Accepted Manuscript

Gating modifier toxin interactions with ion channels and bilayers: Is the trimolecular complex real?

Akello J. Agwa, Sónia T. Henriques, Christina I. Schroeder

PII: S0028-3908(17)30141-7 DOI: 10.1016/j.neuropharm.2017.04.004 Reference: NP 6660

To appear in: Neuropharmacology

Received Date: 14 February 2017 Revised Date: 31 March 2017 Accepted Date: 5 April 2017

Please cite this article as: Agwa, A.J., Henriques, Só.T., Schroeder, C.I., Gating modifier toxin interactions with ion channels and lipid bilayers: Is the trimolecular complex real?, Neuropharmacology (2017), doi: 10.1016/j.neuropharm.2017.04.004.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT Gating modifier toxin interactions with ion channels and lipid bilayers: is the trimolecular complex real?

Akello J. Agwa, Sónia T. Henriques*, Christina I. Schroeder*

Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland,

4072, Australia

*Corresponding Authors:

Dr Christina I. Schroeder

Institute for Molecular Bioscience

The University of Queensland

QLD, 4072, Australia Tel: +61 7 334 62021 MANUSCRIPT E-mail: [email protected]

Dr Sónia Troeira Henriques

Institute for Molecular Bioscience

The University of Queensland

QLD, 4072, Australia Tel: +61 7 334 62026ACCEPTED E-mail: [email protected]

1 ACCEPTED MANUSCRIPT Abstract

Spider peptide toxins have attracted attention because of their ability to target voltage- gated ion channels, which are involved in several pathologies including chronic pain and some cardiovascular conditions. A class of these peptides acts by modulating the gating mechanism of voltage-gated ion channels and are thus called gating modifier toxins (GMTs).

In addition to their interactions with voltage-gated ion channels, some GMTs have affinity for lipid bilayers. This review discusses the potential importance of the cell membrane on the mode of action of GMTs. We propose that peptide–membrane interactions can anchor GMTs at the cell surface, thereby increasing GMT concentration in the vicinity of the channel binding site. We also propose that modulating peptide–membrane interactions might be useful for increasing the therapeutic potential of spider toxins. Furthermore, we explore the advantages and limitations of the methodologies currently used to examine peptide- membrane interactions. Although GMT–lipid membrane binding does not appear to be a requirement for the activity of all GMTs, it is an MANUSCRIPT important feature, and future studies with GMTs should consider the trimolecular peptide–lipid membrane–channel complex.

ACCEPTED

2 ACCEPTED MANUSCRIPT Highlights:

• Spider toxins are excellent tools to study voltage-gated ion channel pharmacology

• Some gating modifier toxins have shown affinity for lipid membranes

• The role of the membrane in spider toxin activity can guide rational drug design

• The cell membrane, ion channel and spider toxins should be studied as a complex

Keywords: Disulfide-rich peptides, model membranes, pain, rational drug design, surface plasmon resonance, sodium channel

Abbreviations

Cryo-EM, cryo-electron microscopy; DOPG, 1,2-dioleoyl-sn -glycero-3-phosphoglycerol;

DHPC, 1,2-diheptanoyl-sn -glycero-3-; GMT, gating modifier toxin; GUV,

giant unilamellar vesicle; ICK, inhibitory cystine knot; K voltage-gated potassium channel; MANUSCRIPTV, LUV, large unilamellar vesicle; MSP, membrane scaffold protein; Na V, voltage-gated sodium channel; NMR, nuclear magnetic resonance; PC, ; PE, ; PG, phosphatidylgycerol; PI, ; PS, ; POPC, 1-palmitoyl-2-oleoyl-sn -glycero-3-phosphocholine; POPE, 1- palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine; POPG, 1-palmitoyl-2-oleoyl-sn - glycero-3-glycerol; POPS, 1-palmitoyl-2-oleoyl-sn -glycero-3-phospho-L-serine; SAR, structure activity relationship; SMA, stearic maleic acid; SPR, surface plasmon resonance; SUV, small unilamellarACCEPTED vesicle; VSD, voltage sensor domain

3 ACCEPTED MANUSCRIPT 1. Introduction

Voltage-gated ion channels are transmembrane proteins that are involved in almost all aspects of human physiology and many of these channels have been validated as therapeutic targets for conditions such as neuropathic and chronic pain, irritable bowel syndrome, cardiovascular disease and epilepsy (Catterall, 2012; Dib-Hajj et al., 2010; Kwong and Carr,

2015; Lewis et al., 2015; Osteen et al., 2016). The development of pharmaceutical modulators for voltage-gated ion channels is therefore an area of great interest (Ahuja et al.,

2015; Flinspach et al., 2017; Murray et al., 2015b; Revell et al., 2013; Shcherbatko et al.,

2016). Peptide toxins extracted from animal venoms are a particularly important class of ion channel inhibitors because they have shown a naturally high potency and relative selectivity for a range of therapeutically relevant voltage-gated ion channels (Kwong and Carr, 2015;

Lewis et al., 2015; Osteen et al., 2016).

Gating modifier toxins (GMTs) are a class of venom peptides that derive their name from the mechanism by which the toxins modify the kinetiMANUSCRIPTcs and gating behavior of voltage-gated ion channels (Catterall et al., 2007). GMTs can be extracted from the venoms of various animals including spiders, cone snails, scorpions and snakes (Catterall et al., 2007); however, this review will focus on spider-derived GMTs. Some GMTs have been shown to bind to lipid membranes (Table 1) (Henriques et al., 2016; Jung et al., 2010; Lau et al., 2016;

Milescu et al., 2007), and a correlation has been drawn between GMT lipid binding and the affinity and inhibitory potency of some of these toxins to voltage-gated ion channels (Agwa et al., 2017; HenriquesACCEPTED et al., 2016; Lau et al., 2016; Lee et al., 2004; Revell Phillips et al., 2005; Salari et al., 2016). GMTs are therefore excellent tools to study the emerging trimolecular complex involving peptide ligands, voltage-gated ion channels, and the lipid membranes surrounding these channels. This review examines the interplay between GMT structure, interactions with voltage-gated ion channels and interactions with lipid membranes.

4 ACCEPTED MANUSCRIPT 1.1. Gating modifier toxin structure

GMTs typically contain an inhibitory cysteine knot (ICK) motif characterized by two– three anti-parallel β-sheets stabilized by three disulfide bridges with Cys1–Cys4, Cys2–Cys5 and Cys3–Cys6 connectivity (King et al., 2002; Pallaghy et al., 1994), as illustrated in Figure

1A by ProTx-II, SgTx-1, Vstx-1 and HwTx-IV. Notably, anti-parallel beta sheets are not always present in the secondary structures of the GMTs (e.g. ProTx-II and HwTx-IV, Figure

1A) due to variation in the size of the interconnecting loops (Henriques et al., 2016; Lee et al., 2004; Minassian et al., 2013); however the conserved cysteine knot connectivity renders stability to these types of peptides (Herzig and King, 2015; Pallaghy et al., 1994). An additional conserved structural feature is the folding of these peptides to form an amphipathic surface profile consisting of a hydrophobic patch surrounded by a charged ring of amino acid residues (Figure 1B) (Jung et al., 2005; King et al., 2002; Takahashi et al., 2000a; Xiao et al.,

2008b). The amino acid residues forming the hydrophobic patch and charged ring include Trp, Tyr and Lys, which are often involved in pepti MANUSCRIPTde– binding (Killian and von Heijne, 2000) (for sequences, see Figure 1C), therefore the conserved hydrophobic patch and charged ring may confer GMTs with affinity for lipid membranes in addition to interactions with the voltage-gated ion channels.

1.2. Voltage-gated ion channel structures: a basis for understanding the mechanism

of action of gating modifier toxins Voltage-gatedACCEPTED ion channels share similar structural features thus the structures of voltage- gated potassium (K V) and voltage-gated sodium (Na V) channels will be reviewed as

representatives of this diverse family of transmembrane proteins. K Vs and Na Vs comprise four domains (domain I–domain IV) (Catterall, 2010; Noda et al., 1986; Noda and Numa,

1987) with each domain comprising six transmembrane helical segments (S1–S6) (Figure

5 ACCEPTED MANUSCRIPT

2A). Whereas each KV domain exists as a homotetramer, Na Vs possess intra- and extra-

cellular loops of various lengths and post-translational modifications that connect the four

domains to form a heterotetramer (Figure 2A–C) (Catterall, 2010; Guy and Seetharamulu,

1986; Jiang et al., 2003; Long et al., 2005; Papazian et al., 1987; Payandeh et al., 2012). In

each domain, segments S1–S4 are classified as the voltage sensor domain (VSD) whereas

segments S5 and S6 form the pore domain (Figure 2B,C) (Catterall, 2010; Guy and

Seetharamulu, 1986; Jiang et al., 2003; Long et al., 2005; Papazian et al., 1987; Payandeh et

al., 2012).

During channel activation, cations on the extracellular side of the membrane create a positive potential, which is detected by a portion of the VSD known as the paddle motif

(S3b–S4) (Figure 2D) (Armstrong and Bezanilla, 1974; Catterall, 2010). The VSDs of domain I–domain III control the movement of the paddle motif such that the paddle moves to open an activation gate and allow the inward conduction of ions through the pore (S5–S6 of domain I–domain IV). Following channel activation MANUSCRIPT some of the voltage-gated ion channels can self-inactivate when the VSD of domain IV closes an inactivation gate to prevent further entry of ions (Bezanilla, 2000; Bezanilla, 2002; Bosmans et al., 2008; Hille, 2001). GMTs alter the kinetics, gating mechanisms and relative stability of the closed, open and inactivated states of the channels by modulating the motion of the paddle motif as it translocates through the membrane (Bosmans and Swartz, 2010; Ruta et al., 2003; Swartz and MacKinnon, 1995,

1997a, b). ACCEPTED 1.3. Specific interactions between gating modifier toxins and voltage-gated ion

channels

The pharmacological effect of GMTs on voltage-gated ion channels has been studied using structure activity relationship (SAR) studies using peptide analogues and novel

6 ACCEPTED MANUSCRIPT techniques like multiple attribute positional scanning (Lau et al., 2016; Li et al., 2004; Liu et al., 2012; Minassian et al., 2013; Murray et al., 2015b; Murray et al., 2016; Murray et al.,

2015c; Revell et al., 2013; Wang et al., 2004).

The pharmacology of GMTs has also been examined using voltage-gated ion channel

chimeras where components of one channel are introduced onto a second channel (Bosmans

et al., 2008; Liu et al., 2013; Murray et al., 2015a). Bosmans et al. (2008), designed a chimera

channel consisting of the rat potassium channel, KV2.1, and each of the paddle motifs from

domain I–domain IV of the rat sodium channel, Na V1.2 (Bosmans et al., 2008). This study

demonstrated that the mechanism of action of GMTs is dictated by the site on the voltage-

gated ion channels to which the peptides bind. The authors found that GMTs able to bind to

both domain II and domain IV (e.g. ProTx-II and ProTx-I), impaired channel activation,

whereas GMTs (e.g. SgTx-1) that had affinity only for domain IV, impaired channel

inactivation (Bosmans et al., 2008). GMT mechanism of action may however be more complex and peptide specific as was evident when MANUSCRIPT ProTx-II inhibited both activation and

inactivation of Na V1.7 even though the peptide is not selective to domain IV (Xiao et al.,

2010).

In addition to intra-channel promiscuity, mutagenesis studies on voltage-gated ion

channels show that GMTs bind to similar restricted residue positions on different channels

(Lau et al., 2016; Li-Smerin and Swartz, 1998; Revell Phillips et al., 2005; Swartz and

MacKinnon, 1997b; Xiao et al., 2010). These restricted positions comprise conserved hydrophobic andACCEPTED acidic residues on the different voltage-gated ion channels, which aligns well with the overall positive charge and hydrophobic patch on the structures of the GMTs

(Figure 3A) (Li-Smerin and Swartz, 1998; Schmalhofer et al., 2008; Swartz and MacKinnon,

1997b; Xiao et al., 2010). For instance, HaTx-1 is a K V2.1 inhibitor that forms interactions with the hydrophobic Ile273 and Phe274 and the anionic Glu277 located on transmembrane

7 ACCEPTED MANUSCRIPT portion of S3b on domain II of the channel (Swartz and MacKinnon, 1997b). Likewise,

ProTx-II forms key interactions with Phe813 on the distal transmembrane portion of S3b, and with Glu818 on the extracellular S3–S4 loop on domain II of Na V1.7 (Schmalhofer et al.,

2008; Xiao et al., 2010), whereas HwTx-IV, another Na V1.7 inhibitor, interacts with Glu811 on the distal transmembrane portion of S3b, and with Leu814, Asp816 and Glu818 on the extracellular S3–S4 loop on domain II (Xiao et al., 2008a; Xiao et al., 2010; Xiao et al.,

2011).

In silico docking studies on HwTx-IV and HnTx-IV, two Na V1.7 inhibitors, led to the proposition that the GMTs recognize the 3D-conformation on their targets rather than specific residues. This hypothesis resulted from the observation that both toxins inserted in the groove between the loops of S1–S2 and S3–S4 on Na V1.7 domain II (Cai et al., 2015;

Minassian et al., 2013). This binding mode places the peptides in a cleft between S2–S3 and

S1–S4 allowing GMTs to bind to both the voltage-gated ion channels and the lipid membrane (Figure 3B) (Cai et al., 2015; Lau et al., 2016; Mi nassianMANUSCRIPT et al., 2013).

2. Gating modifier toxins and lipid membrane interactions

2.1. VsTx-1: the lipid membrane can act as an anchor for gating modifier toxins in

their interactions with voltage-gated ion channels

Studies on VsTx-1 contributed to the initial evidence that some GMTs interact with lipid bilayers (LeeACCEPTED and MacKinnon, 2004). The authors noted that VsTx-1 had weaker affinity for KVAP, a bacterial potassium channel, when KVAP had been extracted from the membrane using detergent, compared to when the channel was embedded in the membrane

(Lee and MacKinnon, 2004). When VsTx-1 was incubated with model lipid membranes and centrifuged in a spin-down assay, the GMT was observed to co-precipitate with the model

8 ACCEPTED MANUSCRIPT membranes comprising 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE)/1-

palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (POPG) (3:1 molar ratio) (see

Table 1). The spectral properties of Trp residues in peptides is dependent on the local

environment (aqueous or surrounded by lipid) of the peptide, therefore if the Trp residues are

inserted into model membranes, the fluorescence emission spectrum of the peptide is

expected to show a change in the quantum yield as well as a shift towards shorter

wavelengths (blue shift) (Henriques et al., 2012; Lakowicz, 2006). Studies following the

VsTx-1 Trp fluorescence emission upon titration with POPE/POPG model membranes

showed both an increase in the fluorescence quantum yield and a blue shift of the

fluorescence emission spectra, suggesting that the Trp residues of VsTx-1 insert into the lipid

membranes. Similar results were obtained when the GMT was studied in model membranes

composed of 1-palmitoyl-2-oleoyl-sn -glycero-3-phosphocholine (POPC)/POPG (1:1 molar

ratio) (Lee and MacKinnon, 2004; Mihailescu et al., 2014). Together, these results suggested

that the discrepancy in GMT–KVAP binding with MANUSCRIPT and without membranes was because the lipid membranes provide an excess of free energy of partitioning. The study further proposed

a step-wise mechanism of GMT action where the toxin binds first to the membrane then,

being concentrated there, is enabled to increase the frequency and apparent affinity of

interaction with the paddle motif. A recent study also suggested that the ability of VsTx-1 to

interact with the lipid membranes further improves its affinity for the K VAP VSD (Lau et al.,

2016). NMR chemicalACCEPTED shift mapping is a technique used to monitor the local environment of a peptide ligand upon its interaction with its target (Bieri et al., 2011; Lau et al., 2016), and was

used to examine the interaction of VsTx-1 with the K VAP VSD (Lau et al., 2016). The K VAP

VSD was stabilized in a 1,2-diheptanoyl-sn -glycero-3-phosphocholine (DHPC) micelle and

1 15 chemical shift changes observed in the 2D H- N HSQC and 3D HNCO spectra of the K VAP

9 ACCEPTED MANUSCRIPT VSD were studied in the presence and absence of VsTx-1 to identify residues to which the

GMT interacts with. Residues on the extracellular loops of S1–S2 and S3–S4 of the K VAP

VSD showed the largest chemical shift perturbations and were identified as the VsTx-1

binding site (Lau et al., 2016). The importance of the hydrophobic patch of VsTx-1 in

interactions with lipid membranes was confirmed using NMR chemical shift mapping in

DHPC model membranes which showed that the VsTx-1 residues with the largest chemical

shift changes belonged to the conserved amphipathic GMT motif. Furthermore, analogues of

VsTx-1 containing mutations to the hydrophobic patch and charged ring were compared

using centrifuge spin-down assays in POPG model membranes and showed that both the

hydrophobic patch and the charged ring of VsTx-1 were involved in interactions with lipid

membranes (Lau et al., 2016). Complementary in silico docking studies of VsTx-1 to the

VSD of K VAP, outlined that when the GMT binds between loops S1–S2 and S3–S4 of the

VSD, VsTx-1 is oriented to form interactions with the VSD in a way similar to that proposed earlier for HwTx-IV and HnTx-IV (Cai et al., 2015;MANUSCRIPT Minassian et al., 2013). In this spatial orientation, the hydrophobic patch of VsTx-1 is proposed to be placed in a lipid filled cavity

formed between S2–S3 and S1–S4 (Figure 3B) (Lau et al., 2016). This model suggests that

the membrane anchors the peptide in the interaction between VsTx-1 and K VAP, and that the

specific arrangement of the residues of the GMTs is important for both channel and lipid

membrane interactions.

2.2. SgTx-1:ACCEPTED the amphipathic structure of gating modifier toxins promotes electrostatic interactions with the lipid membrane

SgTx-1 is a GMT initially identified as an inhibitor of the rat potassium channel, K V2.1

(Marvin et al., 1999; Wang et al., 2004). There is an overlap between the SgTx-1 amino acid

residues that bind to the channel and those that interact with lipid membranes (Milescu et al.,

10 ACCEPTED MANUSCRIPT 2007; Wang et al., 2004). Specifically, residues Arg3, Tyr4, Leu5, Phe6, Arg22 and Trp30 were initially identified as important for interactions with K V2.1 (Wang et al., 2004), and later as important for interactions with model membranes (Figure 3C) (Milescu et al., 2007).

These findings support a binding mode where overlapping residues on SgTx-1 interact with the membrane and the channel at the same time (Figure 3B, C). An alternative theory is that some GMTs bind to the membrane with a membrane interaction face and to the channel with a distinct channel binding face (Figure 3B, D) (Agwa et al., 2017; Deplazes et al., 2016;

Henriques et al., 2016). Molecular dynamic studies on ProTx-I and ProTx-II suggest that these peptides bind to lipid membranes with a region of the GMT distinct from the channel interaction region on the peptides (Deplazes et al., 2016; Henriques et al., 2016). These findings place ProTx-I and ProTx-II in a more surface exposed location of the channel when compared to the binding mode suggested for HwTx-IV, HnTx-IV and VsTx-1 (Cai et al.,

2015; Lau et al., 2016; Minassian et al., 2013), whilst still allowing for key interactions with channel residues on the extracellular S1–S2 and MANUSCRIPTS3–S4 loops (Figure 3B) (Deplazes et al., 2016; Henriques et al., 2016). This binding mode was also proposed in earlier studies on

HaTx-I (Revell Phillips et al., 2005), and more recently for gHwTx-IV, a mutant of HwTx-IV

(Agwa et al., 2017). Both binding modes suggest that GMTs adopt a shallow location in lipid membranes.

Quenchers of peptide Trp fluorescence such as, acrylamide, bromine and 5- and 16- doxy stearic acid adopt different locations within the lipid bilayer and can be used to determine the depthACCEPTED to which peptides insert into model membranes. Quenchers can also be covalently linked to varying positions on the acyl chains of the forming lipid bilayers to determine the depth of peptide insertion (Henriques et al., 2012; Lakowicz, 2006;

Revell Phillips et al., 2005; Torcato et al., 2013). To determine the depth to which the Trp30 residue of SgTx-1 interacts with the membrane, containing bromine at

11 ACCEPTED MANUSCRIPT different lengths along the acyl chains of the were used in the preparation of model membranes composed of POPC/POPG (1:1 molar ratio) (see Table 1). Bromine was used as a quencher of the Trp fluorescence and the study determined that the fluorescence emission from Trp30 on SgTx-1 was quenched when the bromine was localized at a depth of about 9 Å from the centre of the model membranes (Jung et al., 2010). This shallow location in the membrane has also been seen for VsTx-1 and HaTx-1 in anionic lipid membranes

(Mihailescu et al., 2014; Revell Phillips et al., 2005).

The orientation of specific residues on SgTx-1 during interactions between the GMT and model membranes was further studied using NMR transferred cross-saturation (NMR-

TCS) experiments. In these experiments irradiation was applied to POPC model membranes causing a saturation of the resonances from the phospholipids due to spin diffusion. The saturation was then transferred to the residues on uniformly labelled 2H- 15 N SgTx-1 that form the peptide –lipid vesicle interface (Jung et al., 2010; Nakanishi et al., 2002; Takahashi et al., 2000b). The experiments showed that SgTx-1 MANUSCRIPT residues forming the hydrophobic patch interacted with hydrophobic core of the lipid bilayer and that this orientation placed the GMT

such that residues forming the charged ring would interact with the phospholipid head-groups

of the lipid bilayer (Jung et al., 2010).

Binding of SgTx-1 to the surface of lipid bilayers suggests that GMT –lipid interactions are governed by electrostatic interactions between the peptides and the phospholipid head- groups. The importance of electrostatic interactions is supported by an improved interaction with anionic comparedACCEPTED to zwitterionic model membranes as well as through stronger interactions with lipid bilayers prepared using buffers of low ionic strength compared to buffers of higher ionic strength (Milescu et al., 2007). If GMTs are to be used as pharmacological tools, it is therefore relevant to determine whether the electrostatic interactions between these toxins and the lipid membrane can be exploited to optimize the

12 ACCEPTED MANUSCRIPT inhibitory potency of the peptides at the voltage-gated ion channels.

2.3. ProTx-I and ProTx-II: exploiting electrostatic interactions to increase potency

of gating modifier toxins

ProTx-I and ProTx-II have also been shown to interact preferentially with anionic lipid bilayers via electrostatic interactions (Deplazes et al., 2016; Henriques et al., 2016).

Both studies used surface plasmon resonance (SPR) to examine peptide–membrane binding experiments, where lipid bilayers were immobilized onto a chip, and a range of concentrations of the peptide of interest was injected onto the lipid bilayers such that affinity was calculated from the amount of peptide bound to the model membrane at the end of the association phase (Cooper et al., 2000; Henriques et al., 2012; Henriques et al., 2008). SPR was used for analysis of the binding kinetics of both ProTx-I and ProTx-II on lipid bilayers with varying properties including fluid zwitterionic POPC and anionic POPC/1-palmitoyl-2- oleoyl-sn -glycero-3-phospho-L-serine (POPS) (4:1 MANUSCRIPT molar ratio) (see Table 1). ProTx-II was also studied in several lipid systems including rigid, raft-like lipid bilayers comprising the zwitterionic POPC/ (SM)/Cholesterol (2.4:3.3:4 molar ratio) and anionic

POPC/ceramide-1-phosphate (C1P) (3:2 molar ratio) (Henriques et al., 2016). Both studies determined that the peptides had slower dissociation rates when interacting with anionic compared to zwitterionic lipid bilayers regardless of the fluidity of the model membranes tested (Deplazes et al., 2016; Henriques et al., 2016). The preferential interaction between ProTx-II and anionicACCEPTED lipid bilayers has been shown in additional studies, supporting the theory that electrostatic interactions are important for the interaction of this GMT with lipid bilayers (Smith et al., 2005; Xiao et al., 2008b).

To show that ProTx-I and ProTx-II adopt shallow positions on the lipid bilayers, the ability of the peptides to cause changes in the membrane dipole potential was studied

13 ACCEPTED MANUSCRIPT (Deplazes et al., 2016; Henriques et al., 2016). The membrane dipole potential arises from the sum of the polar head-groups and ester carbonyl regions of the phospholipids and is dependent on the arrangement of the dipoles at the water/lipid interface (Henriques et al.,

2008; Matos et al., 2010). If a peptide perturbs the lipid bilayer, the arrangement of the phospholipid bilayer is disturbed and this can be detected by following changes in the fluorescence excitation properties of di-8-ANNEPS, a membrane dipole potential probe. An increase in membrane dipole potential typically results in a blue shift in the fluorescence excitation maximum of di-8-ANNEPS, whereas if a peptide causes a decrease in the membrane dipole potential, this will be evidenced as a red shift in the fluorescence excitation maximum of di-8-ANNEPS (Henriques et al., 2008; Matos et al., 2010). Neither ProTx-I nor

ProTx-II caused considerable changes in the membrane dipole potential of POPC/POPS (4:1 molar ratio) lipid bilayers (see Table 1), providing further evidence for a shallow positioning of the peptide toxins in the lipid membrane (Deplazes et al., 2016; Henriques et al., 2016). Despite the shallow positioning of ProTx-II MANUSCRIPT in the lipid membrane, a correlation between membrane interactions and potency of ProTx-II at Na V1.7 has been observed

(Henriques et al., 2016). Mutants of ProTx-II were designed to investigate the role of Trp and

Lys amino acid residues in lipid membrane interactions and analogues in which Trp residues

were replaced with Tyr showed both a decrease in lipid affinity and a decrease in the potency

of the peptides on inhibition of Na V1.7, whereas a [Glu17Lys]ProTx-II mutation showed a comparable potency to the wild type sequence of ProTx-II in addition to stronger affinity to lipid membranesACCEPTED (Henriques et al., 2016). A positive correlation between membrane interactions and inhibitory potency of ProTx-II has also been found for the interaction of this peptide with voltage-gated calcium channels (Salari et al., 2016), and for the inhibition of

Na V1.7 by gHwTx-IV (Agwa et al., 2017). These findings raise the interesting possibility that

14 ACCEPTED MANUSCRIPT increasing membrane binding of GMTs can be used as a strategy to increase the potency of these peptides as inhibitors of the voltage-gated ion channels.

2.4. HwTx-IV and Hd1a: membrane binding is not a prerequisite for potent

inhibition of voltage-gated ion channels.

Despite the findings that several GMTs show affinity for lipid membranes, some GMTs

have been shown not to have strong interactions with the lipid membrane (Agwa et al., 2017;

Deplazes et al., 2016; Posokhov et al., 2007; Xiao et al., 2008b). HwTx-IV and Hd1a are two

examples of GMTs with potent inhibitory activity at Na V1.7 that have been shown not to have strong interactions with lipid bilayers (IC 50 values of HwTx-IV and Hd1a are 26 nM and

115 nM, respectively) (Klint et al., 2015; Xiao et al., 2008a). HwTx-IV failed to co- precipitate with anionic POPE/POPG (3:1 molar ratio) and zwitterionic POPC lipid bilayers

(Xiao et al., 2008b); furthermore the peptide showed minimal binding to various lipid bilayers when studied using SPR and upon following MANUSCRIPT Trp emission of the peptide toxin using fluorescence spectroscopy (see Table 1) (Agwa et al., 2017). Similarly, SPR sensorgrams of

Hd1a showed a low binding maximum accompanied by a fast dissociation rate from anionic

POPC/POPS (4:1 molar ratio) and zwitterionic POPC lipid bilayers (Deplazes et al., 2016).

The absence of considerable lipid interactions in both neutrally charged and negatively charged lipid bilayers in these three studies indicate GMTs are capable of potent inhibition of voltage-gate ion channels in the absence of strong membrane binding (Agwa et al., 2017; Deplazes et al.,ACCEPTED 2016; Xiao et al., 2008b). Therefore the shared hydrophobic patch and charged ring of native GMTs do not necessarily facilitate membrane binding.

2.5. Possible effects on selectivity arising from increasing GMT –lipid membrane

binding

15 ACCEPTED MANUSCRIPT A major limitation to the use of GMTs as therapeutic leads is that these toxins show promiscuous interactions with various types and subtypes of the voltage-gated ion channels

(Bosmans et al., 2006; Bosmans and Swartz, 2010; Middleton et al., 2002). Little has been done to delineate the possibility of modulating ion channel subtype selectivity of the GMTs by altering peptide–lipid membrane binding. There are two possible scenarios; firstly, engineering peptides to increase overall lipid membrane binding could result in peptides with even higher voltage-gated ion channel subtype promiscuity. This would occur as a result of peptides accumulating close to and binding to membranes related to undesired transmembrane proteins and leading to unwanted downstream effects. Secondly, increasing the lipid interactions of the GMTs could act to augment their specificity for a voltage-gated ion channel of interest resulting in a more selective as well as a more potent modulator of the voltage-gated ion channel. Though several recent efforts have been directed towards the development of selective analogues of GMTs (Flinspach et al., 2017; Murray et al., 2015b; Murray et al., 2016; Park et al., 2014; Revell etMANUSCRIPT al., 2013; Shcherbatko et al., 2016), the relationship between membrane interactions and selectivity of the GMTs remains to be explored and the methodology required to study these interactions in greater detail is still being developed.

3. Choice of model membranes for the study of gating modifier toxins and their

interactions with the lipid membrane The choice ACCEPTEDof model membranes for the study of the trimolecular complex is important and is dependent on the methodology used. For instance, giant and large unilamellar vesicles

(GUVs and LUVs) are liposomes that have been used in centrifuge spin-down assays and fluorescence spectroscopic experiments because these vesicles have large diameters that create a relatively planar surface for the study of the peptides as they interact with lipids

16 ACCEPTED MANUSCRIPT (Figure 4) (Henriques et al., 2009; Ladokhin et al., 2000). Small unilamellar vesicles (SUVs) are used in SPR experiments because their small diameters create a curvature strain that allows for their fusion and deposition in the form of lipid bilayers on the sensor chips (Figure

4) (Agwa et al., 2017; Deplazes et al., 2016; Henriques et al., 2016; Henriques et al., 2009;

Ladokhin et al., 2000).

Micelles were used to stabilize the K VAP VSD in the recent NMR study on VsTx-1

(Figure 4) (Lau et al., 2016). Disadvantages of micelles are that these spherical aggregates do not provide a planar surface with the distinct water/lipid interface and hydrophobic core that is typical of physiologically relevant lipid bilayers. Furthermore, micelles are formed using detergent and peptide interactions studied using these model membranes may therefore be skewed because of the solubility of the peptides in the detergent (Bayburt and Sligar, 2010;

Dörr et al., 2016).

Recent advances in the formation of nanodiscs make these model membranes a stable system for the study of the trimolecular complex MANUSCRIPT (Dörr et al., 2016; Gao et al., 2016; Shenkarev et al., 2014). Nanodiscs are portions of bilayer phospholipids stabilized by a belt-

like membrane scaffold protein (MSP) or copolymers like stearic maleic acid (SMA) (Figure

4) (Bayburt and Sligar, 2010; Dörr et al., 2016). Transmembrane proteins can be embedded

in nanodiscs allowing for the study of these proteins and their ligands in more physiologically

relevant oligomeric protein states in the absence of detergent (Bayburt and Sligar, 2003,

2010; Dörr et al., 2016; Leitz et al., 2006; Whorton et al., 2007). Furthermore, the size of the nanodiscs can beACCEPTED adjusted by modifying the size of the stabilizing membrane scaffold peptide or copolymer (Bayburt and Sligar, 2010; Dörr et al., 2016). Additionally, nanodiscs allow for

accessibility of the extracellular and intracellular domains of the transmembrane proteins

which allows their immobilization of the nanodisc onto SPR sensor chips and other surfaces

(Gluck et al., 2011).

17 ACCEPTED MANUSCRIPT 3.1. Choice of physiologically relevant lipids for the formation of model membranes

KVs, Na Vs and additional ion channels targeted by GMTs are embedded in axons, which transmit electrical signals (Vabnick et al., 1996; Vabnick et al., 1999); thus the lipids forming the plasma membrane of the axon are relevant to the study of GMT–lipid membrane interactions. Table 2 shows the relative proportions of the lipids of the plasma membrane of a human axon as well as the lipids from the oocytes of Xenopus laevis which are model cells used for electrophysiological measurements on voltage-gated ion channels (DeVries, 1984;

Opekarová and Tanner, 2003; Stith et al., 2000).

Phospholipids containing phosphatidylcholine (PC) head-groups are zwitterionic and are the most abundant type of phospholipid in the plasma membranes of most eukaryotic cells

(Table 2) (Stith et al., 2000; van Meer and de Kroon, 2011). Therefore, model membranes containing PC-phospholipids have been used to study interactions of GMTs with lipid bilayers. Phospholipids containing phosphatidylethanolamine (PE) head-groups are also zwitterionic and though PE-phospholipids are more MANUSCRIPT abundant than PC-phospholipids in the overall membrane composition of human axonal cells (Table 2) (DeVries, 1984), PE- phospholipids are primarily distributed to the inner leaflet of plasma membranes where they are involved in the formation of vesicles for the processes of endocytosis (van Meer and de

Kroon, 2011; van Meer et al., 2008). Therefore, the use of lower concentrations of PE- phospholipids with respect to PC-phospholipids would represent more physiologically relevant model membranes when studying peptides like GMTs, which appear to adopt a shallow positioningACCEPTED in the outer leaflet of the lipid membrane. Phospholipids containing a phosphatidylglycerol (PG) head group are anionic, and found in bacterial cell membranes and therefore studies relating to GMTs that inhibit K VAP, a bacterial potassium channel, can justify the use of this anionic phospholipid in their studies

(Jung et al., 2005; Lau et al., 2016; Lee and MacKinnon, 2004; Mihailescu et al., 2014).

18 ACCEPTED MANUSCRIPT However eukaryotic cells only have PG-phospholipids in the mitochondria (van Meer et al.,

2008) and PG-phospholipids are therefore less physiologically relevant to studies involving

GMTs that inhibit eukaryotic Na Vs and K Vs. Phospholipids containing a phosphatidylserine

(PS) head-group form anionic signaling phospholipids and are mostly found in the inner

leaflet of the membrane bilayer of eukaryotic cells. The presence of PS-phospholipids in the

outer leaflet of a cell membrane is a signal to trigger cell apoptosis (van Meer and de Kroon,

2011). Studies that have used PS- and PG-phospholipids have conferred an artificial ability of

the positively charged GMTs to have a higher affinity for the anionic model membranes

compared to neutral model membranes made from PC-phospholipids and many of these

studies have used this as evidence for electrostatic interactions between the peptide toxins

and the lipid membrane (Agwa et al., 2017; Bae et al., 2016; Berkut et al., 2015; Deplazes et

al., 2016; Henriques et al., 2016; Milescu et al., 2007; Posokhov et al., 2007; Salari et al.,

2016; Smith et al., 2005; Xiao et al., 2008b). This raises the question of whether or not the trimolecular complex involving GMTs, the voltage-gaMANUSCRIPTted ion channels and the lipid membrane is real. The next sub-section examines compelling reasons to continue to

investigate the trimolecular complex.

3.2. Why study the trimolecular complex?

Despite criticism of the use of PS- and PG-phospholipids to show interactions between

the GMTs and the lipid bilayer in the studies thus far, the outer leaflet of eukaryotic cell membranes is notACCEPTED entirely zwitterionic, there are several anionic charged moieties including the sugar moieties of glycosphingolipids (Gennis, 1989; van Meer et al., 2008). Therefore,

the use of phospholipids containing PS and PG head-groups can be used to represent an

artificial anionic contribution to the overall charge of the outer leaflet of the cell membrane.

Secondly, the in silico docking studies of HwTx-IV and HnTx-IV and the recent NMR

19 ACCEPTED MANUSCRIPT analysis of VsTx-1 suggest that these GMTs are located in a portion of the VSD that allows for interactions between the peptide toxins and the lipid membrane and therefore studies should examine the three components of the trimolecular complex concomitantly before a definitive conclusion can be made (Cai et al., 2015; Lau et al., 2016; Minassian et al., 2013).

Additional evidence to support the need for studies on the three components of the trimolecular complex comes from work by Milescu et al. (2009). Milescu and colleagues studied an enzyme, sphingomylelinase D (SMase D), which hydrolyses SM, a zwitterionic lipid, to anionic C1P (Milescu et al., 2009; Palsdottir and Hunte, 2004). The VSDs of various voltage-gated ion channels were studied as chimeras of K V2.1, and it was observed that

SMaseD modified the behavior of each VSD differently (Milescu et al., 2009). To examine

the pharmacology of this phenomenon, the ability of a ProTx-I, VsTx-1 and GxTx-1E to

detect changes in the lipid from SM to C1P was studied. The study showed that each GMT

studied had an apparent increase in affinity to its VSD in the presence of SMaseD.

Furthermore, they determined that when two differen MANUSCRIPTt domains of Na V channels (domain II and domain IV) were compared to each other by studying the same toxin inhibitor, ProTx-I, there was a considerable difference in increase in affinity. This indicated that ProTx-I was capable of detecting different interactions of the lipids to the VSD and highlights the importance of studying the three molecules of the trimolecular complex together (Milescu et al., 2009).

4. Methods usedACCEPTED in the study of the trimolecular complex and outlook for future directions

Many of the methods currently used to study the relationship between GMTs and the lipid

membrane lack the sensitivity required to fully delineate intimate binding events between the

peptides and the lipid membranes to provide us with an overall picture of the potential

20 ACCEPTED MANUSCRIPT molecular interactions taking place. As increasingly sensitive methodologies are being developed including 3D and 4D NMR and cryo-electron microscopy (cryo-EM), we expect a more detailed picture to emerge over the next few years including the complete trimolecular complex and not just the peptide–lipid membrane or peptide–channel interactions. A brief summary of current methodologies, their advantages and disadvantages follows.

4.1. Centrifuge spin-down assays

Centrifuge spin-down assays involve the incubation of peptides with model membranes, spinning down the suspension and using analytical reversed-phase high performance liquid chromatography to quantify the amount of peptide left in the supernatant as a measure of the extent to which the peptide has bound to the lipid membrane (Figure 5A). Spin-down assays have been used in several studies between GMTs and lipid membranes (Lau et al., 2016; Lee and MacKinnon, 2004; Milescu et al., 2007; Smith et al., 2005; Suchyna et al., 2004; Xiao et al., 2008b). The advantage of centrifuge spin-down MANUSCRIPT assays is that they provide a quick screening method to determine whether or not the GMTs are combined with the model membranes in the pellet of the suspension. However, this method assumes that the absence of the peptides in the supernatant following centrifugation is evidence that the peptide is bound to the lipid bilayers. In reality, artificial co-precipitation of peptide molecules with the lipids following centrifugation might occur as a result of liposome disruption and/or aggregation, and not necessarily through interactions of the peptide molecules with stable lipid membranes. ThusACCEPTED using techniques such as fluorescence spectroscopy or SPR to study peptide–lipid interactions would be preferable to avoid artifacts.

4.2. Surface plasmon resonance

21 ACCEPTED MANUSCRIPT More recently, SPR has been used to screen for the presence or absence of interactions between GMTs and model membranes (Agwa et al., 2017; Deplazes et al., 2016; Henriques et al., 2016). SPR experiments involve deposition of lipid bilayers onto the surface of an L1 sensor chip, peptides are then injected onto the lipid bilayer surface and the time-dependent binding of the peptides to the lipid bilayers is measured by an increase in the refractive index on the surface of the sensor chip. Once injection of the peptide is stopped, the dissociation of the peptide from the lipid bilayer can be followed by monitoring the decrease in the refractive index on the surface of the L1 sensor chip (Figure 5B) (Cooper et al., 2000; Henriques et al.,

2012; Henriques et al., 2008; van der Merwe, 2001). The advantages of SPR are that the experiments allow for sensitive, automated, semi high-throughput screening of peptides binding to lipid bilayers. Additionally, the peptides do not require prior labeling, and SPR allows for the quantification affinity and kinetic rates in peptide–lipid bilayer binding. Some disadvantages of this method however include firstly, that SPR does not give information on the orientation of the peptide toxins in relation MANUSCRIPTto the lipid bilayers. Secondly, the kinetic experiments are lengthy, and available instruments only allow the study of a few peptides at a time during affinity assays making the process semi high-throughput. Furthermore, the cost of sensor chips, coupled with experience in our lab showing that some GMTs bind so strongly to the sensor chip that it is impossible to wash them off rendering the sensor chip non-reusable, limits the cost-effectiveness of the SPR technique.

4.3. Fluoro-spectrophotometricACCEPTED techniques So far, all GMTs whose interactions with the lipid membrane have been studied contain one or more conserved Trp residues in their sequence (Figure 1C). As mentioned in section

2.1 and 2.2, the fluorescence emission properties of GMT Trp residues can be followed upon titration of the GMTs with LUVs (Figure 5C), or using quenchers of fluorescence emission to

22 ACCEPTED MANUSCRIPT quantify the depth to which Trp residues on GMTs insert into lipid bilayers (Figure 5D)

(Henriques et al., 2012; Henriques et al., 2009; Lakowicz, 2006; Santos et al., 2003). The advantages of the fluorescence techniques mentioned is that these methods provide for sensitive analysis of the peptide interactions with the model membranes allowing for the use of low concentrations of the samples. Furthermore, because of the intrinsic fluorescent properties of the Trp residues, there is no need to label the peptide toxins for these experiments (Matos et al., 2010). Some disadvantages are that GMTs typically contain more than one Trp residue and it is not possible to distinguish which Trp residue/s is/are responsible for the observed spectral behavior without mutagenesis studies. Furthermore, the exact positioning of non-fluorescent amino acid residues in relation to model membranes cannot be monitored using these techniques.

4.4. Molecular dynamics Molecular dynamics (MD), including atomistic MANUSCRIPT and coarse-grained simulations, have been used to model the binding of GMTs to lipid bilayers (Bemporad et al., 2006; Deplazes et al., 2016; Henriques et al., 2016; Nishizawa, 2011; Nishizawa et al., 2015; Nishizawa and

Nishizawa, 2006, 2007; Wee et al., 2007; Wee et al., 2008). MD studies have recently been used alongside experimental studies and have supported the findings that GMTs, which bind to model membranes are likely to have a shallow location at the membrane surface (Deplazes et al., 2016; Henriques et al., 2016). A limitation of MD studies is that unlike nanodiscs and liposomes that ACCEPTED can be made from complex mixtures of lipids, simulation time constraints have so far, limited MD experiments to model membranes made from only one or two different phospholipids (Bemporad et al., 2006; Deplazes et al., 2016; Henriques et al., 2016;

Nishizawa, 2011; Nishizawa et al., 2015; Nishizawa and Nishizawa, 2006, 2007; Wee et al.,

23 ACCEPTED MANUSCRIPT 2007; Wee et al., 2008). Furthermore, to date, no MD studies have examined all three components of the trimolecular complex simultaneously.

4.5. Nuclear magnetic resonance

NMR allows for the identification of specific amino acid residues that are involved in the

interactions between peptides and model membranes by identifying changes in chemical

shifts and in the intensities of resonances of the residues (Bieri et al., 2011; Jung et al., 2010;

Lau et al., 2016; Nakanishi et al., 2002).

1 1D H NMR spectra have been used to study Hm-3, a Na V inhibitor, in its interaction with

POPC and POPC/1,2-dioleoyl-sn -glycero-3-phosphoglycerol (DOPG) (3:1 molar ratio)

model membranes (Figure 5E) (Berkut et al., 2015). The intensities of the peaks in the amide

region of Hm-3 decreased upon titration with model membranes and the authors correlated

this to an increase in binding between the peptide and the model membranes (Berkut et al., 2015). A disadvantage of this method however MANUSCRIPTis that there was also a broadening of the peaks in the amide region of the 1D 1H NMR spectrum of Hm-3 upon lipid titration, and this presents the possibility that rather than interacting with the membrane, the GMT could have been forming aggregates (Berkut et al., 2015; Kwan et al., 2011).

NMR transferred cross-saturation experiments on SgTx-1 involved the transfer of saturation of the resonances from irradiated POPC SUVs to uniformly labelled 2H-15 N SgTx-

1 by cross relaxation at the interaction interface such that the resonances of amino acids

1 15 interacting withACCEPTED the membrane were identified by loss in intensities on 2D H- N HSQC spectra (Figure 5F) (Jung et al., 2010; Nakanishi et al., 2002; Takahashi et al., 2000b). The

use of transferred cross-saturation enabled the examination of weak interactions between the

peptide toxin and the model membrane, a feature which is beneficial to study the superficial

relationship evident between GMTs and lipid membranes. However, this method cannot be

24 ACCEPTED MANUSCRIPT used for lipid systems that would show stronger interactions with the peptide toxins because the validity of effective transferred cross-saturation from the lipids to the free peptide is dependent on the kinetics of the interaction and requires a fast dissociation rate (Jayalakshmi and Krishna, 2002; Jung et al., 2010; Nakanishi et al., 2002).

NMR chemical shift mapping was recently used to study VsTx-1 and its relationship with

model membranes as well as interactions between the peptide toxin and the K VAP VSD stabilized by micelles (Lau et al., 2016). Chemical shift mapping experiments allow for the identification of peptide toxin amino acid residues interacting with the target (e.g. model membrane or VSD) because the resonances from these residues are expected to show changes to their chemical shifts and intensities (Figure 5G) (Bieri et al., 2011). The experiments use

15 N labelled ligands and unlabeled targets such that the target remains invisible on 15 N HSQC

spectra, but the effect of ligand–target interaction is visible as changes to the spectrum of the

labelled ligand. The advantage of chemical shift mapping over transferred cross-saturation experiments is that NMR chemical shift mapping MANUSCRIPT takes into account the relative strength of interaction between the peptide toxin and the target (model membrane or VSD).

Disadvantages of NMR chemical shift mapping are that the technique requires isotopic

labelled 15 N ligand for the assignment of the spectra in the free state and in the case of strong

interactions, additional assignment of the spectra in the bound state (Bieri et al., 2011).

4.6. Cryo-electron microscopy Cryo-EMACCEPTED involves the use of cryogenic temperatures to obtain images of biological structures using transmission emission microscopy (Figure 5H) (Milne et al., 2013). Several images of the structures in different orientations are obtained, and computational image processing and particle picking is applied to build a 3D model based on an average of the processed images (Milne et al., 2013). A recent study has used cryo-EM to delineate the

25 ACCEPTED MANUSCRIPT binding of DkTx to TRPV1, its target ion channel, embedded in a nanodisc (Gao et al., 2016).

DkTx is a bivalent GMT containing two ICK motifs joined with a seven amino acid long chain linker (Bohlen et al., 2010). The cryo-EM model obtained from the study of DkTx showed the GMT interacting with phospholipids involved in the stabilization of the TRPV1 ion channel (Gao et al., 2016). The use of cryo-EM has some disadvantages, including the requirement for highly specialized skills, expensive equipment, and an inability of this technique to follow the kinetics of peptide–lipid interactions using a single sample. However, the study by Gao et al. (2016), which was the first high-resolution depiction of the trimolecular complex including a GMT, a transmembrame protein and a model membrane in concert, highlights the potential benefits of using cryo-EM in the study of GMT–lipid bilayer interactions in the future.

5. Conclusion This review has examined the concept of aMANUSCRIPT trimolecular relationship between GMTs extracted from spider venoms, voltage-gated ion channels, and the lipid membrane. The

studies thus far suggest that the lipid membrane acts to anchor and orient some GMTs for

optimal interactions with voltage-gated ion channels and that the interactions between GMTs

and the lipid membrane appear to be superficial and primarily electrostatic in nature (Agwa et

al., 2017; Deplazes et al., 2016; Henriques et al., 2016; Jung et al., 2010; Lau et al., 2016;

Lee and MacKinnon, 2004; Posokhov et al., 2007; Revell Phillips et al., 2005). Furthermore, initial evidence ACCEPTEDsuggests that increasing the interactions between these peptide toxins with the membrane can be exploited to optimize the inhibitory potency of the GMTs (Agwa et al.,

2017; Henriques et al., 2016; Revell et al., 2013).

A major drawback to the available information we have on the trimolecular complex is

that most of the studies have studied GMTs and model membranes without the voltage-gated

26 ACCEPTED MANUSCRIPT ion channels. However, evidence from recent studies showing VsTx-1 interacting with KVAP

VSD stabilized in micelles, and the atomistic details of DkTx interacting with phospholipids

that stabilize the TRPV1 ion channel embedded in nanodiscs represent evidence that we have

the technology that will enable the study of each of the three components within the

trimolecular complex at the concurrently in more physiologically valid model membranes

(Gao et al., 2016; Lau et al., 2016).

The trimolecular complex is an intriguing concept, which if proven to be existent and relevant, will contribute to a shift from the typical lock-and-key approach to drug design and allow for the consideration of the role of the lipid membrane in rational drug design.

6. Author contributions and acknowledgments

The authors’ work on gating modifier toxins and lipid membranes is supported by a project grant awarded to C.I.S and S.T.H (APP1080405) from the National Health and Medical Research Council (NHMRC). S.T.H isMANUSCRIPT an Australian Research Council (ARC) Future Fellow (FT150100398) and an Institute for Molecular Bioscience (IMB) Fellow, C.I.S is an ARC Future Fellow (FT160100055) and an IMB Industry Fellow and A.J.A is supported by a University of Queensland International postgraduate student scholarship.

ACCEPTED

27 ACCEPTED MANUSCRIPT 7. References

Agwa, A. J., Lawrence, N., Deplazes, E., Cheneval, O., Chen, R., Craik, D. J., Schroeder, C.

I., Henriques, S. T., 2017. Spider peptide toxin HwTx-IV engineered to bind to lipid membranes has an increased inhibitory potency at human voltage-gated sodium channel hNa V1.7. Biochim. Biophys. Acta 1859, 835-844.

Ahuja, S., Mukund, S., Deng, L., Khakh, K., Chang, E., Ho, H., Shriver, S., Young, C., Lin,

S., Johnson, J. P., Wu, P., Li, J., Coons, M., Tam, C., Brillantes, B., Sampang, H., Mortara,

K., Bowman, K. K., Clark, K. R., Estevez, A., Xie, Z., Verschoof, H., Grimwood, M.,

Dehnhardt, C., Andrez, J.-C., Focken, T., Sutherlin, D. P., Safina, B. S., Starovasnik, M. A.,

Ortwine, D. F., Franke, Y., Cohen, C. J., Hackos, D. H., Koth, C. M., Payandeh, J., 2015.

Structural basis of Na V1.7 inhibition by an isoform-selective small-molecule antagonist.

Science (New York, N.Y.) 350, aac5464.

Armstrong, C. M., Bezanilla, F., 1974. Charge movement associated with the opening and closing of the activation gates of the Na V channels. MANUSCRIPT J. Gen. Physiol. 63, 533-552. Bae, C., Anselmi, C., Kalia, J., Jara-Oseguera, A., Schwieters, C. D., Krepkiy, D., Won Lee,

C., Kim, E. H., Kim, J. I., Faraldo-Gomez, J. D., Swartz, K. J., 2016. Structural insights into

the mechanism of activation of the TRPV1 channel by a membrane-bound tarantula toxin.

eLife 5, e11273.

Bayburt, T. H., Sligar, S. G., 2003. Self-assembly of single integral membrane proteins into

soluble nanoscale phospholipid bilayers. Protein Sci. 12, 2476-2481. Bayburt, T. H., Sligar,ACCEPTED S. G., 2010. Membrane protein assembly into nanodiscs. FEBS Lett. 584, 1721-1727.

Bemporad, D., Sands, Z. A., Wee, C. L., Grottesi, A., Sansom, M. S., 2006. VsTx-1, a

modifier of K V channel gating, localizes to the interfacial region of lipid bilayers.

Biochemistry 45, 11844-11855.

28 ACCEPTED MANUSCRIPT Berkut, A. A., Peigneur, S., Myshkin, M. Y., Paramonov, A. S., Lyukmanova, E. N.,

Arseniev, A. S., Grishin, E. V., Tytgat, J., Shenkarev, Z. O., Vassilevski, A. A., 2015.

Structure of membrane-active toxin from crab spider Heriaeus melloteei suggests parallel evolution of sodium channel gating modifiers in Araneomorphae and Mygalomorphae. J.

Biol. Chem. 290, 492-504.

Bezanilla, F., 2000. The voltage sensor in voltage-dependent ion channels. Physiol. Rev. 80,

555-592.

Bezanilla, F., 2002. Voltage sensor movements. J. Gen. Physiol. 120, 465-473.

Bieri, M., Kwan, A. H., Mobli, M., King, G. F., Mackay, J. P., Gooley, P. R., 2011.

Macromolecular NMR spectroscopy for the non-spectroscopist: beyond macromolecular solution structure determination. FEBS J. 278, 704-715.

Bohlen, C. J., Priel, A., Zhou, S., King, D., Siemens, J., Julius, D., 2010. A bivalent tarantula toxin activates the capsaicin receptor, TRPV1, by targeting the outer pore domain. Cell 141, 834-845. MANUSCRIPT Bosmans, F., Martin-Eauclaire, M.-F., Swartz, K. J., 2008. Deconstructing voltage sensor function and pharmacology in sodium channels. Nature 456, 202-208.

Bosmans, F., Rash, L., Zhu, S., Diochot, S., Lazdunski, M., Escoubas, P., Tytgat, J., 2006.

Four novel tarantula toxins as selective modulators of voltage-gated sodium channel subtypes. Mol. Pharmacol. 69, 419-429.

Bosmans, F., Swartz, K. J., 2010. Targeting sodium channel voltage sensors with spider toxins. Trends Pharmacol.ACCEPTED Sci. 31, 175-182. Cai, T., Luo, J., Meng, E., Ding, J., Liang, S., Wang, S., Liu, Z., 2015. Mapping the interaction site for the tarantula toxin hainantoxin-IV (beta-TRTX-Hn2a) in the voltage sensor module of domain II of voltage-gated sodium channels. Peptides 68, 148-156.

29 ACCEPTED MANUSCRIPT Catterall, W. A., 2010. Ion channel voltage sensors: structure, function, and pathophysiology.

Neuron 67, 915-928.

Catterall, W. A., 2012. Voltage-gated sodium channels at 60: structure, function and pathophysiology. J. Physiol. 590, 2577-2589.

Catterall, W. A., Cestele, S., Yarov-Yarovoy, V., Yu, F. H., Konoki, K., Scheuer, T., 2007.

Voltage-gated ion channels and gating modifier toxins. Toxicon 49, 124-141.

Cooper, M. A., Hansson, A., Löfås, S., Williams, D. H., 2000. A vesicle capture sensor chip for kinetic analysis of interactions with membrane-bound receptors. Anal. Biochem. 277,

196-205.

Deplazes, E., Henriques, S. T., Smith, J. J., King, G. F., Craik, D. J., Mark, A. E., Schroeder,

C. I., 2016. Membrane-binding properties of gating modifier and pore-blocking toxins:

Membrane interaction is not a prerequisite for modification of channel gating. Biochim.

Biophys. Acta 1858, 872-882. DeVries, G. H., 1984. The axonal plasma membrane. MANUSCRIPT Plenum Press, New York. 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.

Dörr, J. M., Scheidelaar, S., Koorengevel, M. C., Dominguez, J. J., Schäfer, M., van Walree,

C. A., Killian, J. A., 2016. The styrene–maleic acid copolymer: a versatile tool in membrane research. Eur. Biophys. J. 45, 3-21.

Flinspach, M., Xu, Q., Piekarz, A. D., Fellows, R., Hagan, R., Gibbs, A., Liu, Y., Neff, R. A., Freedman, J., Eckert,ACCEPTED W. A., Zhou, M., Bonesteel, R., Pennington, M. W., Eddinger, K. A., Yaksh, T. L., Hunter, M., Swanson, R. V., Wickenden, A. D., 2017. Insensitivity to pain induced by a potent selective closed-state Na V1.7 inhibitor. Sci. Rep. 7, 39662.

Gao, Y., Cao, E., Julius, D., Cheng, Y., 2016. TRPV1 structures in nanodiscs reveal mechanisms of ligand and lipid action. Nature 534, 347-351.

30 ACCEPTED MANUSCRIPT Gennis, R. B., 1989. Biomembranes. In: Cantor, C. R., (Ed), Molecular structure and function. Springer Science+Business Media, New York, p. 533.

Gluck, J. M., Koenig, B. W., Willbold, D., 2011. Nanodiscs allow the use of integral membrane proteins as analytes in surface plasmon resonance studies. Anal. Biochem. 408,

46-52.

Guy, H. R., Seetharamulu, P., 1986. Molecular model of the action potential sodium channel.

Proc. Natl. Acad. Sci. USA 83, 508-512.

Henriques, S. T., Deplazes, E., Lawrence, N., Cheneval, O., Chaousis, S., Inserra, M.,

Thongyoo, P., King, G. F., Mark, A. E., Vetter, I., Craik, D. J., Schroeder, C. I., 2016.

Interaction of tarantula venom peptide ProTx-II with lipid membranes is a prerequisite for its inhibition of human voltage-gated sodium channel NaV1.7. J. Biol. Chem. 29, 17049-17065.

Henriques, S. T., Huang, Y. H., Castanho, M. A., Bagatolli, L. A., Sonza, S., Tachedjian, G.,

Daly, N. L., Craik, D. J., 2012. Phosphatidylethanolamine binding is a conserved feature of cyclotide-membrane interactions. J. Biol. Chem. 287MANUSCRIPT, 33629-33643. Henriques, S. T., Pattenden, L. K., Aguilar, M.-I., Castanho, M. A. R. B., 2008. PrP(106-126)

does not interact with membranes under physiological conditions. Biophys. J. 95, 1877-1889.

Henriques, S. T., Pattenden, L. K., Aguilar, M.-I., Castanho, M. A. R. B., 2009. The toxicity

of prion protein fragment PrP(106−126) is not mediated by membrane permeabilization as

shown by a M112W substitution. Biochemistry 48, 4198-4208.

Herzig, V., King, G. F., 2015. The cystine knot is responsible for the exceptional stability of the insecticidal spiderACCEPTED toxin omega-hexatoxin-Hv1a. Toxins 7, 4366-4380. Hille, B., 2001. Ion channels of excitable membranes. In: Inc, S. A., (Ed). Sinauer Associates

Inc, Sunderland, Massachusetts USA, p. 788.

31 ACCEPTED MANUSCRIPT Jayalakshmi, V., Krishna, N. R., 2002. Complete relaxation and conformational exchange matrix (CORCEMA) analysis of intermolecular saturation transfer effects in reversibly forming ligand-receptor complexes. J. Magn. Reson. 155, 106-118.

Jiang, Y., Lee, A., Chen, J., Ruta, V., Cadene, M., Chait, B. T., MacKinnon, R., 2003. X-ray structure of a voltage-dependent K + channel. Nature 423, 33-41.

Jung, H. H., Jung, H. J., Milescu, M., Lee, C. W., Lee, S., Lee, J. Y., Eu, Y.-J., Kim, H. H.,

Swartz, K. J., Kim, J. I., 2010. Structure and orientation of a voltage-sensor toxin in lipid

membranes. Biophys. J. 99, 638-646.

Jung, H. J., Lee, J. Y., Kim, S. H., Eu, Y. J., Shin, S. Y., Milescu, M., Swartz, K. J., Kim, J.

I., 2005. Solution structure and lipid membrane partitioning of VsTx1, an inhibitor of the

KVAP potassium channel. Biochemistry 44, 6015-6023.

Killian, J. A., von Heijne, G., 2000. How proteins adapt to a membrane-water interface.

Trends Biochem. Sci. 25, 429-434. King, G. F., Tedford, H. W., Maggio, F., 2002. Stru MANUSCRIPTcture and function of insecticidal neurotoxins from Australian funnel-web spiders. J. Toxicol.: Toxin Rev. 21, 361-389.

Klint, J. K., Smith, J. J., Vetter, I., Rupasinghe, D. B., Er, S. Y., Senff, S., Herzig, V., Mobli,

M., Lewis, R. J., Bosmans, F., King, G. F., 2015. Seven novel modulators of the analgesic target Na V1.7 uncovered using a high-throughput venom-based discovery approach. Br. J.

Pharmacol. 172, 2445-2458.

Kwan, A. H., Mobli, M., Gooley, P. R., King, G. F., Mackay, J. P., 2011. Macromolecular NMR spectroscopyACCEPTED for the non-spectroscopist. FEBS J. 278, 687-703. Kwong, K., Carr, M. J., 2015. Voltage-gated sodium channels. Curr. Opin. Pharm. 22, 131-

139.

Ladokhin, A. S., Jayasinghe, S., White, S. H., 2000. How to measure and analyze tryptophan

fluorescence in membranes properly, and why bother? Anal. Biochem. 285, 235-245.

32 ACCEPTED MANUSCRIPT Lakowicz, J. R., 2006. Principles of fluorescence spectroscopy. Springer Science+Business

Media, New York, NY.

Lau, C. H. Y., King, G. F., Mobli, M., 2016. Molecular basis of the interaction between gating modifier spider toxins and the voltage sensor of voltage-gated ion channels. Sci. Rep.

6, 34333.

Lee, C. W., Kim, S., Roh, S. H., Endoh, H., Kodera, Y., Maeda, T., Kohno, T., Wang, J. M.,

Swartz, K. J., Kim, J. I., 2004. Solution structure and functional characterization of SGTx1, a modifier of K V2.1 channel gating. Biochemistry 43, 890-897.

Lee, S.-Y., MacKinnon, R., 2004. A membrane-access mechanism of ion channel inhibition by voltage sensor toxins from spider venom. Nature 430, 232-235.

Leitz, A. J., Bayburt, T. H., Barnakov, A. N., Springer, B. A., Sligar, S. G., 2006. Functional reconstitution of Beta2-adrenergic receptors utilizing self-assembling nanodisc technology.

BioTechniques 40, 601-612. Lewis, R. J., Vetter, I., Cardoso, F. C., Inserra, M., MANUSCRIPT King, G., 2015. Does nature do ion channel drug discovery better than us? In: Cox, B., Gosling, M., (Eds), Ion Channel Drug

Discovery. The Royal Society of Chemistry, pp. 297-319.

Li, D., Xiao, Y., Xu, X., Xiong, X., Lu, S., Liu, Z., Zhu, Q., Wang, M., Gu, X., Liang, S.,

2004. Structure-activity relationships of hainantoxin-IV and structure determination of active and inactive sodium channel blockers. J. Biol. Chem. 279, 37734-37740.

Li-Smerin, Y., Swartz, K. J., 1998. Gating modifier toxins reveal a conserved structural motif

2+ + in voltage-gatedACCEPTED Ca and K channels. Proc. Natl. Acad. Sci. USA 95, 8585-8589. Liu, Y., Li, D., Wu, Z., Li, J., Nie, D., Xiang, Y., Liu, Z., 2012. A positively charged surface

patch is important for hainantoxin-IV binding to voltage-gated sodium channels. J. Pept. Sci.

18, 643-649.

33 ACCEPTED MANUSCRIPT Liu, Z., Cai, T., Zhu, Q., Deng, M., Li, J., Zhou, X., Zhang, F., Li, D., Li, J., Liu, Y., Hu, W.,

Liang, S., 2013. Structure and function of hainantoxin-III, a selective antagonist of neuronal tetrodotoxin-sensitive voltage-gated sodium Channels isolated from the chinese bird spider

Ornithoctonus hainana . J. Biol. Chem. 288, 20392-20403.

Long, S. B., Campbell, E. B., Mackinnon, R., 2005. Crystal structure of a mammalian voltage-dependent shaker family K + channel. Science (New York, N.Y.) 309, 897-903.

Marvin, L., De, E., Cosette, P., Gagnon, J., Molle, G., Lange, C., 1999. Isolation, amino acid

sequence and functional assays of SgTx-1. The first toxin purified from the venom of the

spider scodra griseipes. Eur. J. Biochem. 265, 572-579.

Matos, P. M., Franquelim, H. G., Castanho, M. A., Santos, N. C., 2010. Quantitative

assessment of peptide-lipid interactions. Ubiquitous fluorescence methodologies. Biochim.

Biophys. Acta 1798, 1999-2012.

Middleton, R. E., Warren, V. A., Kraus, R. L., Hwang, J. C., Liu, C. J., Dai, G., Brochu, R. M., Kohler, M. G., Gao, Y.-D., Garsky, V. M., Bogus MANUSCRIPTky, M. J., Mehl, J. T., Cohen, C. J., Smith, M. M., 2002. Two tarantula peptides inhibit activation of multiple sodium channels.

Biochemistry 41, 14734-14747.

Mihailescu, M., Krepkiy, D., Milescu, M., Gawrisch, K., Swartz, K. J., White, S., 2014.

Structural interactions of a voltage sensor toxin with lipid membranes. Proc. Natl. Acad. Sci.

USA 111, E5463-E5470.

Milescu, M., Bosmans, F., Lee, S., Alabi, A. A., Kim, J. I., Swartz, K. J., 2009. Interactions between lipids andACCEPTED voltage sensor paddles detected with tarantula toxins. Nat. Struct. Mol. Biol. 16, 1080-1085.

Milescu, M., Vobecky, J., Roh, S. H., Kim, S. H., Jung, H. J., Kim, J. I., Swartz, K. J., 2007.

Tarantula toxins interact with voltage sensors within lipid membranes. J. Gen. Physiol. 130,

497-511.

34 ACCEPTED MANUSCRIPT Milne, J. L., Borgnia, M. J., Bartesaghi, A., Tran, E. E., Earl, L. A., Schauder, D. M.,

Lengyel, J., Pierson, J., Patwardhan, A., Subramaniam, S., 2013. Cryo-electron microscopy: a primer for the non-microscopist. FEBS J. 280, 28-45.

Minassian, N. A., Gibbs, A., Shih, A. Y., Liu, Y., Neff, R. A., Sutton, S. W., Mirzadegan, T.,

Connor, J., Fellows, R., Husovsky, M., Nelson, S., Hunter, M. J., Flinspach, M., Wickenden,

A. D., 2013. Analysis of the structural and molecular basis of voltage-sensitive sodium channel inhibition by the spider toxin huwentoxin-IV (mu-TRTX-Hh2a). J. Biol. Chem. 288,

22707-22720.

Murray, J. K., Biswas, K., Holder, J. R., Zou, A., Ligutti, J., Liu, D., Poppe, L., Andrews, K.

L., Lin, F. F., Meng, S. Y., Moyer, B. D., McDonough, S. I., Miranda, L. P., 2015a.

Sustained inhibition of the Na V1.7 sodium channel by engineered dimers of the domain II

binding peptide GpTx-1. Bioorg. Med. Chem. Lett. 25, 4866-4871.

Murray, J. K., Ligutti, J., Liu, D., Zou, A., Poppe, L., Li, H., Andrews, K. L., Moyer, B. D., McDonough, S. I., Favreau, P., Stöcklin, R., Mirand MANUSCRIPTa, L. P., 2015b. Engineering potent and

selective analogues of GpTx-1, a tarantula venom peptide antagonist of the Na V1.7 sodium

channel. J. Med. Chem. 58, 2299-2314.

Murray, J. K., Long, J., Zou, A., Ligutti, J., Andrews, K. L., Poppe, L., Biswas, K., Moyer, B.

D., McDonough, S. I., Miranda, L. P., 2016. Single residue substitutions that confer voltage-

gated sodium ion channel subtype selectivity in the Na V1.7 inhibitory peptide GpTx-1. J.

Med. Chem. 59, 2704-2717. Murray, J. K., Qian,ACCEPTED Y. X., Liu, B., Elliott, R., Aral, J., Park, C., Zhang, X., Stenkilsson, M., Salyers, K., Rose, M., Li, H., Yu, S., Andrews, K. L., Colombero, A., Werner, J., Gaida, K.,

Sickmier, E. A., Miu, P., Itano, A., McGivern, J., Gegg, C. V., Sullivan, J. K., Miranda, L. P.,

2015c. Pharmaceutical optimization of peptide toxins for ion channel targets: potent,

selective, and long-lived antagonists of K V1.3. J. Med. Chem. 58, 6784-6802.

35 ACCEPTED MANUSCRIPT Nakanishi, T., Miyazawa, M., Sakakura, M., Terasawa, H., Takahashi, H., Shimada, I., 2002.

Determination of the interface of a large protein complex by transferred cross-saturation measurements. J. Mol. Biol. 318, 245-249.

Nishizawa, K., 2011. Atomistic molecular simulation of gating modifier venom peptides–two binding modes and effects of lipid structure. Mechanosensitivity and Mechanotransduction.

Springer, pp. 167-190.

Nishizawa, K., Nishizawa, M., Gnanasambandam, R., Sachs, F., Sukharev, S. I., Suchyna, T.

M., 2015. Effects of Lys to Glu mutations in GsMTx4 on membrane binding, peptide orientation, and self-association propensity, as analyzed by molecular dynamics simulations.

Biochim. Biophys. Acta 1848, 2767-2778.

Nishizawa, M., Nishizawa, K., 2006. Interaction between K + channel gate modifier hanatoxin

and lipid bilayer membranes analyzed by molecular dynamics simulation. Eur. Biophys. J.

35, 373-381. Nishizawa, M., Nishizawa, K., 2007. Molecular dynamMANUSCRIPTics simulations of a stretch-activated channel inhibitor GsMTx4 with lipid membranes: two binding modes and effects of lipid

structure. Biophys. J. 92, 4233-4243.

Noda, M., Ikeda, T., Suzuki, H., Takeshima, H., Takahashi, T., Kuno, M., Numa, S., 1986.

Expression of functional sodium channels from cloned cDNA. Nature 322, 826-828.

Noda, M., Numa, S., 1987. Structure and function of sodium channel. J. Recept. Res. 7, 467-

497. Opekarová, M., ACCEPTEDTanner, W., 2003. Specific lipid requirements of membrane proteins—a putative bottleneck in heterologous expression. Biochim. Biophys. Acta 1610, 11-22.

Osteen, J. D., Herzig, V., Gilchrist, J., Emrick, J. J., Zhang, C., Wang, X., Castro, J., Garcia-

Caraballo, S., Grundy, L., Rychkov, G. Y., Weyer, A. D., Dekan, Z., Undheim, E. A. B.,

Alewood, P., Stucky, C. L., Brierley, S. M., Basbaum, A. I., Bosmans, F., King, G. F., Julius,

36 ACCEPTED MANUSCRIPT

D., 2016. Selective spider toxins reveal a role for the Na V1.1 channel in mechanical pain.

Nature 534, 494-499.

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

Palsdottir, H., Hunte, C., 2004. Lipids in membrane protein structures. Biochim. Biophys.

Acta 1666, 2-18.

Papazian, D. M., Schwarz, T. L., Tempel, B. L., Jan, Y. N., Jan, L. Y., 1987. Cloning of genomic and complementary DNA from Shaker, a putative potassium channel gene from

Drosophila. Science (New York, N.Y.) 237, 749-753.

Park, J. H., Carlin, K. P., Wu, G., Ilyin, V. I., Musza, L. L., Blake, P. R., Kyle, D. J., 2014.

Studies examining the relationship between the chemical structure of protoxin II and its activity on voltage gated sodium channels. J. Med. Chem. 57, 6623-6631. Payandeh, J., Gamal El-Din, T. M., Scheuer, T., ZheMANUSCRIPTng, N., Catterall, W. A., 2012. Crystal structure of a voltage-gated sodium channel in two potentially inactivated states. Nature 486,

135-139.

Posokhov, Y. O., Gottlieb, P. A., Morales, M. J., Sachs, F., Ladokhin, A. S., 2007. Is lipid bilayer binding a common property of inhibitor cysteine knot ion-channel blockers? Biophys.

J. 93, L20-22.

Revell, J. D., Lund, P. E., Linley, J. E., Metcalfe, J., Burmeister, N., Sridharan, S., Jones, C., Jermutus, L., Bednarek,ACCEPTED M. A., 2013. Potency optimization of huwentoxin-IV on hNa V1.7: a neurotoxin TTX-S sodium-channel antagonist from the venom of the Chinese bird-eating

spider Selenocosmia huwena . Peptides 44, 40-46.

Revell Phillips, L., Milescu, M., Li-Smerin, Y., Mindell, J. A., Kim, J. I., Swartz, K. J., 2005.

Voltage-sensor activation with a tarantula toxin as cargo. Nature 436, 857-860.

37 ACCEPTED MANUSCRIPT Ruta, V., Jiang, Y., Lee, A., Chen, J., MacKinnon, R., 2003. Functional analysis of an archaebacterial voltage-dependent K + channel. Nature 422, 180-185.

Salari, A., Vega, B. S., Milescu, L. S., Milescu, M., 2016. Molecular interactions between

tarantula toxins and low-voltage-activated calcium channels. Sci. Rep. 6, 23894.

Santos, N. C., Prieto, M., Castanho, M. A., 2003. Quantifying molecular partition into model

systems of biomembranes: an emphasis on optical spectroscopic methods. Biochim. Biophys.

Acta 1612, 123-135.

Schmalhofer, W. A., Calhoun, J., Burrows, R., Bailey, T., Kohler, M. G., Weinglass, A. B.,

Kaczorowski, G. J., Garcia, M. L., Koltzenburg, M., Priest, B. T., 2008. ProTx-II, a selective

inhibitor of Na V1.7 sodium channels, blocks action potential propagation in nociceptors. Mol.

Pharmacol. 74, 1476-1484.

Shcherbatko, A., Rossi, A., Foletti, D., Zhu, G., Bogin, O., Galindo-Casas, M., Rickert, M.,

Hasa-Moreno, A., Bartevitch, V., Crameri, A., Steiner, A. R., Henningsen, R., Gill, A., Pons, J., Shelton, D. L., Rajpal, A., Strop, P., 2016. En gineeringMANUSCRIPT highly potent and selective microproteins against Na V1.7 sodium channel for treatment of pain. J. Biol. Chem. 291,

13974-13986.

Shenkarev, Z. O., Lyukmanova, E. N., Paramonov, A. S., Panteleev, P. V., Balandin, S. V.,

Shulepko, M. A., Mineev, K. S., Ovchinnikova, T. V., Kirpichnikov, M. P., Arseniev, A. S.,

2014. Lipid-protein nanodiscs offer new perspectives for structural and functional studies of water-soluble membrane-active peptides. Acta naturae 6, 84-94. Smith, J. J., Alphy,ACCEPTED S., Seibert, A. L., Blumenthal, K. M., 2005. Differential phospholipid binding by site 3 and site 4 toxins. Implications for structural variability between voltage- sensitive sodium channel domains. J. Biol. Chem. 280, 11127-11133.

38 ACCEPTED MANUSCRIPT Stith, B. J., Hall, J., Ayres, P., Waggoner, L., Moore, J. D., Shaw, W. A., 2000.

Quantification of major classes of Xenopus phospholipids by high performance liquid

chromatography with evaporative light scattering detection. J. Lipid Res. 41, 1448-1454.

Suchyna, T. M., Tape, S. E., Koeppe, R. E., 2nd, Andersen, O. S., Sachs, F., Gottlieb, P. A.,

2004. Bilayer-dependent inhibition of mechanosensitive channels by neuroactive peptide

enantiomers. Nature 430, 235-240.

Swartz, K. J., MacKinnon, R., 1995. An inhibitor of the K V2.1 potassium channel isolated

from the venom of a Chilean tarantula. Neuron 15, 941-949.

Swartz, K. J., MacKinnon, R., 1997a. Hanatoxin modifies the gating of a voltage-dependent

K+ channel through multiple binding sites. Neuron 18, 665-673.

Swartz, K. J., MacKinnon, R., 1997b. Mapping the receptor site for hanatoxin, a gating modifier of voltage-dependent K + channels. Neuron 18, 675-682.

Takahashi, H., Kim, J. I., Min, H. J., Sato, K., Swartz, K. J., Shimada, I., 2000a. Solution structure of hanatoxin1, a gating modifier of volta MANUSCRIPTge-dependent K + channels: common surface features of gating modifier toxins. J. Mol. Biol. 297, 771-780.

Takahashi, H., Nakanishi, T., Kami, K., Arata, Y., Shimada, I., 2000b. A novel NMR method

for determining the interfaces of large protein-protein complexes. Nat. Struct. Mol. Biol. 7,

220-223.

Torcato, I. M., Huang, Y. H., Franquelim, H. G., Gaspar, D., Craik, D. J., Castanho, M. A.,

Troeira Henriques, S., 2013. Design and characterization of novel , R- BP100 and RW-BP100,ACCEPTED with activity against Gram-negative and Gram-positive bacteria. Biochim. Biophys. Acta 1828, 944-955.

Vabnick, I., Novakovic, S. D., Levinson, S. R., Schachner, M., Shrager, P., 1996. The

clustering of axonal sodium channels during development of the peripheral nervous system.

J. Neurosci. 16, 4914-4922.

39 ACCEPTED MANUSCRIPT Vabnick, I., Trimmer, J. S., Schwarz, T. L., Levinson, S. R., Risal, D., Shrager, P., 1999.

Dynamic potassium channel distributions during axonal development prevent aberrant firing patterns. J. Neurosci. 19, 747-758. van der Merwe, P. A., 2001. Surface plasmon resonance. Oxford University Press: New

York, NY, USA, pp. 137-170. van Meer, G., de Kroon, A. I. P. M., 2011. Lipid map of the mammalian cell. J. Cell Sci. 124,

5-8. van Meer, G., Voelker, D. R., Feigenson, G. W., 2008. Membrane lipids: where they are and how they behave. Nat. Rev. Mol. Cell Biol. 9, 112-124.

Wang, J. M., Roh, S. H., Kim, S., Lee, C. W., Kim, J. I., Swartz, K. J., 2004. Molecular surface of tarantula toxins interacting with voltage sensors in K V channels. J. Gen. Physiol.

123, 455-467.

Wee, C. L., Bemporad, D., Sands, Z. A., Gavaghan, D., Sansom, M. S. P., 2007. SgTx-1, a

KV channel gating-modifier toxin, binds to the interf MANUSCRIPTacial region of lipid bilayers. Biophys. J. 92, L07-L09.

Wee, C. L., Gavaghan, D., Sansom, M. S. P., 2008. Lipid bilayer deformation and the free

energy of interaction of a K V channel gating-modifier toxin. Biophys. J. 95, 3816-3826.

Whorton, M. R., Bokoch, M. P., Rasmussen, S. G. F., Huang, B., Zare, R. N., Kobilka, B.,

Sunahara, R. K., 2007. A monomeric G protein-coupled receptor isolated in a high-density lipoprotein particle efficiently activates its G protein. Proc. Natl. Acad. Sci. 104, 7682-7687. Xiao, Y., Bingham,ACCEPTED J. P., Zhu, W., Moczydlowski, E., Liang, S., Cummins, T. R., 2008a. Tarantula huwentoxin-IV inhibits neuronal sodium channels by binding to receptor site 4 and trapping the domain II voltage sensor in the closed configuration. J. Biol. Chem. 283, 27300-

27313.

40 ACCEPTED MANUSCRIPT Xiao, Y., Blumenthal, K., Jackson, J. O., Liang, S., Cummins, T. R., 2010. The tarantula toxins ProTx-II and huwentoxin-IV differentially interact with human Na V1.7 voltage sensors to inhibit channel activation and inactivation. Mol. Pharmacol. 78, 1124-1134.

Xiao, Y., Jackson, J. O., Liang, S., Cummins, T. R., 2011. Common molecular determinants of tarantula huwentoxin-IV Inhibition of Na + channel voltage aensors in domains II and IV. J.

Biol. Chem. 286, 27301-27310.

Xiao, Y., Luo, X., Kuang, F., Deng, M., Wang, M., Zeng, X., Liang, S., 2008b. Synthesis and characterization of huwentoxin-IV, a neurotoxin inhibiting central neuronal sodium channels.

Toxicon 51, 230-239.

MANUSCRIPT

ACCEPTED

41 ACCEPTED MANUSCRIPT

Table 1. Studies on the binding of gating modifier toxins to lipid membranes lipid types Peptide Experiment Results Reference VsTx-1 NMR chemical shift mapping Hydrophobic residues bind to DHPC (Lau et al., 2016) Centrifuge spin-down Binds preferentially to POPG (Lau et al., 2016) Trp fluorescence emission a,b Shallow position in POPC/POPG lipid vesicles (Mihailescu et al., 2014) Centrifuge spin-down Binds to POPG (Jung et al., 2005) Trp fluorescence emission a,b Does not bind strongly to POPC lipid vesicles (Jung et al., 2005) Trp fluorescence emission a,b Binds to POPE/POPG (Lee and MacKinnon, 2004) SgTx-1 NMR transferred cross- Hydrophobic patch interacts with lipid core and (Jung et al., 2010) saturation charged ring with head groups of POPC lipids Trp fluorescence emission a,b Positioned at a shallow position in POPC/POPG (Milescu et al., 2007) lipid vesicles Centrifuge spin-down Binds to membrane of oocytes (Milescu et al., 2007) Trp fluorescence emission b Binds to POPC/POPS (Posokhov et al., 2007) ProTx-I SPR Binds to POPC/POPS lipid vesicles (Deplazes et al., 2016) a,b MANUSCRIPT Trp fluorescence emission and Adopts a shallow position in POPC/POPS lipid (Deplazes et al., 2016) membrane dipolar potential c vesicles ProTx-II SPR Binds preferentially to POPC/POPS, POPC/POPG (Henriques et al., 2016) and POPC/Cer-P over POPC and POPC/SM/CHOL lipid vesicles Trp fluorescence emission a,b and Adopts a shallow position in POPC/POPS lipid (Henriques et al., 2016) membrane dipolar potential c vesicles FRET membrane depolarization d Binds to membrane of neuroblastoma cells (Henriques et al., 2016) Trp fluorescence emission a No significant interaction with POPC lipid (Salari et al., 2016) vesicles; adopts a shallow position in POPC/POPG vesicles Centrifuge spin-down ACCEPTED No interaction with POPC lipid vesicles; binds to (Xiao et al., 2008b) POPE/POPG lipid vesicles Centrifuge spin-down Binds to POPC/POPS lipid vesicles (Smith et al., 2005) HwTx-IV Centrifuge spin-down No binding to POPC or POPE/POPG lipid (Xiao et al., 2008b)

42 ACCEPTED MANUSCRIPT

vesicles SPR Minimal binding to POPC, POPC/POPS, (Agwa et al., 2017) POPC/SM/CHOL and POPC/C1P/CHOL lipid vesicles Trp fluorescence emission a,b Minimal binding to POPC/POPS lipid vesicles (Agwa et al., 2017) Hd1a SPR Minimal binding to POPC and POPC/POPS lipid (Deplazes et al., 2016) vesicles Trp fluorescence emission a and Adopts a shallow position in POPC/POPS lipid (Deplazes et al., 2016) membrane dipolar potential b vesicles HaTx-I Trp fluorescence emission a Positioned at a shallow position in POPC/POPG (Milescu et al., 2007) lipid vesicles Centrifuge spin-down Binds to membrane of oocytes (Milescu et al., 2007) Trp fluorescence emission a,b Positioned at a shallow position in POPC/POPG (Revell Phillips et al., 2005) lipid vesicles GsMTx-4 Trp fluorescence emission a Binds to anionic and zwitterionic lipid vesicles (Posokhov et al., 2007) Trp fluorescence emission a Binds to lipid vesicles of egg lecithin (Suchyna et al., 2004) Hm-3 1D 1H NMR Binds preferentially to anionicMANUSCRIPT lipid vesicles (Berkut et al., 2015) PaTx-1 Trp fluorescence emission a Preferential binding to POPC/POPG over POPC (Salari et al., 2016) lipid vesicles GsAF-I Trp fluorescence emission a Preferential binding to POPC/POPG over POPC (Salari et al., 2016) lipid vesicles GsAF-II Trp fluorescence emission a No binding to POPC/POPG or POPC lipid (Salari et al., 2016) vesicles aTrp fluorescence emission of peptides is followed in the presence or absence of lipids, an increase in quantum yield and leftward (blue) shift in the emission spectrum indicates that Trp residues are partitioning into more hydrophobic environments bTrp fluorescence emission of peptide/lipid suspension is followed in the presence or absence of a quencher, to monitor the depth to which the Trp residues insert into the lipid vesicles cChanges to the membrane dipolar potential in the presence of a peptide are followed by monitoring the excitation spectrum of di-8- ANEPPS dye ACCEPTED dMembrane depolarization of cell membranes in the presence of peptides is monitored using fluorescence resonance energy transfer between nitrobenzoxadiazol-labelled membranes and DisBaC 2 dye

43 ACCEPTED MANUSCRIPT Table 2. Lipid composition of cell membranes relevant to GMTs Cell Lipid a Proportion b (%) Human Phosphatidylethanolamine 19.8 axon c Phosphatidylcholine 18.7 Sphingomyelin 3.0 Phosphatidylserine 3.0 Phosphatidylinositol 4.4 Cholesterol 25.3 Galactolipid (cerebroside and 25.8 sulfatide) Xenopus Phosphatidylcholine 65 laevis Phosphatidylethanolamine 19 oocyte d Sphingomyelin 5 Phosphatidylinositol 10 Phosphatidylserine 2 Cholesterol Cholesterol/phospholipid (mol/mol (0.6–0.7) aPhospholipids are identified in the table by their head groups bProportions are a percentage of dry weight unless otherwise specified c(DeVries, 1984) d(Stith et al., 2000)

MANUSCRIPT

ACCEPTED

44 ACCEPTED MANUSCRIPT Figure 1. Conserved structural features of ProTx-II (PDB ID: 2N9T), SgTx-1 (PDB ID:

1LA4), VsTx-1 (PDB ID: 2NLN) and HwTx-IV (PDB ID: 2M4X). (A) Ribbon representations oriented to show the conserved ICK motif characterized by three disulfide bridges (yellow) and two–three anti-parallel beta sheets in SgTx-1 and VsTx-1, respectively. ProTx-II and HwTx-IV are examples of GMTs that have the ICK motif but do not have anti-parallel beta sheets. NT and CT are the N-terminal and C-terminal, respectively. (B) Surface representations of the GMTs showing the conserved hydrophobic patch and charged ring coloured by residue type (hydrophobic, green; positively charged, blue; negatively charged, red; polar uncharged, white) (C) Sequence alignment of GMTs mentioned in the review. Sequences are aligned along cysteine residues. Trp, Tyr and Lys residues have been identified as important for lipid membrane interactions and are colored green

(Trp and Tyr) and blue (Lys). Residue sequence numbers according to HwTx-IV, HnTx-IV and

Hd1a are shown as well as species names of the spiders from which the peptides were extracted.

Figure 2. The architecture of voltage-gated ion channels. (A) A schematic of the four domains and six transmembrane segments (S1–S6) of a voltageMANUSCRIPT-gated ion channel coloured to differentiate the voltage sensor domains (white) and the pore domain (red). (B,C) Top and side views of a voltage-gated ion channel identifying the pore domain (red) and the four voltage sensor domains

(VSD I–VSD IV) (white). (D) Side view of VSD II is shown highlighting segments S1–S4. S3–S4

(gold) form the paddle motif that facilitates gating of the voltage-gated ion channels.

Figure 3. The trimolecular complex. (A) Sequences of the S3–S4 regions of K V2.1 and Na V1.7; the residues that are targeted by GMTs are highlighted in green (hydrophobic) and red (anionic).

(B) Two possible GMT–lipid membrane–ion channel binding modes with SgTx-1 (PDB ID: 1LA4) used to represent aACCEPTED binding mode where the GMT is nestled between the S1–S2 and S3–S4 loop (Cai et al., 2015; Lau et al., 2016; Minassian et al., 2013); and ProTx-II (PDB ID: 2N9T) used to represent a binding mode where the GMT interacts with the channel and the membrane whilst interacting mainly with the S3–S4 loop (Agwa et al., 2017; Deplazes et al., 2016; Henriques et al.,

2016; Revell Phillips et al., 2005). (C) The proposed overlapping channel binding and membrane

45 binding region of SgTx-1 (PDB AID:CCEPTED 1LA4) (purple) MANUSCRIPT is shown (Jung et al., 2010; Milescu et al.,

2007). (D) The residues that have been proposed to be relevant for the membrane binding of ProTx-

II (PDB ID: 2N9T) form a membrane binding region (green), which is distinct from the residues that are involved in binding to the channel (gold) (Henriques et al., 2016).

Figure 4. Model membranes. Vesicles commonly used in peptide–membrane interaction studies are composed of a single lipid bilayer and are named small unilamellar vesicles (SUVs), large unilamellar vesicles (LUVs) and giant unilamellar vesicles (GUVs), based on their sizes. An example of a transmembrane protein in a micelle (surfactant shown in red) and a nanodisc are also shown.

Figure 5. Techniques used to examine GMT–lipid bilayer interactions. (A) Centrifuge spin- down assay with representative RP-HPLC analytical traces of the peptide in aqueous solution

(black) compared to the peptide after incubation with liposomes and spin-down (orange) (B)

Surface plasmon resonance sensorgram showing deposition of the lipid bilayer followed by injection of the peptide onto the bilayer (association) MANUSCRIPTand dissociation over time. A representation of a peptide binding to the lipid at different concentrations is shown in the inset. (C) Trp fluorescence emission spectra showing partitioning of peptides into lipid bilayers. The Trp emission spectra shown represent the peptide in aqueous solution (black) and upon insertion in the lipid bilayer when in the presence of liposomes (orange). (D) Stern-Volmer representation of Trp fluorescence emission of a peptide in the presence (orange) or absence (black) of liposomes upon titration with an aqueous soluble quencher. (E) 1D 1H NMR lipid titration experiment showing the amide region of the spectrum of a peptide in aqueous solution (black) compared to the peptide in lipid suspension (orange). (F) NMRACCEPTED transferred cross-saturation experiment showing the 2D 15 N HSQC spectrum of a peptide in the absence of lipid (black) and in the presence of lipid (orange) following irradiation of the lipid and transferred cross-saturation to the peptide. (G) NMR chemical shift mapping experiment showing the 2D 15 N HSQC spectrum of a peptide in the absence of lipid (black) and on lipid titration (orange and red). The arrows show the direction of change in the chemical shifts of

46 the residues. (H) Illustration of a Cryo-EMACCEPTED experim MANUSCRIPTent depicting freezing the sample on a grid prior to the collection of images using transmission emission microscopy. The images are then processed prior to the construction of a 3D-model.

MANUSCRIPT

ACCEPTED

47 ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED

ACCEPTED MANUSCRIPT Highlights:

• Spider toxins are excellent tools to study voltage-gated ion channel

pharmacology

• Some gating modifier toxins have shown affinity for lipid membranes

• The role of the membrane in spider toxin activity can guide rational drug

design

• The cell membrane, ion channel and spider toxins should be studied as a

complex

MANUSCRIPT

ACCEPTED