reversal and liquid junction potentials (Atif et al. 2017). Excised patches were continuously perfused via a gravity-fed double-barrelled glass tube, which contained extracellular bath solution containing (in mM), 140 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 10 HEPES, and 10 D-glucose and titrated to pH 7.4. Glutamate and IVM were dissolved in this extracellular solution. Recording electrodes were pulled from borosilicate glass (G150F-3;

Warner Instruments), coated with a silicone elastomer (Sylgard-184; Dow Corning) and heat- polished to a final tip resistance of 4−15 MΩ when filled with an intracellular solution containing (in mM) 145 CsCl, 2 MgCl2, 2 CaCl2, 10 HEPES, and 5 EGTA, pH 7.4. Stock solutions of L-glutamate were also pH-adjusted to 7.4 with NaOH. A 10 mM stock of IVM (Sigma-Aldrich) was dissolved in 100% DMSO and kept frozen at −20oC. Fresh working stocks of IVM at 5 nM were prepared by dissolving the appropriate quantity directly in extracellular solution. 100%

DMSO when dissolved in extracellular solution alone at the same concentration as was

75 present in working solutions containing 5 nM IVM had no effect on patches excised from cells transfected with GluClRs or from untransfected cells.

Currents were recorded using an Axopatch 200B amplifier (Molecular Devices), filtered at 5 kHz and digitized at 20 kHz using Clampex (pClamp 10 suite, Molecular Devices) via a Digidata

1440A digitizer.

Heteroculture preparation and IPSC recording. Primary neuronal cultures were prepared from cortices of E18 rat embryos, (University of Queensland Biological Services), which were triturated and plated at 100,000 cells per 18-mm poly-D-lysine-coated coverslip in DMEM with 10% fetal bovine serum. After 24 h the entire medium was replaced with Neurobasal medium that included 2% B27 and 1% GlutaMAX supplements. After one week half of this medium was replaced and the neurons were allowed to grow in vitro for 3-5 weeks before introducing the transfected HEK293 cells. Recordings of synaptic currents were done in whole-cell configuration at −70 mV using an Axopatch 200B amplifier (Molecular Devices), filtered at 5 kHz and digitized at 20 kHz using Clampex (pClamp 10 suite, Molecular Devices) via a Digidata 1440A digitizer.

Microscopy. Membrane partitioning and FRAP experiments were performed on an LSM 710 inverted 2-photon confocal microscope equipped with a 40X water immersion objective lens

(1.2 NA/280 µM WD/0.208 µM/pixel and a GFP/Alexa 488 filter set. Untransfected HEK293 cells at a confluency of 50-75% were plated onto 35-mm glass bottom dishes one day prior to experiments. The IVM-bdpy fluorophore was dissolved in 100% DMSO and stored at −20 0C at a concentration of 24 mM.

A Mai Tai eHP 2-photon laser/760-1040 nm was used as an excitation source (power at 920 nm) for the FRAP experiments. The laser and bleaching power were set to 7% and 75%

respectively. Membrane partitioning images were taken for 35 min at 0.5 Hz after replacing the cell medium (extracellular solution) with one containing the IVM-bdpy probe at a concentration of 500 nM. FRAP images were taken at a frequency of 0.5 Hz for 10 min after a

30 min equilibration period in 500 nM IVM-bdpy.

Analysis and statistics. Group data were analysed in SigmaPlot 13.0 using one-way ANOVAs, where p < 0.05 was taken as the significance threshold and expressed as mean ± SEM. Tests for normally distributed data are built into the SigmaPlot software. Confocal images were captured and analysed with Zen 2012 SP2 (Zeiss) software. Oocyte concentration-response data were fit to a Hill equation to obtain an EC50 and Hill co-efficient for each oocyte recording.

These parameters were then averaged across multiple oocyte experiments of the same type.

Single channel recordings were analysed in QuB software. Single channel currents were

idealised and separated into discrete activations by applying critical shut times of 50 ms in

glutamate alone or 50-100 ms in glutamate plus IVM to separate and define single receptor

active periods. Patches yielded between 30-250 individual active periods. Critical shut times

were determined by generating an initial shut duration histogram to continuous data that

included inactive periods corresponding to receptor desensitisation. The selected critical

times eliminated periods of receptor desensitisation while retaining singe receptor active

periods. Mean active period duration and intra-activation open probability was estimated

from each patch. Group means were obtained by averaging across multiple patches for each

recording condition.

Acknowledgments

We would like to thank the Queensland Brain Institute microscope facility for the use and

training associated with the LSMS 710 inverted 2-photon confocal microscope.

Fig 1. Activity of glutamate and IVM at homo- and heteromeric GluClRs. A. Structure of the GluClR (3RIF, (Hibbs and Gouaux 2011)) showing the large extracellular domain with glutamate bound at interfaces between adjacent subunits. Also shown is the transmembrane domain, which consists of four helices (M1-M1) per subunit and IVM bound between the M1 and M3 of adjacent subunits. B. Sequence alignments of selected invertebrate GluClRs subunits showing the IVM binding site, which consists of the first (M1) to third (M3) transmembrane domains. Boxed in red are sites of missense mutations that reduce IVM sensitivity. Abbreviations are H. cont., H. contortus, C. eleg., C elegans, P. xylos., P. xylostella, T. urtic., T. urticae (isoforms 1 and 3) and D. melan., D. melanogaster. C. Example currents obtained in oocytes injected with the β subunit (above) or the α and β subunits (ratio 1:50, below) in response to the indicated glutamate concentrations. D. Group concentration- response data for the indicated GluClRs. Holding potential was –40 mV.

Fig 2. Activity of IVM at homo- and heteromeric GluClRs. (A-C) Representative concentration-response experiments at the indicated IVM concentrations in oocytes. IVM- induced currents were normalised to the response elicited by a saturating concentration of glutamate (5 mM) for oocytes injected with cDNA encoding the α subunit (A) or α and β subunits at a ratio of 1:50 (B) or 1:1 (C). D. Group concentration-response data for the indicated GluClRs. Holding potential was –40 mV.

Fig 3. IVM analogue potency. A. Structures of IVM and the three analogues. Arrows indicate the synthesis reaction steps. MOMO represents a methoxymethyl ether group. A red broken border indicates the change in 5-OH configuration between IVM-2 and IVM-3. B. Example experiments in oocytes expressing α GluClRs that test IVM and analogues, IVM-1 and IVM-3. The IVM or analogue-induced current was normalised to the response elicited by a saturating concentration of glutamate (5 mM). C. Group bar plots comparing the potency between α and αβ (1:50) GluClRs for IVM and the three IVM analogues. *** p < 0.001, * p < 0.05. Numbers within the bars indicate the number of experiments (oocytes). Holding potential was –40 mV.

Fig 4. IPSCs mediated by α and αβ GluClRs. (A-B). A whole-cell recordings from HEK293 cell expressing α GluClRs (A) or αβ GluClRs (B) in co-culture with primary neurons. C. Expanded view of a segment from (A) showing isolated IPSCs that rise rapidly to a peak before decaying back to baseline. D. Expanded view of a segment from (B) showing isolated IPSCs that decay faster than those in (C). E. Inhibition of IPSCs mediated by α GluClRs by picrotoxin. (F-H). Group bar plot showing the mean IPSC decay time constant (F), 10-90% rise times (G) and peak amplitude (H) of α and αβ GluClRs. ** p < 0.01. Holding potential was –70 mV.

Fig 5. Effects of IVM at IPSCs mediated by α and αβ GluClRs. (A-B). A whole-cell recording from a HEK293 cell expressing α GluClRs (A) or αβ GluClRs (B) in co-culture with primary neurons in the continuous presence of 5 nM IVM. Note the steady increase in inward tonic current, which reaches a steady-state plateau at about 17 min in both (A) and (C). C. Expanded view of a segment from (A) showing isolated IPSCs, which have larger peak amplitudes and rise and decay more slowly than the corresponding IPSCs in the absence of IVM. D. Expanded view of a segment from (B) showing isolated IPSCs, which have larger peak amplitudes and rise and decay more slowly than the corresponding IPSCs in the absence of IVM. Note that the peak amplitude is smaller and the rise and decay times are faster than those in (C). E-G. Group bar plot showing the mean IPSC decay time constant (E), 10-90% rise times (F) and peak amplitude (G) of α and αβ GluClRs in the presence and absence of IVM. ** p < 0.01, * p < 0.05, ## p < 0.01. Holding potential was –70 mV.

Fig 6. Single channel properties of β and αβ GluClRs. A. Single receptor currents from patches expressing αβ (1:1) GluClRs at 2 µM and 3 mM glutamate and 2 µM glutamate plus 5 nM IVM. Note the increase in the duration of the active periods in the presence of IVM. B. Transfections of α and β subunits produced two distinct types of single channel activity based on current amplitude (1.2 and 0.7 pA). The 1.2 pA activity was the most prominent, accounting for 90% of the openings. The active durations of both types increase at higher glutamate concentrations and in the presence of IVM. C. Single receptors currents mediated by β GluClRs in the presence of the indicated concentrations of glutamate and in 2 µM glutamate plus 5 nM IVM. Note only one amplitude level (0.4 pA). D. Amplitude histograms for single receptor activity from patches expressing αβ GluClRs showing the two main amplitude levels (left) and β GluClRs (left). E. Single receptor mean active period durations (left) and intra-activation open probability (right) for αβ GluClRs in response to the indicated ligands. F. Single receptor mean active period durations (left) and intra-activation open probability (right) for β GluClRs in response to the indicated ligands. * p < 0.05. Numbers within the bars indicate the number of experiments (patches). Holding potential was –70 mV.

Fig 7. IVM-bdpy fluorescent probe. A. Structures of free bdpy fluorophore and IVM-bdpy. B. The spectral emission properties of IVM-bdpy and the free bdpy fluorophore. C. IVM-bdpy retained activity as shown in the sample oocyte current mediated by α GluClRs. D. Group concentration-response plot (n = 6) for IVM-bdpy at α GluClRs, showing reduced potency relative to IVM.

Fig 8. Membrane partitioning and lateral diffusion of IVM-bdpy. A. Confocal image of a field of HEK293 cells before (left) and after 30 min exposure of 500 nM IVM-bdpy (right). B. Group plot (n = 5) of the change in fluorescence over time for IVM-bdpy (green) the free bdpy fluorophore (black) and the subtraction (IVM-1, yellow). These plots were fitted to exponential functions to yield time constants of 6.5 min for the IVM-bdpy, 8.2 min for the bdpy fluorophore and 6.0 min for IVM-1. C. An example of a FRAP experiment showing a cell after equilibrating in 500 nM IVM-bdpy for 30 min (left), after a circular region (demarcated) was pholobleached with laser energy (middle) and after recovery of the bleached region (right). D. Group plot (n = 5) of the rate of recovery after pholobleaching. The data were fitted to an exponential with a time constant of 1.3 min.

References

Althoff, T., et al. 2014. 'X-ray structures of GluCl in apo states reveal a gating mechanism of Cys-loop receptors', Nature, 512: 333-7. Alvarez, L. I., et al. 2007. 'Drug transfer into target helminth parasites', Trends Parasitol, 23: 97-104. Amanzougaghene, N., et al. 2018. 'Mutations in GluCl associated with ivermectin Field-Resistant Head lice from Senegal', Int J Antimicrob Agents. Atif, M., et al. 2017. 'Effects of glutamate and ivermectin on single glutamate-gated chloride channels of the parasitic nematode H. contortus', PLoS Pathog, 13: e1006663. Bebber, Daniel P., et al. 2013. 'Crop pests and pathogens move polewards in a warming world', Nature Climate Change, 3: 985. Callau-Vazquez, D., et al. 2018. 'Investigation of Agonist Recognition and Channel Properties in a Flatworm Glutamate-Gated Chloride Channel', Biochemistry, 57: 1360-68. Chen, X., et al. 2017. 'Physiological and pharmacological properties of inhibitory postsynaptic currents mediated by alpha5beta1gamma2, alpha5beta2gamma2 and alpha5beta3gamma2 GABAA receptors', Neuropharmacology, 125: 243-53. Cull-Candy, S. G. 1976. 'Two types of extrajunctional L-glutamate receptors in locust muscle fibres', J Physiol, 255: 449-64. Cully, D. F., et al. 1994. 'Cloning of an avermectin-sensitive glutamate-gated chloride channel from Caenorhabditis elegans', Nature, 371: 707-11. Degani-Katzav, N., et al. 2016. 'Subunit stoichiometry and arrangement in a heteromeric glutamate- gated chloride channel', Proc Natl Acad Sci U S A. Degani-Katzav, N., et al. 2017. 'Trapping of ivermectin by a pentameric ligand-gated ion channel upon open-to-closed isomerization', Sci Rep, 7: 42481. Dermauw, W., et al. 2012. 'The cys-loop ligand-gated ion channel gene family of Tetranychus urticae: implications for acaricide toxicology and a novel mutation associated with abamectin resistance', Insect Biochem Mol Biol, 42: 455-65. Dixon, C., et al. 2014. 'GABAA receptor alpha and gamma subunits shape synaptic currents via different mechanisms', J Biol Chem, 289: 5399-411. El-Abdellati, A., et al. 2011. 'Altered avr-14B gene transcription patterns in ivermectin-resistant isolates of the cattle parasites, Cooperia oncophora and Ostertagia ostertagi', Int J Parasitol, 41: 951-7. Fitzpatrick, J. L. 2013. 'Global food security: the impact of veterinary parasites and parasitologists', Vet Parasitol, 195: 233-48. Fu, Z., et al. 2003. 'Functional excitatory synapses in HEK293 cells expressing neuroligin and glutamate receptors', J Neurophysiol, 90: 3950-7. Geary, T. G., et al. 1993. 'Haemonchus contortus: ivermectin-induced paralysis of the pharynx', Exp Parasitol, 77: 88-96. Ghosh, R., et al. 2012. 'Natural variation in a chloride channel subunit confers avermectin resistance in C. elegans', Science, 335: 574-8. Hartiadi, L. Y., et al. 2016. 'High and low GABA sensitivity alpha4beta2delta GABAA receptors are expressed in Xenopus laevis oocytes with divergent stoichiometries', Biochem Pharmacol, 103: 98-108. Hibbs, R. E., and E. Gouaux. 2011. 'Principles of activation and permeation in an anion-selective Cys- loop receptor', Nature, 474: 54-60. Horsberg, T. E. 2012. 'Avermectin use in aquaculture', Curr Pharm Biotechnol, 13: 1095-102. Jia, F., et al. 2008. 'Taurine is a potent activator of extrasynaptic GABA(A) receptors in the thalamus', J Neurosci, 28: 106-15.

Kane, N. S., et al. 2000. 'Drug-resistant Drosophila indicate glutamate-gated chloride channels are targets for the antiparasitics nodulisporic acid and ivermectin', Proc Natl Acad Sci U S A, 97: 13949-54. Kaplan, R. M., and A. N. Vidyashankar. 2012. 'An inconvenient truth: global worming and anthelmintic resistance', Vet Parasitol, 186: 70-8. Kwon, D. H., et al. 2010. 'A point mutation in a glutamate-gated chloride channel confers abamectin resistance in the two-spotted spider mite, Tetranychus urticae Koch', Insect Mol Biol, 19: 583- 91. Laing, R., et al. 2017. 'Ivermectin - Old Drug, New Tricks?', Trends Parasitol, 33: 463-72. Laing, R., et al. 2016. 'Analysis of putative resistance gene loci in UK field populations of Haemonchus contortus after 6years of macrocyclic lactone use', Int J Parasitol, 46: 621-30. Lingwood, D., and K. Simons. 2010. 'Lipid rafts as a membrane-organizing principle', Science, 327: 46- 50. Lynagh, T., and J. W. Lynch. 2010. 'A glycine residue essential for high ivermectin sensitivity in Cys- loop ion channel receptors', Int J Parasitol, 40: 1477-81. Lynagh, T., and J. W. Lynch. 2012. 'Molecular mechanisms of Cys-loop ion channel receptor modulation by ivermectin', Front Mol Neurosci, 5: 60. Marder, E., and J. S. Eisen. 1984. 'Electrically coupled pacemaker neurons respond differently to same physiological inputs and neurotransmitters', J Neurophysiol, 51: 1362-74. Marder, E., and D. Paupardin-Tritsch. 1978. 'The pharmacological properties of some crustacean neuronal acetylcholine, gamma-aminobutyric acid, and L-glutamate responses', J Physiol, 280: 213-36. Martin, R. J., and J. R. Kusel. 1992. 'On the distribution of a fluorescent ivermectin probe (4" 5,7 dimethyl-bodipy proprionylivermectin) in Ascaris membranes', Parasitology, 104 ( Pt 3): 549- 55. McCavera, S., et al. 2009. 'An ivermectin-sensitive glutamate-gated chloride channel from the parasitic nematode Haemonchus contortus', Mol Pharmacol, 75: 1347-55. Mermans, C., et al. 2017. 'A G326E substitution in the glutamate-gated chloride channel 3 (GluCl3) of the two-spotted spider mite Tetranychus urticae abolishes the agonistic activity of macrocyclic lactones', Pest Manag Sci. Michael, B., et al. 2001. 'Comparison of ivermectin, doramectin, selamectin, and eleven intermediates in a nematode larval development assay', J Parasitol, 87: 692-6. Oerke, E. C. 2005. 'Crop losses to pests', Journal of Agricultural Science, 144: 31-43. Osei-Atweneboana, M. Y., et al. 2011. 'Phenotypic evidence of emerging ivermectin resistance in Onchocerca volvulus', PLoS Negl Trop Dis, 5: e998. Paini, D. R., et al. 2016. 'Global threat to agriculture from invasive species', Proc Natl Acad Sci U S A, 113: 7575-9. Portillo, V., et al. 2003. 'Distribution of glutamate-gated chloride channel subunits in the parasitic nematode Haemonchus contortus', J Comp Neurol, 462: 213-22. Rose, H., et al. 2014. 'Exploiting parallels between livestock and wildlife: Predicting the impact of climate change on gastrointestinal nematodes in ruminants', Int J Parasitol Parasites Wildl, 3: 209-19. Scott, S., et al. 2015. 'Correlating structural and energetic changes in glycine receptor activation', J Biol Chem, 290: 5621-34. Sykes, D. A., et al. 2014. 'Observed drug-receptor association rates are governed by membrane affinity: the importance of establishing "micro-pharmacokinetic/pharmacodynamic relationships" at the beta2-adrenoceptor', Mol Pharmacol, 85: 608-17. Tigchelaar, M., et al. 2018. 'Future warming increases probability of globally synchronized maize production shocks', Proc Natl Acad Sci U S A, 115: 6644-49.

Wang, X., et al. 2017. 'Mutations on M3 helix of Plutella xylostella glutamate-gated chloride channel confer unequal resistance to abamectin by two different mechanisms', Insect Biochem Mol Biol, 86: 50-57. Wang, X., et al. 2016. 'A point mutation in the glutamate-gated chloride channel of Plutella xylostella is associated with resistance to abamectin', Insect Mol Biol, 25: 116-25. Williamson, S. M., et al. 2011. 'Candidate anthelmintic resistance-associated gene expression and sequence polymorphisms in a triple-resistant field isolate of Haemonchus contortus', Mol Biochem Parasitol, 180: 99-105. Wolstenholme, A. J. 2012. 'Glutamate-gated chloride channels', J Biol Chem, 287: 40232-8. Wolstenholme, A. J., et al. 2016. 'How do the macrocyclic lactones kill filarial nematode larvae?', Invert Neurosci, 16: 7. Wolstenholme, A. J., and A. T. Rogers. 2005. 'Glutamate-gated chloride channels and the mode of action of the avermectin/milbemycin anthelmintics', Parasitology, 131 Suppl: S85-95. Zhang, Y., et al. 2015. 'Functional reconstitution of glycinergic synapses incorporating defined glycine receptor subunit combinations', Neuropharmacology, 89: 391-7.

Supporting Information

GluClR-mediated inhibitory postsynaptic currents reveal targets for ivermectin and potential mechanisms of ivermectin resistance Atif M, Smith JJ, Estrada-Mondragon A, Xiao X, Salim AA,Capon RJ, Lynch JW and

Keramidas A

Correspondence can be sent to Joseph Lynch or Angelo Keramidas Email: [email protected] or [email protected]

This file includes:

Chemical synthesis of ivermectin analogues, IVM-1, IVM-2, IVM-3 and IVM-bdpy

General Experimental Detail for Chemical Synthesis

NMR experiments were performed on a Bruker Avance DRX600 spectrometer and referenced to residual 1H signals in the deuterated solvents. ESIMS experiments were carried out on an Agilent 1100 series LC/MSD instrument. HR-ESIMS data were acquired on a Bruker micrOTOF mass spectrometer by direct infusion in MeCN at 3 μL/min using sodium formate clusters as an internal calibrant. All HPLC analyses and purifications were performed on Agilent 1100 series LC instruments with corresponding detectors, collectors and software inclusively. All chemicals were purchased from Merck, Sigma- Aldrich or Fluka. Solvents used for general purposes were of at least analytical grade, and solvents used for HPLC were of HPLC grade. Phenomenex Luna C8, 10 µm, 210 × 250 mm column (preparative) columns were used for HPLC separations. Flash column chromatography was performed on silica gel LC60A (40–63 mm, Davisil).

Figure S1. Synthetic scheme for IVM analogs, bdpy and IVM-bdpy. Reagents and conditions: (a) 2% H2SO4, MeOH, rt, 16 h, (b) DMAP, imidazole, TBDMSCl, DCM, rt, 16 h, (c) MOMCl, DIPEA, DCM, rt, 16 h, (d) HF.pyridine, THF, rt, 16 h, (e) DIAD, PPh3, HCOOH, THF, rt, 16 h, (f) NaHCO3, MeOH/THF/H2O, rt, 16 h, (g) (1) NaN3, Acetone/H2O, reflux, 16 h, (2) KOH, MeOH/H2O, rt, 16 h, (h) (COCl)2, cat. DMF, DCM, rt, 3 h, (i)

TEA/DMAP, DCM, rt, 16 h, (j) 1% TsOH/MeOH, rt, 16 h, (k) (1) TFA, DCM, rt, (2) DDQ, rt, (3) Et3N/BF3, Et2O, rt, (l) CuSO4.5H2O, sodium ascorbate, CH3CN, 82 °C, 5 d.

Preparation of IVM-1

To a solution of IVM (100 g, 0.11 moL) in MeOH (1 L) was added concentrated sulfuric acid (20 mL) dropwise, and the reaction was stirred overnight. Thin layer chromatography (TLC) was utilised to monitor the reaction. After removal of the solvent in vacuo, the residue was poured into iced water

and quenched with Na2CO3. The mixture was extracted with EtOAc (3 × 300 mL). The combined organic layers were washed with brine, dried over anhydrous MgSO4, filtered, concentrated in vacuo, and purified with silica gel chromatography (EtOAc/hexane = 1:1) to yield IVM-1 (51 g, 76%). 1H NMR (600

MHz, CDCl3), see Figure S3: δH 5.83 (d, J = 12 Hz, 1H), 5.70 (m, 1H), 5.69 (m, 1H), 5.41 (s, 1H), 5.30 (m, 1H), 5.29 (m, 1H), 4.70 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.66 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.32 (m, 1H), 4.01 (br, 1H), 3.97 (d, J = 6 Hz, 1H), 3.67 (m, 1H), 3.25 (m, 1H), 3.20 (m, 1H), 2.53 (m, 1H), 2.30 (m, 2H), 1.98 (m, 1H), 1.86 (s, 3H), 1.75 (m, 1H), 1.65 (m, 1H), 1.51 (m, 1H), 1.50 (s, 3H), 1.50 (m, 2H), 1.49 (m, 1H), 1.48 (m, 1H), 1.42 (m, 2H), 1.33 (m, 1H), 1.17 (d, J = 7.2 Hz, 3H), 0.96 (t, J = 7.2 Hz, 3H), 0.85 (d, J = 6.6 13 Hz, 3H), 0.81 (m, 1H), 0.79 (d, J = 6.6 Hz, 3H); C NMR (150 MHz, CDCl3), see Figure S4: δC 173.4, 139.7, 138.6, 137.6, 137.0, 124.7, 120.4, 118.1, 117.1, 97.4, 80.2, 79.2, 77.5, 77.2, 68.5, 68.4, 67.7, 67.3, 45.6, 41.3, 40.0, 36.7, 35.7, 35.5, 34.2, 31.2, 28.0, 27.4, 19.9, 19.1, 17.4, 14.6, 12.5, 11.7. HRESIMS m/z + 609.3399 [M + Na] (calcd for C34H50O8Na, 609.3398), see Figure S5.

Preparation of IVM-2

To a solution of IVM-1 (47.0 g, 0.08 moL), 4-dimethylaminopyridine (DMAP, 0.97 g, 0.008 moL) and imidazole (54.5 g, 0.8 moL) in DCM (1 L) at 0 °C was added tert-butyldimethylsilyl chloride (TBDMSCl, 26.6 g, 0.176 moL), and the reaction was allowed to reach ambient temperature and stirred overnight. TLC was utilised to monitor the reaction. After washing with water (300 mL) and brine (300 mL), the organic layer was dried over anhydrous MgSO4, filtered, concentrated in vacuo, and purified with silica 1 gel chromatography (EtOAc/hexane = 1:5), to yield (i) (47.7 g, 85%). H NMR (600 MHz, CDCl3): δH 5.77 (d, J = 12.0 Hz, 1H), 5.71 (m, 1H), 5.69 (m, 1H), 5.32 (m, 1H), 5.32 (m, 1H), 5.26 (m, 1H), 4.66 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.56 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.42 (br, 1H), 3.99 (br, 1H), 3.80 (d, J = 6.0 Hz, 1H), 3.67 (m, 1H), 3.34 (m, 1H), 3.18 (m, 1H), 2.51 (m, 1H), 2.30 (m, 2H), 1.98 (m, 1H), 1.78 (s, 3H), 1.73 (m, 1H), 1.64 (m, 1H), 1.52 (m, 1H), 1.52 (s, 3H), 1.50 (m, 2H), 1.49 (m, 1H), 1.47 (m, 1H), 1.41 (m, 2H), 1.31 (m, 1H), 1.16 (d, J = 7.2 Hz, 3H), 0.94 (t, J = 7.2 Hz, 3H), 0.91 (s, 9H), 0.83 (d, J = 6.6 Hz, 3H), 0.82 13 (m, 1H), 0.78 (d, J = 6.6 Hz, 3H), 0.12 (s, 6H); C NMR (150 MHz, CDCl3): δC 173.7, 140.3, 138.7, 137.3, 136.5, 124.8, 119.3, 117.3, 117.1, 97.4, 80.1, 80.0, 77.6, 77.2, 69.4, 68.6, 67.9, 67.2, 45.7, 41.3, 39.9, 36.6, 35.8, 35.5, 34.2, 31.2, 28.0, 27.4, 25.8, 19.9, 19.2, 17.4, 14.6, 12.5, 11.7, -4.6.

To a solution of (i) (10.0 g, 14 mmoL) and N, N-diisopropylethylamine (DIPEA, 4.43g, 68 mmoL) in DCM (300 mL) at 0 °C was added chloromethyl methyl ether (MOMCl, 2.53 g, 62 mmoL), and the reaction was allowed to reach ambient temperature and stirred for 3 h. TLC was utilised to monitor the reaction. After washing with water (100 mL) and brine (100 mL), the organic layer was dried over anhydrous MgSO4, filtered, concentrated in vacuo, and purified with silica gel chromatography

1 (EtOAc/hexane = 1:5) to yield (ii) (9.7 g, 91%). H NMR (600 MHz, CDCl3): δH 5.78 (m, 1H), 5.71 (m, 1H), 5.70 (m, 1H), 5.32 (m, 1H), 5.26 (m, 1H), 5.18 (m, 1H), 4.67 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.59 (m, 2H), 4.57 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.43 (br, 1H), 4.11 (s, 1H), 3.89 (br, 1H), 3.81 (d, J = 6.0 Hz, 1H), 3.67 (m, 1H), 3.42 (s, 3H), 3.34 (m, 1H), 3.18 (d, J = 8.8 Hz, 1H), 2.53 (m, 1H), 2.31 (m, 2H), 1.98 (m, 1H), 1.78 (s, 3H), 1.74 (m, 1H), 1.64 (m, 1H), 1.53 (m, 1H), 1.51 (s, 3H), 1.51 (m, 2H), 1.49 (m, 1H), 1.47 (m, 1H), 1.42 (m, 2H), 1.31 (m, 1H), 1.16 (d, J = 7.2 Hz, 3H), 0.95 (t, J = 7.2 Hz, 3H), 0.92 (s, 9H), 0.83 (d, J = 13 6.6 Hz, 3H), 0.82 (m, 1H), 0.78 (d, J = 6.6 Hz, 3H), 0.12 (s, 6H); C NMR (150 MHz, CDCl3): δC 173.8, 140.2, 137.5, 137.3, 135.2, 124.7, 119.3, 118.3, 117.3, 97.4, 95.5, 82.6, 80.1, 80.1, 77.2, 69.5, 68.6, 67.9, 67.2, 56.1, 45.7, 41.3, 39.7, 36.8, 35.8, 35.5, 34.3, 31.2, 28.0, 27.4, 25.9, 20.0, 19.6, 17.4, 15.0, 12.5, 11.7, -4.6.

TsOH (2 g) was dissolved in methanol (200 mL), followed by the addition of (ii) (10.8 g, 14.5 mmoL), and the reaction was stirred at ambient temperature overnight. TLC was utilised to monitor the reaction. After removal of solvent in vacuo, the residue was purified by flash column chromatography 1 (EtOAc/hexane = 1:2) to yield IVM-2 (7.9 g, 86%). H NMR (600 MHz, CDCl3), see Figure S6: δH 5.70 (m, 1H), 5.70 (m, 1H), 5.40 (br, 1H), 5.30 (m, 1H), 5.18 (d, J = 12.0 Hz, 1H), 4.68 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.64 (dd, J = 14.0, J = 2.0 Hz, 1H), 4.59 (m, 2H), 4.28 (br, 1H), 4.08 (s, 1H), 3.95 (d, J = 6.0 Hz, 1H), 3.89 (br, 1H), 3.67 (m, 1H), 3.42 (s, 3H), 3.25 (m, 1H), 3.18 (d, J = 8.8 Hz, 1H), 2.52 (m, 1H), 2.38 (d, J = 7.2 Hz, 1H), 2.28 (m, 2H), 1.97 (m, 1H), 1.86 (s, 3H), 1.75 (m, 1H), 1.65 (m, 1H), 1.53 (m, 1H), 1.50 (s, 3H), 1.50 (m, 2H), 1.50 (m, 1H), 1.47 (m, 1H), 1.42 (m, 2H), 1.31 (m, 1H), 1.16 (d, J = 7.2 Hz, 3H), 0.95 (t, J = 7.2 Hz, 3H), 0.84 (d, J = 6.6 Hz, 3H), 0.80 (m, 1H), 0.78 (d, J = 6.6 Hz, 3H); 13C NMR (150 MHz,

CDCl3), see Figure S7: δC 173.6, 139.7, 138.0, 137.7, 135.1, 124.6, 120.4, 118.3, 118.1, 97.6, 95.5, 82.6, 80.2, 79.1, 77.2, 68.5, 68.5, 67.7, 67.2, 56.1, 45.7, 41.3, 39.8, 36.8, 35.8, 35.5, 34.2, 31.2, 28.0, 27.4, + 19.9, 19.5, 17.4, 14.9, 12.5, 11.7. HRESIMS m/z 653.3661 [M + Na] (calcd for C36H54O9Na, 653.3660), see Figure S8.

Preparation of IVM-3

To a solution of IVM-2 (500 mg, 0.8 mmoL), triphenylphosphine (1250 mg, 4.8 mmoL) and formic acid (220 mg, 4.8 mmoL) in THF (30 mL) at ambient temperature was added diisopropyl azodicarboxylate (DIAD, 962 mg, 4.8 mmoL), and the reaction was stirred overnight. TLC was utilised to monitor the reaction. After removal of the solvent in vacuo, the residue was purified by flash column 1 chromatography (EtOAc/hexane = 1:4) to yield (iii) (422 mg, 81%). H NMR (600 MHz, CDCl3): δH 8.18 (s, 1H), 5.89 (m, 1H), 5.72 (m, 1H), 5.71 (m, 1H), 5.62 (br, 1H), 5.40 (br, 1H), 5.35 (m, 1H), 5.18 (m, 1H), 4.60 (d, J = 7.2 Hz, 1H), 4.60 (br, 2H), 4.57 (d, J = 7.2 Hz, 1H), 3.89 (br, 1H), 3.75 (s, 1H), 3.68 (m, 1H), 3.43 (s, 3H), 3.20 (d, J = 9.0 Hz, 1H), 3.08 (m, 1H), 2.53 (m, 1H), 2.32 (m, 1H), 2.27 (m, 1H), 1.97 (m, 1H), 1.78 (s, 3H), 1.77 (m, 1H), 1.64 (m, 1H), 1.54 (m, 1H), 1.51 (m, 1H), 1.50 (s, 3H), 1.50 (m, 2H), 1.47 (m, 1H), 1.43 (m, 2H), 1.34 (m, 1H), 1.16 (d, J = 7.2 Hz, 3H), 0.96 (t, J = 7.2 Hz, 3H), 0.85 (d, J = 6.6 Hz, 13 3H), 0.80 (m, 1H), 0.79 (d, J = 6.6 Hz, 3H); C NMR (150 MHz, CDCl3): δC 173.1, 160.3, 138.2, 138.1, 135.2, 133.2, 124.8, 121.7, 120.1, 118.3, 97.4, 95.5, 83.9, 82.6, 77.3, 75.8, 70.3, 68.6, 68.1, 67.3, 56.2, 46.3, 41.3, 40.0, 36.8, 35.8, 35.5, 34.2, 31.2, 28.0, 27.4, 19.5, 18.9, 17.5, 14.9, 12.5, 11.7.

To a solution of (iii) (345 mg, 0.52 mmoL) in THF/MeOH/H2O (10/5/0.5 mL) at ambient temperature was added sodium bicarbonate (88 mg, 1.05 mmoL), and the reaction was stirred overnight. TLC was utilised to monitor the reaction. The mixture was extracted with DCM (3 × 15 mL). The combined organic layers were washed with brine, dried over anhydrous MgSO4, filtered, concentrated in vacuo, and purified with silica gel chromatography (EtOAc/hexane = 1:2) to yield IVM-3 (313 mg, 95%). 1H

NMR (600 MHz, CDCl3), see Figure S9: δH 5.85 (m, 1H), 5.70 (m, 1H), 5.70 (m, 1H), 5.42 (br, 1H), 5.34 (m, 1H), 5.18 (m, 1H), 4.59 (m, 2H), 4.59 (m, 2H), 4.12 (br, 1H), 4.01 (br, 1H), 3.89 (br, 1H), 3.83 (br, 1H), 3.67 (m, 1H), 3.42 (s, 3H), 3.19 (d, J = 9.0 Hz, 1H), 3.04 (m, 1H), 2.52 (m, 1H), 2.32 (m, 1H), 2.26 (m, 1H), 1.97 (m, 1H), 1.90 (s, 3H), 1.76 (m, 1H), 1.64 (m, 1H), 1.53 (m, 1H), 1.50 (m, 1H), 1.50 (m, 2H), 1.50 (s, 3H), 1.47 (m, 1H), 1.42 (m, 2H), 1.32 (m, 1H), 1.16 (d, J = 7.2 Hz, 3H), 0.95 (t, J = 7.2 Hz, 3H), 13 0.85 (d, J = 6.6 Hz, 3H), 0.80 (m, 1H), 0.78 (d, J = 6.6 Hz, 3H); C NMR (150 MHz, CDCl3), see Figure S10:

δC 173.7, 139.2, 137.8, 137.4, 135.1, 124.8, 120.9, 118.3, 117.0, 97.4, 95.5, 85.7, 82.6, 77.3, 76.0, 69.4, 68.5, 68.2, 67.3, 56.2, 46.5, 41.3, 39.9, 36.9, 35.8, 35.5, 34.2, 31.2, 28.0, 27.4, 19.6, 19.3, 17.5, 14.9, + 12.5, 11.7. HRESIMS m/z 653.3660 [M + Na] (calcd for C36H54O9Na, 653.3660), see Figure S11.

Preparation of bdpy

To a suspension of 4-hydroxybenzaldehyde (3.66 g, 30 mmoL) and potassium carbonate (16 g, 120 mmoL) in acetone (250 mL) was added propargyl bromide (14.3 g, 120 mmoL), and the reaction was stirred at ambient temperature overnight. TLC was utilised to monitor the reaction. After filtration, the filtrate was evaporated in vacuo and the residue was partitioned between EtOAc (200 mL) and water (200 mL). The organic layer was washed with brine, dried over MgSO4, filtered, concentrated in vacuo to yield, and purified with silica gel chromatography (EtOAc/hexane = 1:4) to yield (iv) (4.2 g, 1 88%). H NMR (600 MHz, CDCl3): δH = 9.91 (s, 1H), 7.86 (d, J = 8.0 Hz, 2H), 7.09 (d, J = 8.0 Hz, 2H), 4.78 13 (s, 2H), 2.57 (s, 1H); C NMR (150 MHz, CDCl3): δC = 190.8, 162.4, 131.9, 130.6, 115.2, 77.5, 76.4, 55.9.

To a solution of (iv) (230 mg, 1.44 mmoL) and 3-ethyl-2, 4-dimethylpyrrole (v) (407 mg, 3.3 mmoL) in DCM (15 mL) was added TFA (110 mg, 0.96 mmoL). The dark mixture was stirred at r.t. until total disappearance of (iv) (TLC). Then DDQ (523 mg, 2.3 mmoL) was added, and 5 min later TEA (1170 mg, 11.5 mmoL) and trifluoroborate etherate (2450 mg, 17.25 mmoL) were added successively. TLC was utilised to monitor the reaction. The mixture was filtered through a pad of silica. After removal of solvent in vacuo, the residue was purified by flash column chromatography (EtOAc/hexane = 1:15) to 1 yield bdpy (130 mg, 20.8% over 3 steps). H NMR (600 MHz, CDCl3): δH = 7.19 (d, J = 8.0 Hz, 2H), 7.08 (d, J = 8.0 Hz, 2H), 4.76 (s, 2H), 2.56 (t, J = 2.0 Hz, 1H), 2.53 (s, 6H), 2.30 (q, J = 7.6 2H), 1.33 (s, 6H), 13 0.98 (t, J = 7.6 Hz, 3H); C NMR (150 MHz, CDCl3): δC = 158.0, 153.6, 139.9, 138.4, 132.7, 131.1, 129.5, 128.9, 115.5, 78.1, 75.9, 56.0, 17.1, 14.6, 12.5, 11.8.

Preparation of IVM-bdpy

A solution of acetone/H2O (30 mL/10 mL) of ethyl 4-bromobutyrate (vi) (3.9 g, 20 mmoL) and sodium azide (2.6 g, 40 mmoL) was heated under reflux overnight. Then acetone was removed in vacuo and

94 the residue was partitioned between DCM (30 mL) and water (30 mL). The mixture was extracted with

DCM (3 × 20 mL), and the combined organic extracts were washed with water, brine, dried over MgSO4 and evaporated to dryness to give the azido ester as colourless oil. The obtained azido acid ethyl ester was dissolved in MeOH/water (30 mL/30 mL) solution, cooled to 0 °C and KOH (2.2 g, 40 mmol) was added. The reaction mixture was stirred at room temperature overnight. The methanol was removed in vacuo and the residue was partitioned between DCM (30 mL) and water (30 mL). The aqueous layer was extracted with DCM (2 × 20 mL), acidified to pH 1 with 1N aqueous HCl and extracted with EtOAc

(4 × 30 mL). The combined organic layers were dried over MgSO4 and evaporated to dryness to afford (vii) (2.2 g, 85%) as colorless oil. To a solution of (vii) (500 mg, 3.88 mmoL) in DCM (10 mL) was added oxalyl chloride (1500 mg, 11.6 mmoL) at ambient temperature, followed by the addition of 3 drops of

DMF. The mixture was stirred for another 3 h, and then the solvent was removed under N2 to yield (viii), which was used without further purification.

To a solution of (i) (1360 mg, 1.94 mmoL), TEA (589 mg, 5.82 mmoL) and DMAP (710 mg, 5.82 mmoL) in DCM (20 mL) at ambient temperature was added (viii) (reaction mixture, 3.88 mmoL), and the reaction was stirred overnight. TLC was utilised to monitor the reaction. The mixture was poured into water (25 mL) and extracted with DCM (3 × 20 mL). The combined organic layers were washed with brine, dried over anhydrous MgSO4, filtered, concentrated in vacuo to yield, and purified with silica gel 1 chromatography (EtOAc/hexane = 1:4) to yield (ix) (1200 mg, 76%). H NMR (600 MHz, CDCl3): δH 5.81 (m, 1H), 5.77 (m, 1H), 5.67 (m, 1H), 5.32 (br, 1H), 5.30 (m, 1H), 5.16 (br, 1H), 4.98 (m, 1H), 4.68 (d, J = 7.2 Hz, 1H), 4.58 (d, J = 7.2 Hz, 1H), 4.44 (m, 1H), 4.17 (s, 1H), 3.82 (d, J = 5.6 Hz, 1H), 3.63 (m, 1H), 3.39 (t, J = 6.6 Hz, 2H), 3.37 (br, 1H), 3.19 (d, J = 9.0 Hz, 1H), 2.64 (m, 1H), 2.53 (m, 2H), 2.26 (m, 2H), 1.99 (m, 1H), 1.96 (m, 2H), 1.79 (s, 3H), 1.71 (m, 1H), 1.64 (m, 1H), 1.58 (s, 3H), 1.55 (m, 1H), 1.51 (m, 1H), 1.51 (m, 2H), 1.47 (m, 1H), 1.43 (m, 2H), 1.33 (m, 1H), 1.03 (d, J = 7.2 Hz, 3H), 0.95 (t, J = 7.2 Hz, 3H), 0.93 (s, 9H), 0.85 (d, J = 6.6 Hz, 3H), 0.84 (m, 1H), 0.78 (d, J = 6.6 Hz, 3H), 0.13 (s, 6H); 13C NMR (150

MHz, CDCl3): δC 173.9, 171.8, 141.2, 137.5, 136.2, 134.6, 125.4, 119.0, 117.7, 117.3, 97.5, 80.2, 80.1, 78.9, 76.9, 69.4, 68.7, 67.9, 67.1, 50.6, 45.7, 41.3, 38.9, 36.7, 35.8, 35.5, 34.2, 31.2, 31.1, 28.0, 27.3, 25.9, 24.3, 20.0, 18.9, 18.4, 17.4, 14.7, 12.5, 12.0, -4.6.

TsOH (100 mg) was dissolved in methanol (10 mL), followed by the addition of ix (522 mg, 1.94 mmoL), and the reaction was stirred at ambient temperature overnight. TLC was utilised to monitor the reaction. After removal of solvent in vacuo, the residue was purified by flash column chromatography 1 (EtOAc/hexane = 1:2) to yield (x) (350 mg, 78%). H NMR (600 MHz, CDCl3): δH 5.86 (m, 1H), 5.78 (m, 1H), 5.68 (m, 1H), 5.42 (br, 1H), 5.33 (m, 1H), 5.16 (br, 1H), 4.98 (m, 1H), 4.70 (d, J = 7.2 Hz, 1H), 4.65 (d, J = 7.2 Hz, 1H), 4.29 (d, J = 6.0 Hz, 1H), 3.97 (d, J = 6.6 Hz, 1H), 3.63 (m, 1H), 3.39 (t, J = 6.6 Hz, 2H), 3.27 (m, 1H), 3.19 (d, J = 9.0 Hz, 1H), 2.64 (m, 1H), 2.53 (m, 2H), 2.25 (m, 2H), 1.98 (m, 1H), 1.95 (m, 2H), 1.87 (s, 3H), 1.72 (m, 1H), 1.64 (m, 1H), 1.57 (s, 3H), 1.55 (m, 1H), 1.51 (m, 1H), 1.50 (m, 2H), 1.47 (m, 1H), 1.43 (m, 2H), 1.34 (m, 1H), 1.04 (d, J = 7.2 Hz, 3H), 0.95 (t, J = 7.2 Hz, 3H), 0.85 (d, J = 6.6 Hz, 13 3H), 0.82 (m, 1H), 0.78 (d, J = 6.6 Hz, 3H); C NMR (150 MHz, CDCl3): δC 173.6, 171.8, 140.6, 137.9, 136.7, 134.5, 125.3, 120.1, 118.0, 117.7, 97.5, 80.3, 79.1, 78.8, 76.9, 68.6, 68.4, 67.7, 67.0, 50.6, 45.6, 41.2, 39.0, 36.8, 35.7, 35.5, 34.1, 31.2, 31.1, 28.0, 27.3, 24.3, 19.9, 18.8, 17.4, 14.6, 12.5, 12.0.

A mixture of (x) (69.7 mg, 0.1 mmoL), bpdy (43 mg, 0.1 mmoL), copper sulfate pentahydrate (2.5 mg, 0.01 mmoL) and sodium ascorbate (2.0 mg, 0.01 mmoL) in acetonitrile (10 mL) was heated to 82 °C for 5 days under nitrogen. TLC were utilised to monitor the reaction. After removal of solvent in vacuo, the residue was purified by flash column chromatography (EtOAc/hexane = 1:1) to yield IVM-bdpy (81 1 mg, 71%). H NMR (600 MHz, CD3OD), see Figure S12: δH 8.10 (s, 1H), 7.19 (s, 1H), 7.19 (s, 1H), 5.95 (m, 1H), 5.81 (m, 1H), 5.67 (m, 1H), 5.42 (s, 1H), 5.24 (s, 2H), 5.20 (br, 1H), 5.13 (m, 1H), 5.00 (m, 1H), 4.65 (d, J = 7.2 Hz, 1H), 4.58 (d, J = 7.2 Hz, 1H), 4.52 (t, J = 7.2 Hz, 2H), 4.23 (m, 1H), 3.77 (d, J = 6.0 Hz, 1H), 3.68 (m, 1H), 3.24 (m, 1H), 3.23 (s, 1H), 2.78 (m, 1H), 2.51 (q, J = 7.2 Hz, 2H), 2.46 (s, 3H), 2.34 (q, J = 7.8 Hz, 2H), 2.25 (m, 2H), 2.25 (m, 2H), 2.20 (m, 1H), 1.89 (m, 1H), 1.82 (s, 3H), 1.60 (m, 1H), 1.59 (s, 3H), 1.56 (m, 1H), 1.51 (m, 1H), 1.48 (m, 2H), 1.46 (m, 1H), 1.41 (m, 2H), 1.35 (s, 3H), 1.19 (m, 1H), 1.05 (d, J = 7.2 Hz, 3H), 0.98 (t, J = 7.8 Hz, 3H), 0.93 (t, J = 7.2 Hz, 3H), 0.86 (d, J = 6.6 Hz, 3H), 0.83 (m, 13 1H), 0.78 (d, J = 6.6 Hz, 3H); C NMR (150 MHz, CD3OD), see Figure S13: δC 173.3, 173.1, 160.4, 154.6, 144.8, 142.5, 142.0, 139.7, 137.1, 137.1, 136.2, 134.0, 132.3, 130.8, 129.5, 127.1, 125.4, 121.4, 120.1, 118.9, 116.8, 98.9, 82.1, 81.8, 80.3, 77.9, 70.1, 68.9, 68.7, 68.6, 62.6, 50.5, 47.0, 42.8, 40.3, 37.7, 36.8, 36.8, 35.0, 32.5, 31.6, 29.2, 28.4, 26.7, 19.8, 19.4, 17.8, 17.8, 15.1, 15.1, 13.0, 12.6, 12.5, 12.2. HRESIMS + m/z 1154.6172 [M + Na] (calcd for C64H84BF2N5O10Na, 1154.6172), see Figure S14.

1 Figure S2. H NMR (DMSO-d6, 600 MHz) spectra of IVM

13 Figure S3. C NMR (DMSO-d6, 150 MHz) spectra of IVM

1 Figure S4. H NMR (CDCl3, 600 MHz) spectra of IVM-1

13 Figure S5. C NMR (CDCl3, 150 MHz) spectra of IVM-1

Figure S6. HRMS data for IVM-1

100

1 Figure S7. H NMR (CDCl3, 600 MHz) spectra of IVM-2

101

13 Figure S8. C NMR (CDCl3, 150 MHz) spectra of IVM-2

102

Figure S9. HRMS data for IVM-2

103

1 Figure S10. H NMR (CDCl3, 600 MHz) spectra of IVM-3

13 Figure S11. C NMR (CDCl3, 150 MHz) spectra of IVM-3

104

Figure S12. HRMS data for IVM-3

105

1 Figure S13. H NMR (MeOH-d4, 600 MHz) spectra of IVM-bdpy

106

13 Figure S14. C NMR (MeOH-d4, 150 MHz) spectra of IVM-bdpy

107

Figure S15. HRMS data for IVM-bdpy

108

Effects of glutamate and IVM on single glutamate-gated chloride channels of the

parasitic nematode H. contortus

Background

Many studies have been carried out on many pLGIC family including GlyRs and GABAARs to determine how agonists binds and activates the receptor. However, a detailed study on biophysical properties of GluClRs has not been performed. Previous studies shows that a substitution of G36’A in TM3 domain causes reduced sensitivity to IVM. To identify whether it is the binding of IVM molecule itself or the intrinsic properties of the receptor we employed single channel electrophysiology to determine the kinetic properties of α HcoGluClRs and α G36’A HcoGluClRs.

Statement of contribution by authors

I carried out 40% of all electrophysiological experiments (all the IVM experiments) and 35% of analysis for these electrophysiological experiments. Dr. Argel Estrada-Mondragon did the molecular study and made that figure (Figure 11). Bindi Nguyen performed around 10% electrophysiological experiments (some of the macropatch recordings). Professor Joseph W. Lynch conceptualized the project (50%) and wrote the first draft. Dr. Angelo Keramidas conceptualized the project (50%), performed 50% of all the electrophysiological experiments and wrote the first draft (50%). All the authors contributed equally to subsequent editing of the manuscript.

Contributor Statement of contribution

Author Mohammed Atif (Candidate) Electrophysiological recordings and data analysis (40%).

Figure preparations (20%).

Subsequent draft editing (20%)

Author Argel Estrada-Mondragon Molecular study and Figure 11 (100%).

Subsequent draft editing (20%)

Author Bindi Nguyen Electrophysiological recordings (10%)

Subsequent draft editing (20%)

Author Joseph W. Lynch Writing and editing first draft (50%)

110

Subsequent draft editing (20%)

Author Angelo Keramidas Electrophysiological recordings and data analysis (10%).

Figure preparation (80%)

Writing and editing first draft (50%)

Subsequent draft editing (20%)

111

PLOS Pathogens, October 2017; PMID: 28968469

Effects of glutamate and IVM on single glutamate-gated chloride channels of the parasitic nematode H. contortus

Atif M1, Estrada-Mondragon A1, Nguyen B1, Lynch JW1,2* and Keramidas A1*

1Queensland Brain Institute and 2School of Biomedical Sciences, The University of Queensland,

Brisbane, QLD 4072 Australia

*Correspondence: Angelo Keramidas, Queensland Brain Institute, The University of Queensland,

Brisbane, QLD 4072 Australia E-mail: [email protected]; Joseph Lynch, Queensland Brain

Institute, The University of Queensland, Brisbane, QLD 4072 Australia E-mail: [email protected]

112

Abstract (250 words)

Ivermectin (IVM) is a widely-used anthelmintic that works by binding to and activating glutamate- gated chloride channel receptors (GluClRs) in nematodes. The resulting chloride flux inhibits the pharyngeal muscle cells and motor neurons of nematodes, causing death by paralysis or starvation.

IVM resistance is an emerging problem in many pest species, necessitating the development of novel drugs. However, drug optimisation requires a quantitative understanding of GluClR activation and modulation mechanisms. Here we investigated the biophysical properties of homomeric α (avr-

14b) GluClRs from the parasitic nematode, H. contortus, in the presence of glutamate and IVM. The receptor proved to be highly responsive to low nanomolar concentrations of both compounds.

Analysis of single receptor activations demonstrated that the GluClR oscillates between multiple functional states upon the binding of either ligand. The G36’A mutation in the third transmembrane domain, which was previously thought to hinder access of IVM to its binding site, was found to decrease the duration of active periods and increase receptor desensitisation. On an ensemble macropatch level the mutation gave rise to enhanced current decay and desensitisation rates.

Because these responses were common to both glutamate and IVM, and were observed under conditions where agonist binding sites were likely saturated, we infer that G36’A affects the intrinsic properties of the receptor with no specific effect on IVM binding mechanisms. These unexpected results provide new insights into the activation and modulatory mechanisms of the H. contortus

GluClRs and provide a mechanistic framework upon which the actions of drugs can be reliably interpreted.

113

Author Summary (165 words)

IVM is a gold standard anti-parasitic drug that is used extensively to control invertebrate parasites pest species. The drug targets the glutamate-gated chloride channel receptor (GluClR) found on neurons and muscle cells of these organisms, causing paralysis and death. However, IVM resistance is becoming a serious problem in human and animal health, as well as human food production. We provide the first comprehensive investigation of the functional properties of the GluClR of H. contortus, which is a major parasite in grazing animals, such as sheep and goats. We compared glutamate and IVM induced activity of the wild-type and a mutant GluClR, G36’A, that markedly reduces IVM sensitivity in wild populations of pests. Our data demonstrate that the mutation reduces IVM sensitivity by altering the functional properties of the GluClR rather than specifically affecting the binding of IVM, even though the mutation occurs at the IVM binding site. This study provides a mechanistic framework upon which the actions of new candidate anthelmintic drugs can be interpreted.

114

Introduction

Pentameric ligand gated ion channels (pLGICs) are membrane-bound receptors that facilitate the diffusion of ions across cell membranes in response to the binding of agonists. The glutamate-gated chloride channel receptor (GluClR), first identified in arthropods, such as insects and crustaceans [1-

3], is an anion-selective pLGIC found at neuronal and neuromuscular inhibitory synapses [4]. GluClRs are also present in other major metazoan phyla, including platyhelminths and nematodes [4], but have not yet been identified in vertebrates. GluClRs can exist as homo- or hetero-pentamers [5].

Crystal structures of the homomeric GluClR from the nematode, C. elegans, have recently been determined in ligand-bound [6] and apo [7] states.

The mechanism of agonist activation has been studied extensively in vertebrate pLGIC members, such as the glycine (GlyR) [8], acetylcholine (AChR) [9-11], serotonin (5-HT3R) [12] and

GABAA (GABAAR) [13-15] receptors, as well as ELIC, which is a bacterial pLGIC [16]. A detailed study of the biophysical properties of GluClR activation has not been undertaken, even though GluClRs constitute a major group of pLGICs, many organisms that express them are serious parasitic pests, or vectors for disease transmission and they are a major target for anthelminthic drugs. For instance,

O. volvulus and W. bancrofti are nematodes that cause river blindness (onchocerciasis) and elephantiasis (lymphatic filariasis), respectively, in humans. Another nematode, H. contortus [17] is a serious pathogen in ruminant agricultural animals such as cattle, sheep and goats. The sea lice

(arthropod) species C. rogercresseyi [18] and L. salmonis [19], ravage salmon and trout farming industries worldwide. The cereal cyst nematode H. avenae devastates broad acre cereal crops across temperate wheat-producing regions of the world [20, 21]. A. gambiae is the mosquito vector that transmits malaria in over 90% of world-wide cases [22]. Macrocyclic lactones (MLs) such as ivermectin (IVM), moxidectin, abamectin and emamectin are widely used to control all of these, as well as many other, nematode and arthropod pests [23]. IVM works by activating GluClRs in

115 pharyngeal muscle cells and motor neurons of these organisms, thereby causing death by flaccid paralysis or starvation [24]. Unfortunately, however, IVM resistance is emerging as a serious problem in many pest species [21, 25-28] prompting the need for new generation treatments.

Functional and crystallographic studies have recently delineated the binding pocket of IVM and identified potential residues that IVM interacts with [6, 29, 30]. The main structure of the pocket is formed by first (TM1) and third (TM3) transmembrane domains of adjacent receptor subunits, at the level of the upper leaflet of the cell membrane [6]. Site-directed mutagenesis of transmembrane domains has identified critical residues that drastically affect IVM potency in the avr-14b subunit of

H. contortus, such as TM3-G36’ [29] and TM1-P230 [30], and in the α subunit of C. elegans, such as

TM1-L279 and TM1-F276 [31, 32]. The glutamate binding site and TM3 domain are also sites that harbour mutations in wild ML-resistant strains of C. elegans [26], whereas ML resistance in wild isolates of pest species have been attributed to mutations at, TM3-30’ in P. xylostella [33] and TM3-

36’ in T. urticae [34, 35]. Of particular note, a Gly at the 36’ position is thought to be essential for exquisite IVM [29] and abamectin [33] sensitivity, and larger substitutions at this location were proposed to reduce ML sensitivity by hindering access to its binding site [30, 33, 36]. However, the effects of these mutations are generally evaluated using functional assays that lack the resolution needed to distinguish discrete functional states in the activation process. A detailed mechanistic understanding of how wild-type and mutated receptors respond to glutamate is a prerequisite to understanding how IVM and other modulating ligands affect the receptor. This aim is best achieved through the study of single channel currents mediated by individual receptors [37]. Without a quantitative understanding of activation and modulation mechanisms of the receptor, attempts to design drugs with higher potency and selectivity for the receptor would be intractable.

Four GluClR subunits have been identified in H. contortus [α3A (avr-14a), α3B (avr-14b), GLC-

3, GLC-5, GLC-6, β and α], all of which express on motor neuron commissures [38-41]. However, the

116 native stoichiometric combinations of these subunits is unknown [4]. Here we investigated homomeric receptors comprising the avr-14b subunit, which is also expressed in pharyngeal neurons [38, 39]. We will refer to this subunit as α (avr-14b). In heterologous expression systems, homomeric receptors comprising either α (avr-14b) [29, 42] or α subunits form high affinity IVM binding sites, whereas the β subunit homomers do not [43].

In this study we investigated the biophysical properties of the homomeric α (avr-14b) GluClR from H. contortus as: 1) H. contortus is a major parasitic pest of domestic ruminant animals, 2) IVM is used widely to control H. contortus, 3) IVM resistance has emerged as a major problem in this species [44], and 4) GluClRs comprising or containing the α (avr-14b) subunit are most likely the major biological IVM target in this species [4, 24]. Here we sought to quantify the activation properties of this receptor in the presence of glutamate and IVM, and to explore the mechanism by which the TM3-G36’A mutation reduces IVM sensitivity to a level that is similar to vertebrate GlyRs and GABAARs [29, 45-47].

117

Results

Single channel conductance and current-voltage (i-V) relations of (avr-14b) GluClRs

Single receptor currents (Fig. 1A) were measured and plotted as a function of voltage (Fig. 1B) to determine the single channel conductance of the receptor. Using equation 1 and a mean current amplitude of 1.80 ± 0.03 pA (n = 7, at −70 mV), the estimated single channel conductance of the homomeric GluClR was 22.9 ± 0.3 pS. The i-V was nearly linear (Fig. 1B). The slight inward rectification and relatively small conductance of the homomeric GluClR is very similar to that determined for ternary GABAARs containing, α, β and γ subunits [13, 14]. A recent study has also estimated the current amplitude of the heteromeric GluClR of C. elegans at  90 mV to be ∼ 1.9 pA

[31].

Glutamate concentration-dependent properties of wild-type GluClRs

Single channel activity was recorded in the presence of a broad range of L-glutamate concentrations

(10 mM – 5 nM) to determine the receptor’s sensitivity to glutamate, the active durations of single receptors, the total time spent in conducting configurations (PO) and the shut and open dwell characteristics within each active period. Continuous sweeps of single channel activity, recorded from a patch in the presence of 200 µM glutamate is shown in Fig. 1C. At this and higher concentrations the activity of single receptors occurred as clearly defined periods of variable duration, termed ‘activations’, where the receptor oscillated between conducting and non- conducting configurations. These active periods were interrupted by relatively long intervals of inactivity where the receptor adopted desensitised states. These states are distinct from ligand- bound shut states both structurally [48, 49] and functionally [50]. With few exceptions, desensitised states are much longer-lived than shut states. Mean dwell times of the shut and open durations within activations were generated by plotting histograms and fitting these to mixtures of exponentials (Fig. 1D). We employ widely used and up-to-date terminology here, and it is virtually 118 identical to the nomenclature as used by Changeux in his later publications. The shut dwell data were best described by two exponential components, whereas the open dwell histogram was best fit to three exponential components. The dwell time constants were similar when the receptors were exposed to 10 mM and 1 mM glutamate, but at 200 µM the time constant of the longer shut component increased (S1 Table).

Reducing the glutamate concentration to 30 µM resulted in similar single channel activity

(Fig. 1E). The number of components and the time constants of both dwell histograms (Fig. 1F), appeared little changed, except for a further increase in the time constant of the longer shut component (S1 Table). In addition, the mean duration of the active periods appeared to become shorter as glutamate concentration decreased (S2 Table).

At low glutamate concentrations there appeared to be a transition from mostly tightly grouped to loosely grouped periods of activity and isolated open-shut events. For example, 2 µM glutamate elicited activity that comprised a mixture of isolated open-shut events and activations consisting of openings and shuttings in rapid succession, as with the higher glutamate concentrations. However, these latter more complex activations were more likely to occur in shorter bursts (Fig. 2A). The dwell histograms also exhibited two shut and three open components with similar time constants to those for the higher concentrations of glutamate, but the time constant of the longer shut component continued to increase and the fraction of the longest open time constant diminished (Fig. 2B, S1 Table). 30 nM glutamate elicited activations that occurred as bursts of loosely spaced openings and brief open-shut events (Fig. 2C). Moreover, long stretches of record corresponding to receptor desensitisation were mostly absent. The dwell histograms revealed changes to both shut and open components. Here both shut components increased and the longest open component disappeared (Fig. 2D, S1 Table).

119

As 30 nM, glutamate was effective at eliciting single channel activity the concentration was lowered even further, to 5 nM. Remarkably, even at this concentration GluClRs were activated. Most of the activity comprised simple shut-open events, but the occasional activation of loosely spaced openings was also apparent. In contrast, in the absence of glutamate, receptor openings were extremely rare, brief and essentially negligible (Fig. 3A). These data show that 1) the homomeric

GluClRs are exquisitely sensitive to glutamate and 2) from 2 µM glutamate and below, the activity becomes increasingly simple and brief, likely reflecting an effect consistent with agonist dissociation from partially liganded receptors. The dwell histograms derived from 30 nM and 5 nM glutamate showed distinct differences compared to those of higher concentrations. Here both the longer and briefer shut components increased (Fig. 3B) and the third, longest open component was absent, whereas the remaining two open components decreased (Fig. 3C, S1 Table). The invariant open dwell components for concentrations ≥ 2 µM glutamate are consistent with an optimal degree of ligation for receptor activation, as has been shown for the GlyR [8]. In contrast, the decrease in the remaining two open component time constants at nanomolar concentrations of glutamate is consistent with sub-optimal activation of receptors. We infer that at 30 nM and 5 nM each receptor is bound to fewer ligand molecules than at the higher concentrations, giving rise to openings with briefer lifetimes [51]. We also infer that the steadily increasing longer shut component at ≤ 200 µM glutamate is additional evidence that the receptors are able to activate without all glutamate binding sites being occupied [8].

The total time spent in open states was also compared across glutamate concentrations. This analysis demonstrates that PO increases as a function of glutamate concentration (Fig. 3D).

Consistent with the inference that the receptors are highly sensitive to glutamate, the PO at glutamate concentrations ≥ 10 µM were all > 0.90. PO showed a significant decrease at 2 µM glutamate and dropped to 0.21 at 30 nM and 0.14 at 5 nM (Fig. 3D, S2 Table). A Hill fit to the PO plot

120 revealed a maximum of 0.99, an EC50 of 70 nM and a Hill coefficient of 0.82. The mean duration of activations was also plotted and showed that active durations declined from ∼500 ms to ∼330 ms between 10 mM and 30 µM glutamate. Fitting the data to a Hill equation produced an EC50 of 31

µM, a Hill coefficient of 0.56, a maximum duration of 500 ms and a minimum of ∼80 ms (Fig. 3E).

Macropatch recordings of glutamate-gated currents

The activation properties of homomeric GluClRs were also investigated at an ensemble current level using rapid solution exchange [14, 52] of glutamate onto macropatches. As GluClRs are located at inhibitory synapses, these experiments were carried out to mimic synaptic activation conditions by determining the response of many (~20-100) receptors to rapid glutamate exposure. By avoiding the distorting effects of receptor desensitisation encountered with slower agonist application methods, rapid solution exchange techniques also establish a more accurate ligand concentration − peak current relationship. Peak current was achieved by rapidly applying glutamate for either 50 ms

(5 mM – 20 µM) or 500 ms (10 µM – 0.5 µM, Fig. 4A). These data were fitted to a Hill equation, yielding an EC50 for glutamate of 43 μM and a Hill slope of 0.8 (Fig. 4B). Whole-cell experiments on the same GluClR produced an EC50 for glutamate of ∼15 μM and a Hill slope of ∼1.7 [29]. The 2-3 fold difference in EC50 and Hill slope between whole-cell and macropatch data are consistent with open and desensitised states, which have a higher affinity for ligand, having made a significant contribution to the whole-cell data.

Similar experiments were carried out to determine the relationship between the activation rate of the current and agonist concentration. Normalised examples of these recordings are illustrated in Fig. 4C and the group data are summarised in Fig. 4D. A Hill fit to this plot produced an

−1 EC50 of 0.95 mM and a Hill slope of 1.0. The upper asymptote of the activation plot was ~9000 s , representing the maximum activation rate [52, 53], whereas the lower level was ~10 s−1.

Single channel properties of the G36’A mutation 121

Homomeric GluClRs containing the G36’A mutation exhibit a markedly reduced IVM sensitivity

(EC50) when recorded in whole-cell configuration [29]. However, any changes to the intrinsic properties of the receptor conferred by the mutation have yet to be examined in mechanistic detail.

G36’A-containing receptors were first examined on a single channel level. Applied glutamate elicited a similar current amplitude to wild-type receptors, suggesting the G36’A had no appreciable effect on channel conductance (Fig. 5A). A current amplitude of 1.81 ± 0.02 pA (n = 7) for the mutant at −70 mV yielded a conductance of 23.0 ± 0.2 pS, if it is assumed that under the same recording conditions the reversal potential is similar to that for wild-type.

However, moderate to high (30 µM – 10 mM) concentrations of glutamate revealed two distinct types of activations in the mutant receptor (Fig. 5A), whereas the same concentration of glutamate elicited homogeneous activations in the wild-type receptor (Fig. 5B). The two activation modes mediated by the mutant GluClR were quantified on the basis of PO and duration only for 1 mM glutamate. The analysis revealed a very low PO activation mode of 0.14 ± 0.02 (n = 6) and mean active periods of 1333 ± 222 ms duration and a higher, more wild-type like mode with a PO of 0.87

± 0.05 and a mean active duration of 309 ± 62 ms (Fig. 5C). In contrast, 1 mM glutamate produced a single PO of 0.99 (S2 Table) in the wild-type receptor consistent with fewer shuttings within each activation (Fig. 5D). The two activation modes observed in the mutant receptors were pooled for further analysis for all concentrations where they were apparent so as to determine the net effect of the mutation on PO and active durations, and facilitate a more direct comparison to wild-type receptors. The dwell histograms for the mutant receptor at 1 mM glutamate required two shut and three open components (Fig. 5E), but the longer shut component was substantially increased compared to wild-type receptors (Fig. 5F) and the two longest open components were reduced (S1

Table).

122

Two distinct gating modes were also observed at a moderate (30 µM) glutamate concentration (Fig. 6A), but were difficult to distinguish at a low (2 µM) concentration because the activations became too brief and simple (Fig. 6B). Over the concentration range tested, the mean duration of activations of the mutant receptor were considerably shorter than wild-type with a maximum mean duration of 200 ms, as was the mean PO, which peaked at 0.73 (Fig. 6C, S2 Table).

Figure 6D summarises the dwell component data over the glutamate concentrations that were tested on the mutant receptors. Consistent differences to wild-type receptors include an increase in the long shut component and briefer open components (Fig. 6E). At 2 µM glutamate only one shut component and two open components were resolvable (Fig. 6F, S1 Table).

The G36’A mutation induces faster desensitisation

The briefer active periods exhibited by the G36’A mutant receptors is indicative of accelerated ensemble current deactivation [14, 54] and desensitisation [50]. To investigate whether receptor desensitisation was affected by the G36’A mutation, the long quiescent periods corresponding to desensitisation in single channel records were quantified, then corrected for channel number [50].

Sample recordings for wild-type and mutant receptors are shown in Fig. 7A and 7B, respectively. A saturating concentration of glutamate (10 or 1 mM) was first rapidly applied onto each patch, ensuring that all the receptors in the patch were activated, after which constant agonist perfusion was maintained over the patch for the remainder of the recording. Clearly defined steps corresponding to the single channel amplitude (~2 pA) became apparent as all the receptors desensitised back to baseline. The number of steps was then taken as an estimate of the total number of receptors contained in each recorded patch. Only patches expressing 1-10 steps

(channels) were accepted for analysis.

The long desensitised periods were estimated by plotting shut dwell histograms for the entire record, as is illustrated in Figs 7C and D. The shut events could be divided into two broad

123 components. The briefer component corresponded to shut events within active periods. This component could be subdivided into briefer components (as in Figs, 1, 2, 4 and 5). The longer component represented the mean desensitised lifetime, and it was this component that was corrected for channel number. This method of analysis produced a mean desensitised lifetime for wild-type receptors of 91 s (n = 5), and was used to determine a re-sensitisation transition rate constant (ω, Fig. 7E) of 0.011 s−1. Similarly, the desensitisation rate constant (δ) was estimated from the mean duration of the active periods at saturating glutamate concentrations (500 ms, Fig. 3B) to be 2.00 s−1 [50]. This produced an equilibrium constant (δ/ω) for desensitisation of 182 for wild-type receptors. A similar analysis for the G36’A mutant receptor produced a mean desensitised lifetime of 125 s (n = 8) and an ω of 0.008 s−1. However, a more significant change was estimated for the mean duration of active periods for the mutant, which saturated at 200 ms (S2 Table), producing a

δ value of 5.00 s−1 and an equilibrium constant of 625. Thus, mutant receptors desensitised ~3.4 times more rapidly than wild-type receptors. From this analysis it can be inferred that the G36’A mutation increases the likelihood of the receptors entering desensitised states.

The G36’A mutation induces faster deactivation and desensitisation in ensemble currents

To determine if the estimates of receptor desensitisation reflected current decay and desensitisation in ensemble currents, macropatches expressing wild-type or G36’A mutant GluClRs were exposed to a saturating (3 mM) concentration of glutamate for either ~1 ms or 500 ms. In response to a ~1 ms application, the deactivation phase of macropatch currents mediated by wild- type GluClRs was adequately described by two standard exponential functions with a weighted time constant of 67 ± 4 ms (Fig. 8A, B). The individual time constants (and fractions) are tabulated in

Table 2. To allow comparison with other, better characterised pLGICs, heteromeric α1β GlyRs and

α5β3γ2L GABAARs were tested under similar conditions. A ~1 ms pulse of 3 mM glycine applied to

124

α1β GlyRs elicited macropatch currents that also exhibited a two component decay phase with a weighted mean of time constant of 22 ± 3 ms (Fig. 8A). In contrast, a ~1 ms pulse of 3 mM GABA applied to α5β3γ2L GABAARs activated macropatch currents that decayed considerably more slowly than those of GluClRs. They also deactivated with two components, with a weighted time constant of 275 ± 2 ms (Fig. 8A, B, Table 2). 3 mM glutamate was also rapidly applied for ~1 ms onto patches expressing the G36’A mutant GluClR (Fig. 8C). The weighted time constant from a two component fit was 11 ± 1 ms, which was 2-fold faster than those of the α1β GlyR and over 6-fold faster than the wild-type GluClR (Fig. 8B).

Table 1. Macropatch deactivation and desensitisation

Deactivation, 1 ms, 3 mM agonist (glutamate, glycine or GABA)

Receptor τ1 (ms) A1 τ2 (ms) A2 τw (ms) τact (ms) n

Wild-type 90 ± 6 0.64 ± 0.04 20 ± 4 0.35 ± 0.05 67 ± 4 0.153 ± 0.012 12 GluClR

G36’A 21 ± 3 0.32 ± 0.07 7.3 ± 0.9 0.68 ± 0.07 11 ± 1 0.195 ± 0.017 9 GluClR

α1β 42 ± 4 0.37 ± 0.04 9.3 ± 1.5 0.62 ± 0.04 22 ± 3 0.125 ± 0.016 6 GlyR

α5β3γ2L 275 ± 2 0.59 ± 0.02 41 ± 10 0.41 ± 0.02 181 ± 24 0.133 ± 0.020 10 GABAAR

Desensitisation, 500 ms, 3 mM glutamate

Receptor τ1 (ms) A1 τ2 (ms) A2 τw (ms) n

Wild-type 492 ± 38 − − − − 6 GluClR

G36’A 266 ± 28 0.92 ± 0.02 22 ± 6 0.08 ± 0.01 252 ± 26 10 GluClR

n represents the number of patches.

125

The activation phase of the currents was also measured by fitting 10-100% of the rising phase of the current to eqn. 2. The measurements, summarised in Fig. 8D, demonstrate that currents mediated by all receptors tested activate with similar time constants, which ranged between 0.1 − 0.2 ms. In contrast to the deactivation kinetics, the G36’A mutation had no significant effect on the ability of the receptor to activate upon exposure to glutamate.

Ensemble desensitisation was examined by rapidly applying 3 mM glutamate onto macropatches for a duration of 500 ms (Fig. 8E). Wild-type GluClRs desensitized with single time constant of 492 ± 38 ms, whereas the G36’A mutant receptor required two exponential functions to adequately describe the desensitisation phase of the current (Table 1). The weighted desensitisation time constant for the mutant receptor was 252 ± 26 ms (Fig. 8F). We infer that the number of components that were needed to describe single receptor and ensemble desensitisation is related to modal activation in the mutant receptor. Consistent with this inference, estimates of the mean active durations for both receptors at saturating glutamate match very closely with the time constants of ensemble desensitisation (Tables 1 and S2). Overall, these data demonstrate that the G36’A mutation abbreviates single channel active periods, which manifest as accelerated deactivation and desensitisation in ensemble currents. These alterations to the intrinsic activation properties of the receptor are likely the underlying reasons for the order of magnitude rightward shift in the whole-cell concentration-response relationship for glutamate, reducing its sensitivity

(EC50) from 15 µM to 154 µM. However, studies have also revealed a parallel shift in IVM sensitivity

(EC50), from 40 nM to 1.2 µM in the G36’A-containing receptor [29, 46].

Direct activation by IVM at wild-type and G36’A GluClRs

IVM is both a direct agonist and a potentiator of glutamate responses at the GluClR. In our final set of experiments we wished to test the hypothesis that the changes to the functional properties of 126 the receptors conferred by the G36’A substitution gives rise to the reduced sensitivity to IVM, as it does for glutamate. To test this idea, we recorded single channel currents in the presence of 5 nM

IVM alone (direct activation) or in 5 nM IVM + 2 µM glutamate (potentiation). In both experiment types the receptors opened to an amplitude of 1.8 pA (e.g., Fig. 9A), suggesting that the presence of IVM had little effect on the permeation pathway.

Wild-type receptors exhibited a substantial degree of potentiation and direct activation by

IVM. However, the recordings also revealed that these experiments were not ‘steady state’. We confined our analysis of the steady-state phase of both experiments types (direct activation and potentiation). When membrane patches expressing wild-type receptors were exposed to 5 nM IVM alone, no receptor activity was apparent for the first 41 ± 4 ms. After this initial silent period the activations were initially well separated, but increased in duration for 47 ± 25 s, after which the active durations reached a steady-state equilibrium of almost continuous activity (Fig. 9A) of all the receptors present in each patch (between 1-4 receptors). The mean active duration of the receptors at steady-state was 9.5 ± 2.6 s and had a PO of 0.65 ± 0.07 (Table 2). The shut intervals were best described by three components whereas the open interval histograms required four exponential components for fitting (S3 Table). The presence of additional shut and open components suggests that IVM alone induces activity of greater complexity or exposes state lifetimes that are not easily resolvable when glutamate is present. Receptor desensitisation by IVM alone had a mean lifetime of 536 ± 140 ms (ω = 1.87 s−1). A mean active duration of 9.5 ± 2.6 s (δ = 0.105 s−1) yielded an equilibrium constant of 0.06.

Direct activation of G36’A mutated receptors by 5 nM IVM produced a similar lag time before equilibrium was reached (Fig. 9C). At equilibrium the receptors were active for a mean duration of

46 ± 8 ms and a PO of 0.85 ± 0.01 (Fig. 9D, Table 2). These active periods were much briefer than wild-type receptors when activated by IVM directly (9.5 s). Dwell histograms revealed two shut and

127 three open components, which is less complex than wild-type (S3 Table). Moreover, the mutated receptors desensitised for a mean of 2004 ± 268 ms, yielding a desensitisation equilibrium constant

1 −1 −1 of 41.7 s (ω = 0.499 s and δ = 20.8 s ). The mean active durations and PO data are summarised as bar plots in Fig. 9 E and F, respectively.

Potentiation by IVM at wild-type and G36’A GluClRs

Wild-type receptors activated rapidly upon exposure to 5 nM IVM and 2 µM glutamate. For the first 66 ± 18 s after commencement of the recording, the active periods were well-separated increasing in duration over time (Fig. 10A), until an apparent steady-state equilibrium was reached

(Fig. 10B).

An analysis of the active durations, PO and the dwell time components at equilibrium produced a mean active duration of 15.9 ± 4.2 s and a PO of 0.88 ± 0.02 (Table 2) for potentiation of wild-type currents. The dwell histograms were best described by three shut and three open components (S3 Table). The time constants for the first two shut and open components were similar to those in the presence of low concentrations of glutamate (S2 Table). In contrast, the longest open component was at ~ 200 ms and represented about 40% of the total open intervals. To estimate receptor desensitisation, long stretches of record consisting of single receptor activity were analysed to obtain the main shut component that separated discrete active periods (Fig. 10B). This shut component produced a short-lived mean, after correcting for channel number, of 223 ± 36 ms and thus an ω of 4.48 s−1. Using a mean active duration of 15.9 s (δ = 0.063 s−1), the calculated equilibrium constant for receptor desensitisation was 0.01 in the presence of IVM and glutamate.

128

Table 2. IVM-dependent active durations and open probability (PO)

Active duration (ms) Po n

wild-type GluClR Direct activation 9453 ± 2556 0.65 ± 0.07 6

Potentiation 15940 ± 4210 0.88 ± 0.02 7

G36’A GluClR

Direct activation 48 ± 7* 0.82 ± 0.03* 7

Potentiation 113 ± 25* 0.60 ± 0.04* 6

n represents the number of patches. * p < 0.05 compared to the wild-type value.

Similar experiments were carried out for the G36’A-containing mutant. Here too the active periods initially increased in duration (Fig. 10C), before equilibrating to steady-state activity (Fig. 10D).

However, steady-state activity was not near continuous, as was observed for the wild-type receptors. Instead, individual receptors were active for a mean duration of 113 ± 25 ms and had a

PO of 0.60 ± 0.04 (Table 2). Receptor desensitisation was also unlike that of wild-type receptors. The mean shut lifetime for long stretches of record was 2381 ± 657 ms (ω of 0.420 s−1). This yielded an equilibrium constant for desensitisation of 21.1. As for glutamate-gated activity, IVM produced briefer active periods and induced greater desensitisation in the G36’A GluClRs than wild-type. The summary of the mean active durations and PO is provided in Fig. 10E and F, respectively. The POs were significantly different between direct activation and potentiation for both wild-type and mutant receptors. However, because direct activation by IVM of the mutant receptors produced brief, simple activations, the PO determined for this activity was relatively high.

129

In summary, IVM acted as an agonist and potentiated currents in the presence of glutamate at wild-type and G36’A mutated GluClRs to elicit significantly longer active periods and markedly reduce receptor desensitisation. In addition, the sparse activity at the start of the recordings, which equilibrated to steady-state activity implies that, as with glutamate, additional binding of IVM molecules to each receptor saturates receptor activation.

130

Discussion

Actions of glutamate

The two broad aims of this study were firstly, to examine the activation properties of GluClRs expressed by a parasitic species in the presence of its physiological agonist and secondly, to explore the mechanism of IVM sensitivity. To achieve the first aim glutamate-gated currents were examined over a wide concentration range on single receptor and ensemble levels. The conductance of the receptor channel was determined to be ~23 pS, which is close to that of GABAARs that comprise α,

β and γ subunits [13, 55]. Upon binding to glutamate, wild-type GluClRs activated rapidly (~9000 s−1,

Fig. 4), comparable with the rate of other pLGICs, including the G36’A mutated GluClRs (Fig. 8). The experiments also revealed that wild-type GluClRs were highly responsive even at low nanomolar concentrations of glutamate and exhibited active durations and an open probability that was concentration-dependent. These parameters saturated at ~500 ms and 0.99, respectively (Fig. 3).

Dwell interval analysis of active periods demonstrated that the receptors have multiple components, indicating that each receptor oscillates between multiple functional states during receptor activation [8, 11, 13]. The pattern of dwell components also indicated that at ≥ 2 µM glutamate an optimal number of bound glutamate molecules achieves efficient receptor activation.

This is similar to GlyR activation, whereby fitting data to postulated kinetic schemes it was deduced that three bound glycine molecules are sufficient for optimal activation [8]. The decrease in open dwell times at nanomolar concentrations of glutamate clearly showed that at these concentrations fewer glutamate molecules were bound on average to each receptor [51].

Effects of the G36’A mutation

The G36’D and G36’E mutations have been identified in the ML-resistant agricultural pest mite T. urticae [34, 35]. These mutations occur on different subunit isoforms, suggesting that heteromeric

131

GluClRs containing different substitutions to G36’ could work either individually or synergistically to reduce ML sensitivity [35, 56]. The G36’E mutation is particularly effective at reducing ML sensitivity on its own and homomeric receptors expressed in oocytes demonstrate complete insensitivity to two MLs (abamectin and milbemycine A4) [56]. Our data suggest that the G36’A mutation gives rise to significant functional changes, such as a reduced active duration and an increase in desensitisation of single receptors, which manifest as faster current decay and reduced sensitivity to glutamate and IVM. Whether these functional changes are also present in G36’D or E has yet to be determined. However, given that both substitutions contribute large side groups that are likely negatively charged, it is likely that these too would affect the activation properties of the receptors.

The physico-chemical properties of aspartate and glutamate may also restrict access of IVM to its binding site.

We chose to study the G36’A mutation because it dramatically decreases IVM sensitivity [29] and is located on the TM3 domain where crystallographic data show that it contributes one side of the IVM binding site [6] (Fig. 11). Given its location, it is tempting to hypothesise that the G36’A substitution reduces IVM sensitivity simply by disrupting the binding of IVM. However, the mutation also decreases the EC50 of glutamate [29], which binds at an extracellular domain site that is over 3 nm from the site of the mutation. Another mechanism that could reconcile the parallel decrease in glutamate and IVM sensitivities is that the actions of both ligands reveal changes to the intrinsic activation properties of the receptor conferred by the mutation. Distinguishing between these two possibilities is critical to understanding the mechanism of action of IVM. This is of particular importance given that IVM resistance in H. contortus and other pest species is an emerging concern

[21, 25, 28].

To help distinguish between these two possibilities we analysed glutamate- and IVM-gated currents in wild-type and G36’A mutated receptors. Clear evidence that the G36’A mutation

132 markedly compromised receptor activation was gleaned in the presence of glutamate alone. The mutation gave rise to two distinct and stable modes of activation; one that was similar to wild-type, and another with a much reduced PO (Figs. 5 and 6). The wild-type like mode was briefer than the activations mediated by wild-type receptors over the glutamate concentrations tested and both modes had lower POs than wild-type. When analysed together, the net effect of these modes produced a maximum mean active duration of ~200 ms and a PO of 0.70 (Fig. 6). These parameters underlie the reduced sensitivity to glutamate observed in the G36’A mutated receptors. Indeed, where 2 µM glutamate elicited robust activity in wild-type receptors, it produced only sparse, simple activity in the mutant. These results led to the hypothesis that the mutation impaired receptor desensitisation and ensemble current decay. This was tested at the single receptor level (Fig. 7) and in macropatches (Fig. 8). The single receptor experiments yielded desensitisation equilibrium constants of ~180 and ~625 for wild-type and mutant receptors, respectively, representing a 3.4- fold greater likelihood of adopting desensitised states in the mutant. Ensemble deactivation and desensitisation rates were also much abbreviated in the mutant, producing mean time constants that corresponded well to the mean active durations of single receptors (Fig. 8). It is notable that other pLGICs, such as α1β GlyRs and α1β2γ2 GABAARs have a similar sensitivity to IVM [29, 45] and also exhibit similar rates of current decay [14, 52, 54] to the G36’A mutant. Moreover, pLGICs that exhibit low IVM sensitivity also contribute non-G36’-containing TM3 domains to their IVM binding sites [29, 36, 45].

Actions of IVM

IVM acted as a ligand on its own and synergistically with glutamate to enhance currents elicited by glutamate. It did not affect the single channel conductance even though it binds at a site within the transmembrane segments and is predicted to interact with the pore-lining TM2 domain [6]. At

133

GABAARs it has been demonstrated that 10 nM IVM alone lengthens the durations of single receptor currents without changing single channel conductance [47].

2 µM glutamate applied alone at wild-type GluClRs gave rise to a mean active duration of

~150 ms. When 5 nM IVM and 2 µM glutamate were applied together, current potentiation manifested as prolonged active durations with a mean of ~16 s, representing a two order of magnitude increase. The same combination of IVM and glutamate at G36’A mutated receptors produced active durations with a mean of 113 ms (Fig. 10), compared to a mean of ~11 ms elicited by 2 µM glutamate alone. This also represents an order of magnitude change, but the absolute durations were much briefer than in wild-type receptors. A similar pattern was observed between wild-type and G36’A receptors in the presence of 5 nM IVM alone. The mean duration of active periods for wild-type receptors was ~9.5 s, whereas that for the mutant was a mere 48 ms (Fig. 9).

As for glutamate-gated currents, the active periods of the G36’A mutated receptor were much briefer when activated by IVM alone or in conjunction with glutamate compared to wild-type receptors.

Receptor desensitisation in the presence of IVM was estimated by fitting shut histograms to long periods of record that contained successive single receptor activations (Figs. 9 and 10).

Receptor saturation, where all the receptors in each patch became active, was then used to count active receptors and correct for the desensitisation time constant. This analysis revealed that desensitisation was nearly abolished at wild-type receptors, especially in the presence of IVM and glutamate. The mean lifetimes of desensitised states were between ~220 ms and ~540 ms and yielded equilibrium constants of ~0.01 for IVM plus glutamate and ~0.06 for IVM alone, respectively.

IVM alone induced a mean desensitisation lifetime of 2002 ms and an equilibrium constant of ~42 in the mutant receptors. This represents a significant increase in desensitisation compared to wild- type receptors under the same recording conditions. These data demonstrate that in the presence

134 of each agonist alone and when they are co-applied, the G36’A mutated receptors exhibited briefer active periods and enhanced desensitisation compared to wild-type. Our data strongly support the inference that the loss of sensitivity reported for both agonists [29] is due to the same mechanistic process, and not fundamentally related to IVM binding interactions at the 36’ position. Although we cannot categorically rule out an IVM binding effect our data show that the wild-type and the G36’A mutant receptors are similarly affected even when receptor activation is at saturation throughout the recording. These conditions also correspond to ligand saturation where occupancy of receptors in unbound states is negligible. That this is the case for glutamate (Figs 5 and 6) and IVM (Figs 9 and

10) strongly suggests that both agonists are less efficacious at activating the mutant receptors.

Comparison of glutamate and IVM activation mechanisms

A notable difference between the actions of glutamate and IVM was that the onset and equilibration of currents in the presence of IVM were much slower than observed for glutamate. A lag time of over ~11.5 minutes was apparent between the initiation of channel activity and the time when activations equilibrated to a constant mean duration for both mutant and wild-type receptors.

Indeed, no activity was seen when IVM was applied alone for the first minute or so. Diffusion limited binding rates, calculated for ligands that encounter receptor binding sites directly from aqueous solution, including ligands of similar dimensions to IVM are in the range of ~57 x 109 M1s1 [57].

For instance, the upper limit of the diffusion rate for a small aqueous molecule like glutamate is ~109

M1s1 [58]. The binding energy and correlated structural changes at binding sites can reduce these values by about two orders of magnitude (~106108 M1s1)[57]. These diffusion rates are far too high to account for the lag time observed in the recordings, suggesting the existence of other rate- limiting factors [59]. Structural evidence indicates that IVM binds to an inter-subunit cavity in the upper leaflet of the lipid bilayer [6], as do other highly lipophilic ligands such as neurosteroids [60] and anaesthetics [61]. The IVM binding pocket in GluClRs is likely to be partly occupied by lipid,

135 requiring its displacement by IVM for access to the pocket [6, 7]. Due to its lipophilic nature, IVM is believed to partition into cell membranes [62] where it reaches a high local concentration, consistent with persistent whole-cell currents [29]. Thus, much of the ‘binding energy’ of IVM could be derived from the nonspecific free energy of membrane partitioning, giving rise to a high apparent affinity, whereas the actual ligand-channel interaction could be relatively weak [63]. Our data are consistent with IVM partitioning in the lipid membrane and diffusing to its binding pocket [64], where its concentration would increase to produce current saturation over the course of several minutes in patches of membrane. The increase in the active durations of individual receptors over the initial phase of the recordings and the emergence of a long open time constant at saturation also suggests that multiple IVM molecules bind to each receptor over course of the experiment to produce saturation. Heteromeric α1β2γ2 GABAARs have also been shown to bind multiple IVM molecules, to produce interface-specific potentiation and direct current activation [45].

Structural mechanism of the G36’A mutation

It has been suggested that the flexible ‘hinge’ function of glycine residues found within K+ channels

[65, 66] and pLGICs [67] can serve to isolate protein segments, or even entire domains, from surrounding protein conformational changes during channel activation [66]. According to their respective high resolution molecular structures, the TM3 domain backbones of the α1 GlyR (which contains an endogenous A36’ residue) and the GluClR are closely aligned in the shut state (Fig. 11A).

However, upon IVM binding, the GlyR TM3 undergoes a larger displacement (Fig. 11B). This differential displacement is also observed when the TM3 domains corresponding to shut and IVM- bound states are overlaid from the same receptor (Fig. 11C, D). This strongly suggests that the A36’ residue confers structural rigidity to the TM3. The structural comparisons in Fig. 11 illustrate that the G36’ acts to minimise deformation of the TM3 between state transitions during the conformational activation ‘wave’ of pLGICs [68]. Because the G36’A mutation causes briefer active

136 durations and an increased likelihood of adopting glutamate and IVM-induced desensitised states we conclude that the alanine destabilises open states via reduced backbone flexibility and a larger

TM3 displacement. Functional studies have established that pLGIC activation and desensitization are mediated by structurally distinct sets of conformational changes at the both extracellular- transmembrane domain interface [49, 50] and at the intracellular end of the pore [48]. The difference in IVM-induced TM3 displacement in the wild-type and G36’A mutant GluClRs will cause

TM3 to interact differentially with one or both of these regions, and could thus explain the differential effect of the mutation on desensitization.

Conclusion

The H. contortus α (avr-14b) GluClR is an important biological target for IVM, although IVM resistance is emerging as a problem in this pest species. Here we quantified the effects of glutamate and IVM on these receptors with the aim of understanding the structural and functional bases of their modulatory effects. We found the receptor to be highly responsive to low nanomolar concentrations of both ligands. Dwell interval analysis of active periods demonstrated that the receptor oscillates between multiple functional states during activation by either ligand. However, we also observed that the duration of activations increased with increasing ligation of receptors by either ligand. The G36’A mutation, which was previously thought to hinder access of IVM to its binding site on the receptor, was found to decrease the duration of active periods and increase receptor desensitisation. On an ensemble macropatch level these changes gave rise to enhanced current decay and desensitisation rates. There are two main reasons why we consider these effects are due to impaired channel gating and not impaired IVM binding. First, the impairment to gating was quantitatively similar for the two ligands which bind to structurally distinct sites, and second, the impairment was observed at saturating concentrations of either ligand, thus ruling out a contribution to gating from binding and unbinding events. We infer that G36’A affects the intrinsic

137 properties of the receptor with no specific effect on IVM binding. These results provide new insights into the activation and modulatory mechanism of the GluClR and provide a mechanistic framework upon which the actions of new candidate anthelmintic drugs can be reliably interpreted.

138

Materials and Methods

Cell Culture

HEK AD293 cells (ATCC cell lines, VA USA) were seeded onto poly-D-lysine coated glass coverslips and transfected with cDNAs encoding the GluClR subunit avr-14B (pcDNA 3.1+) of H. contortus using a calcium phosphate-DNA co-precipitate method. The cDNA encoding the CD4 surface antigen was also added to the transfection mixture and acted as a marker of transfected cells. Cells were used for experiments 2-3 days after transfection. The point mutation, TM3-G36’A, was incorporated into the subunit using the QuickChange site-directed mutagenesis method. Successful incorporation of mutation was confirmed by sequencing the mutated DNA.

Patch clamp electrophysiology

KCl, 1 MgCl2, 2 CaCl2, 10 HEPES, and 10 D-glucose and titrated to pH 7.4. The adjacent barrel contained agonist dissolved in this extracellular solution. Glass electrodes were pulled from borosilicate glass (G150F-3; Warner Instruments), coated with a silicone elastomer (Sylgard-184;

Dow Corning) and heat-polished to a final tip resistance of 415 MΩ when filled with an intracellular solution containing (in mM) 145 CsCl, 2 MgCl2, 2 CaCl2, 10 HEPES, and 5 EGTA, pH 7.4. Stock solutions of L-glutamate were also pH-adjusted to 7.4 with NaOH. A 10 mM stock of IVM (Sigma-Aldrich) was dissolved in 100% DMSO and kept frozen at 20oC. Fresh working stocks of IVM at 5 nM were prepared by dissolving the appropriate quantity directly in extracellular solution. 100% DMSO when dissolved in extracellular solution alone at the same concentration as is present in working solutions

139 containing 5 nM IVM had no effect on patches excised from cells transfected with GluClRs or from untransfected cells.

Excised patches were directly perfused with extracellular solution by placing them in front of one barrel of the double-barrelled glass tube. Single channel currents were elicited by exposing the patch continuously to agonist containing solution, flowing through the adjacent barrel. 1-2 glutamate concentrations were applied to most patches for single receptor experiments. A ~1 minute wash with agonist-free extracellular solution was applied between each glutamate application. Because IVM does not readily wash out, either 2 µM glutamate + 5 nM IVM or 5 nM

IVM alone were applied to a given patch. Macropatch currents were elicited by lateral translation of the tube from the agonist free to agonist containing barrel using a piezo-electric stepper

(Siskiyou). This achieved rapid solution exchange (<1 ms). Currents were recorded using an Axopatch

200B amplifier (Molecular Devices), filtered at 5 kHz and digitized at 20 kHz using Clampex (pClamp

10 suite, Molecular Devices) via a Digidata 1440A digitizer.

Data Analysis

The experiments that were carried out can be broadly divided into 1) single receptor currents at steady-state and 2) ensemble currents, which are phasic. The two types or experiments are complimentary and provide different data. Single channel recordings yield information on receptor conductance and functional state complexity (eg. active durations and dwell histograms). The fast application (~1 ms) ensemble measurements mimic synaptic currents.

Single-channel current amplitudes were measured in Clampfit. In current-voltage (i-V) experiments, the amplitude was measured at voltages of, ±70 mV, ±35 mV, ±15 mV and 0 mV. The data were fit to a polynomial function in Sigmaplot (Systat Software) and the reversal potential was read directly from the plots. Single-channel conductance (γ) was calculated from the single-channel amplitude (i) using Ohm’s law: 140

= eqn. 1 𝑖𝑖 𝛾𝛾 𝑉𝑉ℎ𝑜𝑜𝑜𝑜𝑜𝑜−𝑉𝑉𝑙𝑙𝑙𝑙𝑙𝑙−𝑉𝑉𝑟𝑟𝑟𝑟𝑟𝑟 Where Vhold is the holding potential (−70mV), Vljp is the liquid junction potential and Vrev is the reversal potential. Vljp was calculated to be 4.7 mV for the solutions used in the experiments [69].

We confined our analysis to the largest, main conductance level. QuB software was used to analyse the kinetic properties of GluClR activations. Segments of single-channel activity separated by long periods of baseline were selected by eye and idealized into noise-free open and shut events using a temporal resolution of 70 μs. Idealized data were initially fit with a simple activation scheme in which open and shut states were added to a central shut state. This fit was used to determine the critical time (tcrit), which was taken from the shut interval durations and used to divide the idealized segments into clusters (or bursts at < 2 μM glycine) of single receptor activity. Clusters and bursts will be referred to as activations. tcrit applied to single channel records of wild-type activity varied between 5-30 ms for concentrations ≥ 2 µM and 120-180 ms for 30 nM and 5 nM glutamate.

Activation mode analysis for G36’A-containing receptors at 1 mM glutamate required tcrit times of

180-200 ms (low PO) or 15-50 ms (high PO). Pooled data obtained from G36’A-containing receptors were defined using tcrit times of 20-50 ms. Finally, IVM-induced single channel currents were defined using tcrit values of 50-120 ms for both wild-type and mutant receptors. This analysis yielded mean cluster durations and intra-activation open probabilities (PO). All data are presented as mean ± SEM of between 3 and 16 patches. The shut periods that correspond to receptor desensitisation were estimated by generating shut histograms for long stretches of record (several minutes) that exhibited single receptor activations. Receptor desensitisation was modelled as a single transition from an open conducting state (ARo) to a desensitised state (ARd). Where A is the agonist, R is the receptor and the superscripts denote open (o) or desensitised (d). δ denotes the desensitisation rate constant whereas the re-sensitisation (or re-activation) rate constant is denoted by ω. The equilibrium constant for desensitisation is δ/ω.

141

Macropatch currents were analysed by fitting the onset phase of the current to a single exponential of the form:

( ) = (1 ) eqn. 2 −𝑘𝑘𝑜𝑜𝑜𝑜𝑜𝑜𝑡𝑡 𝐼𝐼 𝑡𝑡 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 − 𝑒𝑒 Where I(t) is the current at time t, Imax is the maximum current amplitude and kobs is the pseudo-first order rate constant for current activation. The decay phase of the macropatch currents were fit to two standard exponential functions.

Data are presented as mean ± SEM. Power analysis of our data sets for IVM revealed power levels of 0.91.0. Two-tailed, unpaired t-tests were used to compare wild-type and mutant current parameters in the presence of IVM and p < 0.01 was taken as the significance threshold.

3D Structure alignments of TM3 domains

The alignments of TM3 domains of the GluClR and GlyR were done using the Internal Coordinate

Mechanics software (ICM-Pro Molsoft LLC, San Diego, CA). The α-carbon atoms of the N-terminal residues from TM3-29’ to TM3-36’ were superimposed and used as a fixed reference. The displacement between α-carbon atoms at position TM3-56’ were then measured between two given TM3 domains The structures used for this analysis were, the GluClR in a non-conducting conformation (PDB, 4TNV [7]), the GluClR in complex with IVM (PDB, 3RHW [6]), the α1 GlyR in a non-conducting conformation in complex with strychnine (PDB, 3JAD [70]), and the structure of the

α1 GlyR in complex with IVM (PDB, 3JAF [70]). The final representations were created using the

Pymol Molecular Graphics System, Version 1.3.

142

Fig. 1. Single channel conductance and kinetic properties of GluClRs in response to glutamate activation. A) Sample traces of single channel activity recorded in outside-out patches at indicated holding potentials. Channels were activated by 10 µM glutamate. B) Mean current-voltage relationship averaged from 6 patches. Error bars were smaller than symbol size. Vrev = reversal potential (4.0 mV). C) Examples of single channel activity in response to 200 µM glutamate. In this and all subsequent figures, recordings were performed at 70 mV and channel openings are downward deflections from the baseline. D) Shut and open dwell histograms for data obtained at 200 µM glutamate. The histograms show that the receptors have two shut and three open components. E) Examples of single channel activity recorded in response to 30 µM glutamate, indicating the presence of shorter activations. F) Shut and open dwell histograms for data obtained at 30 µM glutamate, again revealing the presence of two shut and three open components, but the longest shut component is slightly increased.

143

Fig. 2. Kinetic properties of GluClRs at low glutamate concentrations. A) Examples of continuous single channel activity recorded from an outside-out patch in response to 2 µM glutamate. B) Shut and open dwell histograms for data obtained at 2 µM glutamate. The histograms show that despite the reduced length of active periods, the receptors have two shut and three open components, but the longest shut component is increasing with decreasing concentration. C) Examples of single channel activity recorded in response to 30 nM glutamate, indicating the presence of mostly brief activations. D) Shut and open dwell histograms for data obtained at 30 nM glutamate, revealing the presence of two shut and two open components.

144

Fig. 3. Concentration-dependence of glutamate effects on GluClRs. A) Examples of how intra-activation open probability (PO) increased as glutamate concentration was increased. In the absence of glutamate single receptor activity was negligible. Activations of similar duration were selected to facilitate comparison. B) Effect of glutamate concentration on the time constants of the long (red symbols) and short (black symbols) shut-state dwell components. C) Effect of glutamate concentration on time constants of the open-state dwell components. The symbols denote the long (green symbols), intermediate (red symbols) and short (black symbols) time constants Note the disappearance of the longest open component and the reduction in length of the shorter open components at nanomolar glutamate. D) Mean intra-activation open probability (PO) plotted as a function of glutamate concentration. The curve represents a Hill equation fit with an EC50 of 70 nM. E) Mean active period duration plotted as a function of glutamate concentration.

The curve represents a Hill equation fit with an EC50 of 31.2 µM. The data in B-E are means from 3-12 patches (see S2 Table).

145

Fig. 4. Ensemble glutamate-induced activation properties of GluClRs. A) Superimposed recordings revealing the effects of 50 ms (above) or 500 ms (below) applications of indicated glutamate concentrations onto macropatches expressing multiple GluClRs. B) Mean glutamate concentration-response relationship of peak currents as determined by fast agonist application. The curve represents a Hill equation fit with an EC50 of 43 µM. C) Normalised currents showing the concentration dependence of the activation phase of the current. The activation phase of each current trace was fitted to equation 2. D) Mean glutamate concentration-response relationship of the activation rate (kact). The curve represents a Hill equation fit with an EC50 of 0.95 µM. The numbers with arrows in A and C are the glutamate concentrations (in µM) that correspond to the currents. The arrows point to the peak current in A and the corresponding current onset in C. The data in B and D are means from 6-15 patches.

146

Fig. 5. Comparison of the effect of 1 mM glutamate on wild-type and G36’A mutant GluClRs. A) Examples of continuous single channel activity recorded from G36’A mutant GluClRs. Note the emergence of a ‘spiky’ activation mode (red boxes) that is not observed in wild-type GluClRs. Wild-type-like activations are termed ‘mode 1’ or ‘high Po’, whereas spiky activations are termed ‘mode 2’ or ‘low Po’. B) Examples of continuous single channel activity recorded from wild- type GluClRs included for comparison. C) Comparison of mean active durations (upper panel) and Po (lower panel) of low (LPO) and high (HPO) PO events recorded from G36’A mutant GluClRs (n = 6 patches). D) Examples of activations demarcated by a grey bar in A and B. These activations are of the high PO mode for the G36’A mutant (above) and normal mode for wild-type (below). The comparison indicates that there are more numerous open-shut events within the activations of G36’A compared to wild-type. E) Shut and open dwell histograms for data obtained from G36’A mutant GluClRs at 1 mM glutamate. This plot combined LPO and HPO activations of G36’A receptors at 1 mM glutamate. The histograms show that the mutant receptors have two shut and three open components. F) Shut and open dwell histograms for data obtained from wild-type GluClRs at 1 mM glutamate, revealing two shut and three open components.

147

Fig. 6. Concentration-dependence of glutamate effects on G36’A mutant GluClRs. A) Examples of continuous single channel activity recorded at 30 µM glutamate. Note the presence of mode 1 and 2 events. B) Examples of single channel activity recorded in response to 2 µM glutamate, indicating the presence of brief activations only. C) Upper panel: Mean PO plotted as a function of glutamate concentration. The corresponding curve for the wild-type receptor is included as a dashed line. Lower panel: Mean active period duration plotted as a function of glutamate concentration. The corresponding curve for the wild-type receptor is included as a dashed line. Data represent mean from 3-7 patches. D) Upper panel: Effect of glutamate concentration on long (red symbols) and short (black symbols) shut-state dwell components. Lower panel: Effect of glutamate concentration on long (green symbols), intermediate (red symbols) and short (black symbols) open-state dwell components. Data represent mean ± SEM from 4-8 patches. E) Shut and open dwell histograms for data obtained at 30 µM glutamate, revealing two shut and three open components. F) Shut and open dwell histograms for data obtained at 2 µM glutamate, revealing two shut and two open components.

148

Fig. 7. Estimation of desensitisation rate in patches expressing a known number of wild-type or G36’A mutant channels. A) Continuous recording from a patch expressing wild-type receptors in response to rapid application of 1 mM glutamate. 9 channels were present in this patch. B) Corresponding recording from a macropatch expressing G36’A receptors. 6 channels were present in this patch. C) Shut dwell histogram of the activity shown in A. The longer component represented a mean desensitised lifetime of 12780 ms and it was this component that was corrected for channel number (12780 x 9 = 115020 ms or 115 s). D) Shut dwell histogram of the recording shown in B, which yielded a desensitised lifetime of 70 s. E) Desensitisation scheme for calculating equilibrium constant (δ/ω) for desensitisation.

149

Fig. 8. Outside-out macropatch recordings of currents mediated by the indicated receptors. A) Sample recordings from macropatches expressing α (avr-14b) GluClRs, α1β GlyRs and α5β3γ2 GABAARs in response to ~1 ms applications of saturating (3 mM) agonist (glutamate, glycine or GABA). B) Mean weighted current deactivation time constants for the receptors indicated in A. The individual time constants and their relative magnitudes are summarised in Table 1. C) Sample recordings from macropatches expressing wild-type and G36’A mutant GluClRs in response to ~1 ms applications of saturating (3 mM) glutamate. Note the substantial decrease in deactivation time in the mutant. D) Mean activation rates reveal no significant difference between the four receptor isoforms (n = 6-12 patches). E) Sample recordings from macropatches expressing wild-type and G36’A mutant GluClRs in response to 500 ms applications of saturating (3 mM) glutamate. F) Mean desensitisation rates as calculated from n = 6-10 patches. The mutant plot represents a weighted mean of two components.

150

Fig. 9. Direct activation by 5 nM IVM. A) Four segments of recording illustrating the progressive increase (early, intermediate and late) in active durations upon first exposure to 5 nM IVM alone for wild-type GluClRs. B) Steady-state currents mediated by wild-type GluClRs in the continuous presence of 5 nM IVM. C) Four segments of record showing the progressive increase (early, intermediate and late) in active durations upon first exposure to 5 nM IVM alone for G36’A mutant GluClRs. The late phase precedes the steady-state phase. D) Steady-state currents mediated G36’A GluClRs in the continuous presence of 5 nM IVM alone. Note the much briefer activations compared to wild-type. E) Bar plots summarising the mean active durations for wild-type and G36’A mutant GluClRs (n = 6 patches each). F) Bar plots summarising the mean POs for wild-type and G36’A mutant GluClRs (n = 6 patches each). * p< 0.01

151

Fig. 10. Current potentiation in the presence of 2 M glutamate and 5 nM IVM. A) Three segments of record showing the progressive increase (early, intermediate and late) in active durations upon exposure to both ligands for wild-type GluClRs. The late phase precedes the steady-state phase. B) Continuous sweeps of recording showing the equilibrated or steady-state phase of the recordings for wild-type GluClRs. These segments were used to determine receptor desensitisation. C) Three segments of record showing the progressive increase (early, intermediate and late) in active durations upon exposure to both ligands for G36’A mutant GluClRs. D) The steady-state phase of currents mediated by G36’A mutant GluClRs mostly consisted of relatively brief activations, along with the occasional longer activations. E) Summary of the mean active durations for wild-type (n = 7 patches) and G36’A mutant (n = 6 patches) GluClRs. F) Summary of the mean POs for wild-type and G36’A mutant GluClRs. * p < 0.01

152

Fig. 11. Structural representations of TM3 domains in shut and IVM bound configurations. A) TM3 domains of homomeric GluClRs and α1 GlyRs in shut configurations. B) TM3 domains of homomeric GluClRs and α1 GlyRs in IVM- bound configurations. C) TM3 domains of the GluClR in shut and IVM-bound configurations. D) TM3 domains of the α1 GlyRs in shut and IVM-bound configurations. In all cases residues between 29’ and 36’ were fixed as a reference and the extent of displacement (in Angstrom units) was measured at the 56’ position. The pdb files used in the figure were, the GluClR in a shut conformation (PDB, 4TNV), the GluClR in complex with ivermectin (PDB, 3RHW), the α1 GlyR in a shut conformation in complex with strychnine (PDB, 3JAD), and the structure of the α1 GlyR in complex with ivermectin (PDB, 3JAF). IVM is shown as a stick-ball structure in green.

153

References

1. Cull-Candy SG. Two types of extrajunctional L-glutamate receptors in locust muscle fibres. J Physiol. 1976;255(2):449-64. PubMed PMID: 1255528; PubMed Central PMCID: PMCPMC1309257. 2. Gration KA, Clark RB, Usherwood PN. Three types of L-glutamate receptor on junctional membrane of locust muscle fibres. Brain Res. 1979;171(2):360-4. PubMed PMID: 223735. 3. Marder E, Paupardin-Tritsch D. The pharmacological properties of some crustacean neuronal acetylcholine, gamma-aminobutyric acid, and L-glutamate responses. J Physiol. 1978;280:213-36. PubMed PMID: 211227; PubMed Central PMCID: PMCPMC1282656. 4. Wolstenholme AJ. Glutamate-gated chloride channels. J Biol Chem. 2012;287(48):40232-8. doi: 10.1074/jbc.R112.406280. PubMed PMID: 23038250; PubMed Central PMCID: PMCPMC3504739. 5. Cully DF, Vassilatis DK, Liu KK, Paress PS, Van der Ploeg LH, Schaeffer JM, et al. Cloning of an avermectin-sensitive glutamate-gated chloride channel from Caenorhabditis elegans. Nature. 1994;371(6499):707-11. doi: 10.1038/371707a0. PubMed PMID: 7935817. 6. Hibbs RE, Gouaux E. Principles of activation and permeation in an anion-selective Cys-loop receptor. Nature. 2011;474(7349):54-60. doi: 10.1038/nature10139. PubMed PMID: 21572436; PubMed Central PMCID: PMCPMC3160419. 7. Althoff T, Hibbs RE, Banerjee S, Gouaux E. X-ray structures of GluCl in apo states reveal a gating mechanism of Cys-loop receptors. Nature. 2014;512(7514):333-7. Epub 2014/08/22. doi: 10.1038/nature13669. PubMed PMID: 25143115; PubMed Central PMCID: PMC4255919. 8. Burzomato V, Beato M, Groot-Kormelink PJ, Colquhoun D, Sivilotti LG. Single-channel behavior of heteromeric alpha1beta glycine receptors: an attempt to detect a conformational change before the channel opens. J Neurosci. 2004;24(48):10924-40. PubMed PMID: 15574743. 9. Auerbach A. Agonist activation of a nicotinic acetylcholine receptor. Neuropharmacology. 2014. Epub 2014/12/03. doi: 10.1016/j.neuropharm.2014.10.004. PubMed PMID: 25446670. 10. Purohit P, Gupta S, Jadey S, Auerbach A. Functional anatomy of an allosteric protein. Nature communications. 2013;4:2984. Epub 2013/12/20. doi: 10.1038/ncomms3984. PubMed PMID: 24352193. 11. Mukhtasimova N, daCosta CJ, Sine SM. Improved resolution of single channel dwell times reveals mechanisms of binding, priming, and gating in muscle AChR. J Gen Physiol. 2016;148(1):43-63. doi: 10.1085/jgp.201611584. PubMed PMID: 27353445; PubMed Central PMCID: PMCPMC4924934. 12. Corradi J, Gumilar F, Bouzat C. Single-channel kinetic analysis for activation and desensitization of homomeric 5-HT(3)A receptors. Biophysical journal. 2009;97(5):1335-45. Epub 2009/09/02. doi: 10.1016/j.bpj.2009.06.018. PubMed PMID: 19720021; PubMed Central PMCID: PMC2749742. 13. Keramidas A, Harrison NL. The activation mechanism of alpha1beta2gamma2S and alpha3beta3gamma2S GABAA receptors. J Gen Physiol. 2010;135(1):59-75. Epub 2009/12/30. doi: jgp.200910317 [pii] 10.1085/jgp.200910317. PubMed PMID: 20038526; PubMed Central PMCID: PMC2806416. 14. Dixon C, Sah P, Lynch JW, Keramidas A. GABAA receptor alpha and gamma subunits shape synaptic currents via different mechanisms. The Journal of biological chemistry. 2014;289(9):5399-411. Epub 2014/01/16. doi: 10.1074/jbc.M113.514695. PubMed PMID: 24425869; PubMed Central PMCID: PMC3937617. 15. Lema GM, Auerbach A. Modes and models of GABA(A) receptor gating. J Physiol. 2006;572(Pt 1):183- 200. Epub 2006/02/04. doi: 10.1113/jphysiol.2005.099093. PubMed PMID: 16455693; PubMed Central PMCID: PMC1779655. 154

16. Marabelli A, Lape R, Sivilotti L. Mechanism of activation of the prokaryotic channel ELIC by propylamine: a single-channel study. J Gen Physiol. 2015;145(1):23-45. Epub 2014/12/31. doi: 10.1085/jgp.201411234. PubMed PMID: 25548135; PubMed Central PMCID: PMC4278187. 17. Delany NS, Laughton DL, Wolstenholme AJ. Cloning and localisation of an avermectin receptor- related subunit from Haemonchus contortus. Mol Biochem Parasitol. 1998;97(1-2):177-87. PubMed PMID: 9879896. 18. Boxshall GA. Parasitic copepods (Copepoda: Siphonostomatoida) from deep-sea and mid-water fishes. Systematic parasitology. 2000;47(3):173-81. Epub 2000/11/09. PubMed PMID: 11071152. 19. Costello MJ. Ecology of sea lice parasitic on farmed and wild fish. Trends in parasitology. 2006;22(10):475-83. Epub 2006/08/22. doi: 10.1016/j.pt.2006.08.006. PubMed PMID: 16920027. 20. Nicol J, Turner, SJ, Coyne, DL, Nijs, L den, Hockland, S and Tahna, M. Current Nematode Threats to World Agriculture. 2011:21-43. 21. Kaplan RM. Drug resistance in nematodes of veterinary importance: a status report. Trends in parasitology. 2004;20(10):477-81. Epub 2004/09/15. doi: 10.1016/j.pt.2004.08.001. PubMed PMID: 15363441. 22. Meyers JI, Gray M, Kuklinski W, Johnson LB, Snow CD, Black WCt, et al. Characterization of the target of ivermectin, the glutamate-gated chloride channel, from Anopheles gambiae. J Exp Biol. 2015;218(Pt 10):1478-86. doi: 10.1242/jeb.118570. PubMed PMID: 25994631; PubMed Central PMCID: PMCPMC4448665. 23. Omura S. A Splendid Gift from the Earth: The Origins and Impact of the Avermectins (Nobel Lecture). Angew Chem Int Ed Engl. 2016;55(35):10190-209. doi: 10.1002/anie.201602164. PubMed PMID: 27435664. 24. Wolstenholme AJ, Rogers AT. Glutamate-gated chloride channels and the mode of action of the avermectin/milbemycin anthelmintics. Parasitology. 2005;131 Suppl:S85-95. doi: 10.1017/S0031182005008218. PubMed PMID: 16569295. 25. Horsberg TE. Avermectin use in aquaculture. Current pharmaceutical biotechnology. 2012;13(6):1095-102. Epub 2011/11/02. PubMed PMID: 22039799. 26. Ghosh R, Andersen EC, Shapiro JA, Gerke JP, Kruglyak L. Natural variation in a chloride channel subunit confers avermectin resistance in C. elegans. Science. 2012;335(6068):574-8. doi: 10.1126/science.1214318. PubMed PMID: 22301316; PubMed Central PMCID: PMCPMC3273849. 27. Playford MC, Smith AN, Love S, Besier RB, Kluver P, Bailey JN. Prevalence and severity of anthelmintic resistance in ovine gastrointestinal nematodes in Australia (2009-2012). Australian veterinary journal. 2014;92(12):464-71. Epub 2014/11/27. doi: 10.1111/avj.12271. PubMed PMID: 25424758. 28. Osei-Atweneboana MY, Awadzi K, Attah SK, Boakye DA, Gyapong JO, Prichard RK. Phenotypic evidence of emerging ivermectin resistance in Onchocerca volvulus. PLoS Negl Trop Dis. 2011;5(3):e998. doi: 10.1371/journal.pntd.0000998. PubMed PMID: 21468315; PubMed Central PMCID: PMCPMC3066159. 29. Lynagh T, Lynch JW. A glycine residue essential for high ivermectin sensitivity in Cys-loop ion channel receptors. Int J Parasitol. 2010;40(13):1477-81. doi: 10.1016/j.ijpara.2010.07.010. PubMed PMID: 20713056. 30. Lynagh T, Webb TI, Dixon CL, Cromer BA, Lynch JW. Molecular determinants of ivermectin sensitivity at the glycine receptor chloride channel. J Biol Chem. 2011;286(51):43913-24. doi: 10.1074/jbc.M111.262634. PubMed PMID: 22033924; PubMed Central PMCID: PMCPMC3243547. 31. Degani-Katzav N, Gortler R, Gorodetzki L, Paas Y. Subunit stoichiometry and arrangement in a heteromeric glutamate-gated chloride channel. Proceedings of the National Academy of Sciences of the United States of America. 2016. doi: 10.1073/pnas.1423753113. PubMed PMID: 26792524.

155

32. Degani-Katzav N, Gortler R, Weissman M, Paas Y. Mutational Analysis at Intersubunit Interfaces of an Anionic Glutamate Receptor Reveals a Key Interaction Important for Channel Gating by Ivermectin. Frontiers in molecular neuroscience. 2017;10:92. doi: 10.3389/fnmol.2017.00092. PubMed PMID: 28428744; PubMed Central PMCID: PMCPMC5382172. 33. Wang X, Puinean AM, AO OR, Williamson MS, Smelt CLC, Millar NS, et al. Mutations on M3 helix of Plutella xylostella glutamate-gated chloride channel confer unequal resistance to abamectin by two different mechanisms. Insect Biochem Mol Biol. 2017;86:50-7. doi: 10.1016/j.ibmb.2017.05.006. PubMed PMID: 28576654. 34. Kwon DH, Yoon KS, Clark JM, Lee SH. A point mutation in a glutamate-gated chloride channel confers abamectin resistance in the two-spotted spider mite, Tetranychus urticae Koch. Insect Mol Biol. 2010;19(4):583-91. doi: 10.1111/j.1365-2583.2010.01017.x. PubMed PMID: 20522121. 35. Dermauw W, Ilias A, Riga M, Tsagkarakou A, Grbic M, Tirry L, et al. The cys-loop ligand-gated ion channel gene family of Tetranychus urticae: implications for acaricide toxicology and a novel mutation associated with abamectin resistance. Insect Biochem Mol Biol. 2012;42(7):455-65. doi: 10.1016/j.ibmb.2012.03.002. PubMed PMID: 22465149. 36. Lynagh T, Lynch JW. Ivermectin binding sites in human and invertebrate Cys-loop receptors. Trends Pharmacol Sci. 2012;33(8):432-41. doi: 10.1016/j.tips.2012.05.002. PubMed PMID: 22677714. 37. Sivilotti L, Colquhoun D. In praise of single channel kinetics. J Gen Physiol. 2016;148(2):79-88. doi: 10.1085/jgp.201611649. PubMed PMID: 27432998; PubMed Central PMCID: PMCPMC4969800. 38. Portillo V, Jagannathan S, Wolstenholme AJ. Distribution of glutamate-gated chloride channel subunits in the parasitic nematode Haemonchus contortus. J Comp Neurol. 2003;462(2):213-22. doi: 10.1002/cne.10735. PubMed PMID: 12794744. 39. Yates DM, Portillo V, Wolstenholme AJ. The avermectin receptors of Haemonchus contortus and Caenorhabditis elegans. Int J Parasitol. 2003;33(11):1183-93. PubMed PMID: 13678634. 40. Beech R, Levitt N, Cambos M, Zhou S, Forrester SG. Association of ion-channel genotype and macrocyclic lactone sensitivity traits in Haemonchus contortus. Mol Biochem Parasitol. 2010;171(2):74-80. doi: 10.1016/j.molbiopara.2010.02.004. PubMed PMID: 20211658. 41. Glendinning SK, Buckingham SD, Sattelle DB, Wonnacott S, Wolstenholme AJ. Glutamate-gated chloride channels of Haemonchus contortus restore drug sensitivity to ivermectin resistant Caenorhabditis elegans. PLoS One. 2011;6(7):e22390. doi: 10.1371/journal.pone.0022390. PubMed PMID: 21818319; PubMed Central PMCID: PMCPMC3144221. 42. McCavera S, Rogers AT, Yates DM, Woods DJ, Wolstenholme AJ. An ivermectin-sensitive glutamate- gated chloride channel from the parasitic nematode Haemonchus contortus. Mol Pharmacol. 2009;75(6):1347-55. doi: 10.1124/mol.108.053363. PubMed PMID: 19336526; PubMed Central PMCID: PMCPMC2684884. 43. Cheeseman CL, Delany NS, Woods DJ, Wolstenholme AJ. High-affinity ivermectin binding to recombinant subunits of the Haemonchus contortus glutamate-gated chloride channel. Mol Biochem Parasitol. 2001;114(2):161-8. PubMed PMID: 11378196. 44. Kotze AC, Prichard RK. Anthelmintic Resistance in Haemonchus contortus: History, Mechanisms and Diagnosis. Adv Parasitol. 2016;93:397-428. doi: 10.1016/bs.apar.2016.02.012. PubMed PMID: 27238009. 45. Estrada-Mondragon A, Lynch JW. Functional characterization of ivermectin binding sites in alpha1beta2gamma2L GABA(A) receptors. Frontiers in molecular neuroscience. 2015;8:55. doi: 10.3389/fnmol.2015.00055. PubMed PMID: 26441518; PubMed Central PMCID: PMCPMC4585179. 46. Lynagh T, Lynch JW. An improved ivermectin-activated chloride channel receptor for inhibiting electrical activity in defined neuronal populations. The Journal of biological chemistry. 2010;285(20):14890- 156

7. Epub 2010/03/24. doi: 10.1074/jbc.M110.107789. PubMed PMID: 20308070; PubMed Central PMCID: PMC2865309. 47. Adelsberger H, Lepier A, Dudel J. Activation of rat recombinant alpha(1)beta(2)gamma(2S) GABA(A) receptor by the insecticide ivermectin. European journal of pharmacology. 2000;394(2-3):163-70. Epub 2000/04/20. PubMed PMID: 10771281. 48. Gielen M, Thomas P, Smart TG. The desensitization gate of inhibitory Cys-loop receptors. Nature communications. 2015;6:6829. doi: 10.1038/ncomms7829. PubMed PMID: 25891813; PubMed Central PMCID: PMCPMC4410641. 49. Wang Q, Lynch JW. Activation and desensitization induce distinct conformational changes at the extracellular-transmembrane domain interface of the glycine receptor. The Journal of biological chemistry. 2011;286(44):38814-24. doi: 10.1074/jbc.M111.273631. PubMed PMID: 21917927; PubMed Central PMCID: PMCPMC3207466. 50. Keramidas A, Lynch JW. An outline of desensitization in pentameric ligand-gated ion channel receptors. Cell Mol Life Sci. 2013;70(7):1241-53. doi: 10.1007/s00018-012-1133-z. PubMed PMID: 22936353. 51. Jackson MB. Dependence of acetylcholine receptor channel kinetics on agonist concentration in cultured mouse muscle fibres. J Physiol. 1988;397:555-83. PubMed PMID: 2457702; PubMed Central PMCID: PMCPMC1192143. 52. Keramidas A, Kash TL, Harrison NL. The pre-M1 segment of the alpha1 subunit is a transduction element in the activation of the GABAA receptor. J Physiol. 2006;575(Pt 1):11-22. doi: 10.1113/jphysiol.2005.102756. PubMed PMID: 16763005; PubMed Central PMCID: PMCPMC1819431. 53. Maconochie DJ, Zempel JM, Steinbach JH. How quickly can GABAA receptors open? Neuron. 1994;12(1):61-71. PubMed PMID: 8292360. 54. Scott S, Lynch JW, Keramidas A. Correlating structural and energetic changes in glycine receptor activation. The Journal of biological chemistry. 2015. Epub 2015/01/13. doi: 10.1074/jbc.M114.616573. PubMed PMID: 25572390. 55. Hamill OP, Bormann J, Sakmann B. Activation of multiple-conductance state chloride channels in spinal neurones by glycine and GABA. Nature. 1983;305(5937):805-8. Epub 1983/10/02. PubMed PMID: 6314144. 56. Mermans C, Dermauw W, Geibel S, Van Leeuwen T. A G326E substitution in the glutamate-gated chloride channel 3 (GluCl3) of the two-spotted spider mite Tetranychus urticae abolishes the agonistic activity of macrocyclic lactones. Pest Manag Sci. 2017. doi: 10.1002/ps.4677. PubMed PMID: 28736919. 57. Jones MV, Jonas P, Sahara Y, Westbrook GL. Microscopic kinetics and energetics distinguish GABA(A) receptor agonists from antagonists. Biophysical journal. 2001;81(5):2660-70. doi: 10.1016/S0006- 3495(01)75909-7. PubMed PMID: 11606279; PubMed Central PMCID: PMCPMC1301733. 58. Berg OG, von Hippel PH. Diffusion-controlled macromolecular interactions. Annu Rev Biophys Biophys Chem. 1985;14:131-60. doi: 10.1146/annurev.bb.14.060185.001023. PubMed PMID: 3890878. 59. Jadey S, Auerbach A. An integrated catch-and-hold mechanism activates nicotinic acetylcholine receptors. J Gen Physiol. 2012;140(1):17-28. doi: 10.1085/jgp.201210801. PubMed PMID: 22732309; PubMed Central PMCID: PMCPMC3382718. 60. Hosie AM, Wilkins ME, da Silva HM, Smart TG. Endogenous neurosteroids regulate GABAA receptors through two discrete transmembrane sites. Nature. 2006;444(7118):486-9. Epub 2006/11/17. doi: 10.1038/nature05324. PubMed PMID: 17108970. 61. Howard RJ, Trudell JR, Harris RA. Seeking structural specificity: direct modulation of pentameric ligand-gated ion channels by alcohols and general anesthetics. Pharmacological reviews. 2014;66(2):396-412.

157

Epub 2014/02/12. doi: 10.1124/pr.113.007468. PubMed PMID: 24515646; PubMed Central PMCID: PMC3973611. 62. Gonzalez Canga A, Sahagun Prieto AM, Jose Diez Liebana M, Martinez NF, Vega MS, Vieitez JJ. The pharmacokinetics and metabolism of ivermectin in domestic animal species. Vet J. 2009;179(1):25-37. doi: 10.1016/j.tvjl.2007.07.011. PubMed PMID: 17851096. 63. Lee SY, MacKinnon R. A membrane-access mechanism of ion channel inhibition by voltage sensor toxins from spider venom. Nature. 2004;430(6996):232-5. doi: 10.1038/nature02632. PubMed PMID: 15241419. 64. Martin RJ, Kusel JR. On the distribution of a fluorescent ivermectin probe (4" 5,7 dimethyl-bodipy proprionylivermectin) in Ascaris membranes. Parasitology. 1992;104 ( Pt 3):549-55. PubMed PMID: 1641253. 65. Jiang Y, Lee A, Chen J, Cadene M, Chait BT, MacKinnon R. The open pore conformation of potassium channels. Nature. 2002;417(6888):523-6. doi: 10.1038/417523a. PubMed PMID: 12037560. 66. Shealy RT, Murphy AD, Ramarathnam R, Jakobsson E, Subramaniam S. Sequence-function analysis of the K+-selective family of ion channels using a comprehensive alignment and the KcsA channel structure. Biophysical journal. 2003;84(5):2929-42. doi: 10.1016/S0006-3495(03)70020-4. PubMed PMID: 12719225; PubMed Central PMCID: PMCPMC1302856. 67. Purohit P, Auerbach A. Glycine hinges with opposing actions at the acetylcholine receptor-channel transmitter binding site. Molecular pharmacology. 2011;79(3):351-9. doi: 10.1124/mol.110.068767. PubMed PMID: 21115636; PubMed Central PMCID: PMCPMC3061369. 68. Grosman C, Zhou M, Auerbach A. Mapping the conformational wave of acetylcholine receptor channel gating. Nature. 2000;403(6771):773-6. Epub 2000/02/29. doi: 10.1038/35001586. PubMed PMID: 10693806. 69. Barry PH. JPCalc, a software package for calculating liquid junction potential corrections in patch- clamp, intracellular, epithelial and bilayer measurements and for correcting junction potential measurements. J Neurosci Methods. 1994;51(1):107-16. PubMed PMID: 8189746. 70. Du J, Lu W, Wu S, Cheng Y, Gouaux E. Glycine receptor mechanism elucidated by electron cryo- microscopy. Nature. 2015;526(7572):224-9. doi: 10.1038/nature14853. PubMed PMID: 26344198; PubMed Central PMCID: PMCPMC4659708.

Acknowledgement

We would like to thank Antje Blumenthal for critically reading the manuscript.

158

S1 Table. Dwell time components for wild-type and G36’A mutant GluClRs for glutamate. α wild-type GluClR

[E] mM τS1 FS1 τS2 FS2 τO1 FO1 τO2 FO2 τO3 FO3

10 0.59 ± 0.06 88 ± 3 2.0 ± 0.3 16 ± 4 1.70 ± 0.22 24 ± 5 37 ±13 36 ± 9 109 ± 8 40 ±4

1 0.56 ± 0.09 82 ± 9 2.0 ± 0.4 18 ± 19 1.64 ± 0.24 26 ± 9 28 ± 9 35 ± 2 149 ± 28 40 ± 14

0.2 0.53 ± 0.02 80 ± 3 2.6 ± 0.6 20 ± 6 1.59 ± 0.07 22.± 3 16 ± 1 33 ± 6 99 ± 9 45 ± 6

0.03 0.47 ± 0.03 86 ± 5 3.1 ± 0.5 14 ± 6 1.91 ± 0.12 27 ± 5 16 ± 2 33 ± 4 94 ± 4 40 ± 3

0.01 0.50 ± 0.07 80 ± 3 3.8 ± 0.8 20 ± 3 1.92 ± 0.40 22 ± 5 18 ± 4 26 ± 6 107 ± 9 52 ± 5

0.002 0.51 ± 0.03 75 ± 5 6.5 ± 0.4 25 ± 5 1.75 ± 0.16 36 ± 4 18 ± 2 39 ± 5 111 ± 7 26 ± 3

0.00003 1.22 ± 0.19 45 ± 14 34.5 ± 4.3 55 ± 14 1.12 ± 0.20 77 ± 6 13 ± 4 23 ± 6 − −

0.000005 1.22 ± 0.11 43 ± 8 86.3 ± 9.8 57 ± 8 0.75 ± 0.04 78 ± 11 4 ± 1 22 ± 4 − −

α(G36’A) GluClR

10 0.60 ± 0.06 56 ± 7 13.7 ± 1.0 44 ± 7 1.31 ± 0.34 34 ± 7 12 ± 4 34 ± 6 62 ± 20 32 ± 6

1 0.85 ± 0.09 53 ± 6 11.6 ± 2.0 47 ± 5 1.14 ± 0.14 42 ± 6 8 ± 1 29 ± 6 45 ± 11 29 ± 4

0.03 0.57 ± 0.09 58 ± 8 18.0 ± 2 42 ± 8 1.46 ± 0.11 40 ± 8 6 ± 1 32 ± 6 23 ± 2 28 ± 2

0.002 0.54 ± 0.07 − − − 1.10 ± 0.15 11 ± 5 5 ± 1 89 ± 3 − −

Data represent mean ± SEM from 3-12 patches

159

S2 Table. Glutamate-dependent active durations and open probability (PO)

[E] mM Active duration (ms) Po n

wild-type GluClR

10 494 ± 42 0.99 ± 0.03 7

1 437 ± 37 0.99 ± 0.02 7

0.2 374 ± 29 0.98 ± 0.06 12

0.03 328 ± 21 0.95 ± 0.02 7

0.01 206 ± 11 0.93 ± 0.01 5

0.002 146 ± 6.5 0.74 ± 0.02 10

0.00003 102 ± 36 0.21 ± 0.02 3

0.000005 82 ± 21 0.14 ± 0.03 3

G36’A GluClR

10 198 ± 18 0.71 ± 0.06 7

1 188 ± 12 0.66 ± 0.06 8

0.03 114 ± 17 0.41 ± 0.06 3

0.002 11 ± 1 ND 4

n represents the number of patches. ND not determined.

160

S3 Table. Dwell time components for wild-type and G36’A mutant GluClRs for IVM.

τS1 FS1 τS2 FS2 τS3 FS3 τO1 FO1 τO2 FO2 τO3 FO3 τO4 FO4

Direct activation (5 nM IVM)

1.50 ± 6.02 ± 17.5 ± 1.43 ± 5.48 ± 30.5 ± 217 ± Wild-type 60 ± 5 32 ± 5 8 ± 2 50 ± 5 28 ± 2 13 ± 2 9 ± 5 0.22 0.56 0.8 0.19 0.99 4.5 41 GluClRs Potentiation (5 nM IVM + 2 µM glutamate)

1.30 ± 9.62 ± 1.94 ± 17.8 ± 180 ± 82 ± 2 18 ± 2 − − 34 ± 4 29 ± 5 37 ± 5 − − 0.22 3.01 0.43 4.1 18

Direct activation (5 nM IVM)

1.47 ± 36.7 ± 1.83 ± 14.3 ± 52.1 ± 34 ± 8 66 ± 8 − − 35 ± 7 35 ± 8 30 ± 7 − − G36’A 0.27 1.7 0.41 3.7 4.6

GluClRs Potentiation (5 nM IVM + 2 µM glutamate)

1.19 ± 35.8 ± 1.02 ± 8.99 ± 49.7 ± 40 ± 9 60 ± 8 − − 40 ± 6 38 ± 6 22 ± 9 − − 0.20 4.5 0.28 1.37 15.7

Data represent mean ± SEM from 6-7 patches

161

Pharmacological characterization of lindane, fipronil and PTX on two GluClR splice variants of Anopheles gambiae

Introduction

On an average, around 200 million cases of malaria were reported worldwide from the year of 2013-2017. Out of which, on an average 500,000 resulted in death every year according to the WHO world malaria report. Africa, where A. gambiae is the primary vector of malaria, accounted for approximately 90% of these deaths (WHO World Malaria Report, 2014-2018), (Phillips et al. 2017; Ashley et al. 2018). Insecticide-treated mosquito bed nets and indoor residual spraying of pyrethroid-based insecticides forms the current malaria control program. Pyrethroid resistance however is becoming widespread across the African A.gambiae population (Ranson et al. 2011; Trape et al. 2011; Hemingway et al. 2016; Bhatt et al. 2015; Phillips et al. 2017). The mosquito is long-lived, thrives in human environments and prefers human blood (Killeen et al. 2017; White et al. 2014). The efforts to control malaria transmission through vector-targeting interventions with a new mode of action, led to the identification of IVM as a new malaria control candidate. Malaria vectors after having ingested host treated blood meals laced with IVM effectively killed or disabled A.gambiae, by interfering with the feeding patterns, both in and outside lab conditions (Kobylinski et al. 2010; Sylla et al. 2010). A temporary reduction in the proportion of Plasmodium falciparum- infected A.gambiae was seen recently after a mass drug administration of IVM at multiple sites across West Africa (Kobylinski et al. 2011; Alout et al. 2014). An impairment in the flight patterns and inhibition of sporogony of P. falciparum in A. gambiae was noticed at sub-lethal doses of IVM. Clinical trials, involving a combination of artemether-lumefantrine, reduced the likelihood of malaria transmission by killing the mosquitoes (Ouedraogo et al. 2015). These above studies demonstrate that IVM can be a promising tool to control malaria transmission. The majority of research on IVM targets has been done in nematodes or other model organisms (Omura 2008), however thorough functional characterization of GluClRs in mosquito vectors of A. gambiae has not yet been performed. mRNA from the female mosquitoes of A. gambiae were cloned to give four splice isoforms of AgGluClR contrary to a previous prediction of only three isoforms (Megy et al. 2012; Meyers et al. 2015).

Here we characterized the effects of lindane, fipronil, picrotoxin (PTX) and IVM on two splice variants i.e., AgGluClR-a1 and AgGluClR-b from A. gambiae in order to understand the function and pharmacology of GluClRs from this important pathogenic organism. We

163 expressed AgGluClR isoforms in X. laevis oocytes to measure their electrophysiological activity in response to glutamate, IVM, lindane, fipronil and PTX.

Difference between nomenclatures of different splice variants from AgGluClRs.

Table 5.1 compares the names of AgGluClR splice variants clone by Meyers et al., (2015) and the NCBI database.

Table 5.1: The below table depicts the different nomenclature used in a paper by Meyers et al., (2015) and the nucleotide database. The nomenclature in parentheses refers to the nucleotide GenBank accession number.

Nomenclature by Meyers et. al Nomenclature on database AgGluClR a1 AgGluClR b1 (KC902519.1) AgGluClR a2 AgGluClR b2 (KC902520.1) AgGluClR b AgGluClR c (KC902521.1)

AgGluClR c AgGluClR a (KC902518.1)

In order to avoid confusion I will employ the nomenclature from the paper by Meyers et al., (2015).

Results

Glutamate and IVM dose-response relationship of two isoforms of A. gambiae GluClRs:

Although the effect of glutamate and IVM on AgGluClR-a1 and AgGluClR-b was characterised by Meyers et al., (2015), the study lacked a more thorough pharmacological characterization of these two splice variants. In order to address this deficiency, the first set of experiments involved testing the efficacy of glutamate and IVM at AgGluClR-a1 and AgGluClR-b. Glutamate evoked rapidly activating inward currents from homomeric GluClRs comprising either one or the other splice isoform. The current amplitude saturated at 1000 µM glutamate suggesting that the all the channels were completely saturated at this concentration. The glutamate EC50 values for AgGluClR-a1 and AgGluClR-b were similar, being 19 ± 1 µM and 27 ± 1 µM with Hill co-efficient of 3.4 ± 0.3 and 2.1 ± 0.1 (n= 6 and 7 respectively, Figure 5.1-a,b,c, Table 5.2). The above data compared reasonably well with the results of Meyers et al., (2015). A similar concentration response experiment was performed by using IVM; however as IVM is a quasi-irreversible agonist, we employed the protocol 164

whereby increasing IVM concentrations were successively applied to the recorded oocyte at

steady-state current levels. As we did not find any comparable difference in the EC50 values for glutamate in our experiments and the study done by Meyers et al., (2015) on AgGluClR- a1 and AgGluClR-b, we were expecting that we would get a similar set of results for IVM. However, oocytes expressing AgGluClR-a1 and AgGluClR-b both produced slow activating inward currents when subjected to IVM application. This result contradicted the findings from the study of Meyers, where they showed that AgGluCl-a1 was sensitive to IVM whereas

AgGluCl-b was not. AgGluClR-a1 had an EC50 of 145 ± 13 nM and AgGluClR-b had an EC50 of 202 ± 50 nM for IVM which were not statistically significant (Figure 5.1-d,e,f, Table 5.2).

Table 5.2: EC50 values and Hill co-efficient for glutamate and IVM at two splice isoforms of AgGluClRs. ns= not significant, compared with AgGluClR-a1 using unpaired t-test

Receptor Glutamate IVM

subtype LogEC50 (µM) nH N LogEC50 (nM) nH N AgGluClR-a1 19 ± 1 3.4 ± 0.3 6 145 ± 13 1.3 ± 0.1 6 AgGluClR-b 27 ± 1ns 2.1 ± 0.1 7 202 ± 50ns 1.2 ± 0.2 6

Figure 5.1: The above figure depicts the effect of glutamate and IVM on two splice variants of A. gambiae i.e., AgGluClR-a1 (green) and AgGluClR-b (orange). Panels A, B and C shows the effect of glutamate on AgGluClR-a1 and AgGluClR-b and a dose response graph for both the splice isoforms respectively. Panel D, E and F shows cumulative concentration response traces of IVM on AgGluClR- a1 and AgGluClR-b and a dose-response graph that was normalised to their maximum glutamate elicited currents, respectively. All the values mean ± S.E.M (n=6, at least).

165

Differential effects of lindane and fipronil on AgGluClR isoforms

Having established that IVM does indeed activate both AgGluClRs made of either one- splice variant or the other we wanted to see whether insecticides such as lindane and fipronil, which target GluClRs, would have any differential efficacy between these two splice variants. As these splice variants have minimal difference from each other any differential activity that would be observed will be helpful in characterizing a drug that can target GluClRs in a specific manner.

In order to discern the effects of lindane and fipronil on AgGluClR-a1 and AgGluClR-b, we performed a direct inhibition experiment by applying 30 µM lindane or fipronil at both isoforms. In another set of experiments, we determined the effect of co-application of lindane

or fipronil with an EC50 concentration of glutamate at both splice variants of AgGluClRs. A steady-state inhibition relationship was quantitated for lindane and fipronil by measuring the percentage of glutamate current remaining after a closed state drug application. A 2-minute

drug application in a closed state followed by an immediate application of EC50 glutamate was used to determine the extent of inhibition (Figure 5.2-A); this was used to compare the relative magnitude of lindane and fipronil inhibition at AgGluClR-a1 and AgGluClR-b.

Inhibitory profile of lindane, fipronil and PTX

An inhibitory concentration-response experiment for lindane and fipronil in a closed

receptor state were quantitated at their respective EC50 glutamate concentrations. Panels A, and B in figure 5.2 shows the sample inhibitory response traces for lindane at AgGluClR-a1 and AgGluClR-b from a single oocyte, respectively, whereas panel C shows their averaged inhibitory profile (n=6). Glutamate-gated currents mediated by both the splice variants were inhibited by lindane in a dose-dependent manner with an IC50 of 9 µM and 5 µM, respectively (Table 5.3). Notably, lindane was able to completely inhibit the glutamate currents induced by AgGluClR-b whereas only 65% of peak current was blocked at α AgGluClRs-a1 at the maximum lindane concentration (Table 5.3).

Similar experiments were done using fipronil. However, the maximum concentration used was 30 µM as fipronil started to precipitate at a higher concentration in the working solution. Contrary to the inhibitory effect that was seen by lindane, fipronil blocked glutamate

166

currents completely with equal efficacy at both the splice isoforms (AgGluClR-a1= 0.5 µM and AgGluClR-b= 0.9 µM, n=6, Figure 5.2-D, E, F, Table 5.3).

As PTX is an open pore blocker, we had to vary the method by which we measured the inhibitory profile. We employed an open state application where glutamate was co-applied

at its EC50 concentration along with increasing concentrations of PTX. Similar to the effect that was seen with lindane, PTX was also able to block glutamate-mediated currents, however, a differential efficacy was seen between AgGluClR-a1 and AgGluClR-b. PTX blocked approximately 80% of currents mediated by glutamate at AgGluClR-a1 at a maximum concentration of 300 µM whereas complete inhibition of glutamate currents was observed at AgGluClR-b when same concentration was used (Table 5.3). In addition, there was a significant

difference in the IC50 values (Figure 5.2-G, H, I, AgGluClR-a1 = 23.0 ± 3.0 µM, AgGluClR-b = 7.1 ± 0.2 µM, n=6 each).

Table 5.3: The table below depicts the different IC50 values and an approximate percentage of inhibition for lindane, fipronil and PTX at AgGluClR-a1 and AgGluClR-b. All the values are represented as mean ± S.E.M, *p<0.05, ***p<0.001 compared to AgGluClR-a1 using unpaired t-test.

Receptor Lindane Fipronil PTX

subtype LogIC50 % Inhibition n LogIC50 % Inhibition n LogIC50 (µM) % n (µM) (µM) Inhibition AgGluClR-a1 8.0 ± 1.0 65 ± 5 6 0.5 ± 0.1 99 ± 1 6 23.0 ± 3.0 80 ± 9 6 AgGluClR-b 5.8 ± 1.3 ns 99 ± 1* 6 0.8 ± 0.1 ns 99 ± 1 ns 6 7.1 ± 0.2 *** 97 ± 1* 6

Recovery profile from lindane- and fipronil-mediated inhibition

After testing the inhibitory efficacy of lindane, fipronil and PTX, we wanted to see if there would be a difference in their recovery profile. As experiments done in our lab

previously showed that α1, α2, α3 and α1β GlyRs recover to a different extent after lindane and fipronil application (Islam and Lynch 2012), it would be interesting to see whether the same effects would be observed at the GluClRs.

A time-dependent recovery from inhibition by lindane and fipronil was measured at

the EC50 glutamate concentration. Recovery was analysed by bathing the oocyte with an EC50 concentration of glutamate for a period of 20 seconds, separated by 2-minute washes in the

167

Figure 5.2: Mean inhibitory dose response relationship for lindane, fipronil and PTX at AgGluClR-a1 and AgGluClR-b at EC50 glutamate. (A) & (B), sample traces of lindane inhibition at AgGluClR-a1 and AgGluClR-b. (D) & (E) Corresponding traces for fipronil, (G) & (H) effect of PTX. (C), (F) & (I) averaged dose- response relationships for lindane, fipronil and PTX, respectively.

168

extracellular control solution, which was preceded by a 2-minute application of 30 µM of either lindane or fipronil (Figure 5.3A).

A sample recovery trace for lindane at AgGluClRs-a1 and AgGluClRs-b is shown in Figure 5.3A, B respectively, along with the averaged recovery profile for lindane at AgGluClR- -a1 and AgGluClRs-b (Figure 5.3C). A differential efficacy that was observed while characterizing the inhibitory profile of lindane was also seen while characterizing its recovery profile. A maximum recovery of current of around 70% was observed following inhibition by lindane at the AgGluClRs-a1. However, the recovery was much smaller and slower at the β AgGluClRs with a maximum current recovery of about 40% (Figure 5.3C, Table 5.4). A similar trend was seen with fipronil recovery, where AgGluClRs-a1 recovered to about 50% whereas AgGluClRs-b had a smaller recovery of less than 20% (Figure 5.3D,E,F).

Table 5.4: The table below depicts the time dependence recovery from lindane and fipronil induced

inhibition using EC50 glutamate at AgGluClR-a1 and AgGluClR-b. All the values are represented as mean ± S.E.M, *p<0.05, **p<0.01 compared to AgGluClR-a1 using unpaired t-test.

Receptor Lindane Fipronil subtype % Recovery n % Recovery n AgGluClR-a1 71.5 ± 2.1 6 37.3 ± 6.4 6 AgGluClR-b 42.3 ± 3.2* 6 18.1 ± 2.1** 6

Differential efficacy of lindane, fipronil and picrotoxin at inhibiting currents mediated by an

EC50 IVM concentration:

From the above experiments, we know that lindane, fipronil and PTX effectively inhibit currents activated by glutamate. As IVM is an irreversible agonist, we wanted to see whether these three drugs could inhibit the IVM-mediated currents at AgGluClRs-a1 and AgGluClRs-b. The effect of lindane on currents mediated by IVM at AgGluClRs-a1 and AgGluClRs-b is shown in figure 5.4 A, B, C. Among the 3 drugs, lindane was the least efficacious inhibiting currents

by only 20-25% at AgGluClRs-a1 and AgGluClRs-b, with IC50s of 3.0 ± 0.9 µM and 11.0 ± 0.5 µM, respectively (n=6, each, Table 5.5). Fipronil, on the other had was the most effective in inhibiting currents mediated by IVM, inhibiting the currents by approximately 50% or more

for both the splice variants of AgGluClRs (Figure 5.4D, E, F, Table 5.5). The IC50 values for

169

Figure 5.3: Time dependence recovery of lindane and fipronil induced inhibition at EC50 glutamate. A & B, a sample recovery profile for AgGluClRs-a1 and AgGluClRs-b following inhibition by 30 µM lindane. Recovery was monitored by applying glutamate at 2-minute intervals. D & E, similar experiments for fipronil. C & F, averaged recovery profiles for lindane and fipronil respectively (n=6).

170

fipronil at AgGluClRs-a1 and AgGluClRs-b were 0.23 ± 0.1 µM and 0.56 ± 0.1 µM, respectively (n=6, each). PTX was unique in its effect, where it had differential efficacy between the two splice variants. PTX inhibited IVM mediated currents at AgGluClRs-a1 by more than 40% with

an IC50 value of 5.9 ± 1 µM (n=6) whereas the inhibitory effect of PTX was not prominent at

AgGluClRs-b with an IC50 of 16.2 ± 0.5 µM (n=6, Figure 5.4G, H, I, Table 5.5). The order of efficacy of the three drugs is fipronil > PTX > lindane.

Table 5.5: The table below depicts the different IC50 values and an approximate percentage of IVM currents that were inhibited by lindane, fipronil and PTX at AgGluClR-a1 and AgGluClR-b. All the values are represented as mean ± S.E.M, *p<0.05, **p<0.01 compared to AgGluClR-a1 using unpaired t-test. ns = not significant.

Receptor Lindane Fipronil PTX

subtype LogIC50 (µM) % Inhibition n LogIC50 (µM) % Inhibition n LogIC50 % Inhibition n (µM) AgGluClR- 3.0 ± 0.9 17.1 ± 3.4 6 0.23 ± 0.1 50 ± 6 6 5.9 ± 1 19.9 ± 2.3 6 a1 AgGluClR-b 11.0 ± 0.5ns 19.2 ± 5.9ns 6 0.56 ± 0.1 ns 68 ± 10* 6 16 ± 5 ns 45.3 ± 7.3* 6

Discussion and conclusion

In conclusion, it is evident from the inhibitory and recovery profiles of lindane, fipronil and PTX at AgGluClRs-a1 and AgGluClRs-b, that these drugs have different efficacies at the two splice variants. Preliminary characterization of both the splice variants of AgGluClRs against glutamate and IVM proved that they were equally sensitive. Lindane showed differential efficacy between the two splice variants where it only blocked partial glutamate- mediated currents of AgGluClR-a1 in comparison to AgGluClR-b where a total block of currents was seen. Both fipronil and PTX almost completely inhibited glutamate-mediated currents except in case of PTX at AgGluClR-a1 where only 80% of currents were inhibited. Time-dependent recovery profiles were also interesting where neither AgGluClR-a1 nor AgGluClR-b recovered to their original peak amplitude after being inhibited by a single concentration of lindane or fipronil. On the other hand inhibition of IVM-mediated currents by lindane, fipronil and PTX were different compared to the inhibition produced by these drugs on glutamate-mediated currents. Lindane and fipronil were not significant in inhibiting

171

Figure 5.4: Ability of lindane, fipronil and PTX to block currents mediated IVM. (A) & (B), Inhibitory properties of lindane at AgGluClR-a1 and AgGluClR-b. Lindane was the least efficacious drug in inhibiting IVM mediated currents. (D) & (E) Similar experiments done using fipronil at AgGluClR-a1 and AgGluClR-b, fipronil was the most effective in inhibiting IVM invoked currents. (G) & (H) Sample traces for PTX at AgGluClR-a1 and AgGluClR-b. PTX was unique where it was more efficacious at AgGluCl-a1 compared to AgGluClR-b. (C), (F) & (I) averaged inhibitory dose-response relationships for lindane, fipronil and PTX, respectively.

172 the IVM-mediated currents on the two AgGluClRs splice variants, whereas fipronil was significant in inhibiting IVM-mediated currents in comparison to lindane and fipronil. The order of efficacy for inhibiting IVM-mediated currents is fipronil > PTX > lindane.

One of the contributing reasons for this difference could be the molecular determinants of the drug binding sites. (Chen et al. 2006) predicted that lindane and fipronil bind to the 2’ and 6’ pore-lining residues of human β3 homomeric GABAARs. This result was confirmed in D. melanogaster RDL GABAARs, which showed reduced sensitivity to lindane and fipronil at naturally occurring mutations of A2’S and A2’G (Cole et al. 1995; Hosie et al. 1995). The role of 2’ and 6’ in the binding of lindane and fipronil was also supported by the results from (Islam and Lynch 2012) in α1 glycine receptors. However, both splice variants have a serine and a threonine at 2’ and 6’, respectively. Future work should concentrate on why the drugs exhibit altered sensitivities at one isoform compared to the other.

173

References

Alout, H., et al. 2014. 'Evaluation of ivermectin mass drug administration for malaria transmission control across different West African environments', Malar J, 13: 417.

Ashley, E. A., et al. 2018. 'Malaria', Lancet, 391: 1608-21.

Bhatt, S., et al. 2015. 'The effect of malaria control on Plasmodium falciparum in Africa between 2000 and 2015', Nature, 526: 207-11.

Chen, L., et al. 2006. 'Structural model for gamma-aminobutyric acid receptor noncompetitive antagonist binding: widely diverse structures fit the same site', Proc Natl Acad Sci U S A, 103: 5185-90.

Cole, L. M., et al. 1995. 'Drosophila GABA-gated chloride channel: modified [3H]EBOB binding site associated with Ala-->Ser or Gly mutants of Rdl subunit', Life Sci, 56: 757-65.

Hemingway, J., et al. 2016. 'Averting a malaria disaster: will insecticide resistance derail malaria control?', Lancet, 387: 1785-8.

Hosie, A. M., et al. 1995. 'Actions of the insecticide fipronil, on dieldrin-sensitive and- resistant GABA receptors of Drosophila melanogaster', Br J Pharmacol, 115: 909-12.

Islam, R., and J. W. Lynch. 2012. 'Mechanism of action of the insecticides, lindane and fipronil, on glycine receptor chloride channels', Br J Pharmacol, 165: 2707-20.

Killeen, Gerry F, et al. 2017. 'Going beyond personal protection against mosquito bites to eliminate malaria transmission: population suppression of malaria vectors that exploit both human and animal blood', BMJ Global Health, 2: e000198.

Kobylinski, K. C., et al. 2010. 'The effect of oral anthelmintics on the survivorship and re-feeding frequency of anthropophilic mosquito disease vectors', Acta Trop, 116: 119-26.

Kobylinski, K. C., et al. 2011. 'Ivermectin mass drug administration to humans disrupts malaria parasite transmission in Senegalese villages', Am J Trop Med Hyg, 85: 3-5.

Megy, K., et al. 2012. 'VectorBase: improvements to a bioinformatics resource for invertebrate vector genomics', Nucleic Acids Res, 40: D729-34.

Meyers, J. I., et al. 2015. 'Characterization of the target of ivermectin, the glutamate-gated chloride channel, from Anopheles gambiae', J Exp Biol, 218: 1478-86.

Omura, S. 2008. 'Ivermectin: 25 years and still going strong', Int J Antimicrob Agents, 31: 91-8.

174

Ouedraogo, A. L., et al. 2015. 'Efficacy and safety of the mosquitocidal drug ivermectin to prevent malaria transmission after treatment: a double-blind, randomized, clinical trial', Clin Infect Dis, 60: 357-65.

Phillips, M. A., et al. 2017. 'Malaria', Nat Rev Dis Primers, 3: 17050.

Ranson, H., et al. 2011. 'Pyrethroid resistance in African anopheline mosquitoes: what are the implications for malaria control?', Trends Parasitol, 27: 91-8.

Sylla, M., et al. 2010. 'Mass drug administration of ivermectin in south-eastern Senegal reduces the survivorship of wild-caught, blood fed malaria vectors', Malar J, 9: 365.

Trape, J. F., et al. 2011. 'Malaria morbidity and pyrethroid resistance after the introduction of insecticide-treated bednets and artemisinin-based combination therapies: a longitudinal study', Lancet Infect Dis, 11: 925-32.

White, N. J., et al. 2014. 'Malaria', Lancet, 383: 723-35.

175

Other projects as part of my thesis

Discovering new drugs selectively targeting α5 containing GABAARs

Introduction

This project aimed at screening a library of fractions that were isolated from Australian and Antarctic marine invertebrates and algae in order to discover a novel lead compound that is

selective for α5 GABAARs. The crude extracts were first tested using ToxFinder. ToxFinder is an equipment to perform high-throughput screening (HTS). In ToxFinder, we use a yellow fluorescent protein (YFP)-based anion influx assay. This system involves transfecting HEK 293 cells with the desired receptor isoform along with the YFP I52’L. The I52’L mutation imparts high sensitivity to iodide (I-) quench (Gilbert et al. 2009). Drugs dissolved in sodium iodide (NaI) solution results in quenching of the YFP based on the properties of the drug either potentiating or inhibiting the fluorescence (Refer chapter 2 – high-throughput yellow fluorescent protein (YFP)-based anion influx assay).

As the extract quantity was minimal, we limited our screening only on α5β3γ2L GABAAR. The reason for choosing this subtype is its prevalence in hippocampus (Sieghart and Sperk 2002; Sur et al. 1998). Hippocampus is a region of brain that is associated with learning and

memory (Chambers et al. 2003). α5 containing GABAARs account for 20% of total GABAARs in the hippocampus, implicating its role in learning and memory (Sur et al. 1998). Both positive

and negative modulators of α5 GABAARs have the potential to be clinically useful, unlike other

GABAAR subtypes (Soh and Lynch 2015). Professor Robert Capon (Institute of Molecular Bioscience, The University of Queensland, Australia) kindly provide the compound library.

Compounds that show selectivity for α5 GABAARs can be used as leads to design drugs that

selectively target α5 GABAARs. After finding potential hits from ToxFinder they were re- assessed using automated Patchliner. This device permits complete exchange of the solution bathing the cells with 25 μL of new solution. This later feature was particularly important as the Capon lab provided us with only minute amounts of extracts and pure compounds (Refer chapter 2 – automated electrophysiology).

177

Results:

Preliminary marine extract screening

This project initially screened approximately 2800 (Figure 6.1 and 6.2) crude marine

extracts on the ToxFinder at α5β3γ2L GABAARs, by a previous member of our lab (Ming Soh). This screening yielded 36 hits that were further partitioned, triturated and fractionated using different solvents. This process gave us about 180 compounds that were partitions, triturates and fractions of these 36 hits.

Figure 6.1: Flow chart showing the partition and trituration of ethanol crude extracts. Based on the polarity, crude extracts were partitioned into n-butanol and H2O soluble portions, numbered as -1 and -2 respectively. Where the (-1) n-butanol partition was further triturated into hexane (-1-1), dichloromethane (DCM) (-1-2), methanol (MeOH) (-1-3) and H2O (-1-4) soluble triturates. The (-2) H2O soluble partition was triturated into aqueous methanol (Aq. MeOH) (-2-1) and H2O (-2-2) soluble triturates.

Screening of partitions and triturates

Partitioning and trituration of crude extracts using n-butanol and water-soluble parts yielded 36 compounds (Figure 6.2). These 36 compounds were further partitioned and triturated to give 180 compounds (Figure 6.1). The 180 partitions and triturates were tested on the

178

ToxFinder using a 384 well plate, at the α5β3γ2L GABAARs in triplicates (one well per partition/triturate per screen).

Figure 6.2: Summary of marine extract screening

Cells were pre-incubated with a partition or triturate at a final concentration of 4 µg/µL that were dissolved in 5.3% DMSO. During experiments, each well was subjected to

EC20 GABA dissolved in NaI solution that were already pre-treated with a partition or a triturate for half an hour. No significant GABA quench was observed in control experiments where only 5.3% DMSO was applied to the cells. We were looking for partitions or triturates

that either potentiated or inhibited the EC20 GABA quench. The criteria for selecting a partition or a triturate was that it should modulate the of EC20 GABA quench by at least 50%. Out of the 180 compounds that belonged to 36 crude extracts, compounds belonging to 32 crude

extracts returned with non-active partition/triturate i.e., their GABA EC20 quench was less

179

than 50%). 14 compounds belonging to 4 crude extracts had at least one active partition/triturate, and the results listed in table 6.1.

The four crude extracts that were identified had active partitions in them and are CMB-01871, -02130, -03405, and -03208 (Refer to appendices, section 2 for all the traces from these extracts on the Patchliner). These four crude extracts, after partitioning and trituration, gave 14 compounds (fractions). The fractioned products from CMB-02130 (Genus: Theonella) and CMB-03405 (Genus: Axinella) i.e., CMB-02130-1-2, CMB-02130-1-3 and CMB-03405-1-3

respectively were the most significant inhibitors of α5β3γ2L GABAAR mediated currents. Further partitioning and trituration of fractions from CMB-02130-1-3 and CMB-03405-1-3 yielded 19 sub-fractions (Refer to appendices, section 2 for the traces from these sub- fractions on the Patchliner).

After the initial screening on the ToxFinder, the crude extracts were further subjected to partitioning and trituration. The partitions and triturates of 32 crude extracts that did not show any substantial activity. This could be due to two reasons; first, ToxFinder screens compounds quickly but not accurately. Second, there might have been synergistic effect between the active compounds in the whole extract, which was lost once the compounds were partitioned and triturated. Although this was unlikely, the fractions from the pure compounds were re-tested using automated patch clamp (Patchliner) method. Only one pure compound was successfully identified from the all the partitions, triturates, fractions and sub- fractions due to scarcity of the crude extracts. This pure compound was isolated from the fraction CMB-02130-1-2 and identified as Aurantoside A (Figure 6.3). This compound has shown cytotoxicity against P388 and L1210 leukaemia cell lines and potent antifungal activity against Aspergillus fumigatus and Candida albicans (Matsunaga et al. 1991; Sata et al. 1999). Due to its limited availability we were only able to screen this compound against one subtype

of GABAARs i.e., at α5β3γ2L. Even though the effect of Aurantoside A was prominent (IC50 0.19 ± 0.01 µM) (Figure 6.3), we decided not to go ahead with the screening on other subtypes due to obvious reasons of availability and time invested in the screening.

180

Table 6.1: Percentage modulation by the active marine extracts, partitions and triturates from the 14 fractions belonging to four crude extracts screened on ToxFinder. Numbers represent the

modulation of GABA EC20 quench by the extracts. Partitions and triturates that modulated the GABA

EC20 quench by >50% are shown.  Indicates the partition/triturate either were inactive or had no significant effect (The complete list of compounds tested on the ToxFinder is detailed in appendices, section 3.2).

Percentage modulation of EC20 GABA-mediated quench (%) CMB-code Crude n-BuOH n-Hexane DCM MeOH Aq. MeOH extract (-1) (-1-1) (-1-2) (-1-3) (-2-1)

CMB-02130 98.0 92.0 76.6 70.9 96.1 58.6 CMB-01871 76.6 68.8 89.2 63.3 98.9 87.8 CMB-03208 89.2  65.6 97.4 91.2  CMB-03405 58.6 56.9    

Figure 6.3: A) Concentration response of Aurantoside A against α5β3γ2L GABAARs. A sample trace from automated patchliner recording of a single cell shows the inhibiting effect of increasing concentrations of Aurantoside A on the GABA EC20-induced current. B) Chemical structure of Aurantoside A. C) Inhibitory dose-response graph for Aurantoside A (n=4).

181

Quinazoline alkaloids against GABAAR

Apart from testing the fractions and sub-fractions, quinazoline alkaloids derived from

Penicillium sp. was provided by Prof. Robert Capon were also tested on α5β3γ2L GABAARs. These two compounds, named as CMB-TSF015-C02 and –C03, showed significant inhibition

of α5β3γ2L GABAARs.

These pure compounds were identified as Fumiquinazolines. Fumiquinazolines consist of a pyrazino [2, 1-b] quinazoline-3, 6-dione core which is linked with an indole moiety. An anticancer agent, Idelalisib has structural similarity to that of the quinazolinone scaffold (Resende et al. 2018). Alantrypinone (+)-1, a compound having a structural resemblance to fumiquinazoline C (Kuriyama et al. 2004) was shown to inhibit currents induced by GABA in cockroach neurons (Watanabe et al. 2009). Out of the two compounds, CMB-TSF015-C03 was more potent towards α5β3γ2L GABAARs and identified as Fumiquinazoline J, a complete dose

response on α5β3γ2L GABAARs shown below Figure 6.4.

Figure 6.4: A) Concentration response of Fumiquinazoline J against α5β3γ2L GABAARs. A sample trace from whole cell patch clamp recording of a single HEK 293 cell shows the inhibiting effect of increasing concentrations of Fumiquinazoline J on the GABA EC20-induced current. B) Chemical structure of Fumiquinazoline J. C) Inhibitory dose-response graph for Fumiquinazoline J (n=6).

182

Effect of GT-002 on IPSCs mediated by different GABAAR isoforms

As a part of my thesis, I decided to screen a drug GT-002 kindly provided by Dr. Mogens

Nielsen, CSO, Gabather, on GABAAR isoforms. Gabather is a small company based in Sweden, where they are trying to develop novel drug candidates for disorders related to central nervous system (CNS). Gabather has previously done in vitro and in vivo experiments on

GABAARs using two drugs developed by their company i.e., GT-001 and GT-006, both of which showed promising results in models of psychosis and anxiety without having any side effects by targeting α5 containing GABAARs. Now their focus is GT-002 as one of the potential antipsychotic drug. A study done in collaboration with Gabather and University of New South

Wales (UNSW), Australia, has shown that GT-002 has positive allosteric effect on α5 containing GABAARs. It is different from other positive allosteric modulators in that it does not affect neuronal populations containing other than α5 containing GABAARs thereby minimizing sedative and convulsive effects. Apart from that, in vitro studies have shown it to be a highly selective drug for α5 containing GABAARs. Therefore, the aim here was to see if

GT-002 would have differential effect if we substituted three different β isoforms i.e., β1, β2 and β3 in α5βxγ2L GABAARs. For this purpose we used the co-culture method wherein we allowed the growth of cortical neurons from E18 pregnant rat embryos on coverslips with

HEK293 cells transfected with the desired GABAAR isoform and measured the spontaneously evoked GABA mediated inhibitory postsynaptic currents (IPSCs). Only the cells that were transfected with GABAARs subunits in co-culture with neurons produced synaptic currents consisting of fast rise to peak, which was followed by a slower exponential decay back to baseline.

We tested the efficacy of GT-002 at two different concentrations (10 nM and 1 µM) at

α5β1γ2L, α5β2γ2L and α5β3γ2L GABAARs (Figure 6.5). Application of either 10 nM or 1 µM GT-

002 at α5β1γ2L had no significant effect on peak amplitudes, rise times, or their decay times (n=5), when compared with the control numbers (Figure 6.5 & 6.6) (Table 6.2). By contrast, applying GT-002 at a concentration of 10 nM to cells expressing α5β2γ2L, increased the rate of decay from 100 ms to approximately 200 ms. Increasing the concentration of drug applied from 10 nM to 1 µM did not increase the decay rates any further (n=5) (Figure 6.5 & 6.6) (Table 6.2). Other kinetic parameters such as peak amplitude and rise time did not change

183

significantly. Application of 10 nM and 1 µM GT-002 at α5β3γ2L had some visible effect on its decay rates but were not statistically significant.

The results suggests that GT-002 is a specific modulator for α5β2γ2L and that substituting a single subunit of β with a different isoform can result in loss of activity of GT-

002. This could be useful in treating disorders that are known to be involve α5β2γ2L GABAARs, thereby eliminating any side effects that arise from the drugs that target other β receptor isoform.

Figure 6.5: Properties of spontaneous IPSCs recorded from heterosynapses incorporating

α5β1γ2L, α5β2γ2L and α5β3γ2L GABAARs. (A) Mean IPSC decay time constants from cells expressing the indicated isoform. (B) Mean IPSC peak amplitudes from cells expressing indicated isoforms. (C) Mean 10-90% rise times from cells expressing the indicated isoforms. All the values are averaged from 5 cells with at least 10-100 events. Values were tested for significance using one-way ANOVA, where *p<0.05 and ns= not significant.

184

Table 6.2: Effect of 10 nM and 1 µM GT-002 on IPSC parameters for three different GABAAR isoforms.

GABAAR Rise time (ms) Decay time constant (ms) Amplitude (pA) isoform Control GT-002 GT-002 Control GT-002 GT-002 Control GT-002 GT-002 10 nM 1 µM 10 nM 1 µM 10 nM 1 µM

α5β1γ2L 3.4 ± 0.2 3.2 ± 0.3 2.9 ± 0.3 108 ± 13 128 ± 10 121 ± 20 104 ± 21 98 ± 14 99 ± 16

α5β2γ2L 6.1 ± 0.7 5.1 ± 0.3 5.4 ± 0.7 128±9 198±11* 192±13* 426±64 330±50 329±55

α5β3γ2L 4.7 ± 0.7 4.4 ± 0.2 3.9 ± 0.3 206±33 175±25 165±45 271±121 264±94 244±52 Figure 6.6: Representative recordings of IPSCs from HEK293 cells expressing each isoform. (A1), (B1) & (C1) Sample traces showing the effect of GT-002 at two different concentrations on IPSCs recorded

from HEK293 on α5β1γ2L, α5β2γ2L and α5β3γ2L GABAARs, respectively. (A2), (B2) & (C2) Averaged (from

10 to 100 events) normalised IPSCs from individual cells expressing α5β1γ2L, α5β2γ2L and α5β3γ2L GABAARs, respectively.

185

Discussion

An increasing numbers of drug discovery studies are discovering natural compounds that target CNS ion channel receptors. Some of these studies have been fruitful, such as the

discovery a subunit-selective α3 GlyR modulator for treatment of chronic inflammatory pain

(Balansa et al. 2010). Drugs that specifically target α5β3γ2L GABAARs have potential to be useful in treating many CNS-related disorders such as Down syndrome, post-stroke recovery and Alzheimer’s (Soh and Lynch 2015).

Screening of fractions and sub-fractions using ToxFinder and Patchliner, we were able to isolate aurantoside A from the marine sponge Theonella sp. (code CMB-02130) (Fusetani et al. 1990; Fusetani et al. 1991; Matsunaga et al. 1991). Generally, aurantosides are halogenated hydrocarbons from sponges and were previously only found in species of Manihinea Pulitzer-Finali and Theonella Gray (Hall et al. 2010). Aurantosides comprises a number of compounds, ranging from aurantoside A to K. Aurantoside A and B have been shown to possess cytotoxic activity (Matsunaga et al. 1991; Sata et al. 1999). Aurantoside J and K showed potential antifungal activity (Angawi et al. 2011; Kumar et al. 2012). However, as the cytotoxic effect of aurantoside A is known we decided to discontinue further testing of this isolated compound, as it would have meant investing more time without a fruitful result.

In addition to the aurantosides, other products derived from Penicillium sp. i.e.,

quinazoline alkaloids were also tested on α5β3γ2L GABAARs. The two compounds i.e., CMB- TSF015-C02 and –C03 were alantrypinone and fumiquinazoline analogues, respectively. GABA induced currents in cockroach neurons were inhibited by alantrypinone and their derivatives, whereas derivatives of fumiquinazolines (Fumiquinazoline F) possess antifungal activity (Li et al. 2012). Although the fumiquinazoline derivative had prominent activity because of its complicated structure, we did not see it as a viable therapeutic option and therefore decided

not to screen it on different GABAAR subtypes.

Apart from all the preliminary screening from the ToxFinder and Patchliner, I also

tested a drug, GT-002, on three different GABAAR isoforms. This drug has shown selective modulation at α5 containing GABAARs. β subunits have been shown to have an important role

in mediating modulatory effects of the anxiolytic drug, etifoxine, on GABAARs (Hamon et al.

2003). Our results show that the effects of GT-002 on GABAARs is determined by the identity

186

of the β subunit, which played a major role in potentiating the spontaneously evoked IPSCs.

As GT-002 specifically targets α5β2γ2L GABAARs, it could be used as a candidate lead in developing drugs that target diseases involving this specific subunit.

To summarise, a number of CNS related disorders including autism, schizophrenia,

cognitive impairment in Down syndrome and Alzheimer’s disease have α5 containing

GABAARs as their targets. Therefore, it is important for us to continue the search for drugs

that specifically target GABAARs. This chapter has provided some insights relevant to these goals.

187

References:

Angawi, R. F., et al. 2011. 'Aurantoside J: a new tetramic acid glycoside from Theonella swinhoei. Insights into the antifungal potential of aurantosides', Mar Drugs, 9: 2809- 17. Balansa, W., et al. 2010. 'Ircinialactams: subunit-selective glycine receptor modulators from Australian sponges of the family Irciniidae', Bioorg Med Chem, 18: 2912-9. Chambers, M. S., et al. 2003. 'Identification of a novel, selective GABA(A) alpha5 receptor inverse agonist which enhances cognition', J Med Chem, 46: 2227-40. Fusetani, Nobuhiro, et al. 1990. 'Bioactive marine metabolites. 33. Cyclotheonamides, potent thrombin inhibitors, from a marine sponge Theonella sp', Journal of the American Chemical Society, 112: 7053-54. Fusetani, Nobuhiro, et al. 1991. 'Orbiculamide A: a novel cytotoxic cyclic peptide from a marine sponge Theonella sp', Journal of the American Chemical Society, 113: 7811-12. Gilbert, D. F., et al. 2009. 'High Throughput Techniques for Discovering New Glycine Receptor Modulators and their Binding Sites', Front Mol Neurosci, 2: 17. Hall, Kathryn A., et al. 2010. 'The mystery of some sciophilous sponges in the Indo-West Pacific'. Hamon, A., et al. 2003. 'The modulatory effects of the anxiolytic etifoxine on GABA(A) receptors are mediated by the beta subunit', Neuropharmacology, 45: 293-303. Kumar, Rohitesh, et al. 2012. 'Aurantoside K, a new antifungal tetramic acid glycoside from a Fijian marine sponge of the genus Melophlus', Marine drugs, 10: 200-08. Kuriyama, T., et al. 2004. 'Receptor assay-guided isolation of anti-GABAergic insecticidal alkaloids from a fungal culture', J Agric Food Chem, 52: 3884-7. Li, Xiao-Jun, et al. 2012. 'Metabolites from Aspergillus fumigatus, an Endophytic Fungus Associated with Melia azedarach, and Their Antifungal, Antifeedant, and Toxic Activities', Journal of Agricultural and Food Chemistry, 60: 3424-31. Matsunaga, Shigeki, et al. 1991. 'Aurantosides A and B: cytotoxic tetramic acid glycosides from the marine sponge Theonella sp', Journal of the American Chemical Society, 113: 9690- 92. Resende, Disp, et al. 2018. 'Chemistry of the fumiquinazolines and structurally related alkaloids', Nat Prod Rep.

188

Sata, N. U., et al. 1999. 'Aurantosides D, E, and F: new antifungal tetramic acid glycosides from the marine sponge Siliquariaspongia japonica', J Nat Prod, 62: 969-71. Sieghart, W., and G. Sperk. 2002. 'Subunit composition, distribution and function of GABA(A) receptor subtypes', Curr Top Med Chem, 2: 795-816. Soh, M. S., and J. W. Lynch. 2015. 'Selective Modulators of alpha5-Containing GABAA Receptors and their Therapeutic Significance', Curr Drug Targets, 16: 735-46. Sur, C., et al. 1998. 'Rat and human hippocampal alpha5 subunit-containing gamma- aminobutyric AcidA receptors have alpha5 beta3 gamma2 pharmacological characteristics', Mol Pharmacol, 54: 928-33. Watanabe, T., et al. 2009. 'Alantrypinone and its derivatives: synthesis and antagonist activity toward insect GABA receptors', Bioorg Med Chem, 17: 94-110.

189

General discussion

a) The β Subunit: a potential cause of IVM resistance in H. contortus

The parasitic nematode H. contortus has a major detrimental impact on primary production animals including sheep, cattle and goats. Previous studies of GluClRs from C. elegans have demonstrated that receptors comprising different subunit isoforms exhibit different sensitivities to glutamate and IVM (Cully et al. 1994). However, there is no evidence of β subunit of HcoGluClR involved in IVM resistance in field isolates. In order to characterize the different agonist sensitivities of the different receptor isoforms, fixed quantities of α homomer or β homomer and αβ heteromers of H. contortus GluClRs subunits were injected into oocytes from X. laevis.

Glutamate concentration responses for α homomeric, β homomeric and αβ

heteromeric (1:1 and 1:50) HcoGluClRs did not show a marked difference in their EC50s (20-

40 µM), except for the β homomer, which was significantly less sensitive to glutamate (EC50= 394 µM). As an additional means of determining whether β subunits were incorporated in the

heteromer, we also used IVM. The α homomers were extremely sensitive towards IVM (EC50= 20 nM), whereas the β homomers were insensitive. αβ (1:50) heteromers on the other hand showed moderate IVM sensitivity which was intermediate between those of the α and β

homomers (EC50= approx. 150 nM). IVM showed a trend of decreasing sensitivity from α, to αβ (1:50) and finally β homomers. These results gave an indication that IVM sensitivity was a function of the subunit composition of the receptors, with higher incorporation of β subunit corresponding to lower sensitivity to IVM. This brings us to our hypothesis, that the subunit stoichiometry of receptor determines its sensitivity towards glutamate and IVM.

The functional properties of IPSCs mediated by native GluClRs in invertebrates’ neurons or muscle cells have never been reported. For the first time using heterocultures of neurons and transfected HEK293 cells, we characterized the properties of recombinant IPSCs mediated by α and αβ GluClRs in response to spontaneous presynaptic glutamate release. Using this technique, we defined the IPSC kinetics mediated by both GluClR isoforms, how different drugs effect the IPSCs mediated by different GluClR isoforms, and how the posttranslational modifications and mutations that impart IVM resistance to the organism affect the formation and function of synapses. In general, produced an increase in IPSC peak

191

amplitudes and a slowing of IPSC kinetics, in particular the decay rates. A dbeecreased sensitivity might be due to intrinsic activation properties of β – containing receptors.

Single channel data on αβ and β HcoGluClRs revealed that mean active durations are shortened when the concentration of β is increased thereby rendering the receptors less sensitive to potentiation by IVM when compared to α GluClRs.

The time course of action of IVM was attributed to its interactions with cell membranes. In order to see how IVM accumulates and diffuses in the membrane a fluorescent-tagged IVM (IVM-bdpy) was synthesised. In these experiments, the IVM-bdpy needed about 18 minutes to equilibrate in the whole cell membrane, which was similar to the tonic current equilibrium time in our IPSC recordings where IVM-mediated changes take about 17 minutes to plateau. Apart from this, using FRAP we measured the slow lateral diffusion rate of IVM that contributes to the association rates of IVM to the binding sites at GluClRs. The data suggests that rather than inducing a slow conformational rearrangement to the receptor by directly binding from the aqueous portion the major pathway for IVM to reach its target site at GluClRs is membrane partitioning and lateral diffusion.

Finally using structure activity relationship, we identified important receptor-drug interactions that determine IVM analogue potency at α and αβ GluClRs that may be useful in refining the design of anthelmintic drugs so that they target specific GluClR isoforms.

To summarise, based on the above results we conclude that the HcoGluCl β subunit confers IVM resistance without altering glutamate sensitivity when incorporated into heteromeric receptors with the α subunit. This is achieved by reducing the durations of active periods of single receptors and decay times of IPSCs. Based on our data, drug design strategies should be targeted towards increasing the active durations of single receptors and slowing the decay rates of IPSCs in synaptic isoforms of GluClRs

G36’A affects the intrinsic properties of the receptor and not IVM binding.

A decreased sensitivity for glutamate at the mutated α homomeric, TM3-G36’A GluClRs prompted us to investigate the action of IVM towards this mutation. IVM was able to activate these receptors directly as well as in combination with glutamate without having an effect on the channel conductance. 2 µM glutamate alone at the WT receptors elicited single

192

receptor active periods that had a mean duration of ~ 150 ms but when co-applied with 5 nM IVM gave rise to prolonged active periods of ~ 16 s. Even though the same combination of agonists was applied towards the G36’A mutated receptors the mean active durations were about 125 ms which were small in comparison to the WT receptors. However, they were large compared to glutamate 2 µM alone (~11ms) on the G36’A receptors.

Direct activation by 5nM IVM alone had similar patterns on the WT and the G36’A receptors, with the WT receptors having a mean activation duration of ~ 6 s and the mutant with just 50 ms.

In contrast, the active periods of the G36’A mutated receptor were much briefer when activated by IVM alone or when co-applied with glutamate when compared with WT receptors. Receptor desensitisation was also examined in WT and TM3-G36’A mutant receptors. In order to correct the desensitisation time constant, the number of active receptors present in the patch were counted by saturating them with glutamate (Atif et al. 2017). Based on the above analysis we inferred that the WT receptors exhibit little desensitisation, especially when glutamate and IVM were co-applied. At WT receptors, glutamate plus IVM induced a mean desensitised lifetime of ~ 200 ms with an equilibrium constant of 0.01 whereas IVM alone had a mean desensitised lifetime of ~ 550 ms, which gave an equilibrium constant of ~0.1 ms.. On the other hand, the TM3-G36’A mutation enhanced receptor desensitisation with a mean desensitised lifetime of 1964 ms and an equilibrium constant of ~39 ms in IVM alone. Whereas a co-application of glutamate and IVM to TM3- G36’A produced a mean desensitised lifetime of 1786 ms and an equilibrium constant of ~14.3 ms

In summary, the above experiments show that be it direct activation or potentiation by IVM, the G36’A mutant receptors exhibited shorter active durations and increased desensitisation in comparison to the WT receptors for both agonists (glutamate and IVM). From these data, we infer that the G36’A mutation affects the intrinsic activation properties of the receptor and these alterations underlie the reduced sensitivity to IVM at the G36’A mutant GluClRs. This directly contradicts the established paradigm which proposes that the G36’A mutation hinders IVM from accessing its binding site (Lynagh and Lynch 2010). This provides an important new insight into understanding and distinguishing the binding and

193 activation mechanisms of IVM. This in turn will help to define exactly how new IVM analogues affect the receptor binding and activation, and will thus help in tailoring of new anthelminthic compounds to particular GluClR isoforms.

Management strategies to delay development of resistance:

Pasture management and isolation of animal populations are two of the management strategies that are currently employed in primary production with an exclusive purpose of preventing infestation and/or keeping the level of infestation pressure to a bare minimum. These strategies would thereby decrease the dependence on anthelmintics, consequently delaying the development of resistance. The following comprise some of the strategies used to delay resistance.

Correct use of anthelmintics

No matter how effective a drug is at its target, each time a helminth population is exposed to anthelmintics they pose a risk of developing resistance. The periodic use of anthelmintics belonging to the same class and/or repeated under-dosing increases the risk of developing anthelmintic resistance (Sargison 2011). In order to delay the development of resistance, a policy involving rotation of anthelmintic classes has been recommended. One of the contributing factors towards increased selection pressure is excess deworming which leads to the use of unwanted treatment. The best way to ensure that there is no selection pressure is pasture management and appropriate treatment in order to maintain low refugia (De Graef et al. 2013).

Refugia

One of the important tools advocated to slow down the progress of anthelmintic resistance is “Refugia” (van Wyk 2001). Refugia refers to an area where the population of organisms can escape a period of hostile conditions. In this case, it means that the parasites in refugia have not been subjected to an anthelmintic. This includes parasites present as free- living stages in environment, at any lifecycle stage in the host that are stubborn to anthelmintic treatment and in untreated individuals (Fleming et al. 2006; van Wyk 2001). A strategy that aims to slow the rate of development of anthelmintic resistance is maintenance of a population of parasites in refugia. This method of maintaining a population of parasites

194

in refugia has been demonstrated to slow down the development of anthelmintic resistance in sheep model (Pech et al. 2009).

Different anthelmintic products

Apart from the methods mentioned above, another strategy to delay development of resistance might include the use of a combination of active anthelmintic products containing either two or more molecules having similar activity but a different mechanism of action. This method was suggested to delay the development of resistance to anthelmintics that were indicated in modelling studies and some field studies (Learmount et al. 2012; Leathwick 2012; Leathwick et al. 2012).

Other methods

Pasture management forms the other management measures aimed to reduce helminth infestation e.g. removal of animal droppings from pasture to decrease the level of infective larvae, preventing high degree of infestation, decreasing stocking densities or improving the drainage of pastures to minimised the risk of liver fluke infestations (Sargison 2011). Apart from these methods, there are other biological methods under development (Heckendorn et al. 2006; Hertzberg and Sager 2006; Waller et al. 2006), one of which is to select genetically less susceptible livestock (Stear et al. 2007). b) Differential effects of lindane, fipronil and PTX at AgGluClR-a1 and AgGluClR-b.

The original motivation for this study came from a study by (Meyers et al. 2015) that reported that AgGluClR-a1 and AgGluClR-b have different sensitivities to IVM. We were interested in understanding the underlying mechanism given our finding that IVM exhibits different sensitivities at different HcoGluClRs isoform. Given their higher amino acid sequence homology, we considered the residues responsible for the differential IVM sensitivity may be easier to characterize in the AgGluClR splice variants. First, we needed to perform a thorough characterization of glutamate and IVM responses on the two splice variants of AgGluClRs.

Pharmacological characterization revealed an interesting finding; AgGluClR-a1 and AgGluClR-b were both equally sensitive to IVM contrary to what was reported by (Meyers et al. 2015). However, there was differential efficacy for IVM, where the AgGluClR-a1 was more sensitive than the AgGluClR-b. As GluClRs are known targets for insecticides, it would be

195

interesting to determine if some of the known insecticides such as lindane and fipronil had differential affinities towards one of the splice variants of GluClRs from A. gambiae. This information would help us in refining drug discovery strategies to better target specific GluClR isoforms of A. gambiae.

GlyRs present in the cultured cerebellar granule cells were inhibited by lindane with an IC50 of 5 µM (Vale et al. 2003), whereas a complete block by (300 µM lindane) at GlyRs isoforms expressed in HEK293 cells was shown by (Islam and Lynch 2012). It was also shown

to have partial (50%) antagonistic activity against a number of recombinant GABAARs with IC50 in a range of 1-2 µM (Aspinwall et al. 1997; Maskell et al. 2001). We found that lindane

inhibited glutamate mediated currents with an IC50 of 8 µM and 6 µM against AgGluClR-a1 and AgGluClR-b, respectively, however, the extent with which it inhibited glutamate currents differed between the two splice variants. At AgGluClR-b, a complete block of glutamate currents was observed, however, a partial block of ̴65% was seen at AgGluClR-a1. The two

splice variants were extremely sensitive to fipronil, with IC50s in a range of 0.5-0.8 µM. These

values were similar to what was reported for GlyRs (Islam and Lynch 2012) and GABAARs (Ikeda et al. 2001; Li and Akk 2008). Unlike lindane, fipronil did not possess differential potency at the two isoforms and completely blocked glutamate activated currents in both. PTX on the other hand was somewhat similar to lindane in its action, where a complete block was seen at AgGluClR-b and only 80% block on AgGluClR-a1.

Time-dependent recovery experiments helped us to confirm that both lindane and fipronil exhibit slow recovery rates, suggesting that possibly a low (nM) concentration of drugs might be accumulating at the binding sites and producing the significant inhibitory effects.

IVM produces an extremely slow and irreversible current as shown in neuronal preparations of HcoGluClRs, as shown in chapter 3. Lindane and fipronil have been shown to bind in the GlyR channel pore (Islam and Lynch 2012). This is consistent with a mutation in the pore lining the TM2 segment (S2’A) in M. domestica GluClRs that enhanced lindane and fipronil sensitivity (Hirata et al. 2008). As these binding positions are close to IVM binding site, any data on how lindane, fipronil and PTX inhibiting IVM mediated currents in these two AgGluClR splice variants would help to better understand how these drugs act. Lindane was the least sensitive at inhibiting IVM mediated currents, followed by PTX, whereas as fipronil

196

was the most effective at blocking IVM currents. These differential effects of lindane, fipronil and PTX at inhibiting both glutamate and IVM provides a starting point for understanding how these drugs act and help us in refining drugs that are used as insecticides. c) Other projects as part of my thesis

Effective treatment for present or emerging clinical conditions is the primary aim of drug discovery. The drug development program takes approximately 16 years from identification of a lead molecule to its clinical use and is divided into two main stages namely preclinical research and clinical trials. The cost for identifying a single drug and developing it ranges in the billions of dollars (Roses 2008). Modulation of a physiological pathway or a specific protein involved in a clinical condition is usually the starting point for drug discovery. Then comes the validation of the modulatory protein or the pathway that may lead to a potential therapy. However, the first step in this pipeline is the process of identifying lead compounds and developing these using screening methods. High-throughput screening that utilises automated screening machines is one of such approaches to screen through libraries of compounds all of which are made up of diverse chemical structures.

Our focus was the preclinical stage where we aimed to discover novel molecules that

specifically target α5 containing GABAARs. (Soh and Lynch 2015) identified this target receptor to be involved in a number of clinical conditions such as post-stroke recovery, cognitive impairments in Down syndrome and Alzheimer’s, schizophrenia and autism. Drugs targeting this receptor subtype should minimise the side effects due to non-selective receptor

modulation of other GABAARs isoforms. Using a high-throughput YFP based assay, in a short time duration, we initially screened a large number of natural compounds derived from Australian and Antarctic marine sponge and algae on the target receptor expressed in mammalian HEK293 cells (Balansa et al. 2010; Gilbert et al. 2009). A number of hits were filtered in order to prioritise the ones with largest effect, best selectivity profile and appropriate chemistry, for further assessment using Nanion Patchliner, an automated planar chip patch clamp device. Reconfirming the hits in secondary electrophysiological assays took some time and employed relatively large quantities of the compounds. This incomplete project is currently stalled due to insufficient quantities of hit compounds. Further supplies

197

are required to determine the chemical structure, potency and selectivity of promising hit compounds.

Nevertheless, our high-throughput screening has identified some modulators

targeting α5 containing GABAARs that might be helpful in future drug development for this receptor. In particular, we identified a quinazoline alkaloid that demonstrated very high potency for α5 containing GABAARs. As this compound was able to inhibit GABAAR mediated currents in the nanomolar range, we hypothesised that this compound has the potential to be developed and improved further as a lead to therapeutic development and a probe to

investigate the functional properties of α5 containing GABAARs.

Despite the large number of clinical conditions for which α5 GABAAR modulation has been implicated, only a few α5 GABAARs selective modulators have been identified based on synthetically-derived structures (Soh and Lynch 2015). Subsequently only one has managed to make it to clinical trials. Originally known as RG-1662, Basmisanil is currently in phase 2 of clinical trials, to treat cognitive impairments in adults and adolescents with Down syndrome.

In short, our study has identified new GABAAR modulators through natural product screening

that might be useful as clinical leads and/or pharmacological probes of GABAARs.

Drug discovery strategies:

The field of drug discovery and design has seen remarkable advances despite it being a complex and an interdisciplinary process. Natural products have been out of the spotlight in favour of artificial chemical structures over the last five decades in the drug discovery paradigm as sources for new drugs (Harvey et al. 2015). Most of these artificial structures were often based on clinically effective drugs with similar structural scaffolds or integrated with chemical motifs from clinically successful compounds. Compounds generated using this click chemistry often failed to yield much needed effective new therapies and the rate of success of these compounds in clinical trial was extremely poor (Kola and Landis 2004). The limited structural diversity inherent in the conventional combinatorial libraries was a major reason for the failure of artificial chemical structures (Butler et al. 2014; Carter 2011; Harvey et al. 2015). This has prompted renewed interest in the development of leads and drugs from natural sources, as they possess the structural diversity not usually found in synthetically based corporate screening libraries (Abet et al. 2014). Ever since the natural products gained

198

traction in drug discovery and development, a total of 25 natural product-derived drugs (NPDDs) were launched in the market during 2008 - 2013 (Butler et al. 2014). These drugs were sourced from fungus, sponge, plants and bacteria metabolites, and were approved for treating a number of disorders ranging from diabetes to cancer (Butler et al. 2014).

As the resupply of raw materials in NPDDs is usually the rate-limiting factor (Abet et al. 2014), more recently, hybrid techniques have been developed whereby combinatorial libraries incorporate the broader chemical diversity of natural products (Lachance et al. 2012; van Hattum and Waldmann 2014). Recent advances include the use of ligand-directed

chemistry to design semisynthetic biosensors for GABAARs, by coupling chemical labelling with a biomolecular fluorescence quenching and recovery (BFQR) system (Yamaura et al.

2016). The potential for generating new generations of α5 GABAARs-specific modulators, as well as other therapeutic drugs looks promising, when coupled with complementary approaches such as fragment-based discovery (Wang et al. 2015), diversity-oriented synthesis (O' Connor et al. 2012), and dynamic combinatorial chemistry (Li et al. 2013).

However, as most of the drugs lack selectivity, the two most sought-after properties for drug molecules is extremely high affinity and selectivity. For most drugs, selectivity has been difficult to achieve, as the various targets are often large families of proteins with closely-related structures and functions (Kawasaki and Freire 2011). Selectivity of drugs during lead optimization can be maximised in two ways: improving the affinity towards the targets than to off-target molecules by chemical modification of the lead compound: and/or lowering the affinity of the hit compound towards off-target molecules using chemical modification (Kawasaki and Freire 2011). A combination of both the mechanisms can help in achieving maximal selectivity. Apart from this, binding affinity can be improved by introduction of nonpolar functionalities that fill a binding cavity (Kawasaki and Freire 2011). All these methods surely will help in better optimization of drug molecules in the near future.

199

References

Abet, V., et al. 2014. 'Biased and unbiased strategies to identify biologically active small molecules', Bioorg Med Chem, 22: 4474-89. Aspinwall, L. S., et al. 1997. 'The interactions of hexachlorocyclohexane isomers with human gamma-aminobutyric acid(A) receptors expressed in Xenopus oocytes', J Pharmacol Exp Ther, 282: 1557-64. Atif, M., et al. 2017. 'Effects of glutamate and ivermectin on single glutamate-gated chloride channels of the parasitic nematode H. contortus', PLoS Pathog, 13: e1006663. Balansa, W., et al. 2010. 'Ircinialactams: subunit-selective glycine receptor modulators from Australian sponges of the family Irciniidae', Bioorg Med Chem, 18: 2912-9. Butler, M. S., et al. 2014. 'Natural product and natural product derived drugs in clinical trials', Nat Prod Rep, 31: 1612-61. Carter, G. T. 2011. 'Natural products and Pharma 2011: strategic changes spur new opportunities', Nat Prod Rep, 28: 1783-9. Cully, D. F., et al. 1994. 'Cloning of an avermectin-sensitive glutamate-gated chloride channel from Caenorhabditis elegans', Nature, 371: 707-11. De Graef, Jessie, et al. 2013. 'Anthelmintic resistance of gastrointestinal cattle nematodes', Vlaams Diergeneesk. Tijdschr., 82: 113-23. Fleming, S. A., et al. 2006. 'Anthelmintic resistance of gastrointestinal parasites in small ruminants', J Vet Intern Med, 20: 435-44. Gilbert, D. F., et al. 2009. 'High Throughput Techniques for Discovering New Glycine Receptor Modulators and their Binding Sites', Front Mol Neurosci, 2: 17. Harvey, A. L., et al. 2015. 'The re-emergence of natural products for drug discovery in the genomics era', Nat Rev Drug Discov, 14: 111-29. Heckendorn, F., et al. 2006. 'Effect of sainfoin (Onobrychis viciifolia) silage and hay on established populations of Haemonchus contortus and Cooperia curticei in lambs', Vet Parasitol, 142: 293-300. Hertzberg, H., and H. Sager. 2006. '[Overview of helminth problems in domestic ruminants in Switzerland]', Schweiz Arch Tierheilkd, 148: 511-21.

200

Hirata, K., et al. 2008. 'Role of a serine residue (S278) in the pore-facing region of the housefly L-glutamate-gated chloride channel in determining sensitivity to noncompetitive antagonists', Insect Mol Biol, 17: 341-50. Ikeda, T., et al. 2001. 'Fipronil modulation of gamma-aminobutyric acid(A) receptors in rat dorsal root ganglion neurons', J Pharmacol Exp Ther, 296: 914-21. Islam, R., and J. W. Lynch. 2012. 'Mechanism of action of the insecticides, lindane and fipronil, on glycine receptor chloride channels', Br J Pharmacol, 165: 2707-20. Kawasaki, Y., and E. Freire. 2011. 'Finding a better path to drug selectivity', Drug Discov Today, 16: 985-90. Kola, I., and J. Landis. 2004. 'Can the pharmaceutical industry reduce attrition rates?', Nat Rev Drug Discov, 3: 711-5. Lachance, H., et al. 2012. 'Charting, navigating, and populating natural product chemical space for drug discovery', J Med Chem, 55: 5989-6001. Learmount, J., et al. 2012. 'A computer simulation study to evaluate resistance development with a derquantel-abamectin combination on UK sheep farms', Vet Parasitol, 187: 244-53. Leathwick, D. M. 2012. 'Modelling the benefits of a new class of anthelmintic in combination', Vet Parasitol, 186: 93-100. Leathwick, D. M., et al. 2012. 'Managing anthelmintic resistance--use of a combination anthelmintic and leaving some lambs untreated to slow the development of resistance to ivermectin', Vet Parasitol, 187: 285-94. Li, J., et al. 2013. 'Dynamic combinatorial libraries: from exploring molecular recognition to systems chemistry', J Am Chem Soc, 135: 9222-39. Li, P., and G. Akk. 2008. 'The insecticide fipronil and its metabolite fipronil sulphone inhibit the rat alpha1beta2gamma2L GABA(A) receptor', Br J Pharmacol, 155: 783-94. Lynagh, T., and J. W. Lynch. 2010. 'A glycine residue essential for high ivermectin sensitivity in Cys-loop ion channel receptors', Int J Parasitol, 40: 1477-81. Maskell, P. D., et al. 2001. 'Effects of gamma-HCH and delta-HCH on human recombinant GABA(A) receptors: dependence on GABA(A) receptor subunit combination', Br J Pharmacol, 132: 205-12. Meyers, J. I., et al. 2015. 'Characterization of the target of ivermectin, the glutamate-gated chloride channel, from Anopheles gambiae', J Exp Biol, 218: 1478-86.

201

O' Connor, CJ, et al. 2012. 'Diversity-oriented synthesis: producing chemical tools for dissecting biology', Chem Soc Rev, 41: 4444-56. Pech, CL, et al. 2009. 'The value of refugia in managing anthelmintic resistance: a modelling approach', Australian Agricultural & Resource Economics Society's Annual conference: 1-35. Roses, A. D. 2008. 'Pharmacogenetics in drug discovery and development: a translational perspective', Nat Rev Drug Discov, 7: 807-17. Sargison, N. D. 2011. 'Pharmaceutical control of endoparasitic helminth infections in sheep', Vet Clin North Am Food Anim Pract, 27: 139-56. Soh, M. S., and J. W. Lynch. 2015. 'Selective Modulators of alpha5-Containing GABAA Receptors and their Therapeutic Significance', Curr Drug Targets, 16: 735-46. Stear, M. J., et al. 2007. 'Alternatives to anthelmintics for the control of nematodes in livestock', Parasitology, 134: 139-51. Vale, C., et al. 2003. 'The organochlorine pesticides gamma-hexachlorocyclohexane (lindane), alpha-endosulfan and dieldrin differentially interact with GABA(A) and glycine-gated chloride channels in primary cultures of cerebellar granule cells', Neuroscience, 117: 397-403. van Hattum, H., and H. Waldmann. 2014. 'Biology-oriented synthesis: harnessing the power of evolution', J Am Chem Soc, 136: 11853-9. van Wyk, J. A. 2001. 'Refugia--overlooked as perhaps the most potent factor concerning the development of anthelmintic resistance', Onderstepoort J Vet Res, 68: 55-67. Waller, P. J., et al. 2006. 'Biological control of sheep parasites using Duddingtonia flagrans: trials on commercial farms in Sweden', Acta Vet Scand, 47: 23-32. Wang, T., et al. 2015. 'Fragment-based drug discovery and molecular docking in drug design', Curr Pharm Biotechnol, 16: 11-25. Yamaura, K., et al. 2016. 'Discovery of allosteric modulators for GABAA receptors by ligand- directed chemistry', Nat Chem Biol, 12: 822-30.

202

Appendices Section 1

Effect of Glutamate and IVM at two β homomeric mutants of H. contortus

Our first set of experiments in chapter 3 (GluClRs comprising homo- and heteromeric combinations of α and β subunits of H. contortus) we found that β HcoGluClRs is insensitive to IVM. However, as we do not know the molecular determinants of IVM insensitivity in reason β HcoGluClRs for that, we hypothesised that nonconserved residues in the TMDs could be the cause for this insensitivity as IVM binds at the interface of TM3 and TM1 of two adjacent subunits. We then aligned the TMDs of α, β subunits of H.contortus and α subunit of C.elegans to find any potential residues that could differ between α and β subunits of HcoGluClR. We identified two residues in the TMDs of β subunits; one was in TM1 where we mutated it from an isoleucine to leucine at -18’ position (I-18’L) and the second was a TM3 mutation where we substituted the glutamine with leucine at 39’ position (Q39’L). These mutated β subunits were expressed as homomeric receptors in oocytes for functional investigations. Figure A1: Effect of glutamate on WT and two β HcoGluClR mutants using TEVC. (A) A sample dose- response trace for glutamate on WT β HcoGluClRs (B) A sample dose-response trace for glutamate on TM1 mutation of I-18’L. (C) A sample concentration response trace for glutamate on Q39’L. (D) Glutamate concentration-response graphs for WT, I-18’L and Q39’L (n=6 at least).

204

Figure A2: Effect of IVM on I-18’L. (A) A cumulative dose-response trace for IVM at β I-18’L using TEVC. (B) Concentration-response graph for IVM at β I-18’L (n=6).

205

Table A1: Table showing the EC50 values of glutamate and IVM at two mutants from β homomeric HcoGluClRs (n=6 at least)

H. contortus

Agonist β WT (EC50) β I-18’L (EC50) β Q39’L (EC50)

Glutamate 394 ± 57 µM 16 ± 2 µM 2.8 ± 0.3 mM

Ivermectin >10 µM 344 ± 42 nM > 10 µM

206

Section 2

Discovering new drugs selectively targeting α5 containing GABAARs

Table A5: GABA EC50 values from ToxFinder

Receptor EC50 (µM) Hill slope n

α1β2γ2L 1.0 ± 0.1 1.1 ± 0.1 10

α5β2γ2L 0.4 ± 0.1 1.2 ± 0.1 9

α5β3γ2L 0.2 ± 0.1 1.1 ± 0.1 9

Table A6: GABA EC50 values from whole cell patch clamp recordings

Receptor EC50 (µM) Hill slope n

α1β2γ2L 17.5 ± 0.1 1.4 ± 0.3 6

α5β2γ2L 8.8 ± 0.1 1.9 ± 0.1 6

α5β3γ2L 5.1 ± 0.1 1.6 ± 0.1 6

Figure A3: GABAAR concentration responses. Graph (A) illustrates the GABA concentration- responses recorded in HEK 293 cells expressing α1β2γ2L, α5β2γ2L and α5β3γ2L GABAARs on the ToxFinder. Graph (B) shows the GABA concentration-responses obtained from whole cell patch clamp recordings in α1β2γ2L, α5β2γ2L and α5β3γ2L GABAARs expressed in HEK 293 cells.

208

Table A7: Full list of crude extracts with more than 50% activity on the ToxFinder. The subsequent activities of fractions also included. Numbers in the bracket represent the % extent of potentiation (+) or (-) inhibition compared to control GABA EC20 quench. The fractions were n-butanol (n-BuOH), n-hexane, dichloromethane (DCM), methanol (MeOH) and aqueous methanol (Aq. MeOH, refer to main text figure 6.1). Green shade: potentiation; blue shade: inhibition; red shade: cell death; white shade: no activity or no significant activity.

Percentage Modulation (%) CMB-code Crude n-BuOH n-Hexane DCM MeOH Aq. MeOH extract (-1) (-1-1) (-1-2) (-1-3) (-2-1)

CMB-02130 +98.0 +92.0 +76.6 +70.9 +96.1 +58.6 CMB-03706 +83.1 +82.6 +82.0 +73.8 CMB-01871 +76.6 +68.8 +89.2 +63.3 +98.9 +87.8 CMB-03208 +89.2 +65.6 +97.4 +91.2 CMB-02736 +71.8 +69.6 +53.7 +60.6 +56.1 CMB-01699 +52.3 +53.2 +57.6 +61.9 CMB-03217 +50.9 +65.2 +57.2 CMB-03552 +72.5 +66.4 +73.0 +69.9 +56.4 CMB-02999 +68.9 +67.6 +63.0 +73.9 CMB-02031 +75.4 +81.6 +57.4 +73.8 +54.5 CMB-02968 +78.9 +60.5 +83.2 +70.1 +64.9 +82.5 CMB-02671 -35.5 -12.0 CMB-03456 +55.7 +52.4 +55.2 +53.3 +49.9 CMB-02733 +72.0 +75.8 +63.2 +59.3 CMB-02628 +45.3 -22.0 +53.2 +49.1 CMB-01900 +50.0 +60.8 -1.5 +73.2 CMB-01534 +21.1 CMB-02633 +28.9 CMB-01565 +27.3 CMB-03157 +22.5 CMB-03242 +69.3 +68.2 +72.8 +75.1 +69.4 CMB-02996 +70.0 +69.7 +64.8 +66.7 CMB-03210 +65.8 +63.5 +72.8 +61.0 +61.5

209

CMB-01700 +60.0 +55.1 +63.5 +61.9 CMB-03405 +58.6 +56.9 CMB-02571 +45.3 -10.8 -16.4 +53.1 +53.4 CMB-02758 +50.1 -7.0 +54.8 +63.1 +55.7 +54.3 CMB-03077 +71.1 +72.2 +67.8 +61.5 CMB-03247 +50.3 +62.1 CMB-03457 +62.5 +61.7 +59.3 +72.5 CMB-01845 +65.9 +61.9 +67.8 +59.8 CMB-03251 +53.5 +56.9 CMB-02709 -42.7 -13.1 CMB-03216 +35.3 CMB-03357 +62.3 +70.4 +58.3 +65.0 CMB-02997 +45.9

Table A8: Different fractions tested on the Patchliner.

-1-1 (Figure A3-A) CMB-01871

-1-2 (Figure A3-B)

CMB-03405 -1-2 (Figure A4-A)

-1-3 (Figure A4-B)

CMB-03208 -1-1 (Figure A5-A)

-2-1 (Figure A5-B)

CMB-02130 -1-3 (Figure A5-C)

210

Figure A4: (A) A sample dose response trace for GABA at α5β3γ2L GABAARs using the Patchliner. (B) Concentration response graph for α5β3γ2L GABAARs (n=6).

211

Figure A5: Traces from the fractions belonging to the CMB-01871 group. (A) A sample trace depicting the effect of n-Hexane (-1-1) fraction at α5β3γ2L GABAARs when co-applied at a concentration of 0.2 µg/µL with 1 µM GABA. Three control applications of GABA EC20 were done in order to determine the quality of the cell. (B) Similar experiment done with the DCM (-1-2) fraction at α5β3γ2L GABAARs.

212

Figure A6: Traces from the fractions belonging to the CMB-03405 group. (A) A sample trace from the Patchliner showing the effect of the DCM (-1-2) fraction at α5β3γ2L GABAARs when co-applied at a concentration of 0.2 µg/µL with 1 µM GABA. Three control applications of GABA EC20 were done in order to determine the quality of the cell. (B) Similar experiment done with methanol (-1-3) fraction.

213

Figure A7: Traces from the fractions belonging to the CMB-03208 and CMB-02130 group. (A) & (B) Traces from the Patchliner showing the effect of n-Hexane (-1-1) and aqueous methanol (-2-1) fraction from the CMB-03208 group, respectively, at α5β3γ2L GABAARs when co-applied at a concentration of 0.2 µg/µL with 1 µM GABA. Three control applications of 1 µM GABA were done in order to determine the quality of the cell. (B) Inhibitory effect of CMB-02130-1-2 (DCM fraction) at 0.2 µg/µL along with 1 µM GABA.

214

Table A9: Table depicts the products of further partitioning from the fractions belonging to the CMB-02130-1-3 and CMB-03405-1-3 groups.

CMB-02130-1-3 CMB-03405-1-3 -1 -2 -3 -4 -5 -6 -7 -8 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11

Figure A8: Bar graph showing the different sub-fractions from the CMB-02130-1-3 and CMB- 03405-1-3 groups and their effects on α5β3γ2L GABAARs when normalised to 1 µM GABA. All the sub-fractions were tested on the Patchliner. Out the 19 sub-fractions only 9 were significant when compared to 1 µM GABA (highlighted in either green or blue corresponding to CMB-02130-1-3 or CMB-03405-1-3 fractions respectively). All the values were averaged from at least five different experiments. *p<0.05, **p<0.01, using One-way ANOVA.

215

Figure A9: Traces from the sub-fractions belonging to fractions CMB-02130-1-3. (A), (B) & (C) Traces from the Patchliner showing the effect CMB-02130-1-3-2, CMB-02130-1-3-6 and CMB-02130-1-3-8 at α5β3γ2L GABAARs when co-applied at a concentration of 0.2 µg/µL with 1 µM GABA, respectively.

216

Figure A10: Traces from the sub-fractions belonging to fractions CMB-03405-1-3. (A), (B) & (C) Traces from the Patchliner showing the effect CMB-03405-1-3-1, CMB-03405-1-3-5 and CMB-03405-1-3-6 at α5β3γ2L GABAARs when co-applied at a concentration of 0.2 µg/µL with 1 µM GABA, respectively.

217

Figure A11: Traces from the sub-fractions belonging to fractions CMB-03405-1-3. (A), (B) & (C) Traces from the Patchliner showing the effect CMB-03405-1-3-7, CMB-03405-1-3-9 and CMB-03405-1-3-11 at α5β3γ2L GABAARs when co-applied at a concentration of 0.2 µg/µL with 1 µM GABA, respectively.

218

Animal Ethics Approval Certificate for the use of oocytes from Xenopus Laevis

219

UQ Research and Innovation Director, Research Management Office Nicole Thompson 20-Apr-2016 Animal Ethics Approval Certificate Please check all details below and inform the Animal Welfare Unit within 10 working days if anything is incorrect.

Activity Details Chief Investigator: Professor Joseph Lynch, Queensland Brain Institute Title: Investigating the effects of new drugs for pain, stroke and anxiety on neuronal ion channels AEC Approval Number: QBI/AIBN/087/16/NHMRC/ARC Previous AEC Number: Approval Duration: 20- Apr-2016 to 20-Apr -2019 Funding Body: ARC, NHMRC Group: Anatomical Biosciences Other Staff/Students: Lachlan Rash, Natalie Saez, Irene Chassagnon, Angelo Keramidas, Mohammed Atif, Jessie Er, Linlin Ma, Ming Soh, Ben Cristofori-Armstrong, Argel Estrada, Nela Durisic, Glenn King, Jennifer Smith, Sharifun Islam Location(s): St Lucia Bldg 75 - AIBN

Summary Subspecies Strain Class Gender Source Approved Remaining Amphibians Xenopus Laevis Adults Female 58 58 ( breeder ) Amphibians Marsabit Clawed Adults Female 22 22 Frog (Xenopus borealis) (breed)

Permits

Provisos Pest Species (Xenopus frogs) Proviso: a) Xenopus frogs must have permanent identification marks and be housed according to the DNR permit regulations b) The CI must ensure that the toads are well identified and accurate records are kept of their use.

The CI is required to ensure that no animal is used for oocyte collection more than 6 times.

Approval Details

Description Amount Balance

Amphibians (Marsabit Clawed Frog (Xenopus borealis) (breed), Female, Adults, ) 20 Apr 2016 Initial approval 22 22 Amphibians (Xenopus Laevis (breeder), Female, Adults, ) 20 Apr 2016 Initial approval 58 58

220

Animal Welfare Unit Cumbrae-Stewart Building +61 7 336 52925 (Enquiries ) [email protected] UQ Research and Innovation Research Road +61 7 334 68710 (Enquiries ) uq.edu.au/research The University of Queensland Brisbane Qld 4072 Australia +61 7 336 52713 (Coordinator ) P 1 f 2

Please note the animal numbers supplied on this certificate are the total allocated for the approval duration Please use this Approval Number: 1 . When ordering animals from Animal Breeding Houses 2 . For labelling of all animal cages or holding areas. In addition please include on the label, Chief Investigator's name and contact phone number. 3 . When you need to communicate with this office about the project.

It is a condition of this approval that all project animal details be made available to Animal House OIC. ( UAEC Ruling 14/12/2001)

The Chief Investigator takes responsibility for ensuring all legislative, regulatory and compliance objectives are satisfied for this project. This certificate supercedes all preceeding certificates for this project (i.e. those certificates dated before 20-Apr-2016)

221

Animal Ethics Approval Certificate for the use of rat embryos

222

Office of Research Ethics Director Nicole Shively 21-Mar-2018 Animal Ethics Approval Certificate Please check all details below and inform the Animal Ethics Unit within 10 working days if anything is incorrect.

Activity Details Chief Investigator: Professor Joseph Lynch, Queensland Brain Institute Title: Generating artificial synapses for the purpose of evaluating new drugs for neurological disorders AEC Approval Number: QBI/142/16/NHMRC/ARC Previous AEC Number: QBI/203/13/ARC Approval Duration: 10- Jun-2016 to 10-Jun- 2019 Funding Body: ARC Group: Anatomical Biosciences Other Staff/Students: Nela Durisic, Maleeha Waqar, Dejan Gagoski, Edmund Cotter, Angelo Keramidas, Parnayan Syed, Xiumin Chen, Mohammed Atif, Trish Hitchcock Location(s): St Lucia Bldg 79 - Queensland Brain Institute

Summary Subspecies Strain Class Gender Source Approved Remaining Rats - non Wistar Adults Female Institutional 150 113 genetically Breeding Colony modified Rats - non Wistar Prenatal / Embryo Unknown Institutional 2400 1900 genetically Breeding Colony modified

Permits

Provisos

Approval Details

Description Amount Balance

Rats - non genetically modified (Wistar, Female, Adults, Institutional Breeding Colony) 2 Jun 2016 Initial approval 150 150 31 Dec 2016 Use in 2016 (from 2017 MAR; AEMAR25405) -17 133 31 Dec 2017 Use in 2017 (from 2018 MAR; AEMAR38049) -20 113 Rats - non genetically modified (Wistar, Unknown, Prenatal / Embryo, Institutional Breeding Colony) 2 Jun 2016 Initial approval 2400 2400 31 Dec 2016 Use in 2016 (from 2017 MAR; AEMAR25405) -229 2171 31 Dec 2017 Use in 2017 (from 2018 MAR; AEMAR38049) -271 1900

223

Animal Ethics Unit Cumbrae-Stewart Building +61 7 336 52925 (Enquiries ) [email protected] Office of Research Ethics Research Road +61 7 334 68710 (Enquiries ) uq.edu.au/research The University of Queensland St Lucia Qld 4072 Australia +61 7 336 52713 (Coordinator ) P 1 f 2

It is a condition of this approval that all project animal details be made available to Animal House OIC. ( UAEC Ruling 14/12/2001)

224