Inhibiting Voltage Gated Sodium Channel 1.7 with Spider-Venom Peptides

Developing novel analgesics to treat chronic pain: Inhibiting voltage gated sodium channel 1.7 with spider-venom peptides

Darshani Buddhika Rupasinghe Arachchilage BSc in Biochemistry and Microbiology (Honors)

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in February 2014 Institute for Molecular Bioscience

ABSTRACT

Chronic pain is responsible for great physical, mental and economic loss. While there are diverse methods available for managing pain, such as NSAID, opioids and local anesthetics, common issues associated such as addiction and tolerance highlights the unmet demand for a novel analgesic aimed at a novel target.

Numerous reports on individuals suffering from loss of function mutations in SCN9A marker, which leads to complete inability to feel pain, yet otherwise normal physiology, has led to the recognition of voltage gated sodium channel 1.7 (NaV1.7) as a prime pain target. Other reports of gain of function mutations that lead to conditions such as paroxysmal extreme pain disorder and primary erythromelalgia confirms the validity of this target. However the presence of nine closely related subtypes of voltage gated sodium channels,

NaV1.1-NaV1.9 introduces complications as any cross inhibition may lead to detrimental results. For example inhibition of NaV1.5 subtype could result in cardiac arrhythmia to complete cardiac arrest and death. Therefor it is of utmost importance that any pharmaceutical agent used to inhibit NaV1.7 is highly specific for NaV1.7 and display no cross reactivity on other voltage gated channels.

Spiders are one of the most successful terrestrial predators; their venoms have evolved over many millions of years to selectively and potently inhibit nervous system targets in order to rapidly paralyze their prey. Spider-venom peptidomes have been identified as one of the largest known libraries of compounds with very high potency and specificity towards nervous system targets.

Spider toxins listed in the Arachnoserver database have been classified into 12 families of sodium channel toxins (NaSpTx), denoted NaSpTx Families 1–12, based on the level of sequence conservation and intercystine spacing. NaSpTx family 2 is the largest toxin family; it comprises 34 peptides of theraphosid origin.

This thesis primarily aims at discovering novel inhibitors of NaV1.7. However the intriguing fact that people suffering form congenital indifference to pain also suffers form anosmia, has lead to efforts of understanding the role of NaV1.7 in olfaction. This study has been

ii completed with the finding that NaV1.7 is located along axons of olfactory neurones. These findings have been published and research article is attached as chapter 4.

During the initial discovery phase of this project, an assay guided fractionation method was used to identify NaV1.7 inhibitors form six spider venoms. This resulted in nine fully sequenced peptides where four of these belonged to NaSpTx family 2. This included the well-known β/ω-TRTX-Tp1a (Protoxin1). Three novel peptides discovered were named β- TRTX-Pe1a, µ-TRTX-Pe1b, and U-TRTX-Pa1a.

A complete alanine scan of β/ω-TRTX-Tp1a on NaV1.7 was conducted using a Xenopus laevis oocyte based tethered toxin method. Residues that were important for NaV1.7 binding were mapped onto the three-dimensional structure of β/ω-TRTX-Tp1a. Three novel toxin peptides β-TRTX-Pe1a, β/µ-TRTX-Pe1b, and U-TRTX-Pa1a were recombinantly expressed using an Escherichia coli (E. coli) based system. Channel modulating activities of these toxins were assessed using a Xenopus laevis oocyte based two-electrode voltage clamp system (TEVC). µ-TRTX-Pe1b demonstrated a high degree of selectivity for NaV1.7. It is about 20 times more selective over NaV1.2, which is abundant in the central nervous system, and appears to mildly potentate NaV1.5, which is the cardiac subtype. This is despite the complete set of active residues uncovered in β/ω-

TRTX-Tp1a alanine scan being fully conserved in β/µ-TRTX-Pe1b. β-TRTX-Pe1a and U1-

TRTX-Pa1a have shown very weak NaV1.7 inhibition, hence not investigated in great detail.

The introduction of the mutation S35 to the sequence of β/µ-TRTX-Pe1b has resulted in the β/µ-TRTX-Pe1b-S35, which has shown increased NaV1.7 activity that is almost comparable to β/ω-TRTX-Tp1a. Structure activity relations (SAR) of β/ω-TRTX-Tp1a and β/µ-TRTX-Pe1b-S35 are currently under investigation. It is highly promising that better understanding of residues responsible for the high potency of β/ω-TRTX-Tp1a and higher

NaV selectivity of β/µ-TRTX-Pe1b-S35 would lead to engineering a superior NaV1.7 inhibiting analgesic agent.

This thesis has widened our understanding of NaSpTx family 2 peptides, and residues, which may be important in achieving NaV1.7 selectivity by inhibitor peptides.

iii

Declaration by author

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

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

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the General Award Rules of The University of Queensland, immediately made available for research and study in accordance with the Copyright Act 1968.

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

Publications during candidature

Peer reviewed papers

Rupasinghe, D.B., Knapp, O., Blomster, L.V., Schmid, A., Adams, D.J., King, G.F., and

Ruitenberg, M.J. (2012) Localization of NaV1.7 in the normal and injured rodent olfactory system indicates a critical role in olfaction, pheromone sensing and immune function. Channels.

Klint, J.K., Senff, S., Rupasinghe, D.B., Er, S.Y., Herzig, V., Nicholson, G.M., and King, G.F. (2012) Spider-venom peptides that target voltage-gated sodium channels: pharmacological tools and potential therapeutic leads. Toxicon

Conference Abstracts

Rupasinghe, D.B., Knapp, O., Blomster, L.V., Schmid, A., Adams, D.J., King, G.F., and

Ruitenberg, M.J. (2011) Localization of NaV1.7 in the normal and injured rodent olfactory system indicates a critical role in olfaction, pheromone sensing and immune function. 4th Venoms to Drugs Meeting, Heron Island, Australia, May 15-20 (poster).

Darshani b. Rupasinghe, Norelle Daly, Mehdi Mobli, Junhong Gui, Michael N. Nitabach and Glenn F. King (2013) A tethered toxin approach to understand the NaV1.7 binding pharmacophore of Protoxin-1. Chemistry and Structural biology symposium, Institute for molecular Bioscience, University of Queensland, Brisbane, Australia, November 7th (poster).

v Publications included in this thesis

Introduction: Selected sections of the review indicated below that I coauthored are incorporates into the introduction of this thesis, especially the section on NaSpTx family 1, which was written exclusively by me.

Contributor Statement of contribution Rupasinghe, D.B., (Candidate) Designed experiments (50%) Wrote and edited paper (35%) Klint, J.K Designed experiments (25%) Wrote and edited paper (30%) Senff, S Wrote and edited paper (10%) Er, S.Y Wrote and edited paper (10%) Herzig, V Edited the manuscript Nicholson, G.M. Edited the manuscript King, G.F. Wrote and Edited the paper (15%) Designed experiments (25%)

Chapter 3: The complete published paper is attached as the Chapter 3

Rupasinghe, D.B., Knapp, O., Blomster, L.V., Schmid, A., Adams, D.J., King, G.F., and

Ruitenberg, M.J. (2012) Localization of NaV1.7 in the normal and injured rodent olfactory system indicates a critical role in olfaction, pheromone sensing and immune function. Channels.

Contributor Statement of contribution Rupasinghe, D.B (Candidate) Designed experiments (70%) Wrote the paper (80%) Knapp, O Designed experiments (10%) Wrote the paper (5%) Blomster, L.V., Designed experiments (5%) Schmid, A., Designed experiments (5%) Adams, D.J Edited the manuscript King, G.F Edited the manuscript Wrote the paper (5%) Ruitenberg, M.J Designed experiments (10%) Wrote the paper (10%) Edited the manuscript

vi Contributions by others to this thesis

Chapter two: ProTx1 was chemically synthesized by Mr Zoltan Dekan. The structure was determined using 2D homonuclear NMR spectroscopy in collaboration with Dr Mehdi Mobli (Centre for Advanced Imaging, UQ) and Prof. Norelle Daly (James Cook University). A/Prof. Frank Bosmans and John Gilchrist performed the kinetics experiments for the

NaV1.7/KV2.1 chimeric constructs.

Chapter three: The venom based discovery process is a collaboration with the laboratory of Prof Richard Lewis at the Institute for Molecular Bioscience, The University of Queensland. Dr Irina Vetter conducted all FLIPR assays.

Chapter four: (Incorporated paper) Dr Linda V. Blomster, and Dr Anina Schmid were involved in the preliminary findings that olfactory tissue contain NaV1.7 and it is detectable by immunohistochemical methods. Dr Oliver Knapp was involved in checking the NaV subtype specificity of the anti-NaV1.7 antibody used for localizations.

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

None.

vii

Acknowledgement

I would like to express my deepest gratitude to my supervisor Professor Glenn King for introducing me to the exciting world of spider toxins, for allowing me to challenge myself in new techniques and, most importantly the trust that was placed in my work. I would also like to thank my co-supervisor Dr. Susan Rowland for her continuous role as a mentor.

I would also like to thank A/Prof Frank Bosmans for allowing me the opportunity to work in his lab. The trip to United States and the four weeks of electrophysiology training with A/Prof Bosmans made me realize what is it about science that I really enjoy doing. I should also thank Dr Mark Ruitenburg and Prof Matt Cooper for their invaluable contributions as my PhD committee members. A special thanks goes to Dr Ruitenburg for his support during NaV1.7 localization studies. I also thank Prof Matt Cooper and Prof Matt Sweet for allowing me to work in their labs. Very special thanks for Dr. Amanda Carozzi for her invaluable support throughout my candidature.

I also thank Dr Lachlan Rash and Dr Mehdi Mobli, for all the valuable discussions we have had over the years specially on the NaV1.7 meetings. I would also like to thank present and past king group members for their support and friendship. Radha, Jessie, Sandy for our long ‘meaning-of-life’ conversations; Natalie, Jonas, David, Ravi, Carus, Julie, Sebastian, Ben, Irene, Eivind, Evelyn, Nausad, Bruno, Niraj, Maria, Volker, Maggie for their friendship and support in the last five years.

Very special thanks to Anthony, who had to deal with all the ups and downs of my PhD. I thank Jenny and Mike for their support with everything last four years. I also thank my dearest sister for her love, and trust in me. Last but not least, none of this work would have been possible without the love and support of my parents. Thank you for your unshaken trust and belief in me, even at times I was deep in self-doubt.

viii

Keywords

Chronic pain, analgesic, NaV1.7, structure activity relations, pharmacophore, alanine scan mutagenesis, tethered toxins, two electrode voltage clamp, three-dimensional structure, electrophysiology.

Australian and New Zealand Standard Research Classifications (ANZSRC) ANZSCR code: 030406, Proteins and peptides 40% ANZSCR code: 101501, Basic Pharmacology 35% ANZSCR code: 060110, Receptors and Membrane Biology 25%

Fields of Research (FoR) Classification FoR code: 1115, Pharmacology and Pharmaceutical Sciences 50% FoR code: 0304, Medicinal and Bimolecular Chemistry 50%

ix Table of Content

CHAPTER 1: INTRODUCTION 1.1 Chronic pain 1 1.2 Categories of pain and current treatments 2 1.3 Chronic pain, and current management strategies 2 1.4 What are the likely pain targets and where to find them? 3 1.5 Voltage gated sodium channels (Nav) 4

1.5.1 General function of NaV channels 4

1.5.2 NaV structure 4

1.6 NaV channel subtypes 6

1.7 NaV channel localization and channelopathies 6 1.8 Congenital indifference to pain 8

1.9 Spider venom peptides as NaV-targeted analgesic drugs 9 1.10 Mechanisms of toxin activity 12 1.11 NaSpTx family 2 13 1.12 Research aims 18

CHAPTER 2: STRUCTURE-ACTIVITY RELATIONSHIPS OF ProTx1:

ACHIEVING BETTER SPECIFICITY FOR NaV1.7 2.1 introduction 19 2.2 materials and methods 21 2.2.1 TEVC for tethered-toxin-based ProTx1 alanine scan 21 2.2.2 Oocyte surface protein assay 21 2.2.3 Electrophysiological assay of toxin activity 21

2.2.4 Generation of NaV1.7/Kv1.1 chimeras 22

2.2.5 TEVC recordings using NaV1.7/Kv1.1 chimeras 23 2.3 Results 24

2.3.1 ProTx1 activity on NaV channels 24

2.3.2 Identifying the ProTx1 pharmacophore for NaV1.7 25

2.3.3 Structure of ProTx1: comparison of NaV1.7 and

NaV1.2 pharmacophores 28 2.3.4 ProTx1 interacts primarily with the domain II and

domain IV paddle motifs of hNaV1.7 30 2.3.5 Design and characterization of ProTx1 mutants 31

x 2.4 Discussion 35

CHAPTER 3: DISCOVERY AND CHARACTERIZATION OF A NOVEL NaV1.7 INHIBITOR, β/µ-TRTX-Pe1b 3.1 Introduction 42 3.2 Materials and methods 45 3.2.1 Fractionation and purification of spider venoms 45 3.2.1.1 RP-HPLC 45 3.2.1.2 Ion exchange HPLC 45 3.2.1.3 Desalting by reverse phase HPLC 45 3.2.2 FLIPR assays 45 3.2.3 Reduction of disulfide bonds and alkylation of cysteine residues 46 3.2.4 Mass spectrometry (MALDI-TOF and MADLI-ISD) 46 3.2.5 Recombinant expression of µ-TRTX-Pe1b 47 3.2.5.1 Transformation 47 3.2.5.2 Recombinant expression, purification and cleavage of the fusion protein 47 3.2.5.3 Purification using a solid phase extraction column 48

3.2.5.4 Purification of the cleaved peptide with C4 semi-prep column 48 3.2.6 Two-electrode voltage clamp electrophysiology 48 3.2.6.1 Electrophysiological assay of toxin activity 48 3.2.6.2 Analysis of Electrophysiological data 49 3.2.7 Construction of the β-TRTX-Pe1b homology model 49 3.3 Results 50 3.3.1 Fractionation and purification of crude venoms to identify active peptides 50 3.3.2 Recombinant expression of toxins 56

3.3.3 Pe1a and Pa1a activities on NaV channels 60

3.3.4 Pe1b activities on NaV currents 61 3.3.5 Construction of Pe1b mutants 61 3.3.6 Recombinant expression and activity of Pe1b mutants 64 3.4 Discussion 68

CHAPTER 4: LOCALIZATION OF NaV1.7 IN THE NORMAL AND INJURED RODENT OLFACTORY SYSTEM INDICATES A CRITICAL ROLE IN OLFACTION, PHEROMONE SENSING AND IMMUNE FUNCTION (Incorporated research paper) 70

CHAPTER 5: CONCLUSION AND FUTURE DIRECTIONS 79 REFERENCES 81 APPENDIX 1 91

List of Figures & Tables

Figure 1: The nociceptive pathway 1

Figure 2: Generic structure of mammalian NaV channels 5

Table 1: Function and distribution of voltage gated sodium channels 8

Figure 3: Ion channel modulators form spider venoms 10

Figure 4: Molecular architecture and pharmacology of NaV channels 11

Figure 5: The inhibitory cystine knot (ICK) motif 12

Figure 6: NaSpTx families 1–4, 7 and 10 14

Figure 7: NaSpTx Families 5, 6, 8, 9, 11 and 12 15

Table 2: Subtype selectivity and other selected properties of the12 families of NaSpTx 16

Figure 8: Concentration-response curves for ProTx1 inhibition

of human NaV channels 24

Figure 9: The tethered toxin approach 26

Figure 10: Alanine scanning mutagenesis of ProTx1 27

Figure 11: Structure of ProTx1 highlighting pharmacophores

for NaV1.7 and NaV1.2 29

Figure 12: Identification of hNaV1.7 voltage sensing domains that interact with ProTx1 31

Figure 13: The sequence alignment of ProTx1, µ-TRTX-Pe1a and β-TRTX-Pe1b 32

Figure 14: Sequence alignment of NaSpTx Family 2 peptides 33

Figure 15: Effect of ProTx1 mutants on NaV activity 34

Table 3: NaV channel activity of ProTx1 mutants 34

Figure 16: ProTx1 binding residues of rNaV1.2 37

xii Figure 17: The sequence alignment of all NaV S1-S2 and S3-S4 segments of domains I, II and IV 39

Supplementary Figure 1: Membrane tethering has no effect on

ProTx1 inhibition of NaV channels 40

Supplementary Figure 2: Alanine-scanning mutagenesis of t-ProTx1 41

Supplementary Figure 3: Immunoassay of alanine mutant surface expression 41

Figure 18: Assay guided fractionation and discovery of analgesic drug leads 50

Figure 19: Initial venom screen 51

Figure 20: Fractionation of crude spider venoms using reverse phase HPLC (RP-HPLC) 52

Figure 21: FLIPR-based assay profiles for 12 selected spider-venom fractions 53

Figure 22: Purification of active fractions of the venom 785 54

Table 4: NaV1.7 inhibitor peptides isolated from crude venom 55

Figure 23: Sequence alignment of toxins in NaSpTx family 2 56

Figure 24: The design of the recombinant expression system 57

Figure 25: Recombinant expression of β/µ-TRTX-Pe1b 58

Figure 26: Recombinant expression of β-TRTX-Pe1a, a NaV1.7 active peptide from the spider Pormingochillus everetti 59

Figure 27: Recombinant expression of U1-TRTX-Pa1a, a NaV1.7 active peptide from the spider Pampobetious antinous 60

Figure 28: The current-voltage (IV) relationship of β/µ-TRTX-Pe1b in

NaV1.2, NaV1.5, NaV1.6 and NaV1.7 61

Figure 29: β/µ-TRTX-Pe1b activity on NaV and designing β/µ-TRTX-Pe1b based mutant peptides 63

Figure 30: Recombinant expression of Pe1b-E17K, Pe1b-S20V, Pe1b-L2 and Pe1b-S35 64

Figure 31: NaV activity of the four Pe1b mutant peptides 65

Table 5: Concentration-response data of Pe1b mutant-peptides

on NaV1.2, NaV1.5 and NaV1.7 66 .

xiii List of Abbreviations

1,5‐DAN 1,5‐diaminonaphthalene Amp Ampicillin ASIC Acid sensing ion channel BSA Bovine serum albumin

KV Voltage-gates potassium channels

CaV Voltage-gated calcium

NaV Voltage-gated sodium channels hNaV Human voltage-gated sodium channel rNaV Rat voltage-gated sodium channels CHCA α‐Cyano‐4‐hydroxycinnamic acid CIP Congenital indifference to pain CNS Central nervous system DTT Dithiothreitol EDTA Ethylenediaminetetraacetic acid FLIPR Fluorescent imaging plate reader GL Glomerular layer HPLC High performance liquid chromatography HSAN Heriditary sensory autosomal neuropathy IC50 half maximal inhibitory concentration ICK Inhibitor cystine knot LP Laminar propria MALDI‐TOF Matrix‐assisted laser desorption/ionization-Time of Flight MBP Maltose‐binding protein NMR Nuclear magnetic resonance NSAID Non steroidal anti-inflammatory drug OB Olfactory bulb OE Olfactory epithelium OEC Olfactory ensheathing cells OMP Olfactory marker protein ONL Olfactory nerve layer ORN Olfactory receptor neurone PNS Peripheral nervous system SAR Structure‐activity relationship

xiv TEV Tobacco etch virus TEVC Two‐electrode voltage clamp TFA Trifluoroacetic acid TRP Transient receptor potential TRTX Theraphotoxin VSD voltage sensor domain TTX tetrodotoxin DRG Dorsal root ganglion

NaSpTx NaV targeting spider toxins ProTx1 Protoxin 1/ β/ω-TRTX-Tp1a cRNA coding RNA GPI glycophosphatidylinositol t-toxin membrane tethered toxin ELISA Enzyme-linked immunosorbent assay

xv CHAPTER 1

Introduction

1.1 Chronic pain. Pain is a noxious stimulus that indicates current or possible future tissue damage1. Pain is emotionally detrimental but crucial for survival. Certain noxious stimuli such as injury, temperature extremes, chemicals, and mechanical trauma can be detected by specialized receptors known as primary sensory neurons or nociceptors (Fig. 1)2,3. These stimuli are transmitted to the central nervous system, via the spinal cord, as an action potential through high-threshold unmyelinated C fibres and thinly myelinated Aδ fibres. Afferent neurons synapse in the spinal cord, and the data is transmitted to both the brain for processing (via second and third order sensory neurons) and to motor nerves (via interneurons) for immediate action to avoid possible further damage4. The brain is responsible for interpretation of the initial signal from nociceptors as ‘pain’.

A B 23+,&4%4),')'5%)4&$+,-)#"(6%

D+"#$%

(*+,"C%7*+$%$'B+*$%H%

G% G% G% &:CB8%

6I!68% -E/%

K2"L%% -IJ68%% H%

2".903% ! M68% 3 2".303% 2".303% :6;6;8% -" /09%!% !4!/5$6278% . !,./01)!) !9"3@F$6278% -"./01%!%

!+#&"+A%,'$,*+A%$'B+*$%

C !"#$#%&'"()'%(*$+,-)

2" 303% -"./01%% . 2" 903% !, /01%% . . !4!/5$6278%

-"./09%% :6;6;8% 2".303%

!3?3@9% 6=>2,% <8!,% :!28,% !"#$%&'(#")*+,%%%%%%%%%%%%%

Figure 1: (A) The nociceptive pathway3. Noxious external stimuli detected at the terminus of a nociceptor are converted to an action potential and transmitted to the spinal cord via afferent nerve bundles5. Within the grey matter of the dorsal horn of the spinal cord, the dorsal root ganglion (DRG) neuron synapses with an interneuron that transmits the stimulus to the brain, where the nerve impulse is perceived as ‘pain’. The interneuron also synapses with an efferent neuron for

1 immediate involuntary responses4. (B) Molecular components at the synapse between the central terminations of nociceptors and interneurons/second-order sensory neurons in the spinal cord. (C) Molecular components in the naked terminals of nociceptors. Transient receptor potential (TRP) channels in the nociceptors detect intense stimuli generated by thermal, mechanical and chemical 6 sources . Numerous ion channels including the voltage-gated sodium (NaV) channel NaV1.7 are located along the axon of the DRG neuron, which are important for the amplification and propagation of action potentials.

1.2 Categories of pain Pain is broadly categorized into three classes. Nociceptive pain arises due to noxious stimuli from the external environment5. Inflammatory pain is caused by proinflammatory mediators, which are released at sites of infection or tissue damage7. Neuropathic pain, which is also known as dysfunctional pain, is indicative of dysfunction or damage to the nervous system2. Pain can be further divided, based on its duration, into acute or chronic pain.

Acute pain is temporary, and often ceases to exist once the cause is eliminated. Acute pain can usually be successfully managed with non-steroidal anti-inflammatory drugs (NSAIDs), opioids, local anaesthetics and anticonvulsants. Chronic pain is defined as pain that extends beyond the period of healing following the initial trauma. Mechanisms common to chronic pain include detection of inflammatory mediators by nociceptors, peripheral sensitization and central sensitization2,8,9. Some of these pain states, most particularly neuropathic pain, do not respond well to current analgesics10.

1.3 Chronic pain and current management strategies Management of chronic pain presents a major challenge. Current estimates show that 15– 20% of the world population suffers from chronic pain, and this increases to 50% for adults over 65 years of age11,12. This number is predicted to increase over the next few decades as the proportion of seniors increases due to rising life expectancy. Currently used painkillers are categorized into three classes based on their strength and uses. Non-opioid analgesics such as aspirin, paracetamol and NSAIDs are used at the beginning of the treatment. Where the pain persists or increases, weak opioids such as codeine and tramadol are used. If this stage also fails, strong opioids such as oxycodone, morphine or fentanyl are prescribed in combination with an antidepressant, anticonvulsant or a non- opioid drug13. Unfortunately the long-term use of strong opioids can lead to tolerance,

2 physical dependence, and addiction14. Development of opioid tolerance reduces neuronal activity in the central nervous system through inhibition of voltage gated calcium (CaV) channels and inhibition of adenylyl cyclase15. Moreover, the abovementioned drug combinations have failed to provide adequate relief for patients suffering from a number of conditions including facial neuralgia, AIDS, multiple sclerosis, nerve injury after accidents or surgery, and certain neuropathies. Thus, there is much interest in identifying novel pain targets and analgesics with new modes of action.

1.4 What are the likely pain targets and where to find them? Ion channels and receptors present on peripheral nociceptors and synaptic junctions in the central nervous system (CNS) represent attractive targets for the development of novel analgesics. These include NaV1.7, NaV1.8, CaV2.2, CaV3.2, the GABAB receptor, and acid- sensing ion channels (ASICs) of the peripheral nervous system (PNS), and NaV1.8 and 16-20 CaV3.2 in the CNS . Noting that ongoing peripheral input is important for the maintenance of chronic pain, nociceptors have been identified as a very important target for the development of novel analgesics21. Reduction of hyperexcitability in peripheral nociceptors has been achieved by blocking voltage-gated and ligand-gated ion channels.

Significant success has been achieved with inhibition of NaV and CaV channels and increased expression of NaV1.7 was inhibited with cyclooxygenase inhibitors to relieve inflammatory pain22. In a different study it was shown that antisense-mediated knockdown 23-25 of NaV1.8 attenuates complete Freund’s adjuvant (CFA)-induced hyperalgesia .

Systemic administration of lidocaine, a NaV channel blocker, has yielded significant and long lasting pain relief for patients suffering from chronic pain due to postherpetic neuralgia, painful diabetic neuropathy, and idiopathic trigeminal neuralgia24. Lacosamide (Schwarz Pharma) and ralfinamide (Newron Pharma) have both been shown to inhibit 26 tetrodotoxin (TTX)-resistant (R) and TTX-sensitive (S) NaV channels . Even though lacosamide showed efficacy in Phase 3 clinical trials, there have been high rates of adverse events and dropout in trials with patients suffering from diabetic neuropathic pain. Ralfinamide has shown good tolerability in a mixed population of neuropathic pain patients 26-28 in Phase 2 clinical trials . NaV channels on nociceptors appear to be a prominent target for the development of novel analgesic drugs.

3 1.5 Voltage gated sodium channels (NaV)

1.5.1 General function of NaV channels

NaV channels are transmembrane proteins that cause the rapid depolarization required to initiate an action potential in excitable cells29. The single pore forming α-subunit is made up of four domains denoted I–IV, where each domain consists of six transmembrane (TM) segments (S1–S6) joined via intracellular and extracellular loops (Fig. 2)30-32. TM segments S1–S4 of every domain form a voltage sensor domain (Fig. 2B, orange), while the S5–S6 segments and the intervening loop come together in a cyclic arrangement to form the ion selectivity filter and the channel pore. The S4 segment of each voltage sensor contains a positively charged residue in every third position (Arg or Lys), and hence it functions as the primary sensor of changes in membrane potential30. This segment is responsible for the initiation of voltage dependent activation (i.e., the transition form the closed to open state) by responding to membrane depolarization and causing the channel to undergo a conformational change that allows selective influx of Na+ ions through the pore32. The extracellular linker between S5 and S6 of each domain re-enters the membrane to form the ion selectivity filter of the external face of the pore. The intracellular linker between domain III and domain IV, carrying residues IFMT (isoleucine, phenylalanine, methionine, threonine), is believed to fold into the ion conducting pore, and lead to the fast inactivation (i.e., the transition form the open to inactivated state) of the 30,32 33 channel . NaV channels produce rapidly depolarizing currents in excitable cells and the typical current profile is demonstrated below (Fig. 2C).

1.5.2 NaV structure

No three dimensional structure is currently available for any vertebrate NaV channel.

However, the crystal structure of a bacterial NaV channel (NaVAb) from Arcobacter butzleri was solved in 2011 at 2.7 Å resolution; the channel appears to be in pre-open confirmation with all four voltage sensor domains in the activated state and the ion-conducting pore 34 module closed . The primary difference between bacterial (NaChBac) and vertebrate NaV channels is that the bacterial channels are homotetramers, much like voltage-gated potassium (Kv) channels (Fig. 2A). Due to this difference, bacterial NaV channels lack the fast inactivation gates found in vertebrate channels. This leads to the assumption that bacterial channels employ an inactivation mechanism similar to that of slow inactivation of 35 vertebrate NaV channels . Later work revealed two structures of NaChBacs in the inactivated state, which has provided important insights into channel gating mechanisms34.

Structures of two more NaChBacs became available in 2012. The first was that of NaVRh,

4 a NaChBac from the marine bacterium Rickettsiales sp., solved at 3.05 Å resolution in the inactivated state36. A few months later, a 4 Å resolution structure was published of a NaChBac from the marine bacterium Magnetococcus sp. (strain MC-1); this structure is thought to represent the open channel conformation37.

Figure 2: Generic structure of mammalian NaV channels. (A) Eukaryotic NaV channels contain four homologous but non-identical domains (I–IV), each consisting of six TM segments (S1–S6). The S1-S4 segments within each domain forms a voltage sensor domain (VSD), while the S5–S6 segments and the re-entrant loop of the four domains come together to form the ion pore and selectivity filter. The intracellular loop between domains III and IV forms the inactivation loop30. (B)

3D structure of the NaV channel from Arcobacter butzleri obtained in a closed-pore and activated VSD conformation at 2.7 Å resolution. The central pore is shown in green while the four VSDs are shown in orange. (C) Classic NaV channel current profile. When the membrane is polarized to – 90 mV, the channel is inactive. Upon depolarization (shift of membrane potential to +30 mV), the channel becomes activated. Sodium ions enter the cell through the NaV channel pore. Within a few ms the inactivation loop enters the pore and the channel inactivates. When the membrane is repolarized to –50 mV, the channel partially recovers from inactivation, yet no current flows38. Further repolarization to -90 mV would return the channel to its inactive (closed) state. (A), was adapted from a pervious review which has been incorporated as part of the introduction32.

1.6 NaV channel subtypes

Vertebrates contain nine different NaV channel α-subunits, denoted NaV1.1– NaV1.9, that share 50% identity within their TM and extracellular domains29. All these subtypes, with the exception of NaV1.9, have been functionally expressed. They are often characterized by their sensitivity to TTX. NaV1.5, NaV1.8 and NaV1.9 are TTX resistant, whereas all other 29,33 subtypes are TTX sensitive . NaV1.1, NaV1.2, NaV1.3 and NaV1.7 are the most closely related channels based on the evolutionary distance measured using sequence alignments. The genes encoding these channels are all located within the chromosomal region 2q23-24.

NaV1.5, NaV1.8 and NaV1.9 are also closely related. They share 64% sequence identity 29 and display some level of TTX resistance . NaV1.4 is primarily expressed in skeletal muscle, while NaV1.6 is highly abundant in nodes of Ranvier within both the CNS and PNS. These two channels are not closely related to either of the major two groups29.

In mammals, the α-subunit is also associated with one or two smaller auxiliary subunits

(NaVβ1–NaVβ4) of approximately 33–36 kDa that are required for normal kinetics and voltage-dependence of gating but are not required for ion flux, ionic selectivity, and pharmacological modulation33.

1.7 NaV channel localization and channelopathies

NaV1.1 is expressed primarily in soma and apical dendrites of purkinje cells of the cerebellum39. Inherited loss-of-function point mutations and deletions of this channel cause 40,41 febrile seizures and severe myoclonic epilepsy during infancy . NaV1.2 is primarily 42 found in unmyelinated and premyelinated axons of the CNS . Various NaV1.2 mutations are known to be associated with familial autism, neonatal epilepsy, late onset episodic 43,44 ataxia, myoclonus, and multiple sclerosis . NaV1.1 and NaV1.2 are not restricted to the

CNS, and there are reports of NaV1.1 and NaV1.2 in sensory neurons of the dorsal root 45 ganglia . NaV1.3 is found in both the CNS and PNS, but is mostly expressed during embryonic and early prenatal life. However, NaV1.3 expression is up-regulated following 29 nerve injury . NaV1.4 is expressed in skeletal muscle and loss-of-function mutations are known to cause hyperkalemia periodic paralysis, muscle weakness and paralysis46. 47 NaV1.5 is found predominantly in cardiac myocytes , however some evidence shows its

6 presence in denervated skeletal muscles48 and certain brain neurons49. Intriguingly,

NaV1.5 has also been found to be involved in regulation of endosome acidification in primary human monocyte derived macrophages50, hence it plays a significant role in 51 defense against mycobacterial infections . NaV1.5 mutations can lead to long QT 52 syndrome, Brugada syndrome and atrial fibrillation . NaV1.5 shows partial TTX- 53 resistance, with an IC50 of 1.17 µM . NaV1.6 is widespread throughout both the CNS and PNS. It has been found in soma and dendrites of cerebellum, cerebral cortex and hippocampus output neurons54. It is also found in nodes of Ranvier in the CNS and 55 sensory and motor neurons in PNS . Mutations in NaV1.6, such as in domain II S4-S5, 56 can cause cerebellar ataxia . NaV1.7 is only found in the PNS. Loss-of-function mutations 17 lead to congenital indifference to pain (CIP) with anosmia . NaV1.8 is primarily expressed 57 in the PNS, and it shows promise as a target for analgesic drugs . NaV1.8 is a TTX-R 29 channel (IC50 of 60 mM for TTX) . Substitution of Ser356 with an aromatic residue 58 removed the TTX resistance of NaV1.8 . NaV1.9 is responsible for the persistent sodium current in small diameter DRG neurons59. It has been identified as a potential analgesic 29 target due to its preferential expression in c-type DRG neurons . NaV1.9 is TTX-R with an

IC50 of 40 mM. Table 1 summarizes information about the gene, localization, and channelopathies associated with each of the mammalian NaV channel subtypes.

Table 1: Function and distribution of voltage gated sodium channels29. Subunit Gene Function Localization Diseases caused by name mutations

NaV1.1 SCN1A Fast kinetics, TTX- Soma and apical Epilepsy S (Kd ~ 5 µM) dendrites CNS and PNS

NaV1.2 SCN2A Fast kinetics, TTX- Unmyelinated Epilepsy S axons of CNS and PNS

NaV1.3 SCN3A Fast kinetics, TTX- CNS and PNS in Increased expression following S embryonic tissue nerve injury

NaV1.4 SCN4A Fast kinetics, TTX- Skeletal muscles Myotonia, periodic paralysis S

NaV1.5 SCN5A Fast kinetics, TTX- Cardiac muscles Cardiac arrhythmia, long QT, R Brugada (Kd ~ 1µM)

NaV1.6 SCN8A Fast kinetics, TTX- Nodes of Ranvier Cerebral atropathy, cerebellar S in CNS and PNS ataxia, mental retardation

NaV1.7 SCN9A Fast kinetics, TTX- PNS Altered pain sensitivity, S anosmia

NaV1.8 SCN10A Slow kinetics, TTX- PNS Altered pain sensation, R paresthesias (Kd ~ 30 µM)

NaV1.9 SCN11A Slow kinetics, TTX- PNS Familial episodic pain R

NaX SCN7A Activated through Heart, Uterus Not reported altered Na+ concentration60

1.8 Congenital indifference to pain A case was reported in 1932 of an individual who was insensitive to pain61. This condition was later categorized as “congenital indifference to pain” (CIP)62. CIP patients have a normal sense of touch, vibration, joint position sense, and temperature perception, however they fail to identify noxious stimuli as unpleasant or painful62. In addition to 8 absence of pain perception, CIP patients also display complete or partial loss of smell (anosmia)17. CIP is an extremely rare condition with only about 30 reported cases worldwide. DNA samples from individuals with CIP were studied using traditional positional cloning, and all individuals were found to contain multiple loss-of-function mutations in the gene region SCN9A that encodes the pore-forming α-subunit of NaV1.7 (Fig. 2). Thus far, there are ~10 characterized SCN9A mutations that cause CIP17,63. In striking contrast to CIP, gain-of-function mutations in SCN9A are responsible for two painful inherited neuropathies, paroxysmal extreme pain disorder and primary erythromelalgia64-66. In addition, it was recently shown that single nucleotide polymorphisms in SCN9A correlate with altered pain perception in human populations67. Thus, there is a strong link between pain and level of NaV1.7 activity, making it a prominent target in the search for novel analgesics.

1.9 Spider venom peptides as NaV-targeted analgesic drugs Venoms have been diversified over millions of years of evolution in order to gain an advantage in defense, prey capture, or competitor deterrence. Most venoms are a highly complex mixture of ligands that affect a variety of pharmacological targets. Venoms include both non-peptide and peptide toxins. Non-peptide toxins are usually orally active. Peptide toxins however need to be delivered into the soft tissues of the targeted animal using an ‘envenomation apparatus’68. Mollusks and arachnids depend on their venom to rapidly paralyze prey68. These venoms typically contain a large number of peptide neurotoxins that target a variety of ion channels, including NaV channels. Indeed, four of the seven pharmacological sites on vertebrate NaV channels are defined by the binding of mollusk or arachnid toxins. NaV channel modulation is one of the major pharmacologies of spider venoms69,70 and the spider toxins responsible have proved to be valuable tools for 33,71 understanding the structure and function of NaV channels . More than 44,500 extant species of spiders have been described72, with an even greater number remaining to be characterised73. This makes spiders the largest group of terrestrial predators, with the possible exception of predatory beetles. One of the major contributors to the evolutionary success of spiders is their ability to produce complex venom for predation and predator deterrence74. Spider venoms are complex chemical cocktails but the major components are small, disulfide-rich peptides. Since a single venom can contain as many as 1000 peptides, it has been conservatively estimated that >10 million bioactive peptides are likely to be present in the venoms of spiders75 with only 0.01% of this diversity having been characterised. The spider-venom peptides that have been described

9 to date are listed in Arachnoserver, a manually curated database that provides detailed information about proteinaceous toxins from spiders76,77.

Voltage gated sodium (NaV) channels play a key role in the rising phase of the action 29 potential in excitable cells . In contrast with vertebrates, insects express only a single NaV 33 channels subtype and consequently they are extremely sensitive to NaV channel modulators, as underlined by the fact that several of the most successful classes of chemical insecticides are NaV channel modulators (e.g., pyrethroids, indoxacarb, dihydropyrazoles, N-alkylamides, and DTT)78-80. One third of the 186 ion channel modulators listed in Arachnoserver target NaV channels (Fig. 3), which is perhaps not surprising as these toxins allow spiders to rapidly incapacitate their prey. NaV channel toxins are found in a taxonomically diverse range of spiders, which suggests that this pharmacology was recruited at a very early stage in venom gland evolution. Moreover the ubiquity of NaV channel toxins in spider venoms suggests that these peptides might be useful insecticides for the control of arthropod pests81,82.

Figure 3: Ion channel modulators form spider venoms. Molecular targets of the 186 ion channel toxins listed in Arachnoserver (www.arachnoserver.org; accessed February 12, 2012)77.

The majority of toxins target voltage-gated ion channels (CaV, NaV, KV), followed by calcium- activated potassium (KCa) channels, mechanosensitive ion channels (MSC), transient receptor potential (TRP) channels, and acid-sensing ion channels (ASICs). Note that some toxins target more than one type of channel, and therefore the cumulative number of toxins in the histogram is >186.

In addition to prey capture, spiders also use their venom to deter predators, which can include vertebrates. Moreover, even though the vast majority of spiders prey primarily on

10 invertebrates, there is no evolutionary selection pressure to prevent spider toxins acting on vertebrate ion channels. It is therefore not surprising that many spider-venom peptides 33,70,83 have been found to modulate the activity of vertebrate NaV channels .

Figure 4: Molecular architecture and pharmacology of NaV channels. Coloured regions represent neurotoxin receptor sites. The grey circles represent the outer (EEDD) and inner (DEKA) rings of amino acid residues that form the ion-selectivity filter and constitute the proposed neurotoxin receptor site 1 for the water-soluble guanidinium toxins TTX and saxitoxin. Some m- conotoxin binding sites overlap with those of TTX and are omitted for clarity. Nevertheless, we refer to sites 1a and 1b to distinguish between the different molecular determinants for µ-conotoxin and TTX/saxitoxin binding. In the case of receptor sites 3 and 4, only epitopes where mutagenesis of key residues leads to more than a five-fold decrease in toxin binding affinity are highlighted. Note that TM segment 6 in domain I includes binding epitopes for both site 2 and site 5 toxins. The known binding sites for 8 of the 12 families of spider-venom NaV toxins (NaSpTxs) are indicated. “F” denotes toxin family (e.g., F2 indicates NaSpTx family 2). While the δ-amaurobitoxins in family 10 have been shown to bind to neurotoxin receptor site 4, peptides in this family have variable pharmacology and some may bind at other sites.

11 Spider venom peptides typically comprise 31–50 amino acid residues and 3–4 disulfide bonds77. Three of the disulfide bonds form an inhibitor cystine knot (ICK) motif (Fig. 5)77,83, in which one of the disulfides penetrates a loop formed by the other two disulfides and the intervening sections of the polypeptide backbone. The ICK motif is responsible for the extreme thermal, chemical and enzymatic stability observed for such toxins82,84.

Figure 5: The inhibitory cystine knot (ICK) motif. A. A three dimensional representation of the motif. B. A schematic representation of the ICK motif, that consisted with two conserved antiparallel β-sheets and a third β-sheet that may or may not exist in every ICK motif85.

A larger number of spider venom components are also active on vertebrate VGSCs. Many toxins show VGSC subtype selectivity, while some other toxins like δ-HXTX-Ar1a are nonspecific and lethal. Subtype selectivity of toxins have number of implications, including the design of VGSC inhibitors as analgesics, treatment of arrhythmia, treatment of various pain disorders and ion channelopathies.

1.10 Mechanisms of toxin activity Venom peptides modulate channel activity by two main mechanisms: they either physically occlude the channel pore or they bind to an allosteric site that induces a conformational change in the channel that alters the equilibrium between the open, closed and inactivated 86 states . All known NaV channel toxins from spider venoms are allosteric modulators, also known as “gating modifiers”. However they modulate channel activity in quite different ways. For example the lethal toxin form the Sydney funnel-web spider Atrax robustus (δ- 87,88 HXTX-Ar1a) inhibits inactivation of both insect and vertebrate NaV channels . Some other toxins shift the activation voltage to the left or to the right of a current-voltage

12 distribution. When the activation voltage shifts to the left (to more negative potentials), increasing the probability of the channel opening closer to the resting potential. When a toxin shifts the activation voltage to the right, the channel opens at more positive voltages, and the opening event become less and less likely. Some toxins bind and hold the S3-S4 paddle motif in open state of the channel89.

1.11 NaSpTx family 2 We have conducted a full survey to classify all available spider-venom peptides that target

NaV channels form Arachnoserver, based on sequence homology and disulfide bond configurations. When this survey was completed there were 12 distinct families of spider- venom peptides identified (Fig 6 and Fig 7). This thesis primarily uses the rational nomenclature recently developed for spider-venom peptides90 in order to facilitate identification of orthologs and paralogs, but for convenience we also provide original names of toxins where appropriate. The term NaScTx is employed for scorpion toxins that 91 target NaV channels , and here we introduce the term NaSpTx for spider toxins that target this channel.

Figure 6: NaSpTx families 1–4, 7 and 10. The consensus sequence is shown above each alignment, with the disulfide bond connectivity indicated in blue. Red asterisks denote C-terminal amidation. Strictly conserved residues are highlighted in bold colour while semi-conserved residues are highlighted in a lighter colour. Toxin names are based on the rational nomenclature devised for spider-venom peptides, as used in ArachnoServer and UniProt90. The structure of a representative member from each NaSpTx family is shown adjacent to each alignment: family 1, µ- TRTX-Hh2a (PDB ID: 1MB6); family 2, κ-TRTX-Gr1a (PDB ID: 1D1H); family 3, κ-TRTX-Gr2a (PDB ID: 1LUP); family 4, δ-HXTX-Hv1a (PDB ID: 1VTX); family 7, κ-TRTX-Gr3a (PDB ID: 1S6X); family 10, δ-AMATX-Pl1b (PDB ID: 1V91). β strands are depicted in cyan, helices are red, and disulfide bridges are yellow. All spider photos were taken by Bastian Rast.

Figure 7: NaSpTx families 5, 6, 8, 9, 11 and 12. The consensus sequence is shown above each alignment, with the disulfide bond connectivity indicated in blue. Dotted lines represent predicted disulfide bond connectivities that have not been experimentally validated. Red asterisks denote C terminal amidation. Strictly conserved residues are highlighted in bold colour while semi-conserved residues are highlighted in a lighter colour. Toxin names are based on the rational nomenclature devised for spider-venom peptides90, as used in ArachnoServer and UniProt. The structure of a representative member of F5 is shown adjacent to the alignment (β-HXTX Mg1a; PDB ID: 2GX1). Disulfide bridges are coloured yellow. No structures are available for any member of the other NaSpTx Families depicted in this figure. Spider photos are by Bastian Rast (M. gigas and P. reidyi) and Aloys Staudt (H. melloteei).

15 Table 2 summarizes some of the features of each NaSpTx family, including an indication of their subtype preferences. Notably, peptides that potently block NaV1.5 tend to be lethal in rodents, consistent with the critical role of this subtype in cardiac function. This highlights the need to pinpoint the residues on these peptides that mediate their selectivity and potency in order to be able to rationally engineer improved versions that preferentially bind NaV1.7. This will be the major challenge in progressing these peptides to the clinic. On a positive note, however, peptides have already been isolated from Families 1–3 that potently block NaV1.7 (IC50 ≤ 51 nM) and that have at least some selectivity over other NaV subtypes.

Table 2: Subtype selectivity and other selected properties of the 12 families of NaSpTx.

NaV channel subtype High NaSpT Other ion resolutio x TTX-S TTX-R Lethal channel n 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 family targets structure1 6 a b b b 2b c d 1 407 0.6 42 288 72 - 26 - 1.1 - Yes CaV/KV Yes e f f g 2 - - - - 447 - 51 27 85 - No CaV/KV/TRPV1 Yes h i h i h h h h k k 3 41 230 103 330 79 26 0.3 146 30.2 27.6 Yes KV/CaV Yes 4 ------Yes - Yes 5 - 1200l ------Yes - Yes 6 ------Yes - No m m m 7 - - - - 32 130 130 - - - - KV Yes 8 - 0.08n ------Yes - No o 9 - - - 155 ------KV No 10 ------N0 - Yes 11 ------No - No 12 ------No - No

For each NaSpTx family, the IC50 value (in nM) is given for the peptide that most potently modulates the activity of each NaV subtype. Columns labeled “TTX-S” and “TTX-R” indicate potency of peptides on TTX-S and TTX-R NaV channels, usually tested in rat dorsal root ganglion neurons. The column labeled “Lethal” refers to the toxicity of peptides in rodents. aβ-TRTX- Cm1b92, bβ-TRTX-Ps1a 92, cµ-TRTX-Hh2a93, dµ-TRTX-Hhn2b94, eβ-TRTX-Cm2a92, fβ/ω -TRTX- Tp1a95, gµ-TRTX-Cj1a96, hβ/ω-TRTX-Tp2a95, iβ-TRTX-Gr1a97, jβ-TRTX-Gr1b97, kκ-TRTX-Cj2a98, lβ - HXTX-Mg1a99, mδ-TRTX-Cj1a100, nµ-CNTX-Pn1a101, oµ-TMTX-Hme1b102, p“high-resolution structures” are defined as NMR-derived solution structures with a backbone RMSD <0.5 Å.

16 Family 2 is the largest family of NaSpTxs with 34 members, all originating from theraphosid venoms. Of these, 18 peptides were isolated from the Chinese tarantula Chilobrachys jingzhao. Family 2 peptides range in size from 33 to 41 residues and they each contain three disulfide bonds that form an ICK motif. Currently, there is experimental proof of NaV channel activity for only four of the 34 peptides in this family. However, they are very interesting from a therapeutic perspective as one member of this family, β/ω-

TRTX-Tp1a (ProTx1), is one of the most potent blockers of human NaV1.7 reported to 95,103 date, with an IC50 of 51 nM .

κ-TRTX-Cj1a (Jingzhaotoxin-11) is a weak modulator of NaV currents in cardiac myocytes

(i.e., mainly TTX-R NaV1.5), causing a significant reduction in peak NaV current amplitude 104 and a slowing of NaV current inactivation (IC50 = 1280 nM) . It also inhibits KV2.1 with moderate potency (IC50 = 740 nM), with even weaker block of KV4.1 and KV4.2 and no 104 effect on KV1.1, KV1.2, KV1.3, KV1.4 and KV3.1 . A peptide with identical sequence has been isolated from the venom of the related theraphosid spider Plesiophrictus guangxiensis. It has been proposed that κ-TRTX-Cj1a binds to the extracellular S3–S4 linker region of channel domain IV (i.e., neurotoxin receptor site 3; see Fig. 4)104, which is a common target of gating modifier toxins such as κ-TRTX-Gr1a (Hanatoxin-1) in the same family105.

β-TRTX-Cm2a (Ceratotoxin-3) has moderate affinity towards NaV1.5 (IC50 = 447 nM), and at a concentration of 2 mM it weakly blocks NaV1.8 (45% reduction in current) but has no 92 activity against NaV1.1–1.4 . In striking contrast with κ-TRTX-Cj1a and β-TRTX-Cm2a, µ-

TRTX-Cj1a (Jingzhaotoxin-34) is a potent blocker of TTX-S NaV channels in rat dorsal root 106 ganglion (DRG) neurons (IC50 = 85 nM) but it has no effect on TTX-R NaV currents . Thus, residues that are conserved in κ-TRTX-Cj1a and β-TRTX-Cm2a, but which are absent from µ-TRTX-Cj1a, should provide some insight into residues that are important for interaction of this family of peptides with NaV1.5.

Six toxins from family 2 have confirmed activity on voltage-gated calcium (CaV) channels, with 12 toxins active against KV channels. Nevertheless, the high level of sequence identity within this family and the potency of the known family 2 NaV modulators described above indicate that a thorough functional analysis of all toxins in this family could reveal many more toxins acting on NaV channels. However, from a therapeutic perspective, the promiscuous activity of several family members on multiple types of voltage-gated

17 channels indicates that it will be critical to carefully determine the selectivity of any family 2 peptides that are being considered as therapeutic leads96.

1.12. Research aims

The long term goal of this research is to develop potent and selective blockers of hNaV1.7 that can be used as leads for the development novel analgesics for treatment of chronic pain, and to understand the impact that such drugs might have on olfaction. The specific aims of my PhD thesis project were as follows:

Aim 1: Examine the localization of NaV1.7 in the rat olfactory system.

Aim 2: Characterize novel hNaV1.7 inhibitors from NaSpTx family 2.

Aim 3: Determine the pharmacophore of the most promising peptide and optimise its affinity and selectivity for hNaV1.7.

18 CHAPTER 2

STRUCTURE-ACTIVITY RELATIONSHIPS OF ProTx1: ACHIEVING BETTER

SPECIFICITY FOR NaV1.7

2.1 Introduction Spiders are one of the most successful terrestrial predators. Their venoms have evolved over many millions of years to selectively and potently inhibit nervous system targets in order to rapidly paralyze their prey107. The spider-venom peptidome has been identified as one of the largest known libraries of compounds with high potency and specificity towards nervous-system targets.

The majority of components in spider venoms are small, disulfide-rich peptide toxins that typically comprise 31–50 amino acid residues and 3–4 disulfide bonds107. In many cases, three of these disulfide bonds form an inhibitor cystine knot (ICK) motif in which one of the disulfide bonds penetrates a loop formed by the other two disulfides and the intervening sections of the polypeptide backbone85,107. The ICK motif is responsible for the extreme thermal, chemical and enzymatic stability observed for such toxins85,107.

Spider toxins listed in the Arachnoserver database76,77 have been classified into 12 families of sodium channel toxins (NaSpTx), denoted NaSpTx Families 1–12, based on the level of sequence conservation and intercystine spacing32. NaSpTx family 2 is the largest toxin family; it comprises 34 peptides of theraphosid origin, most of which are 33– 41 residues in length and contain three disulfide bonds that form an ICK motif.

Experimental data for NaV channel activity is only available for four of these toxins, however this group includes β/ω-TRTX-Tp1a, one of the most potent NaV1.7 inhibitors reported to date95.

β/ω-TRTX-Tp1a, commonly known as ProTx1, was initially isolated from the Peruvian green-velvet tarantula Thrixopelma pruriens. Native ProTx1 inhibits NaV1.7 and NaV1.8 95 expressed in Xenopus laevis oocytes with an IC50 of 51 nM and 27 nM, respectively .

Furthermore, 730 nM ProTx1 causes 73–83% inhibition of NaV1.2 and NaV1.5. ProTx1

19 also inhibits the human voltage gated calcium (CaV) channel subtype 3.1 (hCaV3.1) (IC50 = 108 200 nM), while having essentially no effect on hCaV3.2 (IC50 >31 µM) .

The high potency of ProTx1 on NaV1.7 has evoked interest in this molecule as the starting point for development of a NaV1.7-directed analgesic. However, its lack of specificity over 108 other neuronal ion channels, such as NaV1.6 and CaV3.1 , has frustrated these efforts.

In this study we obtained IC50 values for ProTx1 inhibition of NaV1.2, NaV1.5 and NaV1.6. We also determined the pharmacophore on ProTx1 that mediates its interaction with

NaV1.7 and mapped the key functional residues onto the three-dimensional structure of the toxin that we determined using NMR spectroscopy. Furthermore we used a

KV2.1/NaV1.7 chimeric channel to identify ProTx1 binding sites on NaV1.7. Finally, we used all of this information to rationally design several ProTx1 mutants in an attempt to obtain improved specificity for NaV1.7.

20 2.2 Materials and methods 2.2.1 TEVC for tethered-toxin-based ProTx1 alanine scan

The gene construct for hNaV1.7 (a kind gift from Dr Frank Bosmans, John Hopkins University, Baltimore, USA) was linearized with NotI restriction enzyme and cRNA was produced by in vitro transcription reaction with T7 RNA polymerase (mMessenger mMachine kit, Ambione, USA). The tethered-toxin (t-toxin) gene constructs for the 28 ProTx1 mutants and the wild-type toxin (a kind gift from A/Prof Michael Nitabach, Yale University, New Haven, USA) was also linearized with Not1 restriction enzyme and cRNA was produced by in vitro transcription reaction with SP6 RNA polymerase (mMessenger mMachine kit, Ambion, USA). Concentrations of both cRNA stocks were determined using a NanoDrop 2000 (Thermo Scientific, USA). cRNA stocks with concentrations of 0.8–1.0 µg/µl were used for injections mixed at 1:1 ratio. Each oocyte was injected with about 55 nL of mixed cRNA.

Currents within the range of 200–2000 nA were considered to be a positive response, whereas currents below 100 nA were considered to indicate insufficient sodium channels on the surface or inhibition of channels by t-ProTx1. Each mutant was tested in at least two different batches of oocytes, each containing 8–12 oocytes.

2.2.2 Oocyte surface protein assay Surface protein assays were performed 2 days after cRNA injection. For surface labelling, oocytes were blocked for 1 h in ND96 with 3% bovine serum albumin (BSA) at RT on shaker, then treated with 1 µg/ml rat monoclonal anti-Myc antibody (1:500, Cell Signaling), in 3% BSA for 1 h at RT, washed 3 x 5 min at RT in ND96, and incubated with horseradish peroxidase (HRP)-coupled secondary antibody (1:2000, Pierce), in 3% BSA for 1 h at RT. Cells were extensively washed (3 x 5 min in ND96 with 0.1% Tween 20) and transferred to ND96 solution without BSA. Individual oocytes were placed in 65 µl 1-Step Ultra TMB- ELISA solution (Thermo Scientific, USA) in a 96-well culture plate, incubated at RT for 15 min, and the reaction stopped by adding 25 µl of 2 M H2SO4. Chemiluminescence was quantitated using a Bio-Rad Model 680Microplate Reader (Bio-Rad, USA).

2.2.3 Electrophysiological assay of toxin activity Oocytes were surgically removed from Xenopus laevis frogs and treated with 30 mg of collagenase (Sigma type1) in 20 mL of calcium-free ND96 (96 mM NaCl, 2 mM KCl, 2 mM

MgCl2, 5 mM HEPES, pH 7.4) for 60–90 min. Oocytes were washed with calcium-free

21 ND96, three times or until brown, cloudy medium was completely removed. Healthy- looking oocytes with clearly defined animal and vegetal poles, with each pole of uniform color and elastic membrane, were selected and allowed to rest about 1 h in modified ND96 medium (96 mM NaCl, 2 mM KCl, 1 mM CaCl2, 2 mM MgCl2, 5 mM HEPES, 5 mM pyruvic acid, 50 µg/mL gentamicin, 5% horse serum, pH 7.4). Selected oocytes were injected with 41 nL of about 200–800 ng/µL cRNA synthesized using a mMessage mMachine cRNA transcription kit (Life Technologies), and purified using an RNeasy Mini Kit (Qiagen). Oocytes were incubated at 17°C, media was changed every 24 h, and recordings were conducted using TEVC electrophysiology (Axoclamp 900A amplifier, Molecular Devices, Sunnyvale, CA, USA) 2–5 days after theRNA injection. Two standard glass microelectrodes with 0.5–2 MΩ resistance when filled with 3 M KCl solution were used for recordings. Capillary glass for oocyte injections and electrodes was purchased from SDR Clinical Technology (Sydney, Australia).

Data acquisition and analysis was performed using pCLAMP software, version 10 (Molecular Devices). All recordings were conducted at room temperature in ND96 containing 0.1% fatty acid-free BSA (Sigma). A data set was acquired to determine the current-voltage relationship (I-V curve) for the channel in the absence of any channel modulators. Oocytes were clamped at –90 mV, and data was sampled at 20,000 Hz.

Inward NaV currents were invoked by applying 50 ms depolarizing steps from –70 mV to +45 mV with 5 mV increments, while holding the oocyte at –90 mV. Then the toxin was added, and the current was allowed to equilibrate. A test pulse to –10 mV every 10 s was used to detect the equilibration status before acquiring a full I-V curve.

2.2.4 Generation of NaV1.7/Kv1.1 chimeras Chimeras were generated using sequential polymerase chain reactions (PCRs) with 109,110 111 KV2.1Δ7 and hNaV1.7 as template. The KV2.1Δ7 construct contains seven point mutations in the outer vestibule, rendering the channel sensitive to agitoxin-2, a pore- blocking toxin from scorpion venom112. The DNA sequence of all constructs and mutants was confirmed by automated DNA sequencing and cRNA was synthesized using T7 polymerase (mMessage mMachine kit, Ambion) after linearizing the DNA with appropriate restriction enzymes.

22 2.2.5 TEVC recordings using NaV1.7/Kv1.1 chimeras Channel constructs were expressed in Xenopus oocytes and studied following 1–2 days incubation after cRNA injection (incubated at 17°C in 96 mM NaCl, 2 mM KCl, 5 mM

HEPES, 1 mM MgCl2 and 1.8 mM CaCl2, 50 µg/ml gentamycin, pH 7.6) using TEVC recording techniques (OC-725C, Warner Instruments) with a 150 µl recording chamber. Data were filtered at 2 kHz and digitized at 20 kHz using pClamp software (Molecular

Devices). Microelectrode resistances were 0.3–1 MΩ when filled with 3 M KCl. For KV channel experiments, the external recording solution contained 50 mM KCl, 50mM NaCl, 5 mM HEPES, 1 mM MgCl2 and 0.3 mM CaCl2, pH 7.6. All experiments were performed at room temperature (~22 °C). Leak and background conductance, identified by blocking the channel with agitoxin-2, were subtracted for all KV currents. All chemicals were obtained from Sigma-Aldrich.

Voltage–activation relationships were obtained by measuring tail currents for KV channels. Occupancy of closed or resting channels by toxins was examined using negative holding voltages where open probability was very low, and the fraction of unbound channels was estimated using depolarizations that are too weak to open toxin-bound channels110,113-115. After addition of toxin to the recording chamber, equilibration between the toxin and the channel was monitored using weak depolarizations elicited at 5–10 s intervals. For all channels, voltage–activation relationships were recorded in the absence and presence of toxin. Off-line data analysis was performed using Clampfit (Axon), Origin 7.5 (Originlab) and Microsoft Solver (Microsoft Excel) to calculate apparent Kd values.

23 2.3 Results

2.3.1 ProTx1 activity on NaV channels

ProTx1 is a potent inhibitor of many NaV channel subtypes. In order to obtain better NaV subtype selectivity data, we used two-electrode voltage clamp (TEVC) electrophysiology to examine the effect of the toxin on human NaV1.2, NaV1.5, NaV1.6, and NaV1.7 expressed in Xenopus laevis oocytes. We found that chemically-synthesised ProTx1 inhibited NaV1.7 95 with an IC50 of 76 nM (Fig. 8), which is close to the reported value of 51 nM , thus validating the activity of the synthetic toxin. ProTx1 inhibited NaV1.2 and NaV1.6 with IC50 values of 98 nM and 17 nM, respectively (Fig. 8). NaV1.6 is present in nodes of Ranvier on myelinated nerves within both the CNS and PNS54 and therefore inhibition of this channel could have detrimental consequences, thus reducing the toxin's desirability as a drug lead.

ProTx1 was also found to inhibit NaV1.5, the dominant NaV channel in cardiac muscle, with an IC50 ~ 340 nM. Although this is about 4.5-fold higher than its IC50 for inhibition of

NaV1.7, inhibition of NaV1.5 might be a significant off-target liability as inhibition of this channel is likely to cause cardiac failure. The ability of ProTx1 to inhibit NaV1.1, NaV1.3, and NaV1.9 remains to be determined.

NaV1.2- IC50 ~ 98 nM 1.00 NaV1.4- IC50 ~ 57 nM

NaV1.5- IC50 ~ 338 nM 0.75 NaV1.6- IC50 ~ 17 nM max

NaV1.7- IC50 ~ 76 nM

I / 0.50

0.25

-9.5 -9.0 -8.5 -8.0 -7.5 -7.0 -6.5 -6.0 -5.5 -5.0 -4.5 -4.0

log[ProTx-1], M

Figure 8: Concentration-response curves for ProTx1 inhibition of human NaV channels.

ProTx1 potently inhibited NaV1.2, NaV1.4 and NaV1.7 heterologously expressed in Xenopus oocytes, with very similar IC50 values of 98 nM, 57 nM and 76 nM, respectively. ProTx1 inhibited

NaV1.6 more potently, with IC50 ~17 nM. ProTx1 also inhibited NaV1.5, albeit more weakly, with an

IC50 of about 340 nM. Error Bars show SEM. Data obtained using TEVC (n = 2).

2.3.2 Identifying the ProTx1 pharmacophore for NaV1.7 Understanding the nature of peptide-target interactions is fundamental to the development of a peptide drug lead. Identification of the ProTx1-NaV1.7 interaction surface (pharmacophore) is crucial both in terms of describing the toxin activity and enabling rational design of mutants with improved potency and/or selectivity. In this study, a tethered toxin (t-toxin)-based alanine scanning strategy was used to determine the NaV1.7 pharmacophore116. Briefly, an expression vector was designed to encode the t-toxin as a fusion protein, containing an N-terminal secretory signal, followed by the toxin sequence, hydrophilic linker (with c-Myc epitope tag), and C-terminal glycophosphatidylinositol (GPI) membrane-anchor targeting sequence (Fig. 9). This construct enables the t-toxins to be secreted but they remain attached to the plasma membrane of the secreting cell via the GPI anchor. The long hydrophilic linker is expected to allow sufficient freedom of motion so that the tethering process does not affect toxin activity. In contrast to traditional peptide- based alanine scanning mutagenesis process, this technique is much faster and cheaper. The peptide-based alanine-scanning process requires every mutant peptide to be chemically synthesised, or produced recombinantly, and folded into the correct conformation, which can be very expensive and time consuming. In the t-toxin approach, the cRNA is transcribed from a plasmid vector, and injected into the oocyte, which uses its own protein expression and folding machinery to produce the final folded toxin.

When tethered ProTx1 (t-ProTx1) was co-expressed in Xenopus oocytes with NaV1.7 channels (using equal quantities of cRNA), complete inhibition of channel current was observed (Fig. 10A). In cases where partial currents were observed, the activation voltage was right-shifted in current-voltage (I–V) traces. This shift was identical to that observed for partial inhibition of NaV1.7 currents by untethered ProTx1. Furthermore, Nitabach and colleagues established that both soluble and tethered ProTx1 are capable of binding to 116 NaV1.2 channel and stabilizing the NaV activation gate . This was demonstrated using TEVC by applying a large prepulse (from –100 mV to +100 mV) to oocytes expressing

NaV1.2 and t-ProTx1, and NaV1.2 with soluble ProTx1 added, where currents were fully inhibited by the toxin. The prepulse was applied for varying lengths of time (0 to 1000 ms), and the maximum current amplitude was observed after a subsequent –100 mV to –10 mV test pulse. The magnitude of the current recovered was directly proportional to the duration of the prepulse, and oocytes treated with both soluble and tethered toxin

25 exhibited identical current recovery. This indicates that the tethering process has minimal or no effect on ProTx1 activity.

Figure 9: The tethered toxin approach. (A) The fusion construct contains an N-terminal secretory signal sequence, followed by the toxin sequence, a hydrophilic linker with a c-Myc epitope tag (for antibody-based detection of the t-toxin on cell surface), and a C-terminal GPI membrane-anchor targeting-sequence. (B) Schematic diagram of the tethered toxin expressed on the oocyte surface, and interacting with the channel. The GPI anchor is shown in orange, the linker containing a c-Myc epitope tag is yellow, and the toxin is shown binding to the pore of the channel. Panel A taken from Gui and colleges (2014) publication116.

Expression vectors encoding a panel of 28 ProTx1 mutants were received as a kind gift from Prof. Michael Nitabach (Yale University, USA); this panel included mutations of all non-Ala and non-Cys residues to alanine. Xenopus oocytes were co-injected with cRNA synthesised from each of the mutant expression vectors plus NaV1.7 RNA. Oocytes were incubated at 17°C, and currents were recorded 48 h later. The maximum current amplitude (activated by pulses from –90 to –10 mV) was recorded for every oocyte. Oocytes injected with only NaV1.7 RNA served as the negative control, and oocytes injected with cRNA encoding wild-type (WT) ProTx1 served as the positive control. Each mutant was tested in at least two batches of oocytes, and the maximum current amplitude for each oocyte was recorded (Fig. 10A). In order to confirm that tethered toxins were expressed, an ELISA assay was conducted that recognized the c-Myc epitope in the tethered-toxin linker region (See Materials and Methods for details). This assay was conducted with the same oocytes used in the TEVC recordings for consistency. Mean surface signal measured as the absorbance at 450 nm was normalized to that of WT oocytes (Fig. 10B). Oocytes with no

26 injections were used as the negative control, and subtracted from all surface protein values prior to normalization.

NaV1.7 currents were still observed when NaV1.7 was co-expressed with the tethered

ProTx1 mutants L6A, L19A, W27A, V29A, W30A, and D31A, whereas NaV1.7 currents were completely absent when NaV1.7 was co-expressed with WT t-ProTx1 (Fig. 10A). The recovery of channel current was not due to lack of surface expression of these mutant t- toxins (Fig. 10B). L19 is deeply buried in the toxin with little surface exposure (see below) and hence we speculate that recovery of channel current for this t-toxin mutant is due to misfolding of the toxin. Thus, we conclude that the NaV1.7 pharmacophore comprises residues L6, W27A, V29, W30, and D31. Based on the t-toxin alanine scan, residue L6,

W30 and D31 appear to be the most critical residues for block of NaV1.7.

Figures 10: Alanine scanning mutagenesis of ProTx1. (A) Mean of maximum current amplitudes for each mutant in comparison to the NaV1.7 current when a similar quantity of NaV1.7 RNA alone was injected. (B) Levels of surface expression of each mutant, normalized to the expression of WT t-ProTx1 expression (dotted line). Oocytes with no RNA injected ("Uninjected") were used as a negative control.

2.3.3 Structure of ProTx1: comparison of NaV1.7 and NaV1.2 pharmacophores This structure of the chemically synthesised ProTx1 was determined using 2D homonuclear NMR spectroscopy in collaboration with Dr Mehdi Mobli (Centre for Advanced Imaging, UQ) and Prof. Norelle Daly (James Cook University). The toxin forms a canonical ICK motif in with the C1–C4 and C2–C5 disulfide bonds and the intervening sections of polypeptide backbone form a 20-residue loop that is pierced by the C3–C6 disulfide bond.

The ProTx1 pharmacophores for NaV1.2 and NaV1.7 were mapped onto the 3D structure of ProTx1 (Fig. 11). The availability of the NaV1.2 pharmacophore facilitated efforts to identify residues that might be important for selectivity for NaV1.7 over NaV1.2.

NaV1.7 NaV1.2 A

R23 D31 R3 D31 L19 S22

V29 F34 L19

L6 W27 G32 W27 W30 W30 W5 o B 90 C

D31 W30 W30 L19 R3 D31

L6 W5 G32 V29 V29 S22 F34 W27 R23 W27

NaV1.7 ECRYWLGGCSAGQTCCKHLVCSRRHGWCVWDGTFS

NaV1.2 ECRYWLGGCSAGQTCCKHLVCSRRHGWCVWDGTFS

Figure 11: Structure of ProTx1 highlighting pharmacophores for NaV1.7 and NaV1.2. The residues critical for ProTx1 inhibition of NaV1.7 (left panel) are shown in yellow (L6, L19, W27, V29, W30) and red (D31). Note that these residues form a largely hydrophobic cluster on one face of the toxin. L19 is mostly buried. Of the 11 residues in the ProTx1 pharmacophore for NaV1.2, five are shared with the pharmacophore for NaV1.7. Two basic residues in the NaV1.2 pharmacophore, R3 and R23, are exposed on the outer surface of the molecule, facing the opposite direction to the hydrophobic patch. This might be an indication of extra interactions that ProTx1 makes with

NaV1.2, but not with NaV1.7.

29 2.3.4 ProTx1 interacts primarily with the domain II and domain IV paddle motifs of hNaV1.7 The helix-turn-helix motif, which is also the paddle motif of the voltage sensor domain, is know to be the integral component in KV channels, that moves near the protein-lipid interface to drive the activation of the channel by opening the pore117-121. Further, it is also known that certain tarantula toxins, including ProTx1, exert their inhibitory effect on KV, and NaV channels by interacting with VSD paddle motif, thereby modulating the channel 110,113,120,122-127 function . It has been previously established that ProTx1 binds to NaV1.2 VSDs of different domains, with varying affinity114. Bosmans and colleagues (2008) have used a chimeric channel technique, where VSD paddle motifs of NaV1.2 were transplanted on the KV2.1 channel, which is structurally very similar to a single domain of a NaV channel. This resulted in four chimeric channels of rNaV1.2a/KV2.1 for each VSD paddle motif. This experiment was conducted to observe the change in toxin affinity for the chimeric channel, hence to determine individual VSD-ProTx1 affinities in an isolated system114. In this experiment it is of most importance that the toxin in question does not have high affinity toward the KV2.1 channel, which would complicate the binding process and would make it impossible to come to a conclusion of toxin interaction with the chimeric channel.

As mentioned before Bosmans and colleagues uncovered that ProTx1 binds to VSD of domains II and IV, exhibiting minimal affinity towards KV2.1 channel. To understand whether ProTx1 inhibits NaV1.7 opening by interacting with one or more VSDs, effect of

ProTx1 on KV2.1 and the four hNaV1.7/KV2.1 paddle chimeras were examined. It was found that 100 nM of the toxin only slightly influenced wild-type KV2.1 (apparent Kd = 2408 ± 189 nM) whereas the domain II and domain IV chimeras are strongly inhibited with an apparent Kd of 114 ± 12 nM and 199 ± 48 nM, respectively. In addition, the NaV1.7/KV2.1 domain I paddle chimera is also sensitive to ProTx1, albeit at higher concentrations (apparent Kd = 612 ± 98 nM) whereas the domain III chimera is not inhibited at 100 nM (apparent Kd = 10452 ± 836 nM).

Figure 12: 4 Identification of hNaV1.7 voltage sensing domains that interact with ProTx1. (A)

KV2.1 potassium current elicited by a depolarization to 0 lmV clamped back to –60mV from a holding potential of –90 mV before (black) and in the presence of 100 nM ProTx1 (grey). (B)

Apparent affinity of ProTx1 for KV2.1. Concentration dependence for toxin inhibition measured at negative voltages is shown. The apparent equilibrium dissociation constant (Kd) for toxin interaction with KV2.1 was calculated assuming four independent toxin-binding sites per channel, with single occupancy being sufficient to inhibit opening in response to weak depolarizations. (C) Potassium currents elicited by depolarizations near the foot of the voltage–activation curve for hNaV1.7/KV2.1 chimeric constructs. Currents are shown before (black) and in the presence of 100 nM ProTx1. Scale bars are 50 ms (abscissa) and 0.3 µA (ordinate axis).

2.3.5 Design and characterization of ProTx1 mutants In order to increase ProTx1's utility as an analgesic lead, it is essential to increase its selectivity towards NaV1.7 in order to reduce potential off-target effects. Three ProTx1 mutants were designed to investigate the effect of these mutations on NaV1.7 selectivity.

These designs were based on the NaV1.7 and NaV1.2 pharmacophores mapped onto the structure of ProTx1 (Fig. 11), as well as a sequence alignment of ProTx1 with two novel

ProTx1 orthologs (Fig. 13). The three ProTx1 mutants were K17E, ProTx1-NH2 (amidated

C-terminus), and ProTx1-G32A-NH2 (G32A mutation with amidated C-terminus). The K17E mutant was designed before the 3D structure of ProTx1 became available. A sequence alignment of ProTx1 and its orthologs, and the NaV1.7 affinity of each toxin, were analysed (Fig. 13). Notably, the entire ProTx1 pharmacophore for NaV1.7 is conserved in β-TRTX-Pe1b, but this toxin lacks the potency of ProTx1 for NaV1.7. This suggests that additional residues might be involved in the toxin-channel interaction. µ-

TRTX-Pe1a also has high homology to ProTx1, but lower potency against NaV1.7. These

31 toxins appear to contain a high density of acidic residues near C15 and C16. In particular, we wondered whether the charge reversal at position 17 (K17 in ProTx1 and E17 in β- TRTX-Pe1b) might contribute to the loss of potency seen in β-TRTX-Pe1b. In the absence of the ProTx1 structure to identify the topological disposition of this residue in the folded toxin, it was decided to make the K17E mutant synthetically.

ProTx-1 ECRYWLGGCS AGQTCCKHLV CSRRHGWCVW DGTFS 76 nM β-TRTX-Pe1b GECRYWLGGCS KTGDCCEHLS CSPKWHWCVW DGTF 152 nM µ-TRTX-Pe1a GECKYLWGTCS KDEDCCAHLG CNRKHDWCGW DYTF >10 µM

Figure 13: Sequence alignment of ProTx1, µ-TRTX-Pe1a and β-TRTX-Pe1b. The five NaV1.7 residues comprising the ProTx1 pharmacophore are shown in orange (hydrophobic residues) and red (acidic residues). The IC50 values for block of NaV1.7 by each toxin are indicated. All five pharmacophore residues are conserved in β-TRTX-Pe1b, yet the potency toward NaV1.7 is about two-fold less. Four out of five active residues are conserved in µ-TRTX-Pe1a, and it shows about 130-fold loss of activity in comparison to ProTx1.

The K17E mutant was produced by solid-phase peptide synthesis and tested using TEVC electrophysiology. The charge reversal mutation lead to reduced activity against all NaV channel subtypes that were tested (Fig. 15). Activity against NaV1.6 was reduced about

11-fold, while activity against NaV1.7 was 6.7-fold lower. NaV1.2 and NaV1.5 were least affected with about a 3-fold loss of activity (Table 3).

Amidated ProTx1 and amidated ProTx1-G32A mutants were designed after the ProTx1 structure became available. Many toxins within NaSpTx family 2 are amidated (Fig. 14, red asterisks). We therefore wondered whether C-terminal amidation might affect ProTx1 activity. As mentioned previously, F34 is an important part of the NaV1.2 pharmacophore, and G32 appears to be critical for the positioning of this residue in the last turn prior to the C-terminus of the toxin. Thus, we decided to mutate G32 to an alanine, with the expectation that the ProTx1-G32A-NH2 mutant would have compromised affinity for

NaV1.2, but maintain affinity for NaV1.7.

When tested using TEVC electrophysiology, the ProTx1-G32A-NH2 showed reduced activity towards all NaV channels tested. Activity against NaV1.6 was 32-fold lower (IC50 increase from 20 nM to 636 nM). Activity against NaV1.5, NaV1.2 and NaV1.7 was reduced 12-fold, 8-fold, and 5.4-fold, respectively.

32 J.K. Klint et al. / Toxicon 60 (2012) 478–491 481

The ProTx1-NH2 mutant had 14-fold increased activity towards NaV1.2 with an IC50 of ~7 nM for inhibition of this channel. C-terminal amidation led to a 3-fold loss of activity against

NaV1.5 (IC50 ~ 800 nM), and a 2-fold loss of activity against NaV1.6. In contrast, C-terminal amidation had no affect on ProTx1 activity against NaV1.7 (IC50 ~ 77 nM).

Figure 14: Sequence alignment of NaSpTx Family 2 peptides. This is the largest NaSpTx family, and it contains the well-known ProTx1 (β/ω-TRTX-Tp1a) and the close ortholog SGTX1 (κ- TRTX-Scg1a). Peptides that are C-terminally amidated are marked with a red asterisk.

33 Fig. 3. NaSpTx families 1–4, 7 and 10. The consensus sequence is shown above each alignment, with the disulﬁde bond connectivity indicated in blue. Red asterisks denote C-terminal amidation. Strictly conserved residues are highlighted in bold colour while semi-conserved residues are highlighted in a lighter colour. Toxin names are based on the rational nomenclature devised for spider-venom peptides (King et al., 2008b), as used in ArachnoServer and UniProt (King et al., 2008b). The structure of a representative member from each NaSpTx family is shown adjacent to each alignment: Family 1, m-TRTX-Hh2a (PDB ID: 1MB6); Family 2, k-TRTX-Gr1a (PDB ID: 1D1H); Family 3, k-TRTX-Gr2a (PDB ID: 1LUP); Family 4, d-HXTX-Hv1a (PDB ID: 1VTX); Family 7, k-TRTX-Gr3a (PDB ID: 1S6X); Family 10, d-AMATX-Pl1b (PDB ID: 1V91). b strands are depicted in cyan, helices are red, and disulﬁde bridges are yellow. All spider photos were taken by Bastian Rast.

typically resistant to extremes of pH, organic solvents, One member of Family 1 (m/u-TRTX-Hh1a; huwentoxin- high temperatures, and proteases (Saez et al., 2010). Of the 1) was found to be analgesic in a rat model of formalin- 10 families of NaSpTxs whose disulﬁde connectivity has induced pain when administered intrathecally (Chen been determined or which can be conﬁdently predicted et al., 2005), although in another study the same peptide based on homology (see Figs. 3 and 4), nine contain an ICK was reported to be lethal to mice when delivered via the motif. intraperitoneal or intracisternal route (Liang et al., 1993).

Figure 15: Effect of ProTx1 mutants on activity against NaV1.2, NaV1.5, NaV1.6, and NaV1.7. A

K17E mutation reduced the potency of ProTx1 for NaV1.2 and NaV1.5 by approximately 3-fold.

Potency for NaV1.7 was reduced about 7-fold, while potency against NaV1.6 was reduced ~10-fold.

The ProTx1-G32A-NH2 mutant exhibited even greater changes in activity. Activity against NaV1.2 was least affected, with a 8-fold reduction in IC50, while potency against NaV1.6 and NaV1.7 was reduced ~30-fold. C-terminal amidation of ProTx1 significantly improved potency against NaV1.2.

Table 3: NaV channel activity of ProTx1 mutants. IC50 values for inhibition of various NaV channel subtypes by wild-type (WT) ProTx1 and mutants thereof. In columns 3–5, the fold increase in IC50 for each mutant compared to that for WT ProTx1 is shown in parentheses.

Channel WT-ProTx1 K17E G32A-NH2 ProTx1-NH2

NaV1.2 98 nM 304 nM (3.1x) 770 nM (8x) 7 nM

NaV1.5 237 nM 800 nM (3.3x) 3000 nM (12.6x) 800 nM (3.3x)

NaV1.6 20 nM 220 nM (11x) 636 nM (32x) 43 nM (2x)

NaV1.7 67 nM 450 nM (6.7x) 362 nM (5.4x) 77 nM

34 2.4 Discussion The experiments described in this chapter provide data on the potency of ProTx1 and various mutants thereof against NaV1.2, NaV1.4, NaV1.5 and NaV1.6, data that is very important for developing a selectivity profile for ProTx1. The NaV1.7 pharmacophore of

ProTx1 was identified and compared to the NaV1.2 pharmacophore identified by Nitabach 116 and colleagues . The NaV1.7 pharmacophore is primarily composed of hydrophobic residues (L6, W27, V29 and W30) plus one charged residue (D31). Additional mutatagensis data indicated that G32 has a significant role in the ProTx1 interaction with

NaV1.7. Although K17E mutation showed significant changes to the toxin activity, the alanine scan did not reveal this effect (Fig. 10). One explanation could be that K17 is not a critical residue in the ProTx1-NaV1.7 interaction, however the introduction of E, which is a charge reversal mutation introduces rather unfavourable interactions. Thereby displaying substantially changed potency on various NaV. In contrast, the ProTx1 pharmacophore for interaction with NaV1.2 is much larger and consists of 11 residues, with five residues (L19,

W27, V29, W30 and D31) being shared with the NaV1.7 pharmacophore.

The most remarkable feature of this structural comparison was the set of four hydrophobic residues W5, L6, W27 and W30. Together, these residues comprise a very distinguishable hydrophobic protrusion on one side of the molecule. Looking at this patch of hydrophobicity, one might expect all of them to form part of a binding surface. Intriguingly, the mutational studies described above revealed that a W5A mutation has no effect on the ability of ProTx1 to inhibit NaV1.7, while an L6A mutation has very little effect on the ability of ProTx1 to inhibit NaV1.2. However, if one examines the original data from the NaV1.2 tethered-toxin alanine scan, it appears that the L6A mutation was just below the cut-off for selection as a pharmacophore residue (Supplementary Figure 1). Thus, the NaV1.2- ProTx1 interaction appears to be mediated by all four hydrophobic residues, W5, L6, W27 and W30, while the NaV1.7-ProTx1 interaction does not require W5.

The NaV1.7 pharmacophore lacks the ring of charged residues (R3, R23) and polar uncharged residue (S22) (Fig. 11B). These are away from the hydrophobic patch, facing the opposite side of the molecule, around the widest region of the folded peptide. This motif of a hydrophobic patch surrounded by a ring of charged/polar residues has been observed before in other NaSpTx Family 2 toxins such as ω-TRTX-Gr1a (GrTx), κ-TRTX- Gr1a (Hanatoxin-1) and κ-TRTX-Scg1a (SGTx)128-130.

35 A full alanine scan mutagenesis study was conducted to determine the pharmacophore on

κ-TRTX-Scg1a that mediates its interaction with KV2.1; the minimal active surface consists of 11 residues130. Four of these (Y4, L5, F6 and W30) form a hydrophobic patch, while other polar residues form a ring-like assembly concentrically arranged around this hydrophobic patch. Most importantly, alanine mutations of basic residues R3 and R23 led to a dramatic loss of activity, indicating that these residues may be directly interacting with 130 the voltage sensor domain of KV2.1 .

Previous studies revealed that an L19A mutant of κ-TRTX-Scg1a failed to fold properly130. L19 is fully conserved across the NaSpTx family 2 (Fig. 14), and it is completely buried within the ProTx1 structure. An L19A mutation abrogated the ability of t-ProTx1 to inhibit both NaV1.2 and NaV1.7. While ELISA assays indicated that this t-toxin mutant is expressed on the cell surface of oocytes, these assays do not address whether the L19A mutant is properly folded. It therefore seems likely that L19 has an important structural role in ProTx1. Thus, we conclude that an L19A mutation is likely to perturb ProTx1 folding but is unlikely to be part of the ProTx1 pharmacophore for NaV1.2 or NaV1.7.

As discussed previously (Section 2.3.4) ProTx1 binds to voltage sensor domains (VSDs) on channel domains II and IV (DII and DIV). Scanning mutagenesis of the rat NaV1.2 α- subunit (rNaV1.2) VSDs revealed several important residues on the DII VSD for the interaction with ProTx1 (Fig. 16). The mutations L840A, A841V, E844A, E844K, G845A, S847A, S851A, and L854A on DII all changed the apparent affinity of ProTx1 for 114 rNaV1.2 . This sequence of this region in the DII-VSD is identical in rat and human

NaV1.2, and the same region of human NaV1.7 differs from rNaV1.2/hNaV1.2 by only four residues, not inclusive of the residues listed above. The mutations F1610A, E1613A,

L1614A and Y1619A in the DIV-VSD of rNaV1.2 also significantly reduced apparent affinity for ProTx1. Again, the sequence of this region is identical in rNaV1.2 and hNaV1.2 and only two residues are different in hNaV1.7. In NaV1.7, E1613 is conservatively substituted for Asp, while a Lys to Thr substitution at position 1618 is not significant for ProTx1 binding.

Thus, while it may be assumed that the ProTx1 binding surface on NaV1.7 is similar to that of NaV1.2, the precise NaV1.7 VSD residues that bind to ProTx1 remain to be determined.

Figure 16: ProTx1 binding residues of rNaV1.2. Alignment of the S3-S4 segments of the DII and

DIV VSDs of hNaV1.2 and hNaV1.7. Transmembrane residues are marked in black, and the extracellular loops of DII and DIV are marked in red and purple respectively. The rNaV1.2 residues that are involved in ProTx1 binding are marked with blue boxes. Two sequences of NaV1.2 and

NaV1.7 are mostly identical; differences are marked with an asterisk. All ProTx1 binding residues in the DII-VSD of rNaV1.2 DII are conserved in NaV1.7. Three of the four ProTx1 binding residues in the DIV-VSD of rNaV1.2 are conserved with NaV1.7. E1613 is substituted by an Asp residue in the

NaV1.7 DIV-VSD.

In a previous study aimed at understanding the role of the κ-TRTX-Scg1a pharmacophore residues in lipid partitioning, it was found that alanine point mutations of eight residues (Y4, L5, F6, K17, H18, K26, G32, F34) each lost the capacity to partition significantly into lipid membrane130,131. Therefore it was argued that these residues have a role in lipid partitioning prior to binding the channel. However, the authors ruled out the possibility that the toxin exerts its inhibitory effect purely by disrupting lipid packing around the VSD and pore region. This was demonstrated by construction of the D-enantiomer of κ-TRTX-

Scg1a, which showed very similar lipid partitioning capacity, yet no inhibition of KV2.1. On the other hand, removing polar head groups of membrane lipids by converting sphingomyelin to ceramide, diminished the stability of the toxin-channel complex131,132. In conclusion, it was demonstrated that toxins such as κ-TRTX-Scg1a can partition into membranes under physiologically relevant conditions, and interact with the S3-S4 132 transmembrane helices of the VSD to inhibit KV2.1 . Thus, many toxins of the NaSpTx Family 2 likely have two sets of overlapping functional residues that are important for their channel inhibitory function, where one set is important for lipid partitioning, and the second set makes direct protein-protein interactions with the channel.

Two of the ProTx1 mutants constructed here, K17E and G32A, showed different selectivity and potency towards NaV1.2, NaV1.5, NaV1.6 and NaV1.7 relative to wild-type ProTx1. The

K17E charge reversal mutation did not improve ProTx1 specificity towards NaV1.7.

However, the ProTx1-G32A mutant had slightly improved selectivity for NaV1.7 over other

37 NaV channels. The side chain of glycine comprises a single hydrogen atom, and hence it can never contribute to energetically important protein-protein interactions. Therefore, we speculate that G32 plays an important role in the ProTx1-NaV channel interaction by helping to position C-terminal residues that are important for direct interactions with NaV channels.

The S3-S4 extracellular loops in the DII-VSD of NaV1.7 and NaV1.6 are identical, yet the

G32A mutation caused a greater loss of activity against NaV1.6 than NaV1.7. Similarly, though less profound, the K17E mutation also showed a difference in loss of activity between NaV1.6 and NaV1.7. These observations suggest that there are additional regions on the extracellular surface of NaV channels that are required for the interaction with ProTx1.

ProTx1 binds to NaV1.6, NaV1.7 and NaV1.8 with similarly high affinity (IC50 range of 20–90 nM). However the S3-S4 loop in the DII-VSD region of NaV1.8 is substantially different to that of NaV1.2, NaV1.6 and NaV1.7. Therefore it seems that the promiscuous activity of

ProTx1 on NaV might arise from a region with higher conservation, such as the S3/S4 transmembrane segments of the VSD, or perhaps ProTx1 simply interacts with different VSDs on the different channels.

NaSpTx Family 2 toxins, including ProTx1, may use a set of residues to partition into the membrane and interact with highly conserved transmembrane helices S3 and S4, while the ring of charged residues, which are exposed to the solvent, make interactions with the extracellular loop.

38 Figure 17: Sequence alignment of the S1–S2 and S3–S4 segments of domains I, II and IV from each of the nine human NaV channel subtypes.

ProTx1, a member of NaSpTx family 2 (Fig. 14), is a potent NaV1.7 inhibitor with promiscuous activity on a number of other neuronal targets including TRPV1, CaV channels and other NaV channels. This makes it a challenging molecule to develop into an analgesic drug lead. We identified the ProTx1 pharmacophore responsible for interaction with ProTx1, and identified the ProTx1 binding domain on NaV1.7 using a panel of

KV2.1/NaV1.7 chimeric channels. We also designed and tested three ProTx1 mutants in the hope of achieving increased selectivity for NaV1.7. One of these ProTx1 mutants,

G32A-NH2, had increased selectivity for NaV1.7, although this was achieved by slightly compromising toxin potency. The ProTx1 G32A-NH2 mutant therefore provides a solid foundation for development of ProTx1 as an analgesic drug lead.

39 Supplementary figures/ Appendix

Supplementary Figure 1: Membrane tethering has no effect on ProTx1 inhibition of NaV channels. (A) I–V traces for human NaV1.7 expressed in Xenopus oocytes with no toxin (blue), addition of soluble ProTx1 (orange), or co-expressed with t-ProTx1 (green). ProTx1 shifts the activation voltage to more positive potentials. (B) A large voltage prepulse (–100 mV to +100 mV) applied to NaV1.2-expressing oocytes before the test pulse (–100 mV to –10 mV) for a varying duration of time (0–1000 ms). (C) and (D) show recovery of NaV1.2 currents with increasing duration of the prepulse, in the presence of tethered ProTx1. (E) and (F) show the increasing

NaV1.2 recovery with longer prepulse, in the presence of the soluble ProTx1. This indicates that tethering does not alter the mode of action of ProTx1 (NaV1.2 data from Prof. Michael Nitabach, Yale University).

Supplementary Figure 2: Alanine-scanning mutagenesis of t-ProTx1. cRNAs of WT and alanine mutant t-ProTxI were transcribed in vitro and co-injected into Xenopus oocytes with hNaV1.2. 2-4 days after injection, voltage- or mustard oil-induced currents were recorded by two- electrode voltage clamp. Currents were normalized to oocytes in which the channel of interest is co-expressed with a control t-toxin: ACTX=δ-ACTX-Hv1a, an inhibitor of Na+ channel inactivation; PLTX=PLTX-II, an inhibitor of N-type Ca2+ channel133. Asterisks indicate alanine mutants with significantly larger currents compared to oocytes co-expressing the channel of interest with wild- type t-ProTxI (p<0.05, one-way ANOVA with Bonferroni's paired-comparison versus control).

Supplementary Figure 3: Immunoassay of alanine mutant surface expression. cRNAs of WT and mutant t-ProTxI were transcribed in vitro and injected into Xenopus oocytes. After 2-3 days incubation, surface expression of WT and mutant t-ProTx1 were measured by immunodetection of the linker domain epitope tag in living unpermeabilized oocytes and normalized to the surface signal of WT t-ProTx-I.

41 CHAPTER 3

DISCOVERY AND CHARACTERIZATION OF A NOVEL NaV1.7 INHIBITOR, β/µ-TRTX- Pe1b

3.1 Introduction

Chapter 2 revealed how alanine-scanning mutagenesis was successfully employed to uncover structure-activity relationships of protoxin-1. However, an alternative approach to understanding structure-activity relationships, which was employed in the work described in this Chapter, is to investigate the consequences of natural sequence variations found in homologous spider toxins.

Spider venoms are highly complex cocktails of peptides, proteins, small organic molecules, and salts75,85,134. Traditional proteomics-based estimates revealed that one venom could contain up to 1000 unique peptides75, whereas newer data suggests that this number can be as high as 3000135. Due to the incredible speciosity of the spiders (~45,000 extant species have been characterised;72), as a group they provide a vast number of peptides with varying levels of homology to each other. This natural sequence variation can be utilized as a tool to discover sequence variations that alter potency and specificity for selected pharmacological targets.

The discovery of novel bioactive peptides is a complicated process. Traditionally, this has been done via chromatographic fractionation of venoms92,95, but profiling of venom-gland transcriptomes is now being used more commonly135,136. In either case, there is usually a heavy emphasis on the dominant component in the venom or transcriptome, and this component is often screened against a series of ion channels and receptors to identify its natural target. This method can overlook compounds present in minor quantities. In many cases, a combined proteomic and trascriptomic approach is taken to identify full sequences of novel peptides, subject to availability of time and resources135,137. However the traditional discovery process remains an extremely time consuming process.

42 In contrast, assay-guided fractionation presents researchers the opportunity to rapidly screen a range of compounds against one or more selected targets, and follow up only on compounds with activity against these targets. Assay-guided fractionation consists of three major stages138. In the first stage, a panel of venoms is screened against the target of interest to identify venoms with desired activity. The screening assay can be high throughput (HTP) or low throughput. These methods are further discussed in next paragraph. In the second stage, venoms with desired activity ("hit" venoms) are chromatographically fractionated using reverse-phase high performance liquid chromatography (RP-HPLC) and then individual fractions are screened in the target assay. The purity of these fractions can be assessed using mass spectroscopic techniques such as matrix-assisted laser desorption (MALDI). In many cases, an additional orthogonal chromatographic fractionation of the "hit" RP-HPLC fractions will be required to isolate the pure bioactive compound. In the third stage, the purified bioactive peptide is sequenced using a combination of Edman degradation and MS/MS sequencing and, if it is available, reference to a venom-gland transcriptome.

As mentioned above, the venom screens can be conducted using high- or low-throughput methods. Many high-throughput techniques such as the fluorescent imaging plate reader (FLIPR) and automated patch clamp systems have been developed over the last decade, and many compound libraries have been screened very efficiently using these methods139. Although high-throughput techniques provide fast and cost-effective discovery of lead molecules, there is always the risk of false positives and false negatives. In contrast, low throughput techniques typically allow better accuracy, but the number of samples that can be tested in a given time is restricted. Therefore, the most appropriate screening technique needs to be selected based on the size and quality of the compound library tested and the time permissible for the initial screening process.

The advantage of assay-guided fractionation is not limited to the discovery of novel peptides. It can also be utilized as a tool to probe the inherent SAR potential of peptides. When a large number of spider venoms are fractionated and assessed on a single target, it is normal to find a number of homologous peptides amongst the hits. This is particularly true for screens against NaV and CaV channels, which are the dominant pharmacology of spider-venom peptides32,33. Minor sequence variations between homologous peptides can lead to drastic differences in potency and target subtype selectivity. For example, the two spider-venom peptides β-TRTX-Cm1a and β-TRTX-Cm1b only differ by a single amino

43 acid residue (D32 in β-TRTX-Cm1a compared to Y32 in β-TRTX-Cm1b)92. Nevertheless they show substantial difference in activity towards NaV1.3, with β-TRTX-Cm1b having

IC50 of 88 ± 25 nM and β-TRTX-Cm1a showed no observable effect up to a concentration of 2 µM. Therefore the availability of homologous toxins provides a great resource for understanding the target activity and selectivity of venom peptides.

The primary aim of this chapter is to demonstrate the power of assay-guided fractionation in the search for novel toxins and analogues of known toxins. Secondly, this chapter will demonstrate how a homology-model-based mutagenesis approach can be employed to enhance the preferred selectivity profile of a NaV channel toxin in order to obtain a potent and selective NaV1.7 inhibitor as an analgesic drug lead.

44 3.2 Materials and methods

3.2.1 Fractionation and purification of spider venoms

3.2.1.1 RP-HPLC

Crude venom (1 mg) from each spider was fractionated using an analytical C18 RP-HPLC column (Vydac). Buffer A was 0.05% trifluoroacetic acid (TFA) in water and buffer B was 0.05% TFA in 90% acetonitrile. Fractionation was performed using the gradient of 5% solvent B for 5 min, 5-20% solvent B for 5 min, 20-40% solvent B for 40 min, 40-80% solvent B for 4 min, wash at 80% solvent B for 2 min, 80-5% solvent B for 4, and a final 5% solvent B wash for 10 min at a flow rate of 1 ml/min. The total program length was 70 min.

3.2.1.2 Ion exchange HPLC

“NaV1.7-active” RP-HPLC fractions selected on the basis of FLIPR assays were further fractionated using ion exchange HPLC. A polysulfoethyl ion exchange column was used with buffer A (20 mM KH2PO4, 20% acetonitrile, pH 2.7) and buffer B (20 mM KH2PO4, 20% acetonitrile, 1 M KCl). Fractionation was performed using a salt gradient of 0% solvent B for 3 min, 0-45% solvent B for 45 min, 45-100% in 1 min, at 100% for 2 min, 100-0% in 1 min, and at 0% solvent B for 3 min at a flow rate of 1 ml/min The total fractionation time was 55 min.

3.2.1.3 Desalting by reverse phase HPLC

Ion exchange fractions were desalted using RP-HPLC on a Monolithic C18 column (Phenomenex). Buffer A was 0.1% formic acid in water and buffer B was 0.1% formic acid in 90% acetonitrile. Desalting was performed using a gradient of 5–60% solvent B over 8 min at a flow rate of 1 ml/min.

3.2.2 FLIPR assays FLIPR assays were performed by Dr Irina Vetter (Institute for Molecular Bioscience, The University of Queensland). Human neuroblastoma cells, SH-SY5Y (Max Planck Institute for Experimental Medicine, Goettingen, Germany) were incubated in the presence of 5%

CO2 with 15% of fetal bovine serum and 2 mM L-glutamine at 37°C. SH-SY5Y cells were routinely passaged ever 3-5 days at a 1:5 dilution with 0.25% trypsin and EDTA (Gibco). Cells were plated on an uncoated black-walled 96-well imaging plate (Corning, Lowell, MA, 45 USA) approximately 150,000 cells per well density, 48 h prior to the FLIPR assay. At the beginning of the assay cells were treated with Fluo-4 AM (5 µM) in physiological salt solution (PSS; composition NaCl 140 mM, glucose 11.5 mM, KCl 5.9 mM, MgCl2 1.4 mM,

NaH2PO4 1.2 mM, NaHCO3 5 mM, CaCl2 1.8 mM, HEPES 10 mM) with 0.3% BSA and incubated at 37°C for 30 min. At the end of 3 min, cells were treated with PSS and incubated another 10 min to ensure complete de-esterification of the fluorophore. Cells were washed twice with PSS. The 96 well-plate was transferred into the FLIPRTETRA (Molecular Devices, Sunnyvale, CA) fluorescent plate reader. Calcium responses were measured using a cooled CCD camera. The excitation and emission wavelengths of Flou- 4 AM are between 470–495 nm and 515–575 nm respectively. Ten baseline fluorescence readings were taken prior to the addition of the spider venom, or venom fraction. Then, fluorescent readings were taken for 300 s and NaV1.7 potentiator, veratridine, was added. Fluorescent readings were continued for another 300 s to observe the effect of the toxin or venom component in NaV1.7 activity.

3.2.3 Reduction of disulfide bonds and alkylation of cysteine residues Purified peptides were lyophilized and centrifuged at 13,000 rpm for 2–3 min in a 1.5 mL microcentrifuge tube. 50 µL of reduction/alkylation mixture (88% acetonitrile, 10% iodoethanol, 2% triethylphosphine) was placed on to the lid of the tube. The tube was closed carefully and the tube was then placed in a 37°C incubator for 45 min, after which the reduction/alkylation mixture was discarded along with the lid of the tube. The solid peptide at the bottom of the tube was dissolved in 50 µL of water and transferred into a fresh tube for further analysis. The extent of reduction and alkylation was assessed using mass spectrometry.

3.2.4 Mass spectrometry (MALDI-TOF and MADLI-ISD) Peptides masses were determined using MALDI time-of-flight (MALDI-TOF) mass spectrometry on a Model 4700 proteomics analyzer (Applied Biosystems). 1 µL of the sample was mixed with a 0.5 µL of α-cyano-4-hydroxycinnamic acid (CHCA) matrix (5 mg/ml in 50% acetonitrile) in a MALDI plate spot and dried at room temperature. Results were analyzed with Data Explorer® software (Applied Biosystems).

46 3.2.5 Recombinant expression of µ-TRTX-Pe1b 3.2.5.1 Transformation Transformation of E. coli cells with pLic-MBP expression vector was carried out as follows. Competent DH5α cells frozen at –80°C were thawed on ice for about 5 min, then 5 µL of plasmid was dissolved in nuclease-free water. Cells were left on ice for 45 min then heat- shocked at 42°C for 45 s. Cells were then placed on ice for another 5 min. 1 mL of Luria- Bertani (LB) medium was then added into the tube prior to incubation at 37°C for 1 h. The cell culture was then centrifuged at 8000 rpm for 5 min and ~900 µL of supernatant was removed. The pellet was resuspended in the remaining 100 µL of medium, plated on LB/agar + ampicillin (100 µg/ml), and incubated overnight at 37°C. Plasmids were extracted using a QIAprep Spin Miniprep Kit (Qiagen, Australia), and used to transform into E. coli strain BL21 (Invitrogen).

3.2.5.2 Recombinant expression, purification and cleavage of the fusion protein Production of recombinant venom peptides was performed using a modified version of a published protocol 140. Briefly transformed BL21(λDE3) cells were grown in 2 L of Luria-

Bertani (LB) medium to OD600 of 0.8, induced with 250 µM β-D-1-thiogalactopyranoside (IPTG) (Invitrogen), incubated for 24 h at 16°C, and harvested by centrifugation at 4000 rpm for 15 min at 4°C. The cell pellet was resuspended in 40 mM Tris, 400 mM NaCl, pH 8 (TN buffer), then cells were lysed by continuous flow cell disruption (TS Series Benchtop System, Constant Systems Ltd, Northants, UK) at a constant pressure of 25 kPa. The soluble fraction was separated by centrifuging the cell lysate at 15,000 rpm for 30 min at

4°C. The His6-MBP-toxin fusion protein was purified by passing the soluble fraction through an affinity chromatography column containing charged Ni-NTA super flow resin (Pierce, USA). The resin was then washed with 15 mM imidazole in TN buffer in order to remove proteins that were bound nonspecifically. The fusion protein was then eluted with 500 mM imidazole in TN buffer. The fusion protein sample was concentrated by centrifugal filtration (Amicon® Ultra, Millipore). Following concentration, the fusion protein was cleaved by incubation at 30°C for 3 h on a shaker with His6-tagged TEV protease (0.02 mg/mL) in the presence of reduced (GSH) and oxidized (GGSG) glutathione (0.2 mM GSH/0.04 mM GSSG) (Sigma). The TEV protease was produced in-house according to a published protocol 141.

47 3.2.5.3 Purification using a solid phase extraction column Sufficient HPLC solvent B (90% acetonitrile, 0.05% TFA) was added to the cleavage mixture to make a final concentration of 5% solvent B, then the mixture was centrifuged at 5500 rpm and filtered through a 0.45-µm syringe filter to remove MBP precipitates. The toxin was purified away from MBP and uncleaved fusion protein using a syringe solid phase extraction column (C18-LP; Alltech) conditioned with methanol and equilibrated with 5% solvent B. The column was washed with 5% solvent B, and toxin eluted with 50% solvent B. Any remaining components were eluted with 100% solvent B and methanol. Each fraction was lyophilized overnight. The 50% elution was resuspended in HPLC solvent A (water with 0.05% TFA), filtered with a 0.45 µm syringe filter and analysed using RP-HPLC. Toxins were further purified to obtain a single isoform on an analytical grade

C18 Vydac column (4.6 × 250 mm, 5 µm) using a Shimadzu HPLC system (Shimadzu, Kyoto, Japan).

3.2.5.4 Purification of the cleaved peptide with C4 semi-preparative column

When a semi-preparative C4 column (Phenomenex Jupiter, 250 x10 mm, 10 µm) became available, the cleavage mixture could be purified in one step without use of SPE column purification described before. The cleavage mixture was acidified with 0.5% TFA and 5% solvent B was added to equilibrate the mixture to RP-HPLC baseline conditions. The cleavage mixture was centrifuged at 5500 rpm for 15 min to remove precipitated MBP, then the clear supernatant was filtered through a 0.45 µm filter before injection into the HPLC system.

3.2.6 Two-electrode voltage clamp electrophysiology 3.2.6.1 Electrophysiological assay of toxin activity Oocytes were surgically removed form Xenopus laevis frogs and treated with 30 mg of collagenase (Sigma type1) in 20 mL of calcium-free ND96 solution (96 mM NaCl, 2 mM

KCl, 2 mM MgCl2, 5 mM HEPES, pH 7.4) for 1–1.5 h. Oocytes were washed three times with calcium-free ND96, or until the brown cloudy medium was completely removed. Healthy-looking oocytes were selected and allowed to rest about 1 h in modified ND96 medium (96 mM NaCl, 2 mM KCl, 1 mM CaCl2, 2 mM MgCl2, 5 mM HEPES, 5 mM pyruvic acid, 50 µg/ml gentamicin, 5% horse serum, pH 7.4). Selected oocytes were injected with 41 nL of about 200–800 ng/µL cRNA synthesized using a mMessage mMachine cRNA transcription kit (Life Technologies) and purified using an RNeasy Mini Kit (Qiagen). Oocytes were incubated at 17°C, media was changed every 24 h, and recordings were

48 conducted using two-electrode voltage-clamp (TEVC) electrophysiology (Axoclamp 900A amplifier, Molecular Devices, Sunnyvale, CA, USA) 2–4 days after the RNA injection. Two standard glass microelectrodes with 0.5–2 MΩ resistance when filled with 3 M KCl solution were used for recordings. Capillary glass for oocyte injections and electrodes was purchased from SDR Clinical Technology (Sydney, Australia). Data acquisition, and analysis was performed using pCLAMP software version 10 (molecular Devices). All recordings were conducted at room temperature in ND96 containing 0.1% fatty acid-free BSA (Sigma). Oocytes were clamped at –90 mV, and data were sampled at 20,000 Hz.

Inward NaV currents were invoked by applying 50 ms depolarizing steps form –80 mV to +45 mV with 10-mV increments, while holding the oocyte at –90 mV. A test pulse to –10 mV every 10 s was used to detect the presence of NaV channels and reach equilibrium upon toxin additions.

3.2.6.2 Analysis of Electrophysiological data All electrophysiological data were analysed using the software CLAMPEX®, Microsoft Excel (2011), and Prism 6. Voltage-current relationship curves (V-I curves) were plotted for each control trace and test traces with toxin. For each toxin experiment, the maximum current amplitude at a pre-selected voltage (e.g., -20 mV) was normalized to that of the current without the toxin added (0.2% BSA containing ND96 buffer only). This value was used as the fraction of the current remaining after the addition of a toxin dilution (I/Imax) for construction of dose-response curves. The maximum current amplitudes for NaV1.2 and

NaV1.7 experiments were measured at –20 mV, while –30 mV was used for analyzing

NaV1.5 and NaV1.6.

3.2.7 Construction of the β-TRTX-Pe1b homology model A homology model of β-TRTX-Pe1b was constructed based on the 3D structure of β/ω- TRTX-Tp1a (Chapter 2, Sections 2.3.3). A sequence alignment was made using CLUSTALW. MODELLER software was used to build the homology model, which was based on sequence and structure alignments of the peptides.

49 3.3 Results

3.3.1 Fractionation and purification of crude venoms to identify active peptides The workflow for this part of the project is shown schematically in Fig. 18A. In the initial stage, a unique in-house panel of more than 200 tarantula venoms was tested using a high-throughput FLIPR-based assay (Fig. 18B) to isolate venoms that inhibited human

NaV1.7. The hit rate was an astonishing 36%, making spider venoms likely to be the best natural source for discovering novel modulators of this channel. Crude venoms were prioritized for follow-up based on the amount of material available for deconvolution of the venom to isolate pure active compounds.

Figure 18: Assay guided fractionation and discovery of analgesic drug leads. (A). Workflow for isolating and characterizing NaV1.7 blockers from crude spider venoms. The isolation of active peptides from crude venoms involves HPLC separation techniques (reverse-phase and ion- exchange) and FLIPR-based assays. Sequencing involves Edman degradation and tandem MS. A recombinant expression system is then required in order to produce sufficient peptide for determination of the peptide’s structure, pharmacophore, and subtype selectivity. (B). FLIPR- based assay for isolating modulators of NaV1.7. Human neuroblastoma cells (SH-SY5Y) are incubated with venom, venom fractions, or purified peptide for ~5 min while baseline fluorescent readings are taken. Then veratridine, a NaV1.7 activator, is added. If NaV1.7 channels are not + inhibited, Na enters the cell leading to membrane depolarization, which activates endogenous CaV channels. This causes Ca2+ ions to enter the cell and bind to a calcium-sensitive dye (Fluo-4), leading to increase in fluorescence. If the added venom, venom fraction, or purified peptide inhibits + NaV1.7, no Na enters the cell, and there is no change in fluorescence. In contrast, if the venom, venom fraction, or purified peptide contains an activator of NaV1.7, an increase in fluorescence will be seen prior to addition of veratridine.

50 Six spider venoms were chosen for follow-up based on the initial FLIPR screen (Fig. 19). Three of the spiders were of Central and South American origin (Sericopelma sp., Pamphobeteus antinous and Euanthlus pulcherimaklaasi), while the other three were of Asian origin (Ornithoctoninae sp., Phormingochilus everetti and Monocentropus balfouri). Four of the venoms cased a complete block of veratridine-induced fluorescence (Fig. 19A–

D), and therefore might contain blockers of hNaV1.7, whereas two venoms induced a fluorescence response prior to the addition of veratridine (Fig. 19E,F), and therefore might contain activators of hNaV1.7.

Figure 19: FLIPR-based screen of crude spider venoms. (A). Spiders from which venom was sourced for isolation of NaV1.7 modulators: (a) Monocentropus balfouri (#473); (b) Sericopelma sp. (#531); (c) Phormingochilus everetti (#730); (d) Pamphobeteus antinous (#785); (e) Euathlus pulcherrimaklaasi (#662); (f) Ornithoctoneae sp. (#748). (B). FLIPR-based screen of crude spider venoms. The fluorescence response was abrogated when venoms from spiders 473, 531, 662, 730, 748, and 785 were added prior to addition of veratridine, indicating that these venoms most likely contain blockers of the human NaV1.7 channel. Venoms 531 and 748 shows initial bursts of fluorescence prior to the addition of veratridine, indicating that these venoms may contain activators of NaV1.7 as well as inhibitors.

Small portions (about 1mg) of the six selected crude venoms were fractionated using RP- HPLC (Fig. 20), then individual fractions were examined for activity in the FLIPR assay. All venoms produced ~50 different fractions from the first RP-HPLC fractionation, and all six venoms yielded RP-HPLC fractions that blocked hNaV1.7. Overall, 12 RP-HPLC fractions were obtained that contained putative modulators of the hNaV1.7 (Fig. 21).

51 A Venom 531 (Sericopelma spec.) B Venom 785 (Pampobetious antinous) From chiriqui, Panama (female) 1mg (female) 1mg 3000 214nm 100 3000 214nm 100 280nm 280nm

80 80 G G r r

e 2000 e 2000 a a c c d d n n

60 i 60 i e e a a n n b b t t r r

( ( o o % % s s

b b 40 B 40 B A A 1000 ) 1000 )

20 20

0 0 0 0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Time (min) Time (min)

C Venom 473 (Monocentropus balfouri) D Venom 748 (Ornithoctoninae spec) (female) 1mg 'Oranged fringed' (male) 1mg 3000 214nm 100 3000 214nm 100 280nm 280nm

80 80 G G r r e 2000 e 2000 a a c c d d n n

60 i 60 i e e a a n n b b t t r r

( ( o o % % s s

b b 40 B 40 B A A 1000 ) 1000 )

20 20

0 0 0 0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Time (min) Time (min)

E Venom 730 (Pormingochilus everetti) F Venom 662 (Euanthlus pulcherimaklaasi) (female) 1mg (female) 1mg 3000 214nm 100 3000 214nm 100 280nm 280nm

80 80 G G r r

e 2000 e 2000 a a c c d d n n i 60 60 i e e a a n n b b t t r r

( ( o o % % s s

b b B 40 40 B A A ) 1000 1000 )

20 20

0 0 0 0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Time (min) Time (min)

Figure 20: Fractionation of crude spider venoms using RP-HPLC. Chromatograms showing fractionation of crude venom from (A) Monocentropus balfouri (#473); (B) Sericopelma sp. (#531); (C) Phormingochilus everetti (#730); (D) Pamphobeteus antinous (#785); (E) Euathlus pulcherrimaklaasi. (#662); (F) Ornithoctoneae sp. (#748). In each case, peaks corresponding to fractions that were active in the FLIPR-based NaV1.7 assay are coloured.

Figure 21: FLIPR-based assay profiles for the 12 selected spider-venom fractions. Activation of endogenous hNaV1.7 channels in SH-SY5Y human neuroblastoma cells with veratridine leads to a fluorescence response as shown in the control well labelled CTRL-H2O. In contrast, the veratridine-induced fluorescence response was inhibited when cells were pre-incubated with selected venom fractions. This inhibition of the fluorescence signal is similar to that obtained with tetrodotoxin (TTX), a potent inhibitor of NaV1.7 and various other NaV channel subtypes. The two fractions of venom #531 displays increased levels of fluorescence prior to the addition of veratridine, indicating the presence of NaV1.7 agonists.

The five RP-HPLC fractions that caused NaV1.7 inhibition were impure as judged by MS, and consequently they were further fractionated using an orthogonal cation-exchange chromatography (CEX) step in an attempt to obtain pure active peptides. For example, faction 785-20 produced one symmetrical peak in CEX (Fig. 22A), while fraction 785-24 in CEX produced two fractions 785-24-1 and 785-24-2 (Fig. 22B). All three CEX fractions from venom #785 were found to inhibit hNaV1.7 in the FLIPR assay.

Figure 22: Purification of active peptides from venom 785. (A) Chromatogram from CEX separation of RP-HPLC fraction 785-20.The peak labeled in yellow was lyophilized and desalted. The monoisotopic mass (MIM) of the fraction was determined by MALDI-TOF MS to be 4388.40 Da. (B) Chromatogram resulting from CEX separation of RP-HPLC fraction 785-24 which yielded two major fractions 785-24-1 (blue) and 785-24-2 (green). Both fractions were desalted using RPHPLC. MALDI-TOF analysis yielded MIMs of 3937.8 Da and 4037.8 Da for peptides 785- 24-1 and 785-24-2, respectively.

Overall, 14 RP-HPLC fractions were obtained containing putative modulators of the human

NaV1.7 channel. After omitting four fractions that eluted early on rpHPLC and which therefore might be polyamines or other non-peptidic venom components, and one toxin due to purification difficulties, nine purified toxin peptides were sent to Australian Proteome Analysis Facility (APAF) for Edman sequencing; the sequences of the nine peptides are shown in Table 4.

54 Table 4: NaV1.7 inhibitor peptides isolated from crude venom Sequence Mono- NaSpTx name/numbe Sequence Isotopic Family r Mass (Da) Tp1a(662-33) ECRYWLGGCSAGQTCCKHLVCSRRHGWCVWDGTFS 3985.20 Family 2 Pe1b(730-44) GECRYWLGGCSKTGDCCEHLSCSPKWHWCVWDGTF 3974.70 Family 2 Pe1a(730-30) ECKYLWGTCSKDEDCCAHLGCNRKHDWCGWDYTF-NH2 4066.29 Family 2 Pa1a(785-20) ACREMYGGCEKDEDCCKHLGCKPTLNYCAWDGTFV-NH2 3937.60 Family 2 662-29 YCQKWMWTCDSERKCCEGMVCRLWCKKKLW 3825.23 Family 3 748-11 ACLGFGKSCNPSNDQCCKSSSLACSTKHKWCKYEL 3819.87 Family 1 785-24 DDCLGWFRGCDPKNDKCCEGYKCSRRDQWCK--- 4037.00 Family 1 662-19 DCLKFGWKCNPRNDKCCSGLKCGSNHNWCKLH 3800.25 Family 1 ω-TRTX- 785-20-13-2 FECVISCDIKKEGKPCKPKGGKKCSGGWRCK?K 4388.40 Asp1d* *This peptide shows no homology with any of the established NaSpTx families, however it is 90% identical to ω-TRTX-Asp1d, a CaV channel inhibitor. The second last residue could not be determined by Edman degradation. Due to the low activity of this peptide on NaV1.7, no further attempts were made to determine the complete sequence of peptide 785-20-13-2.

Four of the nine NaV1.7-active peptides were isolated from crude venom of the spiders

Euanthlus pulcherimaklaasi (U1-TRTX-Ep1a), Pormingochillus everetti (U1-TRTX-Pe1a and U1-TRTX-Pe1b), and Pampobetious antinous (U1-TRTX-Pa1a) (see Table 4). U1-

TRTX-Ep1a is 100% identical to the known NaV1.7 inhibitor β/ω-TRTX-Tp1a (ProTx1), 95 which potently blocks the human channel with an IC50 of 51 nM . This provided an excellent validation of the FLIPR-based NaV1.7 assay. The other three peptides are close homologs of β/ω-TRTX-Tp1a; all four peptides belong to the NaSpTx Family 2 (Fig. 23), 32 which is the largest family of spider toxins with predicted NaV inhibitory function .

Figure 23: Sequence alignment of toxins in NaSpTx Family 2 32. Toxin names are based on the rational nomenclature devised for spider-venom peptides90. Light green names indicate peptides discovered in this study. Toxins marked in dark green have been well studied95. NaSpTx Family 2 divides into four major groups based on the primary activity of the toxin. The largest group of toxins contains KV channel blockers; while NaV channel blockers make the second largest group inclusive of the two novel toxins Pe1a and Pe1b. NaV1.7-active surface residues (marked in purple) of ProTx1 have been identified previously (Chapter 2, Section 2.3.2). This active site shares some 130 similarities with the KV2.1 active site of SGTx (κ-TRTX-Scg1a) , especially L6 and W30.

3.3.2 Recombinant expression of toxins Once the peptide sequences were determined, they were used to construct E. coli expression vectors to produce the peptides recombinantly. Recombinant expression enables large quantities of each peptide to be obtained for structure-function characterization and it enables isotopic labeling of the peptides for structure determination using multidimensional heteronuclear NMR140,142. A synthetic gene encoding the peptide of

56 interest, with codons optimised for E. coli expression, was cloned into a variant of the pLic- MBP expression vector143. The fusion protein encoded by the plasmid contains a MalE signal sequence (MalEss), that directs the fusion protein to the E. coli periplasm where the 144 cellular machinery for disulfide bond formation is located , a His6 tag for affinity purification, a maltose binding protein (MBP) fusion tag to aid solubility145, a tobacco etch virus (TEV) protease recognition site, and the spider-venom peptide (Fig. 24).

Figure 24: Design of the recombinant expression system. (A). Schematic representation of the coding region of the pLic-MBP expression vector. This IPTG-inducible plasmid contains genetic components encoding the MalESS-His6-MBP-toxin fusion protein with a TEV protease recognition site between the MBP and toxin coding regions. Other major components include the T7 promoter at the beginning of the construct, Lac operator, ribosomal binding site (RBS), KpnI and SspI restriction sites using for cloning in the toxin gene, and the T7 terminator (B). The cellular processes involved in expression of toxins in the periplasm of E. coli. The fusion peptide is expressed within the cytoplasm, then the MalE signal sequence directs the fusion peptide to the periplasm. The signal sequence is cleaved off once the peptide reaches the periplasm via the SecYEG translocase system. The Dsb machinery in the periplasm allows formation and shuffling of disulfide bonds in order to achieve the native fold of the peptide.

Three novel toxins, β-TRTX-Pe1a (hereafter Pe1a), β/µ-TRTX-Pe1b (hereafter Pe1b) and

U1-TRTX-Pa1a (hereafter Pa1a), were successfully produced by expression in the periplasm of E. coli BL21 cells (Figs 25–27). Following TEV protease cleavage the peptides contained an additional non-native N-terminal glycine residue, a vestige of the TEV protease recognition site, making the recombinant peptides 35 residues long140. The recombinant Pe1b folded into a single isoform.

When the recombinant Pe1b was tested on human NaV1.7 heterologously expressed in Xenopus oocytes using TEVC electrophysiology, 100 nM of peptide was found to inhibit nearly 50% of the current, therefore confirming the correct folding of the toxin. Pe1a produced eight isoforms, however only fraction 7 co-eluted with the native toxin (Fig. 26C).

Fraction 7 showed weak NaV1.7 inhibition, with 10 µM toxin inhibiting 25% of the total current (Fig. 26D), while the other isoforms showed no activity on NaV1.7.

Figure 25: Recombinant production of Pe1b. (A). SDS-PAGE gel showing outcome of various purification steps. Lanes are as follows: L, protein size marker; ISF, insoluble fraction following cell disruption; SF, soluble fraction containing the fusion protein; FT, flow through from the Ni-NTA column; W, wash of Ni-NTA column; Elu, first eluate from wash of Ni-NTA column with 500 mM imidazole; Elu3, third eluate from wash of Ni-NTA column with 500 mM imidazole; CFT, flow through during protein concentration; CF, concentrated fraction of His6-MBP-toxin fusion protein; and Tev, His6-MBP after TEV cleavage. (B). HPLC chromatogram resulting from final purification of Pe1b using a C18 analytical column. Black line indicates the absorbance at 214 nm, while the red line shows the absorbance at 280 nm. The dotted line shows the solvent B gradient used for the RP-HPLC purification. Inset is the MALDI-TOF trace obtained for recombinant Pe1b showing expected monoisotopic mass of 4035.65 Da.

58 Retention Time (min)

Figure 26: Recombinant production of Pe1a, a NaV1.7-active peptide from the spider Pormingochillus everetti. (A). SDS-PAGE gel showing outcome of various purification steps. Lanes are as follows: ISF, insoluble fraction following cell disruption; SF, soluble fraction containing the fusion protein; FT, flow through from the Ni-NTA column; W, wash of Ni-NTA column; Elu, first eluate from wash of Ni-NTA column with 500 mM imidazole; Elu3, third eluate from wash of Ni-NTA column with 500 mM imidazole; CFT, flow through during protein concentration; CF, concentrated fraction of His6-MBP-toxin fusion protein; Tev, His6-MBP after TEV cleavage, and L, protein size marker. (B). HPLC chromatogram resulting from final purification of Pe1a using a C18 semi prep column. The left ordinate axis shows the absorbance in arbitrary units (AU), and the right ordinate axis shows the % solvent B (90% acetonitrile and 0.05% TFA). Black line indicates the absorbance at 214 nm, while the red line shows the absorbance at 280 nm. The dotted line shows the solvent B gradient used for the RP-HPLC purification. Pe1a elutes as many isoforms between 42–48% solvent B. (C). HPLC-based purification of various Pe1a isoforms, denoted fractions 1 to 8. Pe1a-F7 has a similar retention time to native Pe1a (orange line). (D). 10 µM recombinant Pe1a inhibits about 25% of the total NaV1.7 current with a small leftward shift of the I-V curve. The ascissa shows voltage steps (mV) ranging form -80 mV to +45 mV, while the ordinate axis shows the maximum current amplitude (nA) at each voltage step.

Figure 27: Recombinant expression of Pa1a, a

NaV1.7-active peptide from the spider Pampobetious antinous (A). SDS-PAGE gel showing the outcome of various purification steps. Lanes are as follows: ISF, insoluble fraction following cell disruption; L, protein size marker; SF, soluble fraction containing the fusion protein; FT, flow through from the Ni-NTA column; W, wash of Ni-NTA column; Elu, first eluate from wash of Ni-NTA column with 500 mM imidazole; Elu3, third eluate from wash of Ni-NTA column with 500 mM imidazole; CFT, flow through during protein concentration; CF, concentrated fraction of

His6-MBP-toxin fusion protein; and Tev, His6-MBP after TEV cleavage. (B). HPLC chromatogram resulting from final purification of Pa1a using a C18 analytical column. The left ordinate axis shows the absorbance (AU), while the right ordinate axis shows the % solvent B. Black and red lines indicates the absorbance at 214 nm and 280 nm, respectively. The dotted line shows the solvent B gradient used for the RP-HPLC purification. Pa1a elutes around 35% solvent B.

3.3.3 Pe1a and Pa1a activities on NaV channels

When tested on human NaV1.7 heterologously expressed in Xenopus oocytes, Pe1a only inhibited about 25% of the total NaV current at 10 µM concentration (Fig. 26D), demonstrating that the peptide is a weak inhibitor of hNaV1.7. However it should be noted that native Pe1a is amidated (Table 4), and the recombinant peptide is not. Lack of a C- terminal amidation may have affected the NaV1.7 activity of the peptide. Pe1a showed no effect on NaV1.2 currents (data not shown).

Pa1a did not display any activity hNaV1.7 expressed in Xenopus oocytes. As for Pe1a, native Pa1a is amidated at the C-terminus. Lack of amidation may be responsible for the difference in NaV1.7 activity between the native and recombinant peptides. Pa1a showed very high sequence identity (69–71%) with many KV channel modulators in NaSpTx family 2 (see Fig. 23, Pa1a sequence marked in light green), and therefore it would be interesting in the future to examine the activity of Pa1a against a range of KV channels.

60 3.3.4 Pe1b activities on NaV currents

Pe1b was tested for its activity against hNaV1.2, hNaV1.5, hNaV1.6 and hNaV1.7 heterologously expressed in Xenopus oocytes (Fig. 28). Cells were held at –80 mV and current traces were evoked by 50-ms depolarizing steps to –10 mV every 10 s. Once the maximum current stabilized (typically ~1 min), a current versus voltage step (IV) trace was acquired. Pe1b was found to be a weak inhibitor of NaV1.2 (IC50 ~ 2.8 µM; Fig.29) with no shift in activation voltage. Pe1b inhibited NaV1.7 most potently (IC50 ~152 nM; Fig. 29), while demonstrating less potency towards NaV1.6 (IC50 ~696 nM; Fig.29). Based on data currently available, Pe1b is a very weak inhibitor of NaV1.5. At 1 nM-10 µM concentrations, the toxin showed only a weak concentration-dependent inhibition of NaV1.5. At a concentration of 10 µM, Pe1b inhibited about 40% of the NaV1.5 current (Fig. 28).

Figure 28: Current-voltage (IV) relationships for Pe1b inhibition of hNaV1.2, hNaV1.5, hNaV1.6 and hNaV1.7. The voltage (mV) is shown on the abscissa and the current (µA) is shown on the ordinate axis. The channel current in the absence of toxin is shown in black, while current after the addition of Pe1b is shown in color. All traces except NaV1.5 show the effect of 1 µM Pe1b. NaV1.5 trace shows the effect of 10 µMPe1b on channel current. Pe1b weakly inhibited NaV1.2, with no shift in activation voltage. Pe1b partially inhibited NaV1.5 at 10 µM with no shift in channel activation voltage. Pe1b inhibited NaV1.6 and NaV1.7 with about a 10 mV shift of the activation voltage in the positive direction. The current-voltage step protocol used is shown at the far right of the figure.

3.3.5 Construction of Pe1b mutants Pe1b and ProTx1 are 70% identical, yet they display significant differences in activity against NaV channels. We therefore wondered whether it might be possible to determine which residues are responsible for the differences in NaV channel selectivity of these two peptides. Although recombinant production of Pe1b adds a glycine at the N-terminus of the peptide, the amino acid residue numbers for native Pe1b are used in this chapter as

61 this allows direct comparison with ProTx1 and other NaSpTx family 2 peptide sequences without causing confusion.

Amino acid residues that are markedly different between ProTx1 and Pe1b were at the primary focus of this investigation. Five main regions of difference were identified (Fig. 29 B, C), which can be summarized as follows: (i) the four residues K11, T12, G13, and D14 within the second loop of Pe1b were significantly different to that of ProTx1. However when these residues were mapped onto the structure of ProTx1 (Fig. 29C, residues marked in cyan), this loop is located far form the NaV1.7 active site of the toxin (residues marked in orange). Therefor these residues were expected to have minimal effect on toxin activity against NaV1.7; (ii) the basic K17 in ProTx1 is substituted with an acidic Glu residue in Pe1b, leading to a charge reversal; (iii) residue V20 in ProTx1 is substituted with a Ser residue in Pe1b; (iv) four residues, P23, K24, W25 and H26, within loop 4 of Pe1b are different to that of ProTx1; and (v) ProTx1 contain a Ser at the C-terminus, which is not present in Pe1b. Based on these differences, I decided to produce four Pe1b mutants to investigate the effect of above-mentioned substitutions (Fig. 29D). These mutant peptides were named Pe1b-E17K, Pe1b-L2, Pe1b-S20V and Pe1b-S35, in contrast to wild-type Pe1b (Pe1b-wt).

Figure 29: Pe1b activity on NaV channels and design of Pe1b mutants. (A). Concentration- response curves for modulation of hNaV1.2 (orange), hNaV1.5 (green), hNaV1.6 (light blue) and hNaV1.7 (deep blue) by Pe1b. Pe1b inhibited NaV1.7 with an IC5o of ~152 nM, which is about 22- fold lower (i.e., more potent) that its activity on hNaV1.2 (IC50 ~ 2.8 µM) and about 5.5-fold more potent that its activity on hNaV1.6 (IC50 ~ 696 nM). Pe1b shows only very weak activity against hNaV1.5, with 10 µM toxin inhibiting ~40% of the current. (B). Sequence alignment of ProTx1 (green) and Pe1b (orange). These toxins have a similar cystine framework, and many identical non-Cys residues. The four major regions of sequence variation within proximity of the ProTx1-

NaV1.7 pharmacophore are marked with grey boxes. (C). The structure of ProTx1. Residues marked in orange indicate the NaV1.7-binding pharmacophore of ProTx1. Residues marked in yellow are not conserved in Pe1b, however located in close proximity to the NaV1.7 binding pharmacophore residues. Hence might provide some clues to the difference in NaV activities of two peptides. Residues marked in cyan, are also not conserved with Pe1b, however these are far from the active site. Therefore effects of mutations in these residues were not examined further. The loop 4 (as indicated on Fig. 29C) residues R23, R24, H25, and G26 are poorly conserved in Pe1b. (D). Sequences of Pe1b mutants designed by substituting the corresponding residue of ProTx1 (indicated in green). Mutants were only designed for residues that are close to the hypothesized

NaV1.7 pharmacophore (highlighted in yellow in panel C), The mutant Pe1b-S35 contains an additional non-native C-terminal serine residue.

63 3.3.6 Recombinant expression and activity of Pe1b mutants All four Pe1b mutants were successfully produced via recombinant expression (Fig. 30). All mutants yielded a single isoform upon final RP-HPLC purification, and they were of very high purity as judged by MALDI-TOF MS. The activity of each of the Pe1b mutants was assessed against hNaV1.2, hNaV1.5 and hNaV1.7 expressed in Xenopus oocytes.

Figure 30: Recombinant expression of Pe1b-E17K, Pe1b-S20V, Pe1b-L2 and Pe1b-S35. Left panels are SDS-PAGE gels showing the outcome of various steps in the purification of each mutant peptide. Right panels are RP-HPLC chromatograms resulting from purification of each

Pe1b mutant using a C4 semi-preparative column. The peaks corresponding to the desired Pe1b mutants are marked with as asterisk. Pe1b-E17K, Pe1b-L2, and Pe1b-S20 eluted at ~45% solvent B, whereas Pe1b-S35 elutes at ~40% solvent B. Black and red traces correspond to absorbance at 214 nm and 280 nm, respectively. The dotted line indicates the gradient of solvent B gradient that was used.

64 ProTx1 is a potent inhibitor of NaV1.7 (IC50 ~50 nM) but it lacks the Nav channel selectivity required for it to be a useful analgesic lead. In contrast, Pe1b shows better NaV channel selectivity, but it lacks potency against NaV1.7 (IC50 ~152 nM). We therefore constructed four Pe1b mutant peptides in order to investigate the effect of each mutation on the potency and selectivity of the toxin. The activity of each mutant peptide against NaV1.2,

NaV1.5 and NaV1.7 is shown in Fig. 31 and the results are summarized in Table 5.

NaV1.2 NaV1.5 NaV1.7

1.0 1.0 1.0 Pe1b wt, IC50 ~152 nM

Pe1b E17K, IC50 ~50 nM 0.8 0.8 0.8 l l l ro ro ro t t t

on 0.6 on 0.6 on 0.6 c c c

I / I / 0.4 I / 0.4 0.4 Pe1b wt, IC50 ~2.8 µM Pe1b wt, IC50 >10 µM 0.2 Pe1b E17K, IC >1 µM 0.2 0.2 Pe1b E17K 50 Pe1b E17K, IC50 ~720 nM 0.0 0.0 0.0 -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 log[Peptide], M log[Peptide], M log[Peptide], M

Pe1b wt, IC50 ~152 nM 1.0 1.0 1.0 Pe1b L2, IC50 ~123 nM 0.8 0.8 0.8 l l l ro ro ro t t t

on 0.6 on 0.6 on 0.6 c c c

I / 0.4 I / 0.4 I / 0.4 Pe1b L2 Pe1b wt, IC50 >10 µM 0.2 Pe1b wt, IC50 ~2.8 µM 0.2 0.2 Pe1b L2, IC50 ~1.2 µM Pe1b L2, IC50 ~200 nM 0.0 0.0 0.0 -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 log[Peptide], M log[Peptide], M log[Peptide], M

Pe1b wt, IC ~152 nM 1.0 1.0 1.0 50 Pe1b S20V, IC50 ~107 nM 0.8 0.8 0.8 l l l ro ro ro t t t

on 0.6 on 0.6 on 0.6 c c c

I / I / 0.4 I / 0.4 0.4 Pe1b wt, IC >10 µM Pe1b wt, IC50 ~2.8 µM 50 Pe1b S20V 0.2 0.2 Pe1b S20V, IC ~10 µM 0.2 Pe1b S20V, IC50 ~660 nM 50

0.0 0.0 0.0 -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 log[Peptide], M log[Peptide], M log[Peptide], M

1.0 1.0 1.0

0.8 0.8 0.8 l l l ro ro ro t t t

on 0.6 on 0.6 on 0.6 c c c

I / 0.4 I / 0.4 I / 0.4 Pe1b S35 Pe1b wt, IC50 ~2.8 µM Pe1b wt, IC50 >10 µM 0.2 0.2 0.2 Pe1b wt, IC50 ~152 nM Pe1b S35, IC50 ~520 nM Pe1b S35, IC50 >10 µM Pe1b S35, IC50 ~80 nM 0.0 0.0 0.0 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 log[Peptide], M log[Peptide], M log[Peptide], M

Figure 31: Activity of Pe1b mutants. The black dotted line in each panel shows the activity of

Pe1b-wt, while the coloured unbroken line (orange for NaV1.2, green for NaV1.5 and blue for

NaV1.7) shows the activity of each Pe1b mutant peptide. (The number of repeats for each

65 concentration point ranges from 1 to 3). IC50 values calculated from each of the concentration- response curves are indicated on the figures. Data were collected via TEVC electrophysiological analysis of hNaV1.2, hNaV1.2, and hNaV1.2 currents in Xenopus oocytes.

Table 5: IC50 values for inhibition of NaV1.2, NaV1.5 and NaV1.7 by wild-type and mutant Pe1b peptides. The table also shows the fold-increase in activity for each mutant compared to the wild- type toxin (Pe1b-wt).

IC50 and the fold increase in activity in comparison to Pe1b-wt

NaV1.2 NaV1.5 NaV1.7

Pe1b-wt 2.8 µM > 10 µM 152 nM

Pe1b-E17K >1 µM 724 nM (13.7x) 50 nM (3x)

Pe1b-L2 200 nM (14x) 1.2 µM (8.3x) 123 nM

Pe1b-S20V 660 nM (4.2x) ~ 10 µM 107 nM (1.4x)

Pe1b-S35 525 nM (5.3x) > 10 µM 80 nM (2x)

The Pe1b-E17K mutant (IC50 50 nM) proved to be about 3-fold more potent against NaV1.7 than native Pe1b (IC50 152 nM). Critically, however, potency against NaV1.2 was not greatly changed due to this mutation (IC50 >1 µM). Unfortunately, potency against NaV1.5

(IC50 ~ 725 nM), was improved about 14-fold compared to the Pe1b-wt. Nevertheless,

Pe1b-17K still has a 15-fold preference for NaV1.7 over NaV1.5.

The Pe1b-L2 mutant showed a small increase in potency against NaV1.7 (IC50 ~ 123 nM) compared to native Pe1b, whereas activity against NaV1.2 was improved dramatically (i.e., the IC50 decreased 14-fold from 2.8 µM to 200 nM). Activity against NaV1.5 was also markedly improved, with Pe1b-L2 exhibiting an IC50 of 1.2 µM against this channel compared to >10 µM for the native toxin.

Pe1b-S20V activity against NaV1.7 (IC50 143 nM) was similar to Pe1b-wt (IC50 152 nM).

Pe1b-S20V potency against NaV1.2 (IC50 660 nM) was about 4.2-fold better than wild-type

Pe1b (IC50 2.8 µM). Potency against NaV1.5 was very poor as for the native toxin. At a concentration of 10 µM, Pe1b-S20V only inhibited ~50% of the NaV1.5 current.

Pe1b-S35 had increased potency against hNaV1.7 in comparison to Pe1b-wt. The laboratory's standard operating protocol for oocyte recordings uses 0.2% BSA in ND96 buffer to treat the recording oocyte before any peptide is added. Inhibition of hNaV1.7 by

66 Pe1B-S35 was comparable to that obtained with Pe1b-wt when recordings were performed in this manner (i.e., IC50 ~ 150 nM). However, when the protocol was modified to use ND96 buffer containing no BSA, Pe1b-S35 had markedly improved activity against

NaV1.7, with an IC50 of 80 nM. This observation was made by accident, and the effect of various BSA concentrations on peptide activity has not been tested for any other peptide described in this thesis.

67 3.4 Discussion Assay-guided fractionation is an efficient technique to rapidly identify peptides that modulate the activity of a specific channel. The FLIPR assay described herein has proved to be a useful tool for screening spider venoms for NaV1.7 inhibitors. Once the individual peptide component responsible for the activity has been identified and sequenced, it is essential to produce the peptide in quantities sufficient for full characterization (i.e., determination of biological activity, three-dimensional structure and structure-activity relationships). An Escherichia coli-based periplasmic expression system was successfully employed here to produce several spider-venom peptides with activity against hNaV1.7. The major advantage of this technique is the protein folding machinery available in E. coli periplasm. All NaSpTx family 2 toxins are disulfide-rich peptides, and therefore it is crucial that the correct disulfide-bond isoform is produced146.

ProTx1 and Pe1b share 70% sequence identity. The 3D structure of ProTx1 is available (Chapter 2, Section 2.3.3), and the Pe1b structure was modeled based on this structure. Both peptides form a hydrophobic patch at a very similar surface of the toxin scaffold, leading to the hypothesis that they share a significant level of similarity in their NaV1.7 binding site.

An E17K charge reversal mutation in Pe1b had little effect on its activity against NaV1.2.

While this mutation increased NaV1.7 potency by 3-fold, it unfortunately also increased potency against NaV1.5 by almost 14-fold. While Pe1b-E17K still has ~15-fold selectivity for NaV1.7 over NaV1.5, we conclude that its improved potency against NaV1.7 has been obtained with too much sacrifice of the selectivity over NaV1.5. Therefore Pe1b-E17K is not a useful analgesic drug lead.

The Pe1b-L2 mutation caused a drastic 16-fold gain of activity against hNaV1.2 in comparison to a 2-fold increase in activity against NaV1.7. It appears that the set of residues (RRGH) that substituted into loop 2 in place of the PKWH sequence in Pe1b-wt contains one or more residues that are important for hNaV1.2 binding. The Pe1b-L2 mutation also caused an 8-fold gain of activity against NaV1.5, making this peptide undesirable as an analgesic drug lead.

An S20V mutation did not cause a significant change in Pe1b activity against NaV1.7, and therefore S20 cannot be an important part of the toxin's NaV1.7 pharmacophore. This is

68 consistent with the pharmacophore that mediates the interaction of ProTx1 with NaV1.7 (see Chapter 2, Section 2.3.2. Although S20 in Pe12b-wt and V20 in ProTx1 are not related to the NaV1.7 activity of either peptide, V20 appears to be one source of the higher potency of ProTx1 against NaV1.2 since introduction of a S20V mutation increased Pe1b potency against NaV1.2 by 4.2-fold (Table 5).

The increased NaV1.7 activity of the Pe1b-S35 mutant is a good indication that the C- terminus of Pe1b is involved in its interaction with NaV1.7. Although this mutation compromised the NaV1.7/NaV1.2 selectivity of the peptide, it is possible that additional, complementary mutations might serve to increase its NaV1.7/NaV1.2 selectivity. Clues to such mutations are available from the ProTx1 pharmacophore mapped using alanine- scanning mutagenesis (Chapter 2, Section 2.3.2 and Supplementary Figure 2)116.

Residues that are not shared between the NaV1.7 and NaV1.2 pharmacophores of ProTx1, such as R3, W5, and F34, might be the most interesting to examine. Alanine mutations of R3, W5 and F34 should provide an indication of whether alterations of these residues can lead to altered NaV1.7/NaV1.2 selectivity.

Furthermore, the effect of BSA in the recording buffer observed with Pe1b-S35 needs to be further investigated. The low potency of the peptide in the presence of BSA may have been due to the peptide binding to BSA and reducing the effective concentration of toxin available for binding with the channel. In future, peptide activity against NaV1.7 should be compared in ND96 buffer with and without BSA.

In conclusion, the four Pe1b mutations studied here have deepened our understanding of the surface of Pe1b that mediates its interaction with NaV1.7 and other therapeutically important NaV subtypes. Moreover, the Pe1b-E17K and Pe1b-S35 analogs have provided a foundation for developing Pe1b into an analgesic lead molecule for therapeutic inhibition of NaV1.7.

69 CHAPTER 4

LOCALIZATION OF NaV1.7 IN THE NORMAL AND INJURED RODENT OLFACTORY SYSTEM INDICATES A CRITICAL ROLE IN OLFACTION, PHEROMONE SENSING AND IMMUNE FUNCTION (Incorporated research paper)

Localization of Nav1.7 in the normal and injured rodent olfactory system indicates a critical role in olfaction, pheromone sensing and immune function

Darshani B. Rupasinghe,1 Oliver Knapp,2 Linda V. Blomster,3 Annina B. Schmid,3 David J. Adams,2 Glenn F. King1,* and Marc J. Ruitenberg3,4,*

1Division of Chemistry and Structural Biology; Institute for Molecular Bioscience; The University of Queensland; St. Lucia, QLD Australia; 2Health Innovations Research Institute; RMIT University; Melbourne, VIC Australia; 3School of Biomedical Sciences; The University of Queensland; St. Lucia, QLD Australia; 4Queensland Brain Institute; The University of Queensland; St. Lucia, QLD Australia

Keywords: sodium channel, pain, inflammation, injury, regeneration, monocyte

Abbreviations: AOB, accessory olfactory bulb; BSA, bovine serum albumin; CIP, congenital insensitivity to pain; DAB, 3,3'-diaminobenzidine; DAPI, 4',6-diamidino-2-phenylindole; EDTA, ethylenediaminetetraacetic acid; GFP, green fluorescent protein; GL, glomerular layer; LP, lamina propria; OB, olfactory bulb; OE, olfactory epithelium; OMP, olfactory marker protein; © 2012ONL, olfactory nerve layer; Landes PBS, phosphate-buffered saline; RT,Bioscience. room temperature; TTX, tetrodotoxin

Loss-of-function mutations in the pore-forming α subunit of the voltage-gated sodium channel 1.7 (Nav1.7) cause

congenital indifference to pain and anosmia. We used immunohistochemical techniques to study Nav1.7 localization in

the rat olfactory system in order to better understand its role in olfaction. We confirm that Nav1.7 is expressed on olfactory

sensory axons and report its presence on vomeronasal axons, indicating an important role for Nav1.7 in transmission

of pheromonal cues. Following neuroepithelial injury, Nav1.7 was transiently expressed by cells of monocytic lineage. These findings support an emerging role for Na 1.7 in immune function. This sodium channel may provide an important Do notv distribute. pharmacological target for treatment of inflammatory injury and inflammatory pain syndromes.

Introduction touch, vibration, joint position sense and temperature perception but fail to identify noxious stimuli as unpleasant or painful. In

Voltage-gated sodium (Nav) channels are transmembrane pro- addition to their inability to detect pain, people with CIP have a teins that provide a current pathway for the rapid depolariza- complete loss of smell (anosmia).14,15 Similarly, mice containing

tion of excitable cells. The pore-forming α subunits are classified an olfactory-specific knockout of Nav1.7 exhibit an absence of 1,2 15 into nine subtypes denoted Nav1.1 to Nav1.9. In recent years, odor-guided behaviors. Thus, it is clear that Nav1.7 plays a criti- 16 the tetrodotoxin (TTX)-sensitive sodium channel Nav1.7 has cal role in the detection of both nociceptive and olfactory cues. emerged as a well-validated analgesic target based on several The primary olfactory pathway in mammals comprises the remarkable human genetic studies. Gain-of-function mutations olfactory epithelium (OE), cranial nerve I and the main olfac-

in the SNC9A gene encoding the α subunit of Nav1.7 cause pain- tory bulb (OB). The vomeronasal organ and accessory olfactory ful inherited neuropathies, including primary erythromelalgia3 bulb (AOB) provide a second level of olfactory detection in cer- and paroxysmal extreme pain disorder,4 whereas truncation and tain vertebrates, including rodents. While there is debate as to loss-of-function mutations in SCN9A result in a congenital insen- whether or not the vomeronasal organ detects all pheromone-like sitivity to all forms of pain or CIP.5,6 Moreover, single nucleotide cues in mammals, and whether its sensory capacity is restricted to polymorphisms in SCN9A correlate with an individual’s level of pheromonal sensing, the vomeronasal system is generally agreed pain perception.7 Thus, there is much interest in developing sub- to be essential for detecting and processing social, behavioral and 17,18 type-selective blockers of Nav1.7 as analgesics for the treatment of reproductive cues. various types of pain.8-13 Primary sensory neurons in the OE detect external olfactory To date, 13 distinct mutations in SCN9A have been mapped stimuli at their dendritic terminals, which causes transmission of in people with CIP.5,6 These individuals have a normal sense of action potentials along their axons. These axons converge into

*Correspondence to: Glenn F. King and Marc J. Ruitenberg; Email: [email protected] and [email protected] Submitted: 11/28/11; Revised: 01/24/12; Accepted: 01/25/12 http://dx.doi.org/10.4161/chan.19484

www.landesbioscience.com Channels 103 Figure 1. Localization of Nav1.7 in rat olfactory epithelium. (A) Immunohistochemical staining for Nav1.7 in the rat OE. The approximate location from

which the photomicrographs were taken is shown in (B) (red boxed area). Note that axonal bundles (nb) are intensely stained (++) for Nav1.7 within

the lamina propria (LP). Co-localization of Nav1.7 (red) and OMP (green) was revealed through double immunofluorescent staining. Note the overlap ©between the2012 individual staining patterns in theLandes merged image (right). (C) Higher power photomicrographBioscience. of an axonal bundle showing localization of Nav1.7 (red, left panel), OMP (green, middle panel), and the merged image (right panel). Scale bars: 100 μm (A) and 12 μm (C).

axonal bundles (fila olfactoria) within the lamina propria under- raised against a unique epitope, was first examined and con- neath the OE. They then pass through the cribriform plate into firmed in vitro via staining of stably transfected mammalian cell

the nerve layer of the OB, where each axon ultimately synapses lines expressing different Nav channel subtypes. The antibody with the dendrites of mitral cells (2nd order neuron) in defined was found to recognize Na 1.7 but not related Na channel sub- Do not distribute.v v 19,20 OB glomeruli, the functional units of olfactory coding. An types including Nav1.4, Nav1.5 and Nav1.6 (see Fig. S1).

axonal bundle projecting out of the OE is generally composed Nav1.7 localizes to axons of primary olfactory sensory neu- of about 2,000–3,000 olfactory sensory neuron axons wrapped rons and central termination zones. We explored the role of

within a tightly connected channel of olfactory ensheathing cells Nav1.7 in olfaction by studying its expression pattern along the (OECs).21 Because of their exposed location, olfactory sensory primary olfactory neuraxis. In the nasal neuroepithelium, we

neurons are constantly replenished from an endogenous stem cell observed prominent (++) staining for Nav1.7 in the fila olfac- compartment. Therefore, the primary olfactory pathway contin- toria of the lamina propria (LP) (Fig. 1A and B). We did not

uously repairs itself (reviewed in ref. 22). detect staining for Nav1.7 on the cell bodies and dendrites of

The precise role of Nav1.7 in olfaction has been unclear, with primary olfactory sensory neurons in the rat, consistent with no descriptions of the presence of this channel in the mammalian recent observations in mice.15,23 These findings indicate a par- 15,23 olfactory system until very recently. We therefore decided to ticularly high abundance of Nav1.7 along the trajectory of the

examine the anatomical localization of Nav1.7 in the rat olfactory olfactory nerve, which was confirmed by co-localization of

system to gain more insight into its role in olfaction. Our results olfactory marker protein (OMP) and Nav1.7 staining (Fig. 1C),

support the view that Nav1.7 plays an essential role in relaying the former being a marker for olfactory sensory neurons and action potentials initiated by the detection of olfactory cues, their projections. including pheromones, between the nasal olfactory neuroepithe- Coronal sections (taken at ~1.5 mm intervals) through the

lium and the main OB, as well as between the vomeronasal organ rostro-caudal axis of the rat OB were next examined for Nav1.7

and the AOB. We also show that within this same system Nav1.7 staining (Fig. 2). The olfactory nerve layer (ONL) within ros- is expressed on infiltrating inflammatory cells and thus may play tral sections of the OB, demonstrated by regions ‘a’, ‘b’, ‘c’ and important roles in monocyte actions at sites of tissue injury. ‘d’ (Fig. 2B), was darkly stained (++, Fig. 2A). The observed intensity was similar to that seen in axonal bundles in the lamina Results propria. The glomerular layer (GL), which is located internal to the ONL and comprises the termination zone of the axons

We examined the localization of Nav1.7 in rat olfactory tissue via from olfactory sensory neurons, was more lightly stained (+; Fig.

standard immunohistochemical and immunofluorescent staining 2A). At lower magnification, Nav1.7 staining within glomeruli procedures (for details, see Materials and Methods). Specificity appeared most intense close to the ONL, gradually losing some

of the commercially available anti-Nav1.7 antibody, reportedly intensity as axons spread throughout the glomerulus. However,

Figure 2. Localization of Nav1.7 in the rat main olfactory bulb and accessory olfactory bulb. (A) Localization of Nav1.7 in four different regions (a–d) of a rostral section of the OB, shown schematically in (B). The four parts in each row, from left to right, show DAB staining followed by immunofluorescent

staining for Nav1.7 and OMP as a marker for axons of olfactory sensory neurons; a merged image of Nav1.7 and OMP staining patterns is shown on

the right. In each DAB-stained section, the ONL is stained darkly (++) for Nav1.7, whereas the glomerular layer (GL) is stained more lightly (+). Double

immunofluorescent staining procedures again show co-localization between Nav1.7 and OMP in both the ONL and GL. (C) Higher-power confocal

photomicrograph (4 μm thick slice) of a single glomerulus, showing co-localization of Nav1.7 and OMP. (D) Schematic representation of a more caudal section through the OB, showing the accessory olfactory bulb (AOB, red boxed area). (E) Note that the AOB is lightly (+) but uniformly stained for both

Nav1.7 and OMP. Scale bars: 200 μm (A and E), 15 μm (C).

when viewed at higher optical resolution using confocal micros- caudal pole, we observed Nav1.7 staining on vomeronasal axons

copy, Nav1.7 staining appeared relatively uniformly distributed in the AOB, consistent with that of ONL (AOB; Fig. 2D and E).

throughout the glomeruli (Fig. 2C). We again observed colocal- Expression of Nav1.7 on inflammatory cell infiltrate in the

ization of Nav1.7 and OMP staining, confirming the presence injured mouse olfactory epithelium. Because of the emerging 24 of Nav1.7 on axons of primary olfactory sensory neurons. At the role of Nav1.7 in immune function, we studied expression of

Figure 3. Degeneration and regeneration of the olfactory epithelium following chemical ablation. (A and B) Representative images showing intact + olfactory epithelium (OE) and lamina propria (LP). Well-organized axonal bundles (nb) within the LP are stained for NaV1.7. Cell bodies of OMP neurons (~9–10 layers) were also be observed in the stratified OE. (C and D) Two days after methimazole injection, a substantial loss of olfactory sensory neurons was evident by the dramatic reduction in OMP+ cell numbers. Remnants of olfactory tissue containing OMP+ neurons had detached from the gfp/+ basal cells layers and the LP (C, white two-headed arrow). Cx3cr1 cells of monocytic lineage (L) were visible within detached epithelial remnants, + some of which appeared to co-express Nav1.7 (D). Additional GFP cells, many with typical macrophage-like morphology, were present in the LP. (E) At 5 d after injury, no OMP staining was visible within the OE, indicating a near complete ablation of mature (OMP+) olfactory sensory neurons following methimazole treatment. Some residual OMP staining was still present in degenerating axon bundles in the LP (white arrowhead). As anticipated,

a clear reduction in staining intensity for Nav1.7 was observed at 2 and 5 d post-injury compared with non-injured controls (D and F vs. B). Note,

however, the prominent staining for NaV1.7 on inflammatory cell infiltrate accumulating almost exclusively within the LP at 5 d-post-injury (F). (M) Immunohistochemical staining for GFP on adjacent sections revealed that cells of monocytic lineage were present in both the LP and OE. At 10 d after + injury, the first OMP cells began to re-appear in the regenerating OE (G, white arrowhead) and little, if any, Nav1.7 immunoreactivity was observed + within olfactory nerve bundles (H). By this stage, most of Nav1.7 cell infiltrate had disappeared and the few remaining cells were primarily observed in close association with blood vessels (black arrow). However, numerous macrophage-like cells were present in OE and LP at this time point (N). By 18 d + after injury, epithelial regeneration was now clearly evident by a substantial increase in OMP cell numbers (I) and the reappearance of Nav1.7 staining gfp/+ 33 in olfactory nerve bundles (J). The number and distribution of Cx3cr1 macrophage-like cells had also returned to normal control levels (O). Quanti- + + tative counts of GFP cells within the olfactory neuroepithelium are shown in (K). Note that the appearance of Nav1.7 cells in the LP precedes the peak of macrophage infiltration into the OE. Scale bars: 50μ m.

this channel in the olfactory pathway after injury. Methimazole to drug-induced cell death within the olfactory neuroepithelium. drug treatment or Triton-X lesions (data not shown) were used to Horizontal basal cells, the neuroepithelial stem cells, remain induce turnover of olfactory sensory neurons and explore expres- largely unaffected and regenerate the OE in time (~1–2 mo).26,27 25 sion of Nav1.7 on inflammatory cell infiltrate (Fig. 3). Treatment We examined Nav1.7 staining along the primary olfactory path- with methimazole causes robust recruitment of macrophages due way at 2, 5, 10 and 18 d after methimazole injection.

106 Channels Volume 6 Issue 2 Figure 4. Expression of Nav1.7 on a subset of cells of the monocytic lineage. (A) Nav1.7 staining can be detected on small round cells at 5 d after exper- gfp/+ imentally-induced death of olfactory sensory neurons in Cx3cr1 mice (arrowheads). (B) Co-expression of GFP was observed for these cells, indicating that they are most likely newly recruited inflammatory monocytes based on their morphology. Note that GFP+ cells with more mature macrophage

morphology do not stain positively for Nav1.7. A merged image is shown in (C). Scale bar: 50 μm.

At 2 d following methimazole treatment, the OE appeared this channel must be present in olfactory tissue and that it must severely damaged and disorganized with some regions completely play a pivotal role in olfaction. Electrophysiological studies have detached from the more basement membrane (compare Fig. 3A revealed the presence of TTX-sensitive Na+ currents in cultured and C). Olfactory sensory neurons were virtually absent from rat olfactory sensory neurons29 and mouse OE tissue slices.15 All

the OE at 5 and 10 d after methimazole treatment, as expected, Nav subtypes except Nav1.5, Nav1.8 and Nav1.9 are TTX-sensitive, although OMP+ remnants of their axons were still detected within but those responsible for Na+ currents in olfactory receptor neu- olfactory nerve fascicles (Fig. 3E and G). Regeneration of the OE rons were unclear until very recently. It is now known that Na 1.7 © 2012 Landes Bioscience.v 23 became evident by a thickening of the cellular layers with time is the predominant Nav channel in ORNs and that this chan- and the reappearance of OMP+ olfactory sensory neurons at 18 d nel is essential for action potential conduction in olfactory nerve after methimazole treatment (Fig. 3I). Epithelial thickness aver- terminals in OB glomeruli.15

aged ~70% of normal control values at this point in time and the We observed little, if any, Nav1.7 expression on the apical den-

intensity of Nav1.7 staining returned to normal (++) levels at 18 d. drites of both rat and mouse primary olfactory sensory neurons, We noted a high abundance of Na 1.7+ inflammatory cells imme- which is in line with other recent observations in mice15,23 and Dov not distribute. diately following injury. Specifically, at 2 d following methimazole suggests that this channel does not play a direct role in odor- + treatment, and even more so after 5 d, Nav1.7 expression was evi- ant detection. Rather, the observed distribution of Nav1.7 within dent on smaller cells with a globular morphology in the lamina the rat olfactory system, in combination with the severe olfactory

propria underneath the OE (Fig. 3D and F). Significantly fewer deficit in people with CIP that lack Nav1.7, indicates that this + Nav1.7 cells were observed at later time points (i.e., ≥10 d) after channel plays a fundamental role in amplifying and propagating methimazole treatment, at which stage immunoreactive cells were action potentials to the OB following the detection of olfactory

mostly present in close association with blood vessels within the cues. This function is similar to the role of Nav1.7 in nociceptive lamina propria (Fig. 3H). As the vasculature of the lamina propria neurons, where it is believed to act as a threshold channel that forms the entry point for circulating leukocytes into the OE, these amplifies pain signals transmitted above a certain level.30

findings suggest that expression of Nav1.7 may be downregulated Publication of two independent reports detailing the role of

as monocyte precursor cells enter, differentiate and transmigrate Nav1.7 in olfaction in mice coincided with preparation of this 15,23 into the neuroepithelium for debris clearance and phagocytosis. paper. We extended the two previous studies on Nav1.7 Such a hypothesis was supported by the fact that numerous mac- expression along the mouse olfactory pathway by showing that rophage-like cells were present within the lesioned olfactory tissue this channel is also present in the rat AOB. The AOB and the gfp/+ 28 from Cx3cr1 mice at all time points investigated (Fig. 3L–O), vomeronasal organ are responsible for pheromone sensing in + 17 and that the appearance of Nav1.7 cells in the lamina propria mammals. Pheromonal cues are detected within the vomero- preceded the peak of macrophage infiltration into the OE (Fig. nasal organ at the base of the nasal cavity and carried via vom-

3K). Subsequent double-staining procedures confirmed Nav1.7 eronasal receptor neurons to the AOB, where this information is expression on a subset of green fluorescent protein (GFP)+ cells processed. An earlier report demonstrated through in situ hybrid- + + of monocytic lineage. GFP Nav1.7 cells lacked the cytoplasmic ization techniques that Nav1.3 is also expressed in the mouse

protrusions more characteristic of mature macrophages, suggest- vomeronasal organ, together with Nav1.1 and Nav1.2, albeit at ing they were indeed newly recruited monocytes (Fig. 4). extremely low levels.31 Subsequent patch-clamp electrophysiology studies showed that the voltage-dependent Na+ currents were Discussion completely inhibited by 2 μM TTX, confirming the presence of

TTX-sensitive Nav channels in rodent vomeronasal organ neu- 32 The fact that people with CIP as a result of loss-of-function rons. The relatively uniform Nav1.7 staining we observed in the

mutations in Nav1.7 also suffer from anosmia indicates that AOB demonstrates that Nav1.7 is present in terminations of the

www.landesbioscience.com Channels 107 vomeronasal receptor neurons and most likely plays a critical role cells. AlexaFluor 488 goat anti-rabbit antibody alone showed no

in pheromone detection. Given that NaV1.3 is primarily confined nonspecific binding (not shown). to neuronal cell bodies, our findings indicate that, similar to the Rat tissue collection and processing. Male, 5 mo-old Sprague- 15,23 main olfactory system, the TTX-sensitive Nav1.7 channel is Dawley rats (n = 5) were deeply anaesthetized via intraperitoneal very likely responsible for the propagation of action potentials injection with sodium pentobarbital (80–100 mg/kg; Virbac), along the vomeronasal axonal pathway. followed by transcardial perfusion with 100 ml of heparinized

Lastly, we noted prominent expression of Nav1.7 on the early saline (0.9% NaCl) and 300 ml of Zamboni’s fixative. The OE inflammatory cell infiltrate present within the lamina propria and OB were then extracted by dissection followed by overnight after injury. The OE normally contains both macrophages post-fixation at 4°C. Samples were subsequently decalcified via and dendritic cells33 and after injury these populations actively overnight immersion incubation in phosphate-buffered saline increase through recruitment of circulating monocytes.34,35 (PBS) containing 250 mM EDTA (EDTA), followed by cryo- + Nav1.7 cellular infiltrate was transiently observed during the protection in graded (10 and 30%) sucrose solution. Next, tissue gfp/+ acute phase of macrophage recruitment. The use of Cx3cr1 samples were snap-frozen in dry ice-cooled isopentane and stored

mice revealed that even though a sharply diminished Nav1.7 at -80°C until further processing. Tissue samples were sectioned immunoreactivity was observed in the sub-acute phase after into 20 μm thick slices on a Leica CM 3050 S cryostat and serial injury (day 5 onwards), there was a continued presence of sections were collected on Superfrost®-Plus glass slides (Menzel monocyte-derived cells within the tissue, suggesting a down- Glaser), air-dried and again stored at -80°C until used. regulation of this channel. These observations are consistent Immunostaining procedures. Slides were air-dried at RT for

with a recent report showing downregulation of Nav1.7 during 40 min followed by several washes in PBS (3 x 5 min each) for re- 24,36 maturation of dendritic cells. The role of Nav1.7 in immune hydration. For peroxidase-based immunohistochemistry, the tis- cell function and differentiation remains largely unclear. sue was first incubated with 500 μL/slide of quenching solution

However, emerging evidence suggests a critical role for Nav1.7 (0.6% H2O2 and 10% methanol in PBS) to block endogenous ©in the maturation 2012 and migration of a subsetLandes of dendritic cells as peroxidase Bioscience. activity within the tissue. Slides were then washed well as the production of pro-inflammatory cytokines.24 These again in PBS and blocked with 2% bovine serum albumin (BSA)

findings warrant further investigation into the role of Nav1.7 and 0.2% Triton-X100 in PBS for 1 h at RT in a humidified in other cells of the monocytic lineage, since therapeutics that chamber. Next, slides were incubated overnight at 4°C with pri-

target Nav1.7 might be useful immune modulators for the treat- mary rabbit polyclonal anti-Nav1.7 antibody (1:200; Millipore, ment of post-injury inflammation and chronic inflammatory AB5390-200UL). The following day, slides were extensively pain, in addition to beingDo effective analgesics. not distribute.washed (3 x 10 min) in PBS to remove unbound antibody followed by incubation with biotinylated goat anti-rabbit second- Materials and Methods ary antibody (1:400; Jackson ImmunoResearch Laboratories, 711-066-152) at RT. Slides were then washed again and incu- Cell culture and immunofluorescence. Chinese hamster ovar- bated with 1:400 peroxidise-conjugated avidin-biotin complex

ian (CHO) cells stably transfected with Nav1.4, Nav1.6 and (ABC; Vector Laboratories, SP-2002) for 1 h at RT. Slides were

Nav1.7 (Genionics AG) were grown on poly-D-lysine coated washed and stained for 3–5 min at RT using 3,3'-diaminobenzi- glass coverslips in F-12 medium (Gibco) supplemented with dine (DAB) as a chromogen (Sigma Life Science, D4293). The

9% fetal bovine serum (Lonza, BW14-502E), 0.9% penicillin/ reaction was stopped via immersion of slides into ddH2O. Slides streptomycin (Gibco, 15070-063), and 100 μg/ml hygromycin B were dehydrated through a graded series of ethanol washes (70– (Invitrogen, 10687-010). HEK-293 cells stably transfected with 100%), cleared in xylene, and coverslipped with DePex mount-

Nav1.5 (Genionics) were grown in Dulbecco’s modified Eagle ing medium (Merck, 13514). medium (Gibco, 12491015), 0.9% penicillin/streptomycin and For double immunofluorescent staining procedures, similar 100 μg/ml zeocin (Invitrogen, R250-01). procedures were followed, except that incubation with quench- Sub-confluent cells were fixed with 10% paraformaldehyde ing solution was omitted. In brief, tissue sections on slides were (Sigma, 158127-100G), permeabilized with 10% Triton-X washed with PBS and blocked with 2% BSA for 1 h. Sections (Sigma, T8532-100ML), then nonspecific binding was blocked were then incubated overnight at 4°C with a mixture of rab-

by incubation with 10% goat serum (Invitrogen, PCN500). Cells bit anti-Nav1.7 antibody (1:200) and goat anti-OMP antibody were incubated for 2 h at room temperature (RT) with a rabbit (1:1,000; Wako, 019-22811). Staining for OMP was used to

anti-Nav1.7 antibody (1:2,000; Millipore, AB5390-200UL) and, visualize primary olfactory sensory neurons and their projec- after additional washing steps, for 1 h with an AlexaFluor 488 tions in the nasal neuroepithelium. The next day, sections were goat anti-rabbit antibody (1:1,000; Invitrogen, A11008). Nuclei washed (3 x 10 min) and incubated with Cy3-conjugated don- were counterstained with 4',6-diamidino-2-phenylindole (DAPI) key anti-rabbit (1:400) (Jackson ImmunoResearch Laboratories, (1:2,000; Invitrogen, D21490). Coverslips were mounted in 711-165-162) and biotin-conjugated donkey anti-goat secondary fluorescent mounting media (DAKO, S3023) and observed with antibodies (1:400; Jackson ImmunoResearch Laboratories, 705- fluorescence microscopy. To control nonspecific binding, anti- 065-003) for 1.5 h at RT. Following this, sections were washed

Nav1.7 antibodies were blocked for 30 min with a 5-fold excess (3 x 10 min) and incubated with Alexa 488 conjugated strep- of its antigen or incubated together with nontransfected HEK tavidin (1:400; Sigma, S11223). Finally sections were washed

108 Channels Volume 6 Issue 2 again (3 x 10 min) and coverslipped with ProLong® Gold anti- Olfactory ablation. Methimazole drug treatment or Triton-X

fade reagent (Invitrogen, P36935) containing DAPI. Nav1.7 lesions were performed as described in references 25 and 37. In staining specificity was verified via omission of the primary brief, mice received two consecutive intraperitoneal injections of antibody and/or incubation with control (pre-immune) rabbit methimazole (Sigma, M8506; 50 mg/kg) on days 0 and 3 fol-

serum. Cross-reactivity with other Nav channels was excluded lowed by perfusion and tissue processing as detailed above 2, 5, gfp/+ 28,33 in separate in vitro experiments (see Fig. S1). Cx3cr1 mice 10 or 18 d following the first injection. were used to visualize Nav1.7 expression on cells of monocytic lineage. For this purpose, immunohistochemical staining for Disclosure of Potential Conflicts of Interest Nav1.7 (as described earlier) was examined in relation to native No potential conflicts of interest were disclosed. GFP fluorescence. Microscopy and image analysis. DAB-stained tissue sections Acknowledgements were scanned using an Aperio scanscope XT. The staining pat- We acknowledge financial support from the Australian Research tern and distribution were analyzed using ImageScope software Council (Discovery Grant DP110103129 to G.F.K.) and The (Aperio). Different levels of staining intensity representing the University of Queensland (Enabling Grant to M.J.R.). D.J.A.

relative abundance of Nav1.7 were documented as: (1) ‘-’ indi- is an ARC Australian Professorial Fellow. We are grateful to cates no staining, defined as not above background levels; (2) ‘+’ Darren Paul for help with confocal imaging, and the Australian indicates light staining that is clearly elevated above background Cancer Research Foundation (ACRF) for access to the ACRF levels; (3) or ‘++’ for darkly stained regions with the highest Cancer Biology Imaging Facility at the Institute for Molecular

abundance of Nav1.7. Photomicrographs of immunofluorescent Bioscience. stainings were taken with an Olympus BX-51 fluorescence microscope and a Zeiss LSM 710 FCS confocal microscope. Greyscale Supplemental Material images were false colored and channels were merged to visualize Supplemental materials may be found here: ©colocalization 2012 using Image J. Landeswww.landesbioscience.com/journals/channels/article/19484 Bioscience. References 10. Momin A, Wood JN. Sensory neuron voltage-gated 20. Wachowiak M, Denk W, Friedrich RW. Functional sodium channels as analgesic drug targets. Curr Opin organization of sensory input to the olfactory bulb 1. Yu FH, Catterall WA. Overview of the voltage-gated Neurobiol 2008; 18:383-8; PMID:18824099; http:// glomerulus analyzed by two-photon calcium imag- sodium channel family. Genome Biol 2003; 4:207; dx.doi.org/10.1016/j.conb.2008.08.017. ing. Proc Natl Acad Sci USA 2004; 101:9097-102; PMID:12620097; http://dx.doi.org/10.1186/gb-2003- 11. McGowan E, Hoyt SB, Li X, Lyons KA, Abbadie C. A PMID:15184670; http://dx.doi.org/10.1073/ 4-3-207. peripherally acting Na(v)1.7 sodium channel blocker pnas.0400438101. 2. King GF, Escoubas P, Nicholson GM. Peptide toxins reverses hyperalgesia and allodynia on rat models of 21. Li Y, Field PM, Raisman G. Olfactory ensheathing that selectively target insect Na(V)Do and Ca(V) channels. not inflammatory distribute. and neuropathic pain. Anesth Analg cells and olfactory nerve fibroblasts maintain con- Channels (Austin) 2008; 2:100-16; PMID:18849658; 2009; 109:951-8; PMID:19690272; http://dx.doi. tinuous open channels for regrowth of olfactory nerve http://dx.doi.org/10.4161/chan.2.2.6022. org/10.1213/ane.0b013e3181b01b02. fibres. Glia 2005; 52:245-51; PMID:15968636; http:// 3. Waxman SG, Dib-Hajj S. Erythermalgia: molecular 12. Priest BT. Future potential and status of selective dx.doi.org/10.1002/glia.20241. basis for an inherited pain syndrome. Trends Mol Med sodium channel blockers for the treatment of pain. 22. Ruitenberg MJ, Vukovic J. Promoting central nervous 2005; 11:555-62; PMID:16278094; http://dx.doi. Curr Opin Drug Discov Devel 2009; 12:682-92; system regeneration: lessons from cranial nerve I. Restor org/10.1016/j.molmed.2005.10.004. PMID:19736626. Neurol Neurosci 2008; 26:183-96; PMID:18820410. 4. Fertleman CR, Baker MD, Parker KA, Moffatt S, 13. Dib-Hajj SD, Black JA, Waxman SG. Voltage-gated 23. Ahn HS, Black JA, Zhao P, Tyrrell L, Waxman SG, Elmslie FV, Abrahamsen B, et al. SCN9A mutations sodium channels: therapeutic targets for pain. Pain Dib-Hajj SD. Nav1.7 is the predominant sodium in paroxysmal extreme pain disorder: allelic variants Med 2009; 10:1260-9; PMID:19818036; http:// channel in rodent olfactory sensory neurons. Mol underlie distinct channel defects and phenotypes. dx.doi.org/10.1111/j.1526-4637.2009.00719.x. Pain 2011; 7:32; PMID:21569247; http://dx.doi. Neuron 2006; 52:767-74; PMID:17145499; http:// 14. Nilsen KB, Nicholas AK, Woods CG, Mellgren SI, org/10.1186/1744-8069-7-32. dx.doi.org/10.1016/j.neuron.2006.10.006. Nebuchennykh M, Aasly J. Two novel SCN9A muta- 24. Kis-Toth K, Hajdu P, Bacskai I, Szilagyi O, Papp 5. Cox JJ, Reimann F, Nicholas AK, Thornton G, Roberts tions causing insensitivity to pain. Pain 2009; 143:155- F, Szanto A, et al. Voltage-gated sodium channel E, Springell K, et al. An SCN9A channelopathy 8; PMID:19304393; http://dx.doi.org/10.1016/j. Nav1.7 maintains the membrane potential and regu- causes congenital inability to experience pain. Nature pain.2009.02.016. lates the activation and chemokine-induced migration 2006; 444:894-8; PMID:17167479; http://dx.doi. 15. Weiss J, Pyrski M, Jacobi E, Bufe B, Willnecker V, of a monocyte-derived dendritic cell subset. J Immunol org/10.1038/nature05413. Schick B, et al. Loss-of-function mutations in sodium 2011; 187:1273-80; PMID:21715690; http://dx.doi. 6. Goldberg YP, MacFarlane J, MacDonald ML, org/10.4049/jimmunol.1003345. channel Nav1.7 cause anosmia. Nature 2011; 472:186- Thompson J, Dube MP, Mattice M, et al. Loss-of- 90; PMID:21441906; http://dx.doi.org/10.1038/ 25. Vukovic J, Marmorstein LY, McLaughlin PJ, Sasaki

function mutations in the Nav1.7 gene underlie con- nature09975. T, Plant GW, Harvey AR, et al. Lack of fibulin-3 genital indifference to pain in multiple human popula- 16. Nassar MA, Levato A, Stirling LC, Wood JN. alters regenerative tissue responses in the primary tions. Clin Genet 2007; 71:311-9; PMID:17470132; Neuropathic pain develops normally in mice lacking olfactory pathway. Matrix Biol 2009; 28:406-15; http://dx.doi.org/10.1111/j.1399-0004.2007.00790.x. both Na(v)1.7 and Na(v)1.8. Mol Pain 2005; 1:24; PMID:19520159; http://dx.doi.org/10.1016/j.mat- 7. Reimann F, Cox JJ, Belfer I, Diatchenko L, Zaykin PMID:16111501; http://dx.doi.org/10.1186/1744- bio.2009.06.001. DV, McHale DP, et al. Pain perception is altered by a 8069-1-24. 26. Sakamoto T, Kondo K, Kashio A, Suzukawa K, nucleotide polymorphism in SCN9A. Proc Natl Acad 17. Brennan PA, Keverne EB. Something in the air? Yamasoba T. Methimazole-induced cell death in rat Sci USA 2010; 107:5148-53; PMID:20212137; http:// New insights into mammalian pheromones. Curr olfactory receptor neurons occurs via apoptosis trig- dx.doi.org/10.1073/pnas.0913181107. Biol 2004; 14:81-9; PMID:14738757; http://dx.doi. gered through mitochondrial cytochrome c-medi- 8. Dib-Hajj SD, Cummins TR, Black JA, Waxman SG. org/10.1016/j.cub.2003.12.052. ated caspase-3 activation pathway. J Neurosci Res From genes to pain: Na 1.7 and human pain disorders. v 18. Mombaerts P. Genes and ligands for odorant, vom- 2007; 85:548-57; PMID:17171702; http://dx.doi. Trends Neurosci 2007; 30:555-63; PMID:17950472; eronasal and taste receptors. Nat Rev Neurosci org/10.1002/jnr.21155. http://dx.doi.org/10.1016/j.tins.2007.08.004. 2004; 5:263-78; PMID:15034552; http://dx.doi. 27. Leung CT, Coulombe PA, Reed RR. Contribution 9. Krafte DS, Bannon AW. Sodium channels and nocicep- org/10.1038/nrn1365. of olfactory neural stem cells to tissue maintenance tion: recent concepts and therapeutic opportunities. 19. Ramón y Cajal S. Histologie du systeme nerveux de and regeneration. Nat Neurosci 2007; 10:720-6; Curr Opin Pharmacol 2008; 8:50-6; PMID:17964852; l’homme et des vertébré. Paris: Maloine 1911. PMID:17468753; http://dx.doi.org/10.1038/nn1882. http://dx.doi.org/10.1016/j.coph.2007.09.007.

www.landesbioscience.com Channels 109 28. Jung S, Aliberti J, Graemmel P, Sunshine MJ, 33. Vukovic J, Blomster LV, Chinnery HR, Weninger W, 36. Zsiros E, Kis-Toth K, Hajdu P, Gaspar R, Bielanska J, Kreutzberg GW, Sher A, et al. Analysis of fractalkine Jung S, McMenamin PG, et al. Bone marrow chimeric Felipe A, et al. Developmental switch of the expression

receptor CX(3)CR1 function by targeted deletion and mice reveal a role for CX3CR1 in maintenance of of ion channels in human dendritic cells. J Immunol green fluorescent protein reporter gene insertion. Mol the monocyte-derived cell population in the olfac- 2009; 183:4483-92; PMID:19748986; http://dx.doi. Cell Biol 2000; 20:4106-14; PMID:10805752; http:// tory neuroepithelium. J Leukoc Biol 2010; 88:645- org/10.4049/jimmunol.0803003. dx.doi.org/10.1128/MCB.20.11.4106-14.2000. 54; PMID:20610801; http://dx.doi.org/10.1189/ 37. Bergman U, Ostergren A, Gustafson AL, Brittebo 29. Trombley PQ, Westbrook GL. Voltage-gated currents jlb.0410194. B. Differential effects of olfactory toxicants on olfac- in identified rat olfactory receptor neurons. J Neurosci 34. Blomster LV, Vukovic J, Hendrickx DA, Jung S, Harvey tory regeneration. Arch Toxicol 2002; 76:104-12;

1991; 11:435-44; PMID:1992011. AR, Filgueira L, et al. CX3CR1 deficiency exacerbates PMID:11914780; http://dx.doi.org/10.1007/s00204- 30. Clare JJ. Targeting voltage-gated sodium chan- neuronal loss and impairs early regenerative responses 002-0321-2. nels for pain therapy. Expert Opin Investig Drugs in the target-ablated olfactory epithelium. Mol Cell 2010; 19:45-62; PMID:20001554; http://dx.doi. Neurosci 2011; 48:236-45; PMID:21871566; http:// org/10.1517/13543780903435340. dx.doi.org/10.1016/j.mcn.2011.08.004. 31. Fieni F, Ghiaroni V, Tirindelli R, Pietra P, Bigiani A. 35. Suzuki Y, Schafer J, Farbman AI. Phagocytic cells in Apical and basal neurones isolated from the mouse the rat olfactory epithelium after bulbectomy. Exp vomeronasal organ differ for voltage-dependent cur- Neurol 1995; 136:225-33; PMID:7498412; http:// rents. J Physiol 2003; 552:425-36; PMID:14561826; dx.doi.org/10.1006/exnr.1995.1099. http://dx.doi.org/10.1113/jphysiol.2003.052035. 32. Ukhanov K, Leinders-Zufall T, Zufall F. Patch-clamp analysis of gene-targeted vomeronasal neurons expressing a defined V1r or V2r receptor: ionic mechanisms underlying persistent firing. J Neurophysiol 2007; 98:2357-69; PMID:17715188; http://dx.doi. org/10.1152/jn.00642.2007. © 2012 Landes Bioscience. Do not distribute.

110 Channels Volume 6 Issue 2 CHAPTER 5

Conclusion and future directions Conclusion

The overall aim of this thesis was to develop potent and selective blockers of hNaV1.7 that can be used as leads for the development novel analgesics for treatment of chronic pain, and to understand the impact that such drugs might have on olfaction. I used an assay- guided fractionation approach to identify several NaV1.7 modulators, including the novel peptide µ-TRTX-Pe1b and the well-known Nav1.7 modulator ProTx1. A recombinant expression system was successfully employed to produce higher quantities of µ-TRTX- Pe1b necessary for SAR studies. ProTx1 was synthetically produced and the structure was determined in collaboration with several groups and individuals, which facilitated the SAR work.

The work presented in this thesis has shown that ProTx1 is a promiscuous molecule that is not suitable as an analgesic drug lead in its native form. However, the ProTx1-G32A-NH2 mutant I constructed has significant promise as an analgesic drug lead. The Pe1b-S35 mutant I engineered has a high level of NaV1.7 inhibitory activity and selectivity over other

NaV channels. Further mutations of this peptide may provide insights into the NaV1.7 pharmacophore of Pe1b and facilitate rational engineering of its subtype selectivity.

The localization of NaV1.7 in the rodent olfactory system was determined using a combination of immunohistochemical and fluorescent localization methods. NaV1.7 was found to be localized on axons of primary olfactory sensory neurons. This indicated that

NaV1.7 is associated with the amplification and propagation of olfactory signals to the olfactory bulb. This finding explains the anosmia and hyposmia observed in patients suffering form CIP. However, as an indirect observation it was noted that monocyte- derived cells of inflammatory infiltrate downregulate NaV1.7 expression following injury.

Although the exact role of NaV1.7 in these monocyte-derived cells are not clear, there is emerging evidence to suggest that NaV1.7 is associated with maturation and migration of certain monocyte-derived cells, and production of pro-inflammatory cytokines. This observation warrants further investigation; it would be worth investigating, for example, whether CIP patients display any abnormal immune responses. Future Directions

79 The structure of Pe1b is currently not available. The 3D structure of Pe1b would be of immense use in future SAR studies as more mutational data becomes available. In addition, testing of Pe1b against the panel of NaV1.7/KV chimeras described in Chapter 2 will enable identification of the voltage sensor domains (VSDs) that Pe1b interacts with. Construction of point mutations in the cognate VSD for Pe1b binding would then enable identification of specific NaV1.7 residues that mediate binding of Pe1b. Knowledge of the interaction sites on both Pe1b and NaV1.7, as well as having access to the 3D structure of Pe1b and a homology model of the cognate VSD, would enable the initiation of docking studies to construct a testable model of the Pe1b-NaV1.7 complex.

The activity of Pe1b-S35 has not yet been tested against potential off-target channels such as hERG and KV and CaV channels. Future work should include testing of Pe1b against a wide range of pharmacologically relevant targets, including other voltage gated ion channels. Testing of the analgesic efficacy of Pe1b in rodent models of chronic pain could follow if the off-target profile of Pe1b turns out to be good. If not, additional engineering will be required to remove off-target activity that might cause undesirable side effects.

This research has also shown that although some peptides belong to the same NaSpTx family, and share significant sequence similarity, there exists sufficient variation within these sequences to achieve a high level of subtype selectivity. The flip side of this, however, is that one has to be very careful when engineering these peptides for improved potency to ensure that the introduced modifications do not alter subtype selectivity, which is clearly governed by very subtle differences in binding to each VSD.

Although NaV1.7 is a very promising analgesic target, the potential side effects associated with therapeutic inhibition of NaV1.7 are not fully understood. For example, it is unclear whether patients with chronic pain would suffer from induced anosmia if they took NaV1.7 inhibitors over a long period of time. Peptide blockers of NaV1.7 that were delivered orally or by subcutaneous injection might never reach the olfactory neurons in sufficient quantity to affect olfaction, but this issue remains to be examined. Clearly, every effort should be made to fully understand the role of NaV1.7 in the human body, which might reveal the best routes of administration, or localized treatment methods, that would minimize undesirable side effects.

REFERENCES

1 Priest, B. T. Future potential and status of selective sodium channel blockers for the treatment of pain. Curr. Opin. Drug Discov. Devel. 12, 682–692 (2009).

2 Costigan, M., Scholz, J. & Woolf, C. J. Neuropathic Pain: A Maladaptive Response of the Nervous System to Damage. Annual Review of Neuroscience 32, 1-32, (2009).

3 Golan, D. E. Principles of pharmacology: the pathophysiologic basis of drug therapy. (Lippincott Williams & Wilkins, 2008).

4 Kuypers, H. G. J. M. Anatomy of the Descending Pathways. (John Wiley & Sons, Inc., 2010).

5 Woolf, C. J. & Ma, Q. F. Nociceptors-noxious stimulus detectors. Neuron 55, 353- 364, (2007).

6 Dhaka, A., Viswanath, V. & Patapoutian, A. TRP ion channels and temperature sensation. Annual Review of Neuroscience 29, 135-161, (2006).

7 Juhl, G. I., Jensen, T. S., Norholt, S. E. & Svensson, P. Central sensitization phenomena after third molar surgery: A quantitative sensory testing study. European Journal of Pain 12, 116-127, (2008).

8 Binshtok, A. M. et al. Nociceptors Are Interleukin-1 beta Sensors. Journal of Neuroscience 28, 14062-14073, (2008).

9 Fields, H. L., Rowbotham, M. & Baron, R. Postherpetic Neuralgia: Irritable Nociceptors and Deafferentation. Neurobiology of Disease 5, 209-227 (1998).

10 Wolfe, M. D., O'Connor, A.B. Management of Neuropathic Pain in Hospitalized Patients. Hospital Medicine Clinics 2, 16 (2013).

11 Manack, A. N., Buse, D. C., Serrano, D., Turkel, C. & Lipton, R. B. Lost Productive Time and Cost Due to Headache in Chronic Migraine and Episodic Migraine: Results from the American Migraine Prevalence and Prevention (AMPP) Study. Headache 50, S34-S34 (2010).

12 Manchikanti, L. et al. Comprehensive Evidence-Based Guidelines for Interventional Techniques in the Management of Chronic Spinal Pain. Pain Physician 12, 699-802 (2009).

13 Sternberg, T. L. Pain Management: A Practical Guide for Clinicians. JAMA: The Journal of the American Medical Association 288, 2616-2617, (2002).

14 Freye, E. & Latasch, L. Development of opioid tolerance - Molecular mechanisms and clinical consequences. Anasthesiologie Intensivmedizin Notfallmedizin Schmerztherapie 38, 14 (2003).

15 Blake, A. D. et al. Molecular regulation of opioid receptors. Receptors & Channels 5, 231-235 (1997).

16 Altier, C. & Zamponi, G. W. Targeting Ca2+ channels to treat pain: T-type versus N- type. Trends in Pharmacological Sciences 25, 465-470, (2004).

17 Cox, J. J. et al. An SCN9A channelopathy causes congenital inability to experience pain. Nature 444, 894–898, (2006).

18 Cummins, T. R., Sheets, P. L. & Waxman, S. G. The roles of sodium channels in nociception: Implications for mechanisms of pain. Pain 131, 243-257, (2007).

19 Delmas, P. Snapshot: Ion Channels and Pain. Cell 134, 2, (2008).

20 Dib-Hajj, S. D. et al. Voltage-gated sodium channels in pain states: Role in pathophysiology and targets for treatment. Brain Research Reviews 60, 65-83, (2009).

21 Gold, M. S. & Gebhart, G. F. Nociceptor sensitization in pain pathogenesis. Nature Medicine 16, 1248-1257, (2010).

22 Gould, H. J. et al. Ibuprofen blocks changes in Na-v 1.7 and 1.8 sodium channels associated with complete Freund's adjuvant-induced inflammation in rat. Journal of Pain 5, 270-280, (2004).

23 Porreca, F. et al. A comparison of the potential role,of the tetrodotoxin-insensitive sodium channels, PN3/SNS and NaN/SNS2, in rat models of chronic pain. Proceedings of the National Academy of Sciences of the United States of America 96, 7640-7644 (1999).

24 Amir, R., Argoff, C. E., Bennett, G. J. & Cummins, T. R. The role of sodium channels in chronic inflammatory and neuropathic pain. Journal of Pain 7, S1-S29, (2006).

25 Wood, J. N., Boorman, J. P., Okuse, K. & Baker, M. D. Voltage-gated sodium channels and pain pathways. Journal of Neurobiology 61, 55-71, (2004).

26 Dray, A. Neuropathic pain: emerging treatments. British Journal of Anaesthesia 101, 48-58, (2008).

27 Shao, P. P. et al. Discovery of a novel class of isoxazoline voltage gated sodium channel blockers. Bioorganic & Medicinal Chemistry Letters 19, 5329-5333, (2009).

28 McGowan, E., Hoyt, S. B., Li, X., Lyons, K. A. & Abbadie, C. A peripherally acting NaV1.7 sodium channel blocker reverses hyperalgesia and allodynia on rat models of inflammatory and neuropathic pain. Anesth. Analg. 109, 951–958, (2009).

29 Catterall, W. A., Goldin, A. L. & Waxman, S. G. International Union of Pharmacology. XLVII. Nomenclature and structure-function relationships of voltage- gated sodium channels. Pharmacological Reviews 57, 397-409, (2005).

30 Catterall, W. A. Structural biology - A 3D view of sodium channels. Nature 409, 988- 991 (2001).

31 Yu, F. H. & Catterall, W. A. Overview of the voltage-gated sodium channel family. Genome Biology 4 (2003).

32 Klint, J. K. et al. Spider-venom peptides that target voltage-gated sodium channels: Pharmacological tools and potential therapeutic leads. Toxicon 60, 478-491, (2012).

33 King, G. F., Escoubas, P. & Nicholson, G. M. Peptide toxins that selectively target insect NaV and CaV channels. Channels 2, 100–116 (2008).

34 Payandeh, J., El-Din, T. M. G., Scheuer, T., Zheng, N. & Catterall, W. A. Crystal structure of a voltage-gated sodium channel in two potentially inactivated states. Nature 486, 135-U166, (2012).

35 Pavlov E, B. C., Winkfein R, Diao C, Dhaliwal P, French RJ. The pore, not cytoplasmic domains, underlies inactivation in a prokaryotic sodium channel. Biophysical Journal 232 (2005).

36 Zhang, X. et al. Crystal structure of an orthologue of the NaChBac voltage-gated sodium channel. Nature 486, 130-U160, (2012).

37 McCusker, E. C. et al. Structure of a bacterial voltage-gated sodium channel pore reveals mechanisms of opening and closing. Nature Communications 3, (2012).

38 Cannon, S. C. & Bean, B. P. Sodium channels gone wild: resurgent current from neuronal and muscle channelopathies. Journal of Clinical Investigation 120, 80-83, (2010).

39 Schaller, K. L. & Caldwell, J. H. Expression and distribution of voltage-gated sodium channels in the cerebellum. Cerebellum 2, 2-9, (2003).

40 Spampanato, J., Escayg, A., Meisler, M. H. & Goldin, A. L. Functional effects of two voltage-gated sodium channel mutations that cause generalized epilepsy with febrile seizures plus type 2. Journal of Neuroscience 21, 7481-7490 (2001).

41 Nabbout, R. et al. Spectrum of SCN1A mutations in severe myoclonic epilepsy of infancy. Neurology 60, 1961-1967 (2003).

42 Kaplan, M. R. et al. Differential control of clustering of the sodium channels NaV1.2 and NaV1.6 at developing CNS nodes of ranvier. Neuron 30, 105-119, (2001).

43 Weiss, L. A., Escayg,A., Kearney,J.A., Trudeau,M., MacDonald,B.T., Mori, M., Reichert, J., Buxbaum, J.D., Meisler, M.H. Sodium channels SCN1A, SCN2A and SCN3A in familial autism. Mol. Psychiatry, 9 (2003).

44 Liao, Y., Anttonen, A-K., Liukkonen, E., Gaily, E., Maljevic, S., Schubert, S., Bellan- Koch, A., Petrou, S., Ahonen,V.E., Lerche, H., Lehesjoki, A-E. SCN2A mutation associated with neonatal epilepsy, late-onset episodic ataxia, myoclonus, and pain. Neurology 75, 5 (2010).

45 Ho, C. & O'Leary, M. E. Single-cell analysis of sodium channel expression in dorsal root ganglion neurons. Molecular and Cellular Neuroscience 46, 159-166, (2011).

46 Jurkat-Rott, K. & Lehmann-Horn, F. Paroxysmal muscle weakness - the familial periodic paralyses. Journal of Neurology 253, 1391-1398, (2006).

47 Rogart, R. B., Cribbs, L. L., Muglia, L. K., Kephart, D. D. & Kaiser, M. W. Molecular- cloning of a putative tetrodotoxin-resistant rat-heart Na+ channel isoform. Proceedings of the National Academy of Sciences of the United States of America 86, 8170-8174, (1989).

48 Kallen, R. G. et al. Primary structure and expression of a sodium-channel characteristic of denervated and immature rat skeletal-muscle. Neuron 4, 233-242, (1990).

49 Hartmann, H. A., Colom, L. V., Sutherland, M. L. & Noebels, J. L. Selective localization of cardiac SCN5A sodium channels in limbic regions of rat brain. Nature Neuroscience 2, 593-595 (1999).

50 Carrithers, M. D. et al. Expression of the voltage-gated sodium channel NaV1.5 in the macrophage late endosome regulates endosomal acidification. Journal of Immunology 178, 7822-7832 (2007).

51 Carrithers, L. M., Hulseberg, P., Sandor, M. & Carrithers, M. D. The human macrophage sodium channel NaV1.5 regulates mycobacteria processing through organelle polarization and localized calcium oscillations. Fems Immunology and Medical Microbiology 63, 319-327, (2011).

52 Abriel, H. Roles and regulation of the cardiac sodium channel NaV1.5: Recent insights from experimental studies. Cardiovascular Research 76, 381-389, (2007).

53 Scornik, F. S. et al. Functional expression of "cardiac-type" NaV1.5 sodium channel in canine intracardiac ganglia. Heart Rhythm 3, 842-850, (2006).

54 Caldwell, J. H., Schaller, K. L., Lasher, R. S., Peles, E. & Levinson, S. R. Sodium channel NaV1.6 is localized at nodes of Ranvier, dendrites, and synapses. Proceedings of the National Academy of Sciences of the United States of America 97, 5616-5620, (2000).

55 Rush, A. M., Dib-Hajj, S. D. & Waxman, S. G. Electrophysiological properties of two axonal sodium channels, NaV1.2 and Nav1.6, expressed in mouse spinal sensory neurones. Journal of Physiology-London 564, 803-815, (2005).

56 Kohrman, D. C., Smith, M. R., Goldin, A. L., Harris, J. & Meisler, M. H. A missense mutation in the sodium channel Scn8a is responsible for cerebellar ataxia in the mouse mutant jolting. Journal of Neuroscience 16, 5993-5999 (1996).

57 Akopian, A. N. et al. The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways. Nature Neuroscience 2, 541-548 (1999).

58 Sivilotti, L., Okuse, K., Akopian, A. N., Moss, S. & Wood, J. N. A single serine residue confers tetrodotoxin insensitivity on the rat sensory-neuron-specific sodium channel SNS. Febs Letters 409, 49-52, (1997).

59 Coste, B., Crest, M. & Delmas, P. Pharmacological dissection and distribution of 2+ NaN/NaV1.9, T-type Ca currents, and mechanically activated cation currents in different populations of DRG neurons. Journal of General Physiology 129, 57-77, (2007).

60 Eijkelkamp, N. et al. Neurological perspectives on voltage-gated sodium channels. Brain 135, 2585-2612, (2012).

61 Dearborn, G. V. A case of congenital general pure analgesia. Journal of Nervous and Mental Disease 75, 612-615 (1932).

62 Nagasako, E. M., Oaklander, A. L. & Dworkin, R. H. Congenital insensitivity to pain: an update. Pain 101, 213-219, (2003).

63 Goldberg, Y. P. et al. Loss-of-function mutations in the Nav1.7 gene underlie congenital indifference to pain in multiple human populations. Clin. Genet. 71, 311– 319, (2007).

64 Waxman, S. G. & Dib-Hajj, S. Erythermalgia: molecular basis for an inherited pain syndrome. Trends Mol. Med. 11, 555–562, (2005).

65 Rush, A. M. et al. A single sodium channel mutation produces hyperor hypoexcitability in different types of neurons. Proceedings of the National Academy of Sciences of the United States of America 103, (2006).

66 Jarecki, B. W., Sheets, P. L., Jackson, J. O. & Cummins, T. R. Paroxysmal extreme

pain disorder mutations within the D3/S4-S5 linker of Nav1.7 cause moderate destabilization of fast inactivation. Journal of Physiology-London 586, 4137-4153, (2008).

67 Reimann, F. et al. Pain perception is altered by a nucleotide polymorphism in SCN9A. Proc. Natl. Acad. Sci. USA 107, 5148–5153, (2010).

68 Lewis, R. J. & Garcia, M. L. Therapeutic potential of venom peptides. Nature Reviews Drug Discovery 2, 790-802, (2003).

69 Grishin, E. Polypeptide neurotoxins from spider venoms. European Journal of Biochemistry 264, 276-280 (1999).

70 Nicholson, G. M. & Little, M. J. Spider neurotoxins targeting voltage-gated sodium channels. Toxin Reviews 24, 315-345, (2005).

71 Bosmans, F. & Swartz, K. J. Targeting voltage sensors in sodium channels with spider toxins. Trends in Pharmacological Sciences 31, 175-182, (2010).

72 Platnick, N. I., Schwendinger, P. J. & Steiner, H. Three new species of the spider genus Liphistius (Araneae, Mesothelae) from Malaysia. American Museum Novitates 3209, 1-13 (1997).

73 Coddington, J. A. & Levi, H. W. in Johnston, R. F. Annual Review of Ecology and Systematics 565-592 (1991).

74 King, G. F. The wonderful world of spiders: Preface to the special Toxicon issue on spider venoms. Toxicon 43, 471-475, (2004).

75 Escoubas, P., Sollod, B. & King, G. F. Venom landscapes: Mining the complexity of spider venoms via a combined cDNA and mass spectrometric approach. Toxicon 47, 650-663, (2006).

76 Wood, D. L. A. et al. ArachnoServer: a database of protein toxins from spiders. BMC Genomics 10, 1-8 (2009).

77 Herzig, V. et al. ArachnoServer 2.0, an updated online resource for spider toxin sequences and structures. Nucleic Acids Research 39, D653-D657, (2011).

78 Soderlund, D. M. & Knipple, D. C. Actions of insecticides on sodium channels. Multiple target sites and site-specific resistance. (1995).

79 Bloomquist, J. R., Kinne, L. P., Deutsch, V. & Simpson, S. F. Mode of action of an insecticidal peptide toxin from the venom of a weaving spider (Diguetia canities). Toxicon 34, 1072-1075, (1996).

80 Zlotkin, E. The insect voltage-gated sodium channel as target of insecticides. Annual Review of Entomology 44, 429-455, (1999).

81 Maggio, F., Sollod, B. L., Tedford, H. W. & King, G. F. Spider toxins and their potential for insect control. (2010).

82 Windley, M. J. et al. Spider-Venom Peptides as Bioinsecticides. Toxins 4, 191-227, (2012).

83 Escoubas, P. & Bosmans, F. Spider peptide toxins as leads for drug development. Expert Opinion on Drug Discovery 2, 823-835, (2007).

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

85 Saez, N. J. et al. Spider-Venom Peptides as Therapeutics. Toxins 2, 2851-2871, (2010).

86 Ekberg, J., Craik, D. J. & Adams, D. J. Conotoxin modulation of voltage-gated sodium channels. International Journal of Biochemistry & Cell Biology 40, 2363- 2368, (2008).

87 Nicholson, G. M., Walsh, R., Little, M. J. & Tyler, M. I. Characterisation of the effects of robustoxin, the lethal neurotoxin from the Sydney funnel-web spider Atrax robustus, on sodium channel activation and inactivation. Pflugers Archiv-European Journal of Physiology 436, 117-126, (1998).

88 Grolleau, F. et al. Electropysiological analysis of the neurotoxic action of a funnel- web spider toxin, delta-Atracotoxin-Hv1a, on insect voltage-gated Na+ channels. Journal of Experimental Biology 204, 711-721 (2001).

89 Milescu, M., Lee, H. C., Bae, C. H., Kim, J. I. & Swartz, K. J. Opening the Shaker K+ channel with hanatoxin. Journal of General Physiology 141, 203-216, (2013).

90 King, G. F., Gentz, M. C., Escoubas, P. & Nicholson, G. M. A rational nomenclature for naming peptide toxins from spiders and other venomous animals. Toxicon 52, 264-276, (2008).

91 Rodriguez de la Vega, R. C. & Possani, L. D. Overview of scorpion toxins specific for Na+ channels and related peptides: biodiversity, structure-function relationships

and evolution. Toxicon : official journal of the International Society on Toxinology 46, 831-844, (2005).

92 Bosmans, F. et al. Four novel tarantula toxins as selective modulators of voltage- gated sodium channel subtypes. Molecular Pharmacology 69, 419-429, (2006).

93 Xiao, X. C. & Liang, S. P. Inhibition of neuronal tetrodotoxin-sensitive Na+ channels by two spider toxins: hainantoxin-III and hainantoxin-IV. European Journal of Pharmacology 477, 1-7, (2003).

94 Xiao, Y. & Liang, S. Inhibition of neuronal tetrodotoxin-sensitive Na+ channels by two spider toxins: Hainantoxin-III and hainantoxin-IV. European Journal of Pharmacology 477, 1-7, (2003).

95 Middleton, R. E. et al. Two tarantula peptides inhibit activation of multiple sodium channels. Biochemistry 41, (2002).

96 Chen, J. et al. Molecular diversity and evolution of cystine knot toxins of the tarantula Chilobrachys jingzhao. Cellular and Molecular Life Sciences 65, 2431- 2444, (2008).

97 Redaelli, E. et al. Target promiscuity and heterogeneous effects of tarantula venom peptides affecting Na+ and K+ ion channels (vol 285, pg 4130, 2010). Journal of Biological Chemistry 285, 13314-13314, (2010).

98 Zeng, X. et al. Isolation and characterization of Jingzhaotoxin-V, a novel neurotoxin from the venom of the spider Chilobrachys jingzhao. Toxicon 49, 388-399, (2007).

99 Corzo, G. et al. Distinct primary structures of the major peptide toxins from the venom of the spider Macrothele gigas that bind to sites 3 and 4 in the sodium channel. Febs Letters 547, 43-50, (2003).

100 Xiao, Y. C. et al. Jingzhaotoxin-I, a novel spider neurotoxin preferentially inhibiting cardiac sodium channel inactivation. Journal of Biological Chemistry 280, 12069- 12076, (2005).

101 Diniz, C. R. et al. The purification and amino-acid-sequence of the lethal neurotoxin tx1 from the venom of the brazilian armed spider phoneutria-nigriventer. Febs Letters 263, 251-253, (1990).

102 Billen, B., Bosmans, F. & Tytgat, J. Animal peptides targeting voltage-activated sodium channels. Current Pharmaceutical Design 14, 2492-2502, doi:10.2174/138161208785777423 (2008).

103 Priest, B. T., Blumenthal, K. M., Smith, J. J., Warren, V. A. & Smith, M. M. ProTx-1 and ProTx-II: Gating modifiers of voltage-gated sodium channels. Toxicon 49, 194- 201, (2007).

104 Liao, Z. et al. Solution structure and functional characterization of Jingzhaotoxin-XI: A novel gating modifier of both potassium and sodium channels. Biochemistry 45, 15591-15600, (2006).

105 Li-Smerin, Y. & Swartz, K. J. Gating modifier toxins reveal a conserved structural motif in voltage-gated Ca2+ and K+ channels. Proceedings of the National Academy of Sciences of the United States of America 95, 8585-8589, (1998).

106 Chen, J. et al. Expression and characterization of jingzhaotoxin-34, a novel neurotoxin from the venom of the tarantula Chilobrachys jingzhao. Peptides 30, 1042-1048, (2009).

107 King, G. F. & Hardy, M. C. in Annual Review of Entomology, Vol 58 Vol. 58 Annual Review of Entomology (ed M. R. Berenbaum) 475-496 (2013).

108 Ohkubo, T., Yamazaki, J. & Kitamura, K. Tarantula Toxin ProTx-I Differentiates 2+ Between Human T-type Voltage-Gated Ca Channels Ca(v)3.1 and CaV3.2. Journal of Pharmacological Sciences 112, 452-458, (2010).

109 Frech, G. C., VanDongen, A. M., Schuster, G., Brown, A. M. & Joho, R. H. A novel potassium channel with delayed rectifier properties isolated from rat brain by expression cloning. Nature 340, 4 (1989).

110 Swartz, K. J., MacKinnon, R. Hanatoxin modifies the gating of a voltage-dependent K+ channel through multiple binding sites. Neuron 18, 8 (1997).

111 Jarecki, B. W., Piekarz, A. D., Jackson, J. O., Cummins, T. R. Human voltage-gated sodium channel mutations that cause inherited neuronal and muscle channelopathies increase resurgent sodium currents. Journal of Clinical Investigation 120, 369-378, (2010).

112 Garcia, M. L., Garcia-Calvo, M., Hidalgo, P., Lee, A. & MacKinnon, R. Purification and characterization of three inhibitors of voltage-dependent K+ channels from Leiurus quinquestriatus var. hebraeus venom. Biochemistry 33, 6 (1994).

113 Phillips, L. R., Milescu, M., Li-Smerin, Y., Mindell, J.A., Kim, J.I., Swartz, K.J. Voltage-sensor activation with a tarantula toxin as cargo. Nature 436, 4 (2005).

114 Bosmans, F., Martin-Eauclaire, M. F. & Swartz, K. J. Deconstructing voltage sensor function and pharmacology in sodium channels. Nature 456, 202-U228, (2008).

115 Bosmans, F., Puopolo, M., Martin-Eauclaire, M. F., Bean, B. P. & Swartz, K. J. . Functional properties and toxin pharmacology of a dorsal root ganglion sodium channel viewed through its voltage sensors. J. Gen. Physiol. 138, 14 (2011).

116 Gui, J., Liu, B., Cao, G., Lipchik, A.M., Perez, M., Dekan, Z., Mobli, M., Daly,N.L., Alewood,P.F., Parker, L.L., King, G.F., Zhou, Y.Z., Jordt, S-E., Nitabach, M.N., . A Tarantula-Venom Peptide Antagonizes the TRPA1 Nociceptor Ion Channel by Binding to the S1–S4 Gating Domain. Current Biology 24, 11 (2014).

117 Jiang, Y., Lee, A., Chen, J., Ruta, V., Cadene, M., Chait, B.T., MacKinnon, R. X-ray structure of a voltage-dependent K+ channel. Nature 1, 9 (2003).

118 Jiang, Y., Ruta, V., Chen, J., Lee, A., MacKinnon, R. The principle of gating charge movement in a voltage-dependent K+ channel. Nature 423, 7 (2003).

119 Ruta, V., Chen, J., MacKinnon, R. Calibrated measurement of gating-charge + arginine displacement in the KVAP voltage-dependent K channel. Cell, 13 (2005).

120 Alabi, A. A., Bahamonde, M.I., Jung, H.J., Kim, J.I., Swartz, K.J. Portability of paddle motif function and pharmacology in voltage sensors. Nature 430, 6 (2007).

121 Chakrapani, S., Cuello, L.G., Cortes, D.M., Perozo, E. . Structural dynamics of an isolated voltage-sensor domain in a lipid bilayer. Structure 16, 12 (2008).

122 Swartz, K. J., MacKinnon, R. . Mapping the receptor site for hanatoxin, a gating modifier of voltage-dependent K+ channels. Neuron 18, 8 (1997).

123 Li-Smerin, Y., Swartz, K.J. . Gating modifier toxins reveal a conserved structural motif in voltage-gated Ca2+ and K+ channels. Proc. Natl. Acad. Sci. U S A, 5 (1998).

124 Li-Smerin, Y., Swartz, K.J. Localization and molecular determinants of the Hanatoxin receptors on the voltage-sensing domains of a K+ channel. J. Gen. Physiol. 115, 12 (2000).

125 Lee HC, W. J., Swartz KJ. Interaction between extracellular Hanatoxin and the resting conformation of the voltage-sensor paddle in KV channels. Neuron 40, 10 (2003).

126 Lee, S. Y., MacKinnon, R. A membrane-access mechanism of ion channel inhibition by voltage sensor toxins from spider venom. Nature 430, 4 (2004).

127 Milescu, M., Vobecky, J., Roh, S.H., Kim, S.H., Jung, H.J., Kim, J.I., Swartz, K.J. Tarantula toxins interact with voltage sensors within lipid membranes. J. Gen. Physiol. 130, 14 (2007).

128 Takahashi, H., Kim, J.I., Min, H.J., Sato, K., Swartz, K.J., Shimada, I. . Solution structure of hanatoxin1, a gating modifier of voltage-dependent K+ channels: common surface features of gating modifier toxins. . J. Mol. Biol. 297, 8 (2000).

129 Takeuchi, K., Park, E.J.,Lee, C.W., Kim, J. I.,Takahashi, H., Swartz, K. J.,Shimada, I. Solution Structure of ω-Grammotoxin SIA, A Gating Modifier of P/Q and N-type Ca2+ Channel. Journal of Molecular Biology 321, 10 (2002).

130 Wang, J. M. et al. Molecular surface of tarantula toxins interacting with voltage sensors in KV channels. Journal of General Physiology 123, 455-467, (2004).

131 Milescu, M. et al. Tarantula toxins interact with voltage sensors within lipid membranes. Journal of General Physiology 130, 497-511, (2007).

132 Milescu, M. et al. Interactions between lipids and voltage sensor paddles detected with tarantula toxins. Nature Structural & Molecular Biology 16, 1080-U1102, (2009).

133 Wu, Y., Cao, G., Pavlicek, B., Luo, X., and Nitabach, M.N. Phase coupling of a circadian neuropeptide with rest/activity rhythms detected using a membrane- tethered spider toxin. PLoS biology 6 (2008).

134 Escoubas, P. et al. Isolation of a tarantula toxin specific for a class of proton-gated Na+ channels. Journal of Biological Chemistry 275, 25116-25121, (2000).

135 Pineda Gonzalez, S. S. Probing the chemical diversity of venom from the Australian funnel-web spider Hadronyche infensa. Doctor of Philosophy thesis, The University of Queensland, (2012).

136 Abdel-Rahmana, M. A., Quintero-Hernandezb,V., Possanib, L.D. Venom proteomic and venomous glands transcriptomic analysis of the Egyptian scorpion Scorpio maurus palmatus (Arachnida: Scorpionidae). Toxicon 74, 14 (2013).

137 Undheim, E. A. et al. A proteomics and transcriptomics investigation of the venom from the barychelid spider Trittame loki (brush-foot trapdoor). Toxins (Basel) 5, 2488-2503, (2013).

138 Vetter, I. et al. Venomics: a new paradigm for natural products-based drug discovery. Amino Acids 40, 15-28, 9 (2011).

139 Perola, E. An Analysis of the Binding Efficiencies of Drugs and Their Leads in Successful Drug Discovery Programs. Journal of Medicinal Chemistry 53, 2986- 2997, (2010).

140 Klint, J. K. et al. Production of Recombinant Disulfide-Rich Venom Peptides for Structural and Functional Analysis via Expression in the Periplasm of E. coli. Plos One 8, (2013).

141 Fang, L. et al. An improved strategy for high-level production of TEV protease in Escherichia coli and its purification and characterization. Protein Expression and Purification 51, 102-109, (2007).

142 Kwan, A. H., Mobli, M., Gooley, P. R., King, G. F. & Mackay, J. P. Macromolecular NMR spectroscopy for the non-spectroscopist. Febs Journal 278, 687-703, (2011).

143 Cabrita, L. D., Dai, W. W. & Bottomley, S. P. A family of E-coli expression vectors for laboratory scale and high throughput soluble protein production. Bmc Biotechnology 6, (2006).

144 Bassford, P. J. Export of the periplasmic maltose-binding protein of Escherichia coli. Journal of Bioenergetics and Biomembranes 22, 401-439, (1990).

145 Kapust, R. B. & Waugh, D. S. Escherichia coli maltose-binding protein is uncommonly effective at promoting the solubility of polypeptides to which it is fused. Protein Science 8, 1668-1674 (1999).

146 Annis, I., Hargittai, B., Barany, G. Disulfide bond formation in peptides. Methods Enzymology 289, 24 (1997).

90 APPENDIX 1 J.K. Klint et al. / Toxicon 60 (2012) 478–491 479

pore (Fig. 2). During action potential generation, the channel cycles through three states: closed, open, and inactivated. The S4 segments contain a positively charged amino acid (Arg or Lys) at every third position; this segment serves as a ‘voltage sensor’ that initiates voltage- dependent activation (i.e., the transition from the closed to open state) by responding to membrane depolarization and causing the channel to undergo a conformational þ change that allows selective influx of Na ions through the pore. Sodium channel inactivation (i.e., the transition from the open to inactivated state) is mediated by a short intracellular loop termed the “inactivation gate” that connects domains III and IV (Fig. 2). In a major break- through, the three-dimensional structure of a bacterial NaV channel in the closed state was recently determined (Payandeh et al., 2011), which should allow molecular Fig. 1. Ion channel modulators from spider venoms. Molecular targets of modelling of the interaction between NaV channels and the 186 ion channel toxins listed in ArachnoServer (www.arachnoserver. ligands that modulate their activity. org; accessed February 12, 2012) (Herzig et al., 2011). The majority of In vertebrates, the a subunits are classified into nine toxins target voltage-gated ion channels (CaV,NaV,andKV), followed by different subtypes, denoted NaV1.1 to NaV1.9, and they are calcium-activated potassium (KCa) channels, mechanosensitive ion channels (MSC), transient receptor potential (TRP) channels, and acid-sensing further characterised by their sensitivity to tetrodotoxin ion channels (ASICs). Note that some toxins target more than one type of (TTX). NaV1.5, NaV1.8 and NaV1.9 are TTX-resistant, whereas channel, and therefore the cumulative number of toxins in the histogram all other subtypes are TTX-sensitive (Catterall et al., 2005; > is 186. King et al., 2008a). A combination of site-directed mutagenesis, binding studies, and electrophysiological approaches have been used to identify at least seven rapidly incapacitate their prey. Na channel toxins are V distinct ligand binding sites in or around the pore region of found in a taxonomically diverse range of spiders, which the NaV channel a subunit, as shown in Fig. 2. The majority suggests that this pharmacology was recruited at a very of these sites have been identified using venom molecules early stage in venom gland evolution. Moreover, the ubiq- and alkaloids as pharmacological probes (King et al., uity of Na channel toxins in spider venoms suggests that V 2008a). Venom peptides modulate channel activity by these peptides might be useful insecticides for the control two main mechanisms: they either physically occlude the of arthropod pests (Maggio et al., 2010; Windley et al., channel pore or they bind to an allosteric site that induces 2012). a conformational change in the channel that alters the In addition to prey capture, spiders also use their equilibrium between the open, closed, and inactivated venom to deter predators, which can include vertebrates. states (Ekberg et al., 2008). Moreover, even though the vast majority of spiders prey All known NaV channel toxins from spider venoms are primarily on invertebrates, there is no evolutionary allosteric modulators, also known as “gating modifiers”. selection pressure to prevent spider toxins acting on However, they modulate channel activity in quite vertebrate ion channels. It is therefore not surprising that different ways. For example, the lethal toxin from the many spider-venom peptides have been found to modu- Sydney funnel-web spider Atrax robustus (d-HXTX-Ar1a) late the activity of vertebrate NaV channels (Nicholson and inhibits the inactivation of both insect and vertebrate NaV Little, 2005; Escoubas and Bosmans, 2007; King et al., channels (Nicholson et al., 1998; Grolleau et al., 2001), 2008a). whereas the m-agatoxins from unrelated American funnel spiders shift the voltage for channel activation to more 2. NaV channels negative potentials, increasing the probability of channel opening at the resting membrane potential (Adams, In both vertebrates and invertebrates, functional NaV 2004). channels are composed primarily of a large pore-forming a subunit (220–260 kDa) whose gating and kinetics are 3. Spider toxins that target NaV channels modified via association with one of four smaller b subunits (33–36 kDa) in the case of vertebrates or the unrelated TipE The primary aim of this review is to survey the range accessory subunit (65 kDa) in the case of invertebrates of peptide sequences and three-dimensional scaffolds (King et al., 2008a). The a subunit is composed of four that have been evolved by spiders for targeting NaV homologous but non-identical domains (designated I–IV) channels, with an emphasis on those peptides that might connected by intracellular linkers (Fig. 2). Each of have therapeutic potential. The ‘Browse by Molecular these domains contains six transmembrane (TM) segments Targets’ ontology in ArachnoServer (Wood et al., 2009; (S1–S6) joined by intracellular or extracellular loops. Herzig et al., 2011) was used to extract all known spider- Membrane re-entrant loops between TM segments S5 and venom peptides with activity on NaV channels (both S6 dip into the TM region of the protein and form the vertebrate and invertebrate). Homologues of these narrow ion-selectivity filter at the extracellular end of the toxins were found by using the BLAST function within 480 J.K. Klint et al. / Toxicon 60 (2012) 478–491

Fig. 2. Molecular architecture and pharmacology of NaV channels. The pore-forming a subunit of NaV channels is comprised of four homologous domains denoted I–IV connected by intracellular linkers. Each of these domains contains six transmembrane (TM) helical segments (labelled 1–6) joined by intra- or extracellular loops. TM segments 5 and 6 from each domain, along with the intervening membrane re-entrant loops (highlighted with a light green box), come together to form the channel pore and the ion-selectivity filter at the extracellular end of the pore. TM segments 1–4 of each domain form the voltage-sensing units, as indicated for domain I; the highly positively charged TM segment 4 is primarily responsible for sensing changes in membrane polarization. Sodium channel inactivation is mediated by a short stretch of hydrophobic residues (the “inactivation gate”; orange balls) in the intracellular linker connecting domains III and IV. Coloured regions represent neurotoxin receptor sites. The grey circles represent the outer (EEDD) and inner (DEKA) rings of amino acid residues that form the ion-selectivity filter and constitute the proposed neurotoxin receptor site 1 for the water-soluble guanidinium toxins TTX and saxitoxin. Some m-conotoxin binding sites overlap with those of TTX and are omitted for clarity. Nevertheless, we refer to sites 1a and 1b to distinguish between the different molecular determinants for m-conotoxin and TTX/saxitoxin binding. In the case of receptor sites 3 and 4, only epitopes where mutagenesis of key residues leads to more than a five-fold decrease in toxin binding affinity are highlighted. Note that TM segment 6 in domain I includes binding epitopes for both site 2 and site 5 toxins.

The known binding sites for 8 of the 12 families of spider-venom NaV toxins (NaSpTxs) are indicated. “F” denotes toxin family (e.g., F2 indicates NaSpTx Family 2). While the d-amaurobitoxins in Family 10 have been shown to bind to neurotoxin receptor site 4, peptides in this family have variable pharmacology and some may bind at other sites. Adapted from Catterall et al. (2007) and King et al. (2008a).

ArachnoServer. CLC sequence viewer 6 (CLC bio, 3.1. NaSpTx family 1 Denmark) was used to align the peptides and generate families of related peptides based on the level of All peptides in Family 1 were isolated from the venom sequence identity and intercysteine spacings. This resul- of tarantulas (Theraphosidae), and they consist of 33–35 ted in 12 different families of spider-venom peptides that amino acid residues with three disulfide bridges that target NaV channels, as outlined in Figs. 3 and 4.The form an inhibitor cystine knot (ICK) motif, defined as an following sections describe the activity data that is antiparallel b sheet stabilised by a cystine knot (Pallaghy currently available for each family. These sections employ et al., 1994). In spider toxins, the b sheet typically the rational nomenclature recently developed for spider- comprises only two b strands although a third N-terminal venom peptides (King et al., 2008b) in order to facilitate strand is sometimes present (King et al., 2002). The identification of orthologs and paralogs, but for conve- cystine knot itself comprises a ring formed by two disul- nience we also provide original names of toxins. The term fides and the intervening sections of the peptide back- NaScTx is employed for scorpion toxins that target NaV bone, with a third disulfide piercing the ring to create channels (Rodriguez de la Vega and Possani, 2005), and a pseudo-knot (Saez et al., 2010). This “knot” converts ICK here we introduce the term NaSpTx for spider toxins that peptides into hyperstable mini-proteins with tremendous target this channel. chemical, thermal, and biological stability. ICK toxins are J.K. Klint et al. / Toxicon 60 (2012) 478–491 481

Fig. 3. NaSpTx families 1–4, 7 and 10. The consensus sequence is shown above each alignment, with the disulfide bond connectivity indicated in blue. Red asterisks denote C-terminal amidation. Strictly conserved residues are highlighted in bold colour while semi-conserved residues are highlighted in a lighter colour. Toxin names are based on the rational nomenclature devised for spider-venom peptides (King et al., 2008b), as used in ArachnoServer and UniProt (King et al., 2008b). The structure of a representative member from each NaSpTx family is shown adjacent to each alignment: Family 1, m-TRTX-Hh2a (PDB ID: 1MB6); Family 2, k-TRTX-Gr1a (PDB ID: 1D1H); Family 3, k-TRTX-Gr2a (PDB ID: 1LUP); Family 4, d-HXTX-Hv1a (PDB ID: 1VTX); Family 7, k-TRTX-Gr3a (PDB ID: 1S6X); Family 10, d-AMATX-Pl1b (PDB ID: 1V91). b strands are depicted in cyan, helices are red, and disulfide bridges are yellow. All spider photos were taken by Bastian Rast. typically resistant to extremes of pH, organic solvents, One member of Family 1 (m/u-TRTX-Hh1a; huwentoxin- high temperatures, and proteases (Saez et al., 2010). Of the 1) was found to be analgesic in a rat model of formalin- 10 families of NaSpTxs whose disulfide connectivity has induced pain when administered intrathecally (Chen been determined or which can be confidently predicted et al., 2005), although in another study the same peptide based on homology (see Figs. 3 and 4), nine contain an ICK was reported to be lethal to mice when delivered via the motif. intraperitoneal or intracisternal route (Liang et al., 1993). 482 J.K. Klint et al. / Toxicon 60 (2012) 478–491

Fig. 4. NaSpTx families 5, 6, 8, 9, 11 and 12. The consensus sequence is shown above each alignment, with the disulfide bond connectivity indicated in blue. Dotted lines represent predicted disulfide bond connectivities that have not been experimentally validated. Red asterisks denote C-terminal amidation. Strictly conserved residues are highlighted in bold colour while semi-conserved residues are highlighted in a lighter colour. Toxin names are based on the rational nomenclature devised for spider-venom peptides (King et al., 2008b), as used in ArachnoServer and UniProt. The structure of a representative member of Family 5 is shown adjacent to the alignment (b-HXTX-Mg1a; PDB ID: 2GX1). Disulfide bridges are coloured yellow. No structures are available for any member of the other NaSpTx families depicted in this figure. Spider photos are by Bastian Rast (M. gigas and P. reidyi) and Aloys Staudt (H. melloteei). J.K. Klint et al. / Toxicon 60 (2012) 478–491 483

Intraperitoneal administration of m-TRTX-Hhn1b (hai- with an IC50 of 51 nM (Middleton et al., 2002; Priest et al., nantoxin-IV) is also lethal to mice with an LD50 of 0.2 mg/kg 2007). (Liu et al., 2003), as is intracerebroventricular (i.c.v.) k-TRTX-Cj1a (Jingzhaotoxin-11) is a weak modulator administration of b-TRTX-Ps1a (phrixotoxin-3), b-TRTX- of NaV currents in cardiac myocytes (i.e., mainly TTX-R Cm1a (ceratotoxin-1), and b-TRTX-Cm1b (ceratotoxin-1) NaV1.5), causing a significant reduction in peak NaV (Bosmans et al., 2006). current amplitude and a slowing of NaV current inactiva- Despite the high level of residue conservation tion (IC50 ¼ 1280 nM) (Liao et al., 2006). It also inhibits throughout this family, the selectivity of the peptides for KV2.1 with moderate potency (IC50 ¼ 740 nM), with even the various NaV channel subtypes varies dramatically. For weaker block of KV4.1 and KV4.2 and no effect on KV1.1, example, b-TRTX-Ps1a potently blocks NaV1.5 with an IC50 KV1.2, KV1.3, KV1.4 and KV3.1 (Liao et al., 2006). A peptide of 72 nM (Bosmans et al., 2006) whereas m-TRTX-Hh2a with identical sequence has been isolated from the venom (huwentoxin-4), which is 62% identical to b-TRTX-Ps1a, of the related theraphosid spider Plesiophrictus guangxien- does not have any effect on NaV1.5 at concentrations up to sis. It has been proposed that k-TRTX-Cj1a binds to the 10 mM(Xiao et al., 2008). Based on homology with other extracellular S3–S4 linker region of channel domain IV (i.e., toxins, peptides in Family 1 are presumed to function as neurotoxin receptor site 3; see Fig. 2)(Liao et al., 2006), gating modifiers (Bosmans et al., 2006). Consistent with which is a common target of gating modifier toxins such as this hypothesis, analysis of the binding of m-TRTX-Hh2a to k-TRTX-Gr1a (Hanatoxin-1) in the same family (Li-Smerin mutants of NaV1.4 revealed that the peptide interacts with and Swartz, 1998). the S3–S4 linker of channel domain II (i.e., neurotoxin b-TRTX-Cm2a (Ceratotoxin-3) has moderate affinity receptor site 4; see Fig. 2), with the blocking effect achieved towards NaV1.5 (IC50 ¼ 447 nM), and at a concentration of by trapping the domain II voltage sensor in the closed 2 mM it weakly blocks NaV1.8 (45% reduction in current) but configuration (Xiao et al., 2008). In contrast, m-TRTX-Hhn1b has no activity against NaV1.1–1.4 (Bosmans et al., 2006). In was shown to have no effect on the activation or inactiva- striking contrast with k-TRTX-Cj1a and b-TRTX-Cm2a, tion kinetics of TTX-S NaV currents in rat dorsal root m-TRTX-Cj1a (Jingzhaotoxin-34) is a potent blocker of TTX-S ganglion (DRG) neurons and therefore it was hypothesised NaV channels in rat dorsal root ganglion (DRG) neurons to function not as a gating modifier but as a pore blocker (IC50 ¼ 85 nM) but it has no effect on TTX-R NaV currents by binding at neurotoxin receptor site 1 (Li et al., 2004). (Chen et al., 2009). Thus, residues that are conserved in However, this would be a very unusual pharmacology for k-TRTX-Cj1a and b-TRTX-Cm2a, but which are absent from NaSpTxs and it remains to be confirmed by site-directed m-TRTX-Cj1a, should provide some insight into residues mutagenesis and competitive binding studies. The func- that are important for interaction of this family of peptides tionally important residues on m-TRTX-Hhn1b have been with NaV1.5. proposed to be K27 (which is conserved in most Family 1 Six toxins from Family 2 have confirmed activity on peptides) and R29 (which is poorly conserved) as muta- voltage-gated calcium (CaV) channels, with 12 toxins tions of these residues to alanine showed reduced activity active against KV channels. Nevertheless, the high level (Li et al., 2004). of sequence identity within this family and the potency of Very little work has been done to examine the activity of the known Family 2 NaV modulators described above Family 1 peptides against insect NaV channels, but the indicates that a thorough functional analysis of all toxins in limited data available indicates that the insecticidal activity this family could reveal many more toxins acting on NaV of these peptides is likely to be highly variable. m-TRTX- channels. However, from a therapeutic perspective, the Hhn1b has no effect when injected into cockroaches (Liu promiscuous activity of several family members on et al., 2003), whereas m/u-TRTX-Hh1a potently inhibits NaV multiple types of voltage-gated channels indicates that it currents in cockroach dorsal unpaired median (DUM) will be critical to carefully determine the selectivity of any neurons (IC50 ¼ 4.8 nM) (Wang et al., 2012). Falling between Family 2 peptides that are being considered as therapeutic these two extremes is m-TRTX-Hhn2b (hainantoxin-I), which leads (Chen et al., 2008). only weakly inhibits insect NaV channels heterologously expressed in Xenopus oocytes (IC50 ¼ 4.3 mM) (Li et al., 2003). 3.3. NaSpTx family 3

This family consists of 14 reported toxins with highly 3.2. NaSpTx family 2 conserved sequences in the C-terminal region. k-TRTX- Ps1b (Phrixotoxin-2) and k-TRTX-Gr2a (GsMTx-2) are Family 2 is the largest family of NaSpTxs with 34 identical peptides identified from different genera of members, all originating from theraphosid venoms. Of tarantulas (Diochot et al., 1999; Oswald et al., 2002); for these, 18 peptides were isolated from the Chinese taran- simplicity, only k-TRTX-Ps1b is given in the alignment tula Chilobrachys jingzhao. Family 2 peptides range in size (Fig. 3). Family 3 peptides are 29–32 residues long, with from 33 to 41 residues and they each contain three three disulfide bonds forming an ICK motif. These peptides disulfide bonds that form an ICK motif. Currently, there is act as gating modifiers by binding to the voltage sensor experimental proof of NaV channel activity for only four of region (S3–S4 linker) of channel domain II (i.e., neurotoxin the 34 peptides in this family. However, they are very receptor site 4; see Fig. 2), thus impeding channel activa- interesting from a therapeutic perspective as one member tion and causing the channels to remain in a resting state of this family, b/u-TRTX-Tp1a (protoxin I), is one of the (Sokolov et al., 2007). b/u-TRTX-Tp2a (protoxin II) and b- most potent blockers of human NaV1.7 reported to date, TRTX-Gr1b (GsAF I) are both selective against the NaV1.7 484 J.K. Klint et al. / Toxicon 60 (2012) 478–491 subtype, an important analgesic target, but their potency DUM neurons (EC50 w820 pM) (Yamaji et al., 2009). differs by >100-fold despite w90% sequence identity Activity on vertebrate NaV subtypes was much more (Priest et al., 2007; Redaelli et al., 2010). b/u-TRTX-Tp2a is modest, perhaps explaining why envenomation by Macro- the most potent blocker of human NaV1.7 reported to date, thele gigas does not appear to be a serious medical problem. with an IC50 of 0.3 nM (Schmalhofer et al., 2008). It also At a concentration of 5 mM, d-HXTX-Mg1a had a significant inhibits activation of human CaV3.1 currents but at much effect on inactivation of rat NaV1.1 and mouse NaV1.6, higher doses (IC50 w1 mM) (Edgerton et al., 2010). b/u- modest effects on rat NaV1.2, NaV1.3 and NaV1.7, and no TRTX-Tp2a is lethal to rats when administered intrave- effect on rat NaV1.4, rat NaV1.5 and human NaV1.8. In nously at doses 1.0 mg/kg or intrathecally at doses contrast with d-HXTX-Mg1a, Family 4 toxins from lethal 0.1 mg/kg (Schmalhofer et al., 2008). This is perhaps not Australian funnel-web spiders, such as d-HXTX-Hv1a, are surprising as even though b/u-TRTX-Tp2a preferentially more potent on vertebrate TTX-sensitive NaV channels than blocks NaV1.7, it also potently inhibits key peripheral NaV insect NaV channels (Nicholson et al., 1996). channel subtypes (e.g., NaV1.5, IC50 ¼ 79 nM) as well as central NaV channel subtypes (e.g., NaV1.2, IC50 ¼ 41 nM) 3.5. NaSpTx family 5 (Schmalhofer et al., 2008). b-TRTX-Gr1a (GrTx1) is 96% identical to b-TRTX-Gr1b, This family of NaSpTxs comprises only two peptides with the only difference being a Gln at position 22 instead isolated from closely related hexathelid spiders, both of the of Arg. However, this single amino acid change causes a vast genus Macrothele. These peptides contain 28–29 residues difference in selectivity. b-TRTX-Gr1a has similar potency with three disulfide bonds forming an ICK motif. b-HXTX- towards NaV1.1, NaV1.2, NaV1.3, NaV1.6, and NaV1.7 (IC50 Mg1a (Magi-5) and b-HXTX-Mr1a (raventoxin III), from range of 230–770 nM) whereas b-TRTX-Gr1b is >5-fold different species of Macrothele, have identical sequences more selective towards NaV1.7 (Redaelli et al., 2010). (Zeng et al., 2003) so only b-HXTX-Mg1a is presented in the Significant variations in NaV channel subtype selectivity alignment. These peptides are lethal to rodents, but their due to small sequence variations has been noted previously insecticidal activity is variable. Intra-thoracic injection of in scorpion-venom peptides (Weinberger et al., 2010). U12-HXTX-Mg1a (Magi-11) in crickets (Gryllus bimaculatus) induced slow-onset paralysis leading to death, while 3.4. NaSpTx family 4 intracranial injection in mice resulted in lacrimation, sweating, and urination followed by paralysis of breathing Of the seven peptides in this family, six were isolated and death 15 min post-injection. b-HXTX-Mg1a/b-HXTX- from the venom of Australian funnel-web spiders (family Mr1a caused reversible paralysis when injected into Hexathelidae), whereas the seventh is found in non-related caterpillars (Spodoptera litura) and had no effect when Australian mouse spiders (family Actinopodidae). These injected into cockroaches (Periplaneta americana), whereas peptides comprise 42–44 residues with four disulfide intraperitoneal injection in mice resulted in excitation, bonds, three of them arranged in an ICK motif. They are gasping, spastic paralysis, and exophthalmos (bulging of characterised by an unusual sequence of three consecutive the eyes) followed by death (Satake et al., 2004). b-HXTX- Cys residues in the centre of the primary structure and Mg1a binds to site 4 on vertebrate NaV channels (see Fig. 2) a disulfide bond that connects the N- and C-terminal Cys and shifts steady-state activation to more hyperpolarised residues. All peptides in this family are lethal to vertebrates, voltages, whereas it has been reported to bind to site 3 on including d-HXTX-Ar1a (robustoxin) from the Sydney insect channels (Corzo et al., 2003, 2007; Zeng et al., 2003). funnel-web spider A. robustus, which was responsible for at least 14 human fatalities prior to the introduction of an 3.6. NaSpTx family 6 effective antivenom in 1980 (Fletcher et al., 1997). Family 4 peptides bind to neurotoxin receptor site 3 (see Fig. 2) and All peptides in this family were isolated from the venom slow inactivation of both vertebrate TTX-sensitive NaV of spiders in the family Ctenidae. Four were isolated from channels and insect para-type NaV channels (Gunning et al., the highly venomous Brazilian armed spider Phoneutria 2003; Nicholson et al., 2004). Symptoms in primates nigriventer, one was isolated from the venom of the closely include immediate disturbances in respiration, blood related Phoneutria reidyi, and one was identified from pressure and heart rate, followed by severe hypotension a venom gland transcriptome of Phoneutria keyserlingi. The and death due to respiratory and circulatory failure within final member of the family was isolated from the venom of a few hours (Mylecharane et al., 1989). With the exception a related ctenid, the South American fishing spider (Ancy- of d-HXTX-Hv1b (Szeto et al., 2000), all Family 4 peptides lometes sp.). These toxins comprise 47–54 residues and exhibit insecticidal activity, causing paralysis and/or contain five disulfide bonds whose connectivity has not lethality in orthopterans, lepidopterans and dipterans been experimentally determined. Of those that have been (Little et al., 1998; Grolleau et al., 2001). tested in vertebrates, all display lethality even at doses as The activity of d-HXTX-Mg1a (Magi-4), the only low as 7.5 pmol/g (Cordeiro et al.,1992). d-CNTX-Pn2a (Tx2- member of this family not isolated from an Australian 6) has a complex pharmacology that results in inhibition of spider, was recently tested against the full range of rodent NaV channel inactivation and a hyperpolarizing shift in the NaV subtypes expressed in oocytes (Yamaji et al., 2009). d- potential required for channel activation (Matavel et al., HXTX-Mg1a was found to be a potent blocker of the 2002). Due to the high sequence identity (80–89%) Drosophila para/TipE NaV channel (IC50 w23 nM), and it between members of this family, they are all assumed to potently slowed NaV channel inactivation in cockroach modulate NaV channels in this manner. An interesting J.K. Klint et al. / Toxicon 60 (2012) 478–491 485 supplementary pharmacology has been observed for caused excitation and spastic paralysis in mice upon i.c.v. d-CNTX-Pn2a, which originated from the observation that administration at a dose of 1.5 mg/mouse, followed by death males bitten by this spider experience a diverse variety of within 5–8 min (Richardson et al., 2006). symptoms including priapism. Although i.c.v. injection of d- Preliminary studies suggested that m-CNTX-Pn1a might CNTX-Pn2a into rats is lethal, rats with depressed erectile target vertebrate CaV channels (dos Santos et al., 1999, function can be normalised by subcutaneous administra- 2002) but more recent studies have shown that it is tion of d-CNTX-Pn2a (Nunes et al., 2008). A minimum dose a state-dependent blocker of rat NaV1.2 that binds in the of only 0.006 mg/kg was required to cause an erection in vicinity of channel site 1 (Martin-Moutot et al., 2006). The mice when injected directly into the corpus cavernosum high level of sequence identity suggests that all members of and, at this dose, no local and systemic toxic effects were this family will have the same pharmacology. observed (Andrade et al., 2008). Thus, d-CNTX-Pn2a has been touted as a lead molecule for the development of new 3.9. NaSpTx family 9 therapeutics for treatment of erectile dysfunction that might have fewer side-effects than current drugs. This family consists of only two paralogous peptides Only limited information is available regarding the isolated from the venom of the crab spider Heriaeus mel- insecticidal activity of members of this family. d-CNTX- loteei. They comprise 37 and 40 residues, with three Pn2a and -Pn2b (Tx2-5) were reported to be insecticidal to disulfide bonds that are predicted to form an ICK motif. both the larval and adult forms of the housefly Musca m-TMTX-Hme1a (Hm-1) has similar potency against NaV1.2, domestica (Cordeiro et al., 1992), however, no LD50 values NaV1.4, NaV1.5, and NaV1.6 (IC50 150–500 nM) but no effect have been reported. against NaV1.8 (Billen et al., 2008). m-TMTX-Hme1b (Hm-2) has similar pharmacology, except that it has very little 3.7. NaSpTx family 7 activity against NaV1.5 (Billen et al., 2008). The insecticidal activity of these toxins has not been examined. All peptides in this family were isolated from the venom of theraphosid spiders. They each contain 33–36 residues 3.10. NaSpTx family 10 with three disulfide bonds that form an ICK motif. k-TRTX- Pg1b (guangxitoxin-1) and k-TRTX-Cj3a (jingzhaotoxin-4), This family comprises venom peptides from the closely isolated from different genera, are identical (Herrington related spider families Amaurobiidae and Agelenidae. All et al., 2006; Chen et al., 2008), so only the sequence of members of Family 10 are 36–38 residues long and contain k-TRTX-Pg1b is presented in the alignment. These peptides four disulfide bonds, three of which form an ICK motif. appear to have promiscuous activity on both KV and NaV Almost all members of family 10 are C-terminally amidated. channels. Although several family members were isolated The m-agatoxins (m-AGTXs) in Family 10 shift the voltage- as KV channel blockers, these peptides have not been tested activation curve to more negative potentials, increasing on NaV channels. The only two family members that have the probability of NaV channel opening at the resting been directly tested on NaV channels displayed significant membrane potential. The resultant membrane depolariza- activity. Both b/k-TRTX-Cj1a (jingzhaotoxin-3) and d-TRTX- tion leads to calcium entry into nerve terminals and Cj1a (jingzhaotoxin-1) inhibit activation of NaV1.5 with an spontaneous release of neurotransmitter into the neuro- IC50 of 350 nM and 32 nM, respectively (Xiao et al., 2005; muscular junction, as well as repetitive action potentials in Rong et al., 2011), It was recently shown via mutational motor neurons (Skinner et al., 1989; Adams, 2004). The studies that b/k-TRTX-Cj1a binds to the S3–S4 loop of d-amaurobitoxins, previously known as d-palutoxins domain II (i.e., neurotoxin receptor site 4; see Fig. 2)in (Corzo et al., 2000), have a quite different mode of action to NaV1.5 (Rong et al., 2011). the m-agatoxins. They bind to site 4 on insect NaV channels Little is known about the insecticidal activity of these (Ferrat et al., 2005) and inhibit channel inactivation, but toxins. Intra-abdominal injection of m-TRTX-Hh1a (Huwen- they have no effect on rat NaV1.2 channels at concentra- toxin-3) into cockroaches induces reversible paralysis that tions up to 10 mM(Ferrat et al., 2005). They are the only lasts for several hours. The effects of other family members spider toxins that have been definitively shown to bind to on insects or heterologously expressed insect NaV channels site 4 on insect NaV channels but which modulate NaV have not been examined. channel inactivation (King et al., 2008a). The only toxic effect in vertebrates reported for any 3.8. NaSpTx family 8 member of Family 10 was a temporary paralysis (lasting not longer than 15 min) of the left legs of a mouse that received This is a small family comprising only three peptides an i.c.v. injection of 486 pmol of d-AMATX-Pl1b (d-palutoxin derived from two species of ctenid spiders. These peptides IT2) (Corzo et al., 2000). However, a wide range of insecti- are easily the longest NaSpTxs, with 76–77 residues and cidal activities have been reported for Family 10 peptides. seven disulfide bonds. They are similar in size to scorpion The d-amaurobitoxins are weakly insecticidal, with LD50 a and b toxins that also target NaV channels (Rodriguez de values of 2347–11029 pmol/g when injected into larvae of la Vega and Possani, 2005), but they have no sequence the moth Spodoptera litura (Corzo et al., 2000). The m-AGTXs homology and contain a larger number of disulfide bonds. from Agelenopsis aperta induced irreversible spastic paral- These toxins are lethal to rodents. m-CNTX-Pn1a (Tx1; ysis when injected into range of insects, with m-AGTX-Aa1d PhTx1) from P. nigriventer has an LD50 of 5.5 pmol/g in mice (m-agatoxin IV) being particularly potent in the housefly M. (Diniz et al., 1990, 2006), while U5-CNTX-Pk1b (PKTx1B) domestica with an LD50 of 30 pmol/g (Skinner et al., 1989). 486 J.K. Klint et al. / Toxicon 60 (2012) 478–491

The m-AGTXs from Hololena curta caused rapid irreversible identify orthologous and paralogous relationships. While flaccid paralysis when injected into house crickets (Acheta the close relationship between toxin classes and spider domestica)(Stapleton et al., 1990). families superficially suggests that spiders recruited particular NaV toxin scaffolds at specific stages during the 3.11. NaSpTx family 11 course of spider evolution, it seems more likely that most of these scaffolds are ancient recruitments and that deep Family 11 comprises a group of three paralogous phylogenetic relationships are currently obscured by the insecticidal peptides isolated from the primitive weaving relatively poor taxonomic sampling of spider venoms. spider Diguetia canities. These toxins are 56–59 residues There are currently w42,700 characterised species of long with four disulfide bonds whose configuration has not spiders, yet venom peptides have been isolated from less been experimentally determined. The most potent of the than 100. three peptides, m-diguetoxin-Dc1a (DTX9.2), causes rapid paralysis of lepidopteran larvae with a PD50 of 380 pmol/g, 4.2. Comparison with NaV channel toxins from other but it is not toxic to mice at doses up to 657 pmol/g venomous animals (Krapcho et al., 1995). Neurophysiological experiments indicate that these toxins are highly specific insect NaV Only two related families of scorpion-venom peptides channel activators that induce hyperexcitability in sensory have been described that target NaV channels; they have insect nerves and repetitive firing in skeletal muscles been separated based on their pharmacological properties (Bloomquist et al., 1996). The fact that the peptides are not and are denoted the a and b families (Gordon et al., 2007). toxic to mice suggests that they are unlikely to have potent These so-called long-chain toxins comprise 58–76 residues effects on vertebrate NaV channels. and they are typically stabilised by four disulfide bonds with a 1–8, 2–5, 3–6, 4–7 connectivity (Rodriguez de la 3.12. NaSpTx family 12 Vega and Possani, 2005). These toxins have a babb topology that is unrelated to any known spider-toxin This family consists of three paralogous insecticidal scaffolds. Like Family 4 NaSpTxs, scorpion a toxins peptides isolated from the hexathelid spider M. gigas. They interact with NaV channel site 3 and delay channel inacti- comprise 38–40 residues with three disulfide bonds that vation, whereas scorpion b toxins bind to site 4 of the are predicted to form an ICK motif. These peptides bind channel and shift the threshold for channel activation to specifically to site 3 on insect NaV channels, but they have more negative potentials. no effect on NaV channels from rat brain synaptosomes and Three classes of NaV channel toxins (Types I–III) have are not toxic to mice when injected i.c.v. at a dose of been isolated from sea anemones. They all have a similar 20 pmol/g (Corzo et al., 2003). pharmacological profile to scorpion a toxins and Family 4 NaSpTxs in that they delay channel inactivation by binding 4. Discussion to site 3 (Bosmans and Tytgat, 2007). However, they have a distinct structural scaffold compared to both NaScTxs and 4.1. Recruitment of NaV toxins into spider venoms NaSpTxs (Norton, 1991). Type 1 and Type II toxins typically comprise 46–49 residues with six cysteine residues Spiders appear to have evolved a significantly larger arranged into three disulfide bridges, whereas Type III repertoire of NaV channel toxins than any other venomous toxins are much shorter, typically comprising 27–32 resi- animal. Since insects only express a single NaV channel dues, but also with three disulfide bonds (Norton, 1991; subtype that plays a crucial role in action potential gener- Honma and Shiomi, 2006; Bosmans and Tytgat, 2007; ation (King et al., 2008a), this channel is an excellent target Moran et al., 2009). In none of the three classes of sea for insect predators wishing to rapidly incapacitate their anemone NaV toxins do the three disulfide bonds form an prey. Thus, it is not surprising that spiders seem to have ICK motif, and none of the 3D scaffolds resemble those of recruited NaV channel toxins into their venoms at a very NaScTxs or NaSpTxs. early stage of venom evolution. Three major classes of NaV channel toxins have been With only a few exceptions, the members within each of isolated from the venoms of marine cone snails: the m-, d-, the 12 different classes of NaSpTxs outlined in this review and mO-conotoxins. The m-conotoxins isolated from fish- were restricted to certain taxonomic families. For example, hunting cone snails are small peptides (w20 residues) all peptides in Families 1 and 2 were isolated from spiders with three disulfide bonds that are often selective for the of the family Theraphosidae, whereas all peptides in Family skeletal muscle NaV1.4 subtype (Ekberg et al., 2008; Wilson 6 originated from spiders in the Ctenidae family. The only et al., 2011). They do not structurally resemble NaV channel exceptions are Family 10, with toxins from the closely toxins from spiders, scorpions or sea anemones. Moreover, related spider families Amaurobiidae and Agelenidae, and in striking contrast with these latter toxins, the m-cono- Family 4, with toxins from the spider families Hexathelidae toxins are pore blockers rather than gating modifiers that and Actinopodidae, which are not closely related but which compete with TTX and saxitoxin for binding to neurotoxin belong both to the suborder Mygalomorphae. This corre- receptor site 1 (see Fig. 2)(Ekberg et al., 2008). lation between NaV channel toxin classes and spider Despite their vastly different primary structures and families can be easily visualised (see Figs. 3 and 4) using the dissimilar modes of action, the d- and mO-conotoxins are rational nomenclature developed for spider-venom evolutionary related members of the O-superfamily with peptides (King et al., 2008b), which makes it easy to similar 3D scaffolds (Heinemann and Leipold, 2007). They J.K. Klint et al. / Toxicon 60 (2012) 478–491 487 comprise 29–31 residues with three conserved disulfide Bosmans, 2007; Fox and Serrano, 2007; Escoubas and bonds that form an ICK motif. Hence, they are structurally King, 2009; Vetter et al., 2011). There are now six FDA- related to those families of NaSpTxs that also contain an ICK approved drugs derived from venom peptides or proteins, motif. The d-conotoxins delay channel inactivation, result- with a further ten in clinical trials and many more in ing in prolongation of action potentials, and hence they various stages of preclinical development (King, 2011). functionally resemble scorpion a toxins and Family 4 Thus, the vast number of NaV channel modulators isolated NaSpTxs. However, unlike these arachnid toxins that bind from spider venoms begs the questions of whether any of to channel site 3, the d-conotoxins are the defining phar- these peptides might be useful therapeutically. macology for neurotoxin receptor site 6 on vertebrate NaV NaV channels are the target of several classes of thera- channels (see Fig. 2). In contrast to the d-conotoxins, peutics such as local anaesthetics (e.g., lidocaine and mO-conotoxins compete with b scorpion toxins for binding ropivacaine) and antiarrhythmics (e.g., mexiletine and fle- to neurotoxin receptor site 4 and they appear to block cainide), as well as anticonvulsants used to treat epilepsy or sodium flow by hindering activation of the voltage sensor bipolar disorder (e.g., lamotrigine, carbamazepine, and in channel domain II, thereby preventing channel opening phenytoin) (Clare et al., 2000; England and de Groot, 2009). (Leipold et al., 2007). Moreover, while most tricyclic antidepressants (e.g., In summary, while several classes of NaSpTxs func- amitriptyline and imipramine) are serotonin–norepineph- tionally resemble NaV channel toxins from other venomous rine reuptake inhibitors, they also inhibit NaV channels animals, their only structural relatives are the d- and (Nicholson et al., 2002; Dick et al., 2007). However, currently, mO-contoxins from the venom of marine cone snails. It is the major pharmaceutical interest in NaV channels stems remarkable that two such taxonomically diverse animals, from their involvement in pain. Indeed, many of the NaV one confined to marine environments and the other channel drugs described above are now used for treating a terrestrial predator, have evolved such similar molecular various types of pain, particularly post-operative and scaffolds for targeting the NaV channels of their prey. neuropathic pain. Normal pain is a key adaptive response that serves to 4.3. NaSpTxs as insecticides and pharmacological tools limit our exposure to potentially damaging or life- threatening events. In contrast, aberrant long-lasting pain The 12 families of NaSpTxs described here present transforms this adaptive response into a debilitating and a range of different pharmacologies with respect to their often poorly managed disease. About 20% of adults suffer binding site on NaV channels and their modulation of from chronic pain, with the incidence increasing to 50% for channel function. For example, some toxin classes delay those older than 65 (Brennan et al., 2007). There are few channel inactivation (e.g., Family 4) while others alter the drugs available for treating chronic pain and most have voltage-dependence of channel activation (e.g., Family 10). limited efficacy, dose-limiting side-effects, and in some Consequently many NaSpTxs have become valuable tools cases the potential for addiction (Carinci and Mao, 2010). for probing the pharmacology and biophysics of NaV Hence, there is much interest in developing novel pain channels (Bosmans and Swartz, 2010). medications that have minimal or no side-effects and that Some NaSpTxs are also promising leads for bio- do not have problems of tolerance and addiction. insecticide development due to their potency against insect Each of the nine NaV channel subtypes has a different pests and minimal activity, if any, in vertebrates (Maggio anatomical distribution that determines their biological et al., 2010; Windley et al., 2012). Many of the peptides in function. For example, NaV1.5 is restricted to the heart Families 10–12, for example, display the type of phyletic where it is critical for the rising phase of the cardiac action selectivity that is desirable for an insecticide. In contrast, potential whereas NaV1.4 is restricted to muscle tissue and many of the peptides in Families 3–6 display a high degree is critical for muscle contraction (England and de Groot, of vertebrate lethality and therefore they would not be 2009; England and Rawson, 2010). The other seven suitable as either pharmacological or insecticidal agents. subtypes are found in the central or peripheral nervous Because venom-derived NaV channel modulators are system. NaV1.7, NaV1.8 and NaV1.9 are selectively expressed genetically encoded, transgenes encoding the peptides can in peripheral sensory neurons where they appear to play be incorporated into crop plants and entomopathogens in a key role in relaying pain signals to the spinal cord (Dib- order to deliver them to targeted insect pests (Tedford et al., Hajj et al., 2010). Of these subtypes, NaV1.7 has emerged 2004; Windley et al., 2012). For example, the insecticidal as an exciting analgesic target based on several remarkable potency of the Autographa californica nuclear polyhedrosis genetic studies. Gain of function mutations in the SNC9A baculovirus was significantly improved by engineering it to gene encoding the pore-forming a subunit of NaV1.7 have express a transgene encoding m-AGTX-Aa1d (m-Aga IV) been shown to cause painful inherited neuropathies from NaSpTx Family 10 (Prikhod’ko et al., 1996). (primary erythromelalgia and paroxysmal extreme pain disorder) (Yang et al., 2004; Fertleman et al., 2006; Estacion 4.4. NaSpTxs as therapeutic leads et al., 2008). In contrast, loss of function mutations in SCN9A result in a congenital indifference to all forms of pain The extremely high stability, specificity, and potency of (Cox et al., 2006; Nilsen et al., 2009). Remarkably, apart venom peptides, which in the case of spiders is the result of from their complete inability to sense pain, loss of smell hundreds of millions of years of evolutionary fine tuning, (anosmia) is the only other sensory impairment in indi- have made them a valuable source of natural products for viduals with this channelopathy (Nilsen et al., 2009; Weiss drug discovery (Lewis and Garcia, 2003; Escoubas and et al., 2011; Rupasinghe et al., 2012). 488 J.K. Klint et al. / Toxicon 60 (2012) 478–491

It is probable that the known analgesic effects of to the clinic, is still rudimentary. Indeed, it is remarkable a number of non-specificNaV channel blockers such as local that we only know the complete NaV selectivity profile for anaesthetics, anticonvulsants, and tricyclic antidepressants a handful of peptides in the 12 families of NaSpTxs. are at least in part mediated through their effects on NaV1.7 Table 1 summarises some of the salient features of each (Nicholson et al., 2002; Dick et al., 2007; Levinson et al., in NaSpTx family, including an indication of their subtype press). However, the non-specific block of NaV channels by preferences. Notably, peptides that potently block NaV1.5 these drugs results in a very narrow therapeutic window, tend to be lethal in rodents, consistent with the critical role with CNS-related side-effects such as dizziness, headaches, of this subtype in cardiac function. This highlights the need and ataxia, as well as potential cardiac side-effects to pinpoint the residues on these peptides that mediate (Cummins and Rush, 2007; Krafte and Bannon, 2008). their selectivity and potency in order to be able to rationally Thus, there is much interest in the development of engineer improved versions that preferentially bind NaV1.7. subtype-selective blockers of NaV1.7 (England and Rawson, This will be the major challenge in progressing these 2010; Zuliani et al., 2010). In this regard, there are several peptides to the clinic. On a positive note, however, peptides reasons why NaSpTxs might be useful therapeutic leads for have already been isolated from Families 1–3 that potently the development of NaV1.7-selective analgesics (Clare, block NaV1.7 (IC50 51 nM) and that have at least some 2010; England and Rawson, 2010; Saez et al., 2010). First, selectivity over other NaV subtypes. Moreover, it is worth most NaSpTxs contain an ICK motif that provides them noting that parenterally administered NaSpTxs are unlikely with tremendous chemical, thermal, and biological to cross the blood–brain barrier and enter the CNS, and stability. Spider-venom ICK peptides are typically stable in therefore off-target activity at centrally located NaV human serum for several days and have half-lives in gastric subtypes is unlikely to be a therapeutic problem. fluid of >12 h (Saez et al., 2010). Notably, an ICK scaffold In conclusion, it is clear that spiders have evolved forms the core of the analgesic drug PrialtÔ (Miljanich, a larger repertoire of NaV channel toxins than any other 2004). Second, because NaSpTxs are gating modifiers venomous animal. These NaSpTxs modulate NaV channels rather then pore blockers they are more likely to be in a variety of different ways ranging from effects on the subtype-selective than small compounds that bind in the voltage-dependence of activation to the kinetics of channel highly conserved pore of the channel. Nevertheless, our inactivation. Many of these peptides have become useful understanding of the molecular basis of the potency and pharmacological tools for studying the pharmacology and selectivity of NaSpTxs, which is critical for their progression biophysics of NaV channels, while others are being

Table 1 Subtype selectivity and other selected properties of the 12 families of NaSpTxs.

NaSpTx NaV channel subtype TTX-S TTX-R Lethal Other ion High-resolution family channel targets structure?p 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 a b b b b c d 1 407 0.6 42 288 72 – 26 – 1.1 – Yes CaV/KV Yes e f f g 2 –– ––447 – 51 27 85 – No CaV/KV/TRPV1 Yes h i h j h h h h k k 341230 103 330 79 26 0.3 146 30.2 27.6 Yes KV,CaV Yes 4 –––––––––– Yes – Yes 5 – 1200l –––––––– Yes – Yes 6 –––––––––– Yes – No m m m 7 –– ––32 130 130 –– – – KV Yes 8 – 0.08n –––––––– Yes – No o 9 –– –155 –––––– – KV No 10 –––––––––– No – Yes 11 –––––––––– No – No 12 –––––––––– No – No

For each NaSpTx family, the IC50 value (in nM) is given for the peptide that most potently modulates the activity of each NaV subtype. Columns labelled “TTX- S” and “TTX-R” indicate potency of peptides on TTX-S and TTX-R NaV channels, usually tested in rat dorsal root ganglion neurons. The column labelled “Lethal” refers to the toxicity of peptides in rodents. a b-TRTX-Cm1b (Bosmans et al., 2006). b b-TRTX-Ps1a (Bosmans et al., 2006). c m-TRTX-Hh2a (Xiao and Liang, 2003b). d m-TRTX-Hhn2b (Xiao and Liang, 2003a). e b-TRTX-Cm2a (Bosmans et al., 2006). f b/u-TRTX-Tp1a (Middleton et al., 2002). g m-TRTX-Cj1a (Chen et al., 2008). h b/u-TRTX-Tp2a (Middleton et al., 2002). i b-TRTX-Gr1a (Redaelli et al., 2010). j b-TRTX-Gr1b (Redaelli et al., 2010). k k-TRTX-Cj2a (Zeng et al., 2007). l b-HXTX-Mg1a (Corzo et al., 2003). m d-TRTX-Cj1a (Xiao et al., 2005). n m-CNTX-Pn1a (Diniz et al., 1990). o m-TMTX-Hme1b (Billen et al., 2008). p We define “high-resolution structures” as NMR-derived solution structures with a backbone RMSD <0.5 Å. J.K. Klint et al. / Toxicon 60 (2012) 478–491 489 developed as bioinsecticides due to their selective action on Clare, J.J., Tate, S.N., Nobbs, M., Romanos, M.A., 2000. Voltage-gated invertebrate Na channels. In addition, many NaSpTxs are sodium channels as therapeutic targets. Drug. Discov. Today. 5, V 506–520. vertebrate-active and those with high potency and selec- Coddington, J.A., Levi, H.W., 1991. Systematics and evolution of spiders tivity for certain NaV subtypes that are localised in (Araneae). Annu. Rev. Ecol. Syst. 22, 565–592. peripheral pain pathways are being used as leads to Cordeiro, M.N., Diniz, C.R., Valentim, A.C., von Eickstedt, V.R., Gilroy, J., Richardson, M., 1992. The purification and amino acid sequences of develop novel classes of analgesics. Finally, it is worth four Tx2 neurotoxins from the venom of the Brazilian ‘armed’ spider noting that the 12 classes of NaSpTxs described here have Phoneutria nigriventer (Keys). FEBS. Lett. 310, 153–156. been derived from only eight of the 110 extant families of Corzo, G., Escoubas, P., Stankiewicz, M., Pelhate, M., Kristensen, C.P., Nakajima, T., 2000. Isolation, synthesis and pharmacological charac- spiders (Platnick, 1997). Thus, it is likely that additional terization of d-palutoxins IT, novel insecticidal toxins from the classes of NaSpTxs will be discovered through better spider Paracoelotes luctuosus (Amaurobiidae). Eur. J. Biochem. 267, taxonomic sampling of the incredible diversity of spider 5783–5795. Corzo, G., Gilles, N., Satake, H., Villegas, E., Dai, L., Nakajima, T., Haupt, J., venoms available. 2003. Distinct primary structures of the major peptide toxins from the venom of the spider Macrothele gigas that bind to sites 3 and 4 in the sodium channel. FEBS. Lett. 547, 43–50. Acknowledgements Corzo, G., Sabo, J.K., Bosmans, F., Billen, B., Villegas, E., Tytgat, J., Norton, R.S., 2007. Solution structure and alanine scan of a spider toxin that affects the activation of mammalian voltage-gated sodium channels. J. Biol. We acknowledge financial support from the Australian Chem. 282, 4643–4652. Research Council (Discovery Grants DP110103129 and Cox, J.J., Reimann, F., Nicholas, A.K., Thornton, G., Roberts, E., Springell, K., DP1095728) and the National Health & Medical Research Karbani, G., Jafri, H., Mannan, J., Raashid, Y., Al-Gazali, L., Hamamy, H., Valente, E.M., Gorman, S., Williams, R., McHale, D.P., Wood, J.N., Council (Project Grant APP1034958). Gribble, F.M., Woods, C.G., 2006. An SCN9A channelopathy causes congenital inability to experience pain. Nature 444, 894–898. fl Cummins, T.R., Rush, A.M., 2007. Voltage-gated sodium channel blockers Con ict of interest for the treatment of neuropathic pain. Expert. Rev. Neurother. 7, 1597–1612. The authors have no competing interest. Dib-Hajj, S.D., Cummins, T.R., Black, J.A., Waxman, S.G., 2010. Sodium channels in normal and pathological pain. Annu. Rev. Neurosci. 33, 325–347. Dick, I.E., Brochu, R.M., Purohit, Y., Kaczorowski, G.J., Martin, W.J., Priest, B.T., References 2007. Sodium channel blockade may contribute to the analgesic efficacy of antidepressants. J. Pain. 8, 315–324. Adams, M.E., 2004. Agatoxins: ion channel specific toxins from the Diniz, C.R., Cordeiro, M.D., Junor, L.R., Kelly, P., Fischer, S., Reimann, F., American funnel web spider, Agelenopsis aperta. Toxicon 43, 509–525. Oliveira, E.B., Richardson, M., 1990. The purification and amino-acid- Andrade, E., Villanova, F., Borra, P., Leite, K., Troncone, L., Cortez, I., sequence of the lethal neurotoxin TX1 from the venom of the Bra- Messina, L., Paranhos, M., Claro, J., Srougi, M., 2008. Penile erection zilian armed spider Phoneutria nigriventer. FEBS. Lett. 263, 251–253. induced in vivo by a purified toxin from the Brazilian spider Pho- Diniz, M.R.V., Theakston, R.D.G., Crampton, J.M., Cordeiro, M.N., neutria nigriventer. BJU. Int. 102, 835–837. Pimenta, A.M.C., De Lima, M.E., Diniz, C.R., 2006. Functional expres- Billen, B., Bosmans, F., Tytgat, J., 2008. Animal peptides targeting voltage- sion and purification of recombinant Tx1, a sodium channel blocker activated sodium channels. Curr. Pharm. Des. 14, 2492–2502. neurotoxin from the venom of the Brazilian “armed” spider, Bloomquist, J.E., Kinne, L.P., Deutsch, V., Simpson, S.F., 1996. Mode of Phoneutria nigriventer. Protein. Expr. Purif. 50, 18–24. action of an insecticidal peptide toxin from the veom of a weaving Diochot, S., Drici, M.D., Moinier, D., Fink, M., Lazdunski, M., 1999. Effects of spider (Diguetia canities). Toxicon 34, 1072–1075. phrixotoxins on the Kv4 family of potassium channels and implica- Bosmans, F., Swartz, K.J., 2010. Targeting voltage sensors in sodium tions for the role of Ito1 in cardiac electrogenesis. Br. J. Pharmacol. channels with spider toxins. Trends. Pharmacol. Sci. 31, 175–182. 126, 251–263. Bosmans, F., Tytgat, J., 2007. Sea anemone venom as a source of insecti- dos Santos, R.G., Diniz, C.R., Cordeiro, M.N., De Lima, M.E., 1999. Binding þ cidal peptides acting on voltage-gated Na channels. Toxicon 49, sites and actions of Tx1, a neurotoxin from the venom of the spider 550–560. Phoneutria nigriventer, in guinea pig ileum. Braz. J. Med. Biol. Res. 32, Bosmans, F., Rash, L., Zhu, S., Diochot, S., Lazdunski, M., Escoubas, P., 1565–1569. Tytgat, J., 2006. Four novel tarantula toxins as selective modulators dos Santos, R.G., Van Renterghem, C., Martin-Moutot, N., Mansuelle, P., of voltage-gated sodium channel subtypes. Mol. Pharmacol. 69, Cordeiro, M.N., Diniz, C.R., Mori, Y., De Lima, M.E., Seagar, M., 2002. 419–429. Phoneutria nigriventer u-phonetoxin IIA blocks the CaV2 family of Brennan, F., Carr, D.B., Cousins, M., 2007. Pain management: a funda- calcium channels and interacts with u-conotoxin-binding sites. J. Biol. mental human right. Anesth. Analg. 105, 205–221. Chem. 277, 13856–13862. Carinci, A.J., Mao, J., 2010. Pain and opioid addiction: what is the Edgerton, G.B., Blumenthal, K.M., Hanck, D.A., 2010. Inhibition of the connection? Curr. Pain. Headache. Rep. 14, 17–21. activation pathway of the T-type calcium channel CaV3.1 by ProTxII. Catterall, W.A., Goldin, A.L., Waxman, S.G., 2005. International Union of Toxicon 56, 624–636. Pharmacology. XLVII. Nomenclature and structure-function relation- Ekberg, J., Craik, D.J., Adams, D.J., 2008. Conotoxin modulation of voltage- ships of voltage-gated sodium channels. Pharmacol. Rev. 57, 397–409. gated sodium channels. Int. J. Biochem. Cell. Biol. 40, 2363–2368. Catterall, W.A., Cestèle, S., Yarov-Yarovoy, V., Yu, F.H., Konoki, K., England, S., de Groot, M.J., 2009. Subtype-selective targeting of voltage- Scheuer, T., 2007. Voltage-gated ion channels and gating modifier gated sodium channels. Br. J. Pharmacol. 158, 1413–1425. þ toxins. Toxicon 49, 124–141. England, S., Rawson, D., 2010. Isoform-selective voltage-gated Na Chen, J.Q., Zhang, Y.Q., Dai, J., Luo, Z.M., Liang, S.P., 2005. Antinociceptive channel modulators as next-generation analgesics. Future. Med. effects of intrathecally administered huwentoxin-I, a selective N-type Chem. 2, 775–790. calcium channel blocker, in the formalin test in conscious rats. Escoubas, P., Bosmans, F., 2007. Spider peptide toxins as leads for drug Toxicon 45, 15–20. development. Expert. Opin. Drug. Discov. 2, 823–835. Chen, J., Deng, M., He, Q., Meng, E., Jiang, L., Liao, Z., Rong, M., Liang, S., Escoubas, P., King, G.F., 2009. Venomics as a drug discovery platform. 2008. Molecular diversity and evolution of cystine knot toxins of the Expert. Rev. Proteomic. 6, 221–224. tarantula Chilobrachys jingzhao. Cell. Mol. Life. Sci. 65, 2431–2444. Escoubas, P., Sollod, B., King, G.F., 2006. Venom landscapes: mining the Chen, J., Zhang, Y., Rong, M., Zhao, L., Jiang, L., Zhang, D., Wang, M., Mao, Y., complexity of spider venoms via a combined cDNA and mass spec- Liang, S.P., 2009. Expression and characterization of jingzhaotoxin-34, trometric approach. Toxicon 47, 650–663. a novel neurotoxin from the venom of the tarantula Chilobrachys Estacion, M., Dib-Hajj, S.D., Benke, P.J., Te Morsche, R.H., Eastman, E.M., jingzhao. Peptides 30, 1042–1048. Macala, L.J., Drenth, J.P., Waxman, S.G., 2008. NaV1.7 gain-of-function Clare, J.J., 2010. Targeting voltage-gated sodium channels for pain therapy. mutations as a continuum: a1632E displays physiological changes Expert. Opin. Investig. Drugs. 19, 45–62. associated with erythromelalgia and paroxysmal extreme pain 490 J.K. Klint et al. / Toxicon 60 (2012) 478–491

disorder mutations and produces symptoms of both disorders. J. Li, D.L., Xiao, Y.C., Xu, X., Xiong, X.Y., Lu, S.Y., Liu, Z.H., Zhu, Q., Wang, M.C., Neurosci. 28, 11079–11088. Gu, X.C., Liang, S.P., 2004. Structure-activity relationships of Ferrat, G., Bosmans, F., Tytgat, J., Pimentel, C., Chagot, B., Gilles, N., hainantoxin-IV and structure determination of active and inactive Nakajima, T., Darbon, H., Corzo, G., 2005. Solution structure of two sodium channel blockers. J. Biol. Chem. 279, 37734–37740. insect-specific spider toxins and their pharmacological interaction Li-Smerin, Y., Swartz, K.J., 1998. Gating modifier toxins reveal a conserved þ þ þ with the insect voltage-gated Na channel. Proteins 59, 368–379. structural motif in voltage-gated Ca2 and K channels. Proc. Natl. Fertleman, C.R., Baker, M.D., Parker, K.A., Moffatt, S., Elmslie, F.V., Acad. Sci. U S A 95, 8585–8589. Abrahamsen, B., Ostman, J., Klugbauer, N., Wood, J.N., Gardiner, R.M., Liang, S.P., Zhang, D.Y., Pan, X., Chen, Q., Zhou, P.A., 1993. Properties and Rees, M., 2006. SCN9A mutations in paroxysmal extreme pain amino acid sequence of huwentoxin-I, a neurotoxin purified from the disorder: allelic variants underlie distinct channel defects and venom of the Chinese bird spider Selenocosmia huwena. Toxicon 31, phenotypes. Neuron 52, 767–774. 969–978. Fletcher, J.I., Chapman, B.E., Mackay, J.P., Howden, M.E., King, G.F., 1997. Liao, Z., Yuan, C., Deng, M., Li, J., Chen, J., Yang, Y., Hu, W., Liang, S., 2006. The structure of versutoxin (d-atracotoxin-Hv1) provides insights into Solution structure and functional characterization of Jingzhaotoxin- the binding of site 3 neurotoxins to the voltage-gated sodium XI: a novel gating modifier of both potassium and sodium channels. channel. Structure 5, 1525–1535. Biochemistry 45, 15591–15600. Fox, J.W., Serrano, S.M.T., 2007. Approaching the golden age of natural Little, M.J., Wilson, H., Zappia, C., Cestèle, S., Tyler, M.I., Martin- product pharmaceuticals from venom libraries: an overview of toxins Eauclaire, M.F., Gordon, D., Nicholson, G.M., 1998. d-atracotoxins from and toxin-derivatives currently involved in therapeutic or diagnostic Australian funnel-web spiders compete with scorpion a-toxin applications. Curr. Pharm. Des. 13, 2927–2934. binding on both rat brain and insect sodium channels. FEBS. Lett. 439, Gordon, D., Karbat, I., Ilan, N., Cohen, L., Kahn, R., Gilles, N., Dong, K., 246–252. Stuhmer, W., Tytgat, J., Gurevitz, M., 2007. The differential preference Liu, Z., Dai, J., Chen, Z., Hu, W., Xiao, Y., Liang, S., 2003. Isolation of scorpion a-toxins for insect or mammalian sodium channels: and characterization of hainantoxin-IV, a novel antagonist of implications for improved insect control. Toxicon 49, 452–472. tetrodotoxin-sensitive sodium channels from the Chinese bird spider Grolleau, F., Stankiewicz, M., Birinyi-Strachan, L., Wang, X.H., Nicholson, G.M., Selenocosmia hainana. Cell. Mol. Life. Sci. 60, 972–978. Pelhate, M., Lapied, B., 2001. Electrophysiological analysis of the neuro- Maggio, F., Sollod, B.L., Tedford, H.W., Herzig, V., King, G.F., 2010. toxic action of a funnel-web spider toxin, d-atracotoxin-HV1a, on insect Spider toxins and their potential for insect control. In: Gilbert, L.I., þ voltage-gated Na channels. J. Exp. Biol. 204, 711–721. Gill, S.S. (Eds.), Insect Pharmacology: Channels, Receptors, Toxins and Gunning, S.J., Chong, Y., Khalife, A.A., Hains, P.G., Broady, K.W., Enzymes. Academic Press, Oxford, pp. 101–123. Nicholson, G.M., 2003. Isolation of d-missulenatoxin-Mb1a, the major Martin-Moutot, N., Mansuelle, P., Alcaraz, G., Dos Santos, R.G., vertebrate-active spider d-toxin from the venom of Missulena bradleyi Cordeiro, M.N., De Lima, M.E., Seagar, M., Van Renterghem, C., 2006. (Actinopodidae). FEBS. Lett. 554, 211–218. Phoneutria nigriventer toxin 1: a novel, state-dependent inhibitor of Heinemann, S.H., Leipold, E., 2007. Conotoxins of the O-superfamily neuronal sodium channels that interacts with m conotoxin binding affecting voltage-gated sodium channels. Cell. Mol. Life. Sci. 64, sites. Mol. Pharmacol. 69, 1931–1937. 1329–1340. Matavel, A., Cruz, J.S., Penaforte, C.L., Araujo, D.A., Kalapothakis, E., Prado, V.F., Herrington, J., Zhou, Y.P., Bugianesi, R.M., Dulski, P.M., Feng, Y., Warren, V.A., Diniz, C.R., Cordeiro, M.N., Beirao, P.S., 2002. Electrophysiological char- Smith, M.M., Kohler, M.G., Garsky, V.M., Sanchez, M., Wagner, M., acterization and molecular identification of the Phoneutria nigriventer Raphaelli, K., Banerjee, P., Ahaghotu, C., Wunderler, D., Priest, B.T., peptide toxin PnTx2-6. FEBS. Lett. 523, 219–223. Mehl, J.T., Garcia, M.L., McManus, O.B., Kaczorowski, G.J., Slaughter, R.S., Middleton, R.E., Warren, V.A., Kraus, R.L., Hwang, J.C., Liu, C.J., Dai, G., 2006. Blockers of the delayed-rectifier potassium current in pancreatic Brochu, R.M., Kohler, M.G., Gao, Y.D., Garsky, V.M., Bogusky, M.J., b-cells enhance glucose-dependent insulin secretion. Diabetes 55, Mehl, J.T., Cohen, C.J., Smith, M.M., 2002. Two tarantula peptides 1034–1042. inhibit activation of multiple sodium channels. Biochemistry 41, Herzig, V., Wood, D.L.A., Newell, F., Chaumeil, P.-A., Kaas, Q., Binford, G.J., 14734–14747. Nicholson, G.M., Gorse, D., King, G.F., 2011. ArachnoServer 2.0, an Miljanich, G.P., 2004. Ziconotide: neuronal calcium channel blocker for updated online resource for spider toxin sequences and structures. treating severe chronic pain. Curr. Med. Chem. 11, 3029–3040. Nucleic. Acids. Res. 39, D653–D657. Moran, Y., Gordon, D., Gurevitz, M., 2009. Sea anemone toxins Honma, T., Shiomi, K., 2006. Peptide toxins in sea anemones: structural affecting voltage-gated sodium channelsdmolecular and evolu- and functional aspects. Mar. Biotechnol. (NY) 8, 1–10. tionary features. Toxicon 54, 1089–1101. King, G.F., 2004. The wonderful world of spiders: preface to the special Mylecharane, E.J., Spence, I., Sheumack, D.D., Claassens, R., Howden, M.E., Toxicon issue on spider venoms. Toxicon 43, 471–476. 1989. Actions of robustoxin, a neurotoxic polypeptide from the King, G.F., 2011. Venoms as a platform for human drugs: translating toxins venom of the male funnel-web spider (Atrax robustus), in anaes- into therapeutics. Expert. Opin. Biol. Ther. 11, 1469–1484. thetized monkeys. Toxicon 27, 481–492. King, G.F., Tedford, H.W., Maggio, F., 2002. Structure and function of Nicholson, G.M., Little, M.J., 2005. Spider neurotoxins targeting voltage- insecticidal neurotoxins from Australian funnel-web spiders. J. gated sodium channels. Toxin. Rev. 24, 315–345. Toxicol.-Toxin Rev. 21, 359–389. Nicholson, G.M., Little, M.J., Tyler, M., Narahashi, T., 1996. Selective King, G.F., Escoubas, P., Nicholson, G.M., 2008a. Peptide toxins that alteration of sodium channel gating by Australian funnel-web spider selectively target insect Nav and Cav channels. Channels 2, 100–116. toxins. Toxicon 34, 1443–1453. King, G.F., Gentz, M.C., Escoubas, P., Nicholson, G.M., 2008b. A rational Nicholson, G.M., Walsh, R., Little, M.J., Tyler, M.I., 1998. Characterisation of nomenclature for naming peptide toxins from spiders and other the effects of robustoxin, the lethal neurotoxin from the Sydney venomous animals. Toxicon 52, 264–276. funnel-web spider Atrax robustus, on sodium channel activation and Krafte, D.S., Bannon, A.W., 2008. Sodium channels and nociception: inactivation. Pflügers Archiv. (Eur. J. Physiol.) 436, 117–126. recent concepts and therapeutic opportunities. Curr. Opin. Pharmacol. Nicholson, G.M., Blanche, T., Mansfield, K., Tran, Y., 2002. Differential þ þ 8, 50–56. blockade of neuronal voltage-gated Na and K channels by antide- Krapcho, K.J., Kral Jr., R.M., Vanwagenen, B.C., Eppler, K.G., Morgan, T.K., pressant drugs. Eur. J. Pharmacol. 452, 35–48. 1995. Characterization and cloning of insecticidal peptides from the Nicholson, G.M., Little, M.J., Birinyi-Strachan, L.C., 2004. Structure and primitive weaving spider Diguetia canities. Insect. Biochem. Mol. Biol. function of d-atracotoxins: lethal neurotoxins targeting the voltage- 25, 991–1000. gated sodium channel. Toxicon 43, 587–599. Leipold, E., DeBie, H., Zorn, S., Borges, A., Olivera, B.M., Terlau, H., Nilsen, K.B., Nicholas, A.K., Woods, C.G., Mellgren, S.I., Nebuchennykh, M., Heinemann, S.H., 2007. mO conotoxins inhibit NaV channels by Aasly, J., 2009. Two novel SCN9A mutations causing insensitivity to interfering with their voltage sensors in domain-2. Channels 1, pain. Pain 143, 155–158. 253–262. Norton, R.S., 1991. Structure and structure-function relationships of sea Levinson, S.R., Luo, S., Henry, M.A., in press. The role of sodium channels anemone proteins that interact with the sodium channel. Toxicon 29, in chronic pain. Muscle Nerve. http://onlinelibrary.wiley.com/doi/10. 1051–1084. 1002/mus.23314/abstract. Nunes, K.P., Costa-Goncalves, A., Lanza, L.F., Cortes, S.F., Cordeiro, M.N., Lewis, R.J., Garcia, M.L., 2003. Therapeutic potential of venom peptides. Richardson, M., Pimenta, A.M., Webb, R.C., Leite, R., De Lima, M.E., Nat. Rev. Drug. Discov. 2, 790–802. 2008. Tx2-6 toxin of the Phoneutria nigriventer spider potentiates rat Li, D., Xiao, Y., Hu, W., Xie, J., Bosmans, F., Tytgat, J., Liang, S., 2003. erectile function. Toxicon 51, 1197–1206. Function and solution structure of hainantoxin-I, a novel insect Oswald, R.E., Suchyna, T.M., McFeeters, R., Gottlieb, P., Sachs, F., 2002. sodium channel inhibitor from the Chinese bird spider Selenocosmia Solution structure of peptide toxins that block mechanosensitive ion hainana. FEBS. Lett. 555, 616–622. channels. J. Biol. Chem. 277, 34443–34450. J.K. Klint et al. / Toxicon 60 (2012) 478–491 491

Pallaghy, P.K., Nielsen, K.J., Craik, D.J., Norton, R.S., 1994. A common Curtatoxins. Neurotoxic insecticidal polypeptides isolated from the structural motif incorporating a cystine knot and a triple-stranded funnel-web spider Hololena curta. J. Biol. Chem. 265, 2054–2059. b-sheet in toxic and inhibitory polypeptides. Protein. Sci. 3, 1833– Szeto, T.H., Birinyi-Strachan, L.C., Smith, R., Connor, M., Christie, M.J., 1839. King, G.F., Nicholson, G.M., 2000. Isolation and pharmacological Payandeh, J., Scheuer, T., Zheng, N., Catterall, W.A., 2011. The crystal characterisation of d-atracotoxin-Hv1b, a vertebrate-selective sodium structure of a voltage-gated sodium channel. Nature 475, 353–358. channel toxin. FEBS. Lett. 470, 293–299. Platnick, N.I., 1997. Advances in Spider Taxonomy, 1992–1995: With Tedford, H.W., Sollod, B.L., Maggio, F., King, G.F., 2004. Australian funnel- Redescriptions 1940–1980. New York Entomological Society & The web spiders: master insecticide chemists. Toxicon 43, 601–618. American Museum of Natural History, New York. updated online Vetter, I., Davis, J.L., Rash, L.D., Anangi, R., Mobli, M., Alewood, P.F., version available at: http://research.amnh.org/entomology/spiders/ Lewis, R.J., King, G.F., 2011. Venomics: a new paradigm for natural catalog/. products-based drug discovery. Amino. Acids. 40, 15–28. Priest, B.T., Blumenthal, K.M., Smith, J.J., Warren, V.A., Smith, M.M., 2007. Wang, M., Rong, M., Xiao, Y., Liang, S., 2012. The effects of huwentoxin-I ProTx-I and ProTx-II: gating modifiers of voltage-gated sodium on the voltage-gated sodium channels of rat hippocampal and cock- channels. Toxicon 49, 194–201. roach dorsal unpaired median neurons. Peptides 34, 19–25. Prikhod’ko, G.G., Robson, M., Warmke, J.W., Cohen, C.J., Smith, M.M., Wang, P., Weinberger, H., Moran, Y., Gordon, D., Turkov, M., Kahn, R., Gurevitz, M., Warren, V., Kaczorowski, G., Van der Ploeg, L.H.T., Miller, L.K., 1996. 2010. Positions under positive selectiondkey for selectivity and Properties of three baculovirus-expressing genes that encode insect- potency of scorpion a-toxins. Mol. Biol. Evol. 27, 1025–1034. selective toxins: m-Aga-IV, As II, and Sh I. Biol. Contr. 7, 236–244. Weiss, J., Pyrski, M., Jacobi, E., Bufe, B., Willnecker, V., Schick, B., Zizzari, P., Redaelli, E., Cassulini, R.R., Silva, D.F., Clement, H., Schiavon, E., Gossage, S.J., Greer, C.A., Leinders-Zufall, T., Woods, C.G., Wood, J.N., Zamudio, F.Z., Odell, G., Arcangeli, A., Clare, J.J., Alagón, A., de la Zufall, F., 2011. Loss-of-function mutations in sodium channel NaV1.7 Vega, R.C.R., Possani, L.D., Wanke, E., 2010. Target promiscuity and cause anosmia. Nature 472, 186–190. þ heterogeneous effects of tarantula venom peptides affecting Na Wilson, M.J., Yoshikami, D., Azam, L., Gajewiak, J., Olivera, B.M., þ and K ion channels. J. Biol. Chem. 285, 4130–4142. Bulaj, G., Zhang, M.M., 2011. m-Conotoxins that differentially block Richardson, M., Pimenta, A.M.C., Bemquerer, M.P.,Santoro, M.M., Beirao, P.S.L., sodium channels NaV1.1 through 1.8 identify those responsible for Lima, M.E., Figueiredo, S.G., Bloch, C., Vasconcelos, E.A.R., Campos, F.A.P., action potentials in sciatic nerve. Proc. Natl. Acad. Sci. U S A 108, Gomes, P.C., Cordeiro, M.N., 2006. Comparison of the partial proteomes of 10302–10307. the venoms of Brazilian spiders of the genus Phoneutria. Comp. Biochem. Windley, M.J., Herzig, V., Dziemborowicz, S.A., Hardy, M.C., King, G.F., Physiol. C. Toxicol. Pharmacol. 142, 173–187. Nicholson, G.M., 2012. Spider-venom peptides as bioinsecticides. Rodriguez de la Vega, R.C., Possani, L.D., 2005. Overview of scorpion Toxins 4, 191–227. þ toxins specificforNa channels and related peptides: biodiver- Wood, D.L., Miljenovic, T., Cai, S., Raven, R.J., Kaas, Q., Escoubas, P., sity, structure-function relationships and evolution. Toxicon 46, Herzig, V., Wilson, D., King, G.F., 2009. ArachnoServer: a database of 831–844. protein toxins from spiders. BMC. Genomics. 10, 375. Rong, M., Chen, J., Tao, H., Wu, Y., Jiang, P., Lu, M., Su, H., Chi, Y., Cai, T., Xiao, Y., Liang, S., 2003a. Inhibition of neuronal tetrodotoxin-sensitive þ Zhao, L., Zeng, X., Xiao, Y., Liang, S., 2011. Molecular basis of Na channels by two spider toxins: hainantoxin-III and hai- the tarantula toxin jingzhaotoxin-III (b-TRTX-Cj1a) interacting nantoxin-IV. Eur. J. Pharmacol. 477, 1–7. with voltage sensors in sodium channel subtype Nav1.5. FASEB. J. 25, Xiao, Y.C., Liang, S.P., 2003b. Purification and characterization of 3177–3185. Hainantoxin-V, a tetrodotoxin-sensitive sodium channel inhibitor Rupasinghe, D., Knapp, O., Blomster, L.V., Schmid, A., Adams, D.J., King, G.F., from the venom of the spider Selenocosmia hainana. Toxicon 41, Ruitenberg, M.J., 2012. Localization of NaV1.7 in the normal and injured 643–650. rodent olfactory system indicates a critical role in olfaction, pheromone Xiao, Y., Tang, J., Hu, W., Xie, J., Maertens, C., Tytgat, J., Liang, S., 2005. sensing and immune function. Channels 6. Jingzhaotoxin-I, a novel spider neurotoxin preferentially inhibiting Saez, N.J., Senff, S., Jensen, J.E., Er, S.Y., Herzig, V., Rash, L.D., King, G.F., cardiac sodium channel inactivation. J. Biol. Chem. 280, 12069–12076. 2010. Spider-venom peptides as therapeutics. Toxins 2, 2851–2871. Xiao, Y., Bingham, J.P., Zhu, W., Moczydlowski, E., Liang, S., Cummins, T.R., Satake, H., Villegas, E., Oshiro, N., Terada, K., Shinada, T., Corzo, G., 2004. 2008. Tarantula huwentoxin-IV inhibits neuronal sodium channels by Rapid and efficient identification of cysteine-rich peptides by random binding to receptor site 4 and trapping the domain II voltage sensor in screening of a venom gland cDNA library from the hexathelid spider the closed configuration. J. Biol. Chem. 283, 27300–27313. Macrothele gigas. Toxicon 44, 149–156. Yamaji, N., Little, M.J., Nishio, H., Billen, B., Villegas, E., Nishiuchi, Y., Schmalhofer, W.A., Calhoun, J., Burrows, R., Bailey, T., Kohler, M.G., Tytgat, J., Nicholson, G.M., Corzo, G., 2009. Synthesis, solution struc- Weinglass, A.B., Kaczorowski, G.J., Garcia, M.L., Koltzenburg, M., ture, and phylum selectivity of a spider d-toxin that slows inactivation Priest, B.T., 2008. ProTx-II, a selective inhibitor of NaV1.7 sodium of specific voltage-gated sodium channel subtypes. J. Biol. Chem. 284, channels, blocks action potential propagation in nociceptors. Mol. 24568–24582. Pharmacol. 74, 1476–1484. Yang, Y., Wang, Y., Li, S., Xu, Z., Li, H., Ma, L., Fan, J., Bu, D., Liu, B., Fan, Z., Skinner, W.S., Adams, M.E., Quistad, G.B., Kataoka, H., Cesarin, B.J., Wu, G., Jin, J., Ding, B., Zhu, X., Shen, Y., 2004. Mutations in SCN9A, Enderlin, F.E., Schooley, D.A., 1989. Purification and characterization encoding a sodium channel a subunit, in patients with primary of 2 classes of neurotoxins from the funnel web spider, Agelenopsis erythermalgia. J. Med. Genet. 41, 171–174. Aperta. J. Biol. Chem. 264, 2150–2155. Zeng, X.Z., Xiao, Q.B., Liang, S.P., 2003. Purification and characterization of Soderlund, D.M., Knipple, D.C., 1995. Actions of insecticides on sodium- raventoxin-I and raventoxin-III, two neurotoxic peptides from the channelsdmultiple-target sites and site-specific resistance. In: venom of the spider Macrothele raveni. Toxicon 41, 651–656. Clark, J.M. (Ed.), 1995. Molecular Action of Insecticides on Ion Zeng, X., Deng, M., Lin, Y., Yuan, C., Pi, J., Liang, S., 2007. Isolation and Channels, vol. 591, pp. 97–108. characterization of Jingzhaotoxin-V, a novel neurotoxin from the Sokolov, S., Kraus, R.L., Scheuer, T., Catterall, W.A., 2007. Inhibiting of venom of the spider Chilobrachys jingzhao. Toxicon 49, 388–399. sodium channel gating by trapping the domain II voltage sensor with Zlotkin, E., 1999. The insect voltage-gated sodium channel as target of protoxin II. Mol. Pharmacol. 73, 1020–1028. insecticides. Annu. Rev. Entomol. 44, 49–55. Stapleton, A., Blankenship, D.T., Ackermann, B.L., Chen, T.M., Gorder, G.W., Zuliani, V., Rivara, M., Fantini, M., Costantino, G., 2010. Sodium channel Manley, G.D., Palfreyman, M.G., Coutant, J.E., Cardin, A.D., 1990. blockers for neuropathic pain. Expert. Opin. Ther. Pat. 20, 755–779.