Chemoselective tarantula toxins report voltage PNAS PLUS activation of wild-type ion channels in live cells

Drew C. Tilleya,1, Kenneth S. Euma,b,1,2, Sebastian Fletcher-Taylora, Daniel C. Austina, Christophe Dupréb, Lilian A. Patrónb, Rita L. Garciac, Kit Lamd, Vladimir Yarov-Yarovoya,d, Bruce E. Cohenc, and Jon T. Sacka,b,3

aDepartment of Physiology and Membrane Biology, University of California, Davis, CA 95616; bNeurobiology Course, Marine Biological Laboratory, Woods Hole, MA 02543; cMolecular Foundry, Lawrence Berkeley National Laboratory, Berkeley, CA 94720; and dDepartment of Biochemistry and Molecular Medicine, University of California, Davis, CA 95616

Edited* by Richard W. Aldrich, The University of Texas at Austin, Austin, TX, and approved September 23, 2014 (received for review May 27, 2014) Electrically excitable cells, such as neurons, exhibit tremendous trical activity to specific channel subtypes a difficult and impre- diversity in their firing patterns, a consequence of the complex cise task. New tools to optically report activity of VGICs are collection of ion channels present in any specific cell. Although needed to advance study of their contributions to electrical numerous methods are capable of measuring cellular electrical signaling. signals, understanding which types of ion channels give rise to We hypothesized that ion channel activity could be reported by these signals remains a significant challenge. Here, we describe imaging ligands whose affinity for channels is state-dependent. exogenous probes which use a novel mechanism to report activity The of many predatory creatures contain voltage sensor- of voltage-gated channels. We have synthesized chemoselective targeting toxins (VSTs) that act by binding selectively to specific derivatives of the tarantula toxin guangxitoxin-1E (GxTX), an in- channel conformations (14–21). Even at doses too low to have hibitory cystine knot peptide that binds selectively to Kv2-type a significant physiological effect, a state-dependent VST will bind voltage gated potassium channels. We find that voltage activation preferentially to a specific ion channel conformation. Thus, the of Kv2.1 channels triggers GxTX dissociation, and thus GxTX bind- propensity of a state-dependent VST to bind an ion channel is ing dynamically marks Kv2 activation. We identify GxTX residues dependent on the channel’s activation status, and tracking VST that can be replaced by thiol- or alkyne-bearing amino acids, with-

localization could reveal spatiotemporal patterns of channel PHYSIOLOGY out disrupting toxin folding or activity, and chemoselectively ligate fluorophores or affinity probes to these sites. We find that GxTX– activity. An intensively studied class of VSTs is from spider venoms. fluorophore conjugates colocalize with Kv2.1 clusters in live cells “ ” and are released from channels activated by voltage stimuli. Kv2.1 Spider VSTs have an inhibitory cystine knot fold that is sta- bilized by bridges. These structurally similar VSTs act activation can be detected with concentrations of probe that have 2+ + + a trivial impact on cellular currents. Chemoselective GxTX mutants upon Ca ,Na , and K channels, some with remarkable spec- conjugated to dendrimeric beads likewise bind live cells expressing ificity for particular channel subtypes (22). VSTs are water-sol- Kv2.1, and the beads are released by channel activation. These uble peptides that partition into the outer leaflet of the plasma optical sensors of conformational change are prototype probes that membrane, where they bind to the extracellular edge of VGIC can indicate when ion channels contribute to electrical signaling. transmembrane segments that form voltage sensors (21, 23–26). This functional requirement to be soluble in water and partition voltage-gated ion channel | | gating modifier | into membranes suggests a delicate balance between a lipophilic fluorescence | allostery Significance lectrical signals traveling along excitable cell membranes Eactivate voltage-gated ion channels (VGICs), which open Electrically excitable cells, such as neurons, exhibit tremendous their pores to create electrical feedback and trigger signaling variation in their patterns of electrical signals. These variations cascades that lead to neurotransmitter release, hormone secre- arise from the collection of ion channels present in any specific tion, gene transcription, and other cellular responses. Many cell, but understanding which ion channels are at the root of maladies result from aberrant VGIC signaling, including epi- particular electrical signals remains a significant challenge. lepsies, cardiac arrhythmias, and pain syndromes (1, 2). VGIC Here, we describe novel probes, derived from a tarantula complements vary with cell type (3), making cell-specific chan- peptide, that are able to report the activity of voltage- nels important targets to selectively modulate pathophysiological gated ion channels in living cells. This technology uses state- electrical signals (4). There are over 80 VGIC pore-forming selective binding to optically monitor the activation of ion subunits encoded in the human genome (5), and many of these channels during cellular electrical signaling. Activity-reporting are coexpressed within the same cell (3), making it difficult to probes based on these prototypes could potentially identify dissect specific channel contributions to electrical signaling. In when endogenous ion channels contribute to electrical sig- neurons, the precise trafficking of VGIC subtypes to distinct cell naling, thus facilitating the identification of ion channel targets regions such as the axon hillock, presynaptic terminals, and distal for therapeutic drug intervention. dendrites allows specific subtypes to control voltage changes in – Author contributions: D.C.T., K.S.E., V.Y.-Y., B.E.C., and J.T.S. designed research; D.C.T., subcellular compartments (6 9). However, the presence of an K.S.E., D.C.A., C.D., L.A.P., and J.T.S. performed research; S.F.-T., R.L.G., K.L., B.E.C., and J.T.S. ion channel subunit in a cell does not indicate whether it is ac- contributed new reagents/analytic tools; D.C.T., K.S.E., D.C.A., C.D., L.A.P., and J.T.S. ana- tivated under physiological conditions (10–13). Determining lyzed data; and D.C.T., K.S.E., V.Y.-Y., B.E.C., and J.T.S. wrote the paper. which ion channels activate under particular physiological and The authors declare no conflict of interest. pathophysiological conditions is a formidable challenge, yet this *This Direct Submission article had a prearranged editor. understanding is crucial for identification of channel subtypes as 1D.C.T. and K.S.E. contributed equally to this work. therapeutic targets to correct pathologies (4). More than three 2Deceased June 22, 2014. decades after the first ion channel was cloned, voltage clamp 3To whom correspondence should be addressed. Email: [email protected]. remains the primary method to determine which ion channels This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. are activated by electrical stimuli, making assignment of elec- 1073/pnas.1406876111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1406876111 PNAS Early Edition | 1of8 Downloaded by guest on September 24, 2021 face to interact with membrane lipids and bind the channel’s This complex biochemistry has made VSTs refractory to labeling transmembrane segment, and a hydrophilic face essential to attempts, as chemical perturbations are likely to disrupt peptide prevent toxin aggregation and retain solubility (Fig. 1A) (27, 28). folding or activity. However, if these chemistry challenges are overcome, the resultant tools could be valuable probes of ion channel activity. We sought to develop a general VST labeling strategy by AB taking advantage of recent advances in bioconjugation chemistry, reporter 1 which have produced a series of novel chemoselective reactions hydrophilic for modification of peptides. Here, we demonstrate that a cystine

solution exposed K knot VST can be modified with fluorophores and affinity probes G in a two-step process involving the incorporation of a small VST chemoselective side chain, which allows the peptide to retain its membrane & 0 activity and permits the attachment of reporter molecules. We find that these conjugates report the localization of Kv2 chan- channel-binding -80 0 80 step / mV nels, and dynamically detect activation of these ion channels in living cells, suggesting an optical means to monitor activation of C 10 20 30 specific channel subtypes during neuronal signaling. EGECGGFWWKCGSGKPACCPKYVCSPKWGLCNFPMP loop 1 loop 2 loop 3 loop 4 Results VSTs Retain Bioactivity After Chemoselective Modification. D 13 E We chose to synthesize VST reporters from the tarantula peptide 15 guangxitoxin-1E (GxTX), a potent modulator that selectively binds Kv2 channels (29, 30). In combination with patch-clamp elec- 17 trophysiology, GxTX pharmacology is currently the most strin- 27 gent test of whether Kv2 channels contribute to electrical 32 signaling (13). In neurons and pancreatic islet cells, GxTX in- 1 hibits delayed rectifier current with minimal effects on other 21 current types and has been recently used to reveal unexpected F contributions of Kv2 channels to electrical signaling (13, 31, 32). G Based on the effects of GxTX on Kv2 channel gating and the behavior of radiolabeled GxTX (30), we suspected GxTX might bind much more weakly when the channels become activated by voltage. To test this, we synthesized fluorescently labeled GxTX to determine if this VST could optically report Kv2 channel lo- calization and activation. aqueous In an initial attempt at generating GxTX variants, we expressed headgroupsps GxTX with an N-terminal hexahistidine tag and a C-terminal protease cleavage site in bacteria (SI Materials and Methods), but lipids this peptide had little activity against Kv2.1 channels (Fig. S1A). H I This suggested that attachment of reporters to side chains, rather 1 1 than the termini, was most likely to result in bioactive GxTX. We turned to chemical synthesis of GxTX to exploit nonnative amino

K

K acids that could serve as chemoselective conjugation points. We

G G synthesized a GxTX variant with its sole methionine replaced by an oxidation-resistant norleucine isostere and found it altered 0 0 Kv2.1 function by producing a strikingly positive shift in activation voltage (Fig. 1B), similar to wild-type GxTX (27, 29, 33). To avoid -80 0 80 -80 0 80 step / mV step / mV methionine oxidation complications, norleucine 35 GxTX was used as the background for further variants. To identify GxTX Fig. 1. Tarantula toxin retains bioactivity following chemoselective conjuga- residues that can be substantially modified without disrupting tion. (A) Schematic of a VST portrays partitioning into a cell membrane (dashed toxin folding or activity, we introduced chemoselective amino line). Star represents a reporter conjugate. (B) GxTX saturably shifts the Kv2.1 acids into a series of variants. Mutagenesis studies of homologous conductance–voltage relation. Whole-cell voltage-clamp recordings from Kv2.1. VSTs (28, 34) guided selection of residues on the hydrophilic face Vehicle (black circles), 100 nM GxTX (red circles), 1 μM GxTX (red triangles), n = 4. + of GxTX (33) that might have minimal impacts on channel Lines are fits of Eq. S1.VehicleVmid = 8 ± 3mV,z = 1.2 ± 0.1e ; 100 nM GxTX + binding. We substituted orthogonally protected cysteine residues Vmid = 81.5 ± 0.4 mV, z = 1.21 ± 0.03e ;1μMGxTXVmid = 72 ± 1mV,z = 1.24 ± + C 0.05e . Values normalized to maximum conductance in vehicle. (C)Sequenceof [Cys(Acm)] at seven hydrophilic sites (Fig. 1 ) and tested for GxTX. Bold indicates residues replaced by Cys(Acm). Black indicates an intra- activity against Kv2.1 channels. Of these seven Cys(Acm) cystine loop or terminus where at least one residue could be replaced with Cys mutants, five acted similar to GxTX at 100-nM concentration, (Acm) yet retain activity against Kv2.1 at 100 nM. Yellow is cystine. (D)Red shifting the voltage midpoint of Kv2.1 conductance to greater ribbon, backbone trace of GxTX (red ribbon); sticks, side chains replaced by Cys than +40 mV (Fig. S1A). When Lys27 was replaced with Cys (Acm). Colored as in C.(E) Surface mesh representation of GxTX. Side chains (Acm), more than 300 nM was required to induce a perceptible colored as in C.(F) Extracellular view of Rosetta model of GxTX (red) bound to gating shift; GxTX with Asn32 replaced by Cys(Acm) was inactive S3b helix of a Kv2.1 channel (gray). (G) Rendering of GxTX-TMR (red) binding to at concentrations up to 1 μM. Regions of GxTX that were tol- voltage sensor paddle of Kv2.1 (gray). Dashed line represents external mem- brane boundary. (H) GxTX-PEG5K retains activity against Kv2.1. Vehicle (black): erant to perturbations were suggested by mapping substitutions + onto a GxTX NMR structure (33) (Fig. 1 D and E). Vmid = 10 ± 1mV,z = 1.17 ± 0.04e ;1μMGxTX-PEG5K(red):Vmid = 64 ± 2mV, + z = 1.4 ± 0.3e ; n = 5. (I) GxTX-dy550 retains activity against Kv2.1. Vehicle We generated a speculative model of GxTX bound to its Kv2.1

(black), 20 nM GxTX-dy550 (red), n = 4. Lines are fits of Eq. 1. Vehicle Vmid = 3.7 ± channel binding site (35) using Rosetta computational methods + + 0.4 mV, z = 1.72 ± 0.02e ;20nMGxTX-dy550Vmid = 55 ± 1mV,z = 1.04 ± 0.05e . (36–39). The lowest energy resultant docked structure served as

2of8 | www.pnas.org/cgi/doi/10.1073/pnas.1406876111 Tilley et al. Downloaded by guest on September 24, 2021 a structural hypothesis for how GxTX interacts with each subunit A B C PNAS PLUS of Kv2.1 (Fig. 1F). Although GxTX stabilizes resting con- formations of the voltage sensor (35), this model represents GxTX binding to what is proposed to be an activated confor- mation of the voltage sensor (40). If the orientation of the toxin toward the channel remains similar in different states, it could be useful for predicting points of attachment for reporter mole- 30 μm` 8 μm 10 μm cules. This docking configuration places intracystine loop 2 most distal from the channel and membrane, suggesting that bulky reporter conjugates attached to loop 2 might be less likely to interfere with channel binding (Fig. 1G). To create probe attachment sites on loop 2, we incorporated chemoselective groups, either an alkyne or a free thiol, for conjugation of fluorescent and affinity probes. To add an al- kyne, we synthesized GxTX with the unnatural amino acid propargylglycine, for labeling with azide-bearing probes using Cu(I)-catalyzed Click reactions. Whereas many Click variations have been reported, Cu(I) salts are known to catalyze the dispro- portionation of thiol compounds (41), and we sought to identify conditions that would neither reduce nor scramble the cystine knot . We found that addition of the water-soluble Cu(I)- chelator bis[(tertbutyltriazoyl)methyl]-[(2-carboxymethyltriazoyl) methyl]-amine (BTTAA) (42) yielded GxTX Click conjugates that retained function against Kv2.1 channels. We attached a mono- functional 5-kDa polyethylene glycol (PEG5K) by its azide group Fig. 2. Fluorescent tarantula toxin reveals Kv localization in live cells. (A) to GxTX Pra13 (Fig. S1 B and D). The product (GxTX-PEG5K) GxTX binds selectively to cells expressing Kv2.1. Images of two cocultured was active against Kv2.1 channels, thereby inhibiting channel CHO-K1 cell lines that express either Kv2.1 at the cell surface or BFP in the opening by shifting the midpoint of conductance to voltages nucleus. (Top) Differential interference contrast image of a mixed culture. PHYSIOLOGY similar to wild-type GxTX (Fig. 1H). (Middle) Fluorescence from GxTX-dy550. (Bottom) Merge of GxTX-dy550 As an alternate strategy for chemoselective conjugation, we (red), with BFP fluorescence (blue). (B) GxTX labels Kv2.1 clusters. Confocal revealed a free “spinster” thiol on the surface of the peptide (43). slice through confluent CHO cells, one of which is expressing EGFP-Kv2.1. (Top) EGFP-Kv2.1 (green). (Middle) GxTX-TMR (red). (Bottom) Merge. Pear- Removal of the Acm protecting group has typically been cata- + son’s overlap coefficient r = 0.80. (C) GxTX labels Kv subunits on cultured lyzed by Ag and facilitated with excess reductant (44), or oxi- hippocampal pyramidal neurons. (Top) EGFP-Kv2.1 (green). (Middle) GxTX- dant (45), but these conditions are incompatible with preserving dy550 (red). (Bottom) Merge. Pearson’s overlap coefficient r = 0.59. existing disulfides while obtaining a reduced surface thiol. To achieve the required mixed redox state, we removed Acm after GxTX folding and disulfide formation by displacing the Acm is due to membrane partitioning versus channel binding inter- + + − group with Ag followed by precipitation of Ag with Cl under actions (27). The colocalization of GxTX with Kv2.1 in cultured acidic conditions, breaking the Ag–S bond without reducing or cells and neurons demonstrates that a strong, specific interaction scrambling the disulfides (Fig. S1C). Maleimide-functionalized occurs between the tarantula peptide and the ion channel voltage fluorophores, including tetramethylrhodamine and dylight550, sensor. This precise colocalization makes GxTX–fluorophores were condensed with the free thiol, and Cys13-fluorophore viable ion channel probes. GxTX conjugates retained activity against Kv2.1 (Fig. 1I). These results demonstrate that variants of GxTX can be folded and Toxin Binding is Dependent on Channel Activation. We hypothesized chemoselectively ligated to reporter molecules while retaining that GxTX binding is coupled to channel activity: when the bioactivity against ion channels. membrane voltage is depolarized, voltage sensors adopt activated conformations and the binding of the inhibitory VST is weakened Fluorescent Toxin Colocalizes with Kv2.1 Channels in Live Cells. To (Fig. 3A). Studies have concluded that the affinity of ion channels determine whether these VST conjugates could identify ion for certain VSTs is voltage dependent and that depolarizing channels in live cells, we tested whether GxTX–fluorophores pulses can release inhibitory VSTs from the channels (18–21), but would colocalize specifically with Kv2.1 ion channels. When two the precise changes in affinity have never been measured. To mammalian CHO-K1 cell lines, one expressing Kv2.1 and the measure the degree to which GxTX affinity depends on voltage other expressing a blue fluorescent protein (BFP), were briefly activation, the fraction of Kv2.1 channels bound by GxTX was + treated with 100 nM of a GxTX–fluorophore conjugate, the deduced from K current at different holding potentials (SI GxTX fluorescence remained associated only with Kv2.1- Materials and Methods). At a holding potential of −100 mV, + expressing cells (Fig. 2A). When CHO-K1 cells were transiently GxTX was found to inhibit K current during brief test pulses transfected with Kv2.1-GFP, the channels appeared in clusters with a half-maximal inhibitory concentration (IC50) of about 2 on the cell surface (Fig. 2B). GxTX localized tightly with these nM (Fig. 3B, black). In cells held at 0 mV, the IC50 rose beyond clusters (Pearson’s overlap coefficient r = 0.796), whereas sur- 100 nM (Fig. 3B, blue). This shift of GxTX affinity with voltage rounding untransfected cells had little GxTX fluorescence. In indicates that GxTX indeed binds to Kv2.1 in a conformation- living neurons, a similar phenomenon was seen (Fig. 2C), al- dependent manner. though the colocalization with Kv2.1-GFP was not as complete To better understand the thermodynamics of GxTX inter- (Pearson’s overlap coefficient r = 0.589), possibly owing to the actions with Kv2.1, we further analyzed the dose dependence of presence of endogenous Kv channels. Whether tarantula VSTs electrophysiological responses. Whereas four tarantula VSTs such as GxTX have a high affinity for channels is a point of active may bind a Kv channel tetramer (Fig. 1F) (48), only one toxin is debate (27, 46, 47). GxTX, like many other tarantula peptides, required to eliminate current at neutral voltage (27). Due to this partitions into the outer leaflet of the plasma membrane, and it 4:1 stoichiometry, conductance at 0 mV is related to the GxTX has been difficult to discern how much of its affinity for channels dissociation constant (Kd) by the fourth power of the probability

Tilley et al. PNAS Early Edition | 3of8 Downloaded by guest on September 24, 2021 will be even weaker to more activated conformations at positive A strong binding weak binding voltages. Whereas the relative affinities of the many closed, open, and inactivated channel conformations could not be dis- tinguished with these experiments, our results clearly indicate +∆V that voltage activation lowers the affinity of GxTX for Kv2.1. To determine whether the voltage-dependent shift in the resting voltage sensors active voltage sensors GxTX Kd was due to a change in the binding rate, dissociation B rate, or both, we measured kinetics of Kv2.1 inhibition and re- 1.0 * covery at different voltages, and calculated microscopic binding–

xam 0.8 dissociation rates. The time course of inhibition and recovery

K 0.6 were fit with a function that extracts a microscopic binding rate

I/ (kon) and dissociation rate (koff)(SI Materials and Methods, Eq.

K 0.4 I S3). When cells were held at −100 mV, GxTX inhibition oc- 0.2 curred faster than at 0 mV (Fig. 3C), consistent with the toxin’s 0.0 weaker affinity at this depolarized voltage. This suggests that 1 10 100 1000 access of the VST to its binding site is compromised when the [GxTX] / nM voltage sensors are activated. At both voltages, the time course of inhibition accelerated as GxTX concentration increased (Fig. CD 3D). Association rate rose linearly with GxTX concentration, 1.0 consistent with its expected first-order dependence (Fig. 3D). k * on

xam 0.1 became 13-fold slower when the voltage was raised to 0 mV

1 - k s/ (Table 1). To determine off accurately, dissociation was mea-

K I 0.5

/ no 0.01 sured after toxin washout. The return of Kv2.1 current after

K

k I GxTX washout was accelerated when long pulses to 0 mV were included in the voltage stimulus protocol (Fig. 3E; SI Materials 0.0 0.001 and Methods). The calculated koff during these 0-mV pulses was 70-fold faster than at −100 mV (Fig. 3F). Thus, toxin binding and 0 50 10 100 1000 dissociation are both affected by voltage, which indicates that time / s [GxTX] / nM GxTX’s state specificity is due to its slow binding and rapid EF* dissociation from voltage-activated channels. Fundamentally, the k k K 1.0 0.1 ratio of off/ on determines the d of a bimolecular process. At

xam both voltages tested, the ratio of koff/kon (calculated from ki- S3 K 1- netics, Eq. ) is within an order of magnitude of d (calculated K 0.01

I/ 0.5 s S2

/ from equilibrium measurements, Eq. ), but the values do not

ffo

K I precisely match. This difference suggests the thermodynamic

k 0.001 model that forms the basis of parameter extraction requires 0.0 further refinement. Our calculations assume that toxin binds to 0.0001 0 500 1000 each of four channel subunits independently and that toxin af- time / s -100mV 0mV finity is static over time. The independence of a tarantula VST binding to Kv2.1’s four subunits has been experimentally vali- Fig. 3. Affinity of toxin for voltage sensors is conformation dependent. (A) dated (16). However, the voltage-dependent dissociation of ta- Schematic of GxTX portrays the variable affinity for the resting and activated rantula VSTs from mutant Kv2.1 channels was found to be conformations of Kv2.1. Upon depolarization, active channels (blue) bind GxTX dependent on pulse duration and frequency (21). Progression (red) with a lower affinity. (B) Normalized dose–response profile of GxTX with Kv2.1 held at −100 mV (black) or 0 mV (blue). Lines are fit of Eq. S2 (solid lines) ± through multiple resting, activated, and inactivated conformations

SEM (dotted lines). Kd (−100 mV) = 12.7 ± 0.9 nM and Kd (0 mV) = 826 ± 139 nM. is expected to result in affinity change over time. Although our *P = 0.0003 at 100 nM. (C) Representative association of 100 nM GxTX to model is oversimplified, it captures the basic features of activation- cells held at −100 mV (black) or 0 mV (blue). Red bar indicates perfusion of driven affinity change. The parameters extracted from kinetic = × 5 ± × 100 nM GxTX. Lines are fits of Eq. S3 where kon −100 mV 3.13 10 6.58 measurements with this model show that channel activation alters 3 -1· −1 = × 4 ± × 3 -1· −1 10 M s and kon 0 mV 4.40 10 1.81 10 M s .(D) Association rates of the rates of association and dissociation in rough agreement with GxTX at −100 mV (black) and 0 mV (blue) determined by Eq. S3. kon −100 mV = 5 -1· −1 4 -1· −1 equilibrium measurements, demonstrating that the toxin has high 3.34 × 10 ± 1.13 M s and kon 0 mV = 2.53 × 10 ± 1.18 M s were de- termined by a linear fit (solid line) ± SEM (dotted lines). *P = 0.005 at 100 nM. (E) affinity for resting channels and low affinity for activated channels. Representative GxTX dissociation from cells held at −100 mV and given a 100-ms 0-mV test pulse every 10 s (black) or given a 5-s 0-mV step every 10 s (blue). Fluorescent Toxin Reports Channel Activation. We next tested – Red bar indicates GxTX perfusion. Lines are fits of Eq. S3 where koff −100 mV = whether we could optically measure channel activity with VST −3 −5 −1 −2 −4 −1 – 2.94 × 10 ± 2.13 × 10 ·s and koff 0 mV = 2.23 × 10 ± 6.16 × 10 ·s .(F) fluorophore conjugates. The optical reporting ability of a GxTX Summary of dissociation rates of GxTX at −100 mV (black) and with 0-mV pulses fluorophore was tested on CHO cells expressing Kv2.1 in whole-cell (blue) determined by Eq. S3. Mean (solid line) ± SEM (dotted lines). *P = 0.005.

Table 1. Experimental and calculated GxTX-norleucine 35 of a subunit being bound (SI Materials and Methods, Eq. S2). affinity for rKv2.1 channels heterologously expressed in CHO When this established analysis (48) is applied to our dose– cells K response data set, it reports a 65-fold shift in the d for GxTX at Parameter −100 mV 0 mV −100 mV compared with 0-mV holding potential (Fig. 3B, ± ± curves). Voltage activation by depolarization to 0 mV induces Kd (nM) 12.7 0.9 826 139 −1 ± × −3 ± × −3 Kv2.1 channels to reach a new distribution between closed, open, koff (s ) 0.84 0.18 10 38 9 10 -1· −1 ± × 3 ± × 3 and inactivated conformations. Toxin-treated channels do not kon (M s )330110 10 25 12 10 k /k (nM) 2.5 ± 1.0 1,500 ± 780 appreciably open at 0 mV (Fig. 1B), suggesting the toxin binding off on

4of8 | www.pnas.org/cgi/doi/10.1073/pnas.1406876111 Tilley et al. Downloaded by guest on September 24, 2021 voltage-clamp fluorometry experiments. GxTX-dy550 was cho- voltage resulted in dynamic changes of cell fluorescence, in- PNAS PLUS sen for these experiments because it was found to have elec- dicating that the fluorescent VST optically reported the change in trophysiological properties similar to GxTX (Fig. 1I and Fig. S2). activation state of the ion channel (Fig. 4 D and E and Movie S1). To optically measure VST binding to voltage-clamped cells, we The rate of fluorescence change (kΔF) varied with voltage and quantitated the fluorescence that remained after application and saturated at positive voltages (Fig. 4F), consistent with GxTX washout of 100 nM GxTX-dy550 (Fig. 4A). When cells were held localization reporting activation of Kv channels rather than the at −100 mV, fluorescence slowly decayed. When cells were membrane voltage. Thus, the activity of VGICs can be reported depolarized to 0 mV to accelerate toxin dissociation, fluores- by state-selective fluorescent VST localization to live cells. cence loss markedly accelerated (Fig. 4B). By fitting an expo- To test whether our VST probe can report channel activation nential function (Eq. S4) to the fluorescence decays, rates of at trace concentrations, which inhibit a negligible fraction of fluorescence change, k−100 mV and k0mV, were extracted (Fig. channels, we measured fluorescence from cells bathed in probe at 4C). The similar kinetics of k−100 mV fluorescence decay and koff a concentration substantially below the Kd of its target VGIC. At of GxTX-dy550 at −100 mV (Fig. S2C) is consistent with fluo- 1 nM, GxTX-dy550 is 30-fold below its measured Kd for Kv2.1, + rescence decay resulting from VST dissociating from channels. A and has little effect on the whole-cell K current (Fig. S2A). At clearly significant difference in fluorescence decay was seen after this inefficacious concentration, fluorescence could be detected voltage change, with k0mVbeing 53-fold faster than k−100 mV.We on Kv2.1-expressing cells. This fluorescence was modulated by conclude that the voltage-driven fluorescence change results epochs of action-potential-like stimuli: 2-ms steps to +40 mV from GxTX dissociation due to Kv2.1 voltage activation. from a holding potential of −80 mV at 100 Hz. During these To test whether Kv2.1 activation can be reversibly reported by stimulus epochs, fluorescence decayed toward a similar baseline; this fluorescent VST, we bathed cells continuously in GxTX- in the periods between stimulus epochs, fluorescence slowly re- dy550 and switched back and forth between voltages. In 10 nM covered (Fig. 4G). These signals suggest that every action po- GxTX, fluorescence was concentrated on the surface of Kv2.1- tential increases the odds of probe dissociation. The mean expressing cells (Fig. 4D) and remained constant over time (Fig. fluorescence decay rate during the initial stimulus epoch was −1 4E) as the VST slowly bound to and dissociated from Kv2.1 kΔF = 0.025 ± 0.005 s (fit with Eq. S4, F0 unconstrained, n = 5 channels. When voltage-clamped cells were held at a constant cells), intermediate to the dissociation rates seen at polarized or voltage, fluorescence intensity also remained constant. Varying depolarized voltages (Fig. 4 C and F and Fig. S2C). This suggests PHYSIOLOGY

* 10µm 1.0 A B 0mV C 1

1-

s/ 0.1

F/F

F

0.5 ∆

k 0.01

0.0 0.001 -50 0 50 -100 0 time / s step / mV D -100s 0s 20s 220s

-100mV -100mV 0mV -100mV 10µm

0mV AP epochs E F 1 G 1.0 1.0

1

-

s/

F/

F/F F 0.1

F

0.5 k

0.01 0.0 0.5 0 200 -100 0 0 100 200 300 time / s step / mV time / s

Fig. 4. Fluorescent tarantula toxin reports Kv channel activity in Kv2.1-expressing CHO cells. (A) Fluorescence from a patch-clamped cell after washout of 100 nM GxTX-dy550. Bright-field (grayscale); GxTX-dy550 (red). (B) GxTX-dy550 fluorescence decay of voltage-clamped cell as in A. Holding potential switch −1 from −100 mV to 0 mV indicated by blue bar. Gray line is fit of Eq. S4; k−100 mV = 0.00399 ± 0.00001 s . Solid blue line is fit of Eq. S4 to initial 20 s at 0 mV; −1 −1 −1 k0mV= 0.166 ± 0.002 s . Dashed blue line is fit of Eq. S5; k0 mV fast = 0.222 ± 0.003 s , k0 mV slow = 0.0407 ± 0.0008 s .(C) Summary of dissociation rates of GxTX-dy550 from cells when held at −100 mV (black) or 0 mV (blue). Mean (solid line) ± SEM (dotted line). *P = 0.008. (D) Voltage-clamped (solid circle) and -unclamped (dashed circle) CHO cells bathed in 10 nM GxTX-dy550. Voltage of clamped cell held at −100 mV and stepped to 0 mV from 0 to 20 s. (E) Dotted black line, fluorescence change of unclamped cell: dashed circle in D. Solid black line, fluorescence change of voltage-clamped cell: solid circle in D. Blue bar −1 indicates depolarization of clamped cell to 0 mV. Blue line is fit of Eq. S4, kΔF 0mV= 0.163 ± 0.002 s . Gray line is fit of GxTX-dy550 reassociation at −100 mV, −1 + kΔF −100 mV = 0.0134 ± 0.0001 s .(F) Summary of kΔF in 10 nM GxTX-dy550. Line is fit of Eq. S6. V1/2 = -25 ± 11 mV, z = 1.5 ± 0.9e , kΔF min = 0.011 ± 0.001 s, kΔF max = 0.15 ± 0.04 s. (G) Solid black line, fluorescence change of voltage-clamped cell in 1 nM GxTX-dy550. Holding potential = −80 mV. Blue bars indicate action-potential-like epochs: 2-ms steps to +40 mV at 100 Hz. Blue lines are fits of Eq. S4 during action-potential epochs; baseline (F0) constrained to the mean −1 of final 10 s of last epoch (0.531 ΔF/F); kΔF = (in chronological order) 0.024 ± 0.005, 0.029 ± 0.001, and 0.071 ± 0.008 s .

Tilley et al. PNAS Early Edition | 5of8 Downloaded by guest on September 24, 2021 that fluorescence change reports VGIC response to many action- We found that bead–cell adhesion required both the VST peptide potential-like stimuli integrated over time. The probe redistributes and a VGIC. Kv2.1 cells did not adhere well to control beads between channels and solution to reach a binding equilibrium without GxTX, and cells lacking Kv2.1 did not adhere well to GxTX dictated by the VGIC activation pattern. In this way, the state- beads (Fig. 5 C and D). We next tested whether channel con- dependent probe reports channel activation in response to action- formation impacted this GxTX-bead–Kv2.1-cell interaction. As studying adhesion of voltage-clamped cells to beads was in- potential-like voltage changes. + feasible, external [K ] was used to change voltage and activate + Channel-Modulating Toxins Mediate Ion Channel Binding to Microbeads. channels. We first confirmed that increasing external [K ] The conformation-specific binding of GxTX–fluorophores to Kv depolarized cells to release GxTX from Kv2.1. Fluorescence decay resulting from GxTX-dy550 dissociation was measured in channels suggests that the VST could guide other visible probes + a control (5 mM) [K ] solution and after addition of a high (180 similarly. To determine if GxTX could imbue very large molecules + + with its binding properties, we tested whether cells expressing Kv2.1 mM) [K ] solution. The high [K ] solution accelerated GxTX E F + channels would bind to 100-μm microbeads coated with GxTX. We dissociation (Fig. 5 and ), consistent with the high [K ] in- chose beads used for one-bead–one-compound (OBOC) combina- creasing resting voltage to activate the Kv2.1 channels. To assess torial drug discovery programs (49–51) to establish whether GxTX- the effects of channel activation on cell adhesion to GxTX-beads, like molecules might be discoverable in an OBOC screen. These Kv2.1-expressing cells were mixed with beads and allowed to adhere. These bead–cell complexes were then incubated in either OBOC beads are composed of a solid-phase polystyrene dendrimer + a control or high [K ] solution. In blinded trials on five separate core with a PEG shell for water solubility and cell compatibility. We days, the fraction of beads with cells bound was reduced by in- used copper(I)-catalyzed azide-alkyne Click cycloaddition to con- + cubation in high [K ] solution (Fig. 5G). Thus, the activity de- jugate an alkyne-functionalized GxTX variant to azide-functional- A pendence of Kv2.1 binding was preserved when GxTX was ized beads (Fig. 5 ), as this chemistry successfully produced attached to beads. The microbead results indicate that voltage bioactive, PEGylated GxTX (Fig. 1E and Fig. S1 B and D). Beads B sensor modulators can be identified by ion channel binding to with a GxTX coating adhered to Kv2.1-expressing cells (Fig. 5 ). OBOC beads. This suggests that new modulators of VGICs could be discovered in OBOC libraries. Taken together, these experi- ments show that VSTs can guide molecules of a wide range of sizes and compositions, from subnanometer fluorophores to A 2 nm 100 μm N N 100-μm beads, and that these conjugates can act as reporters of N3 Cu+ N azide GxTX GxTX VGIC activation. + bead BTTAAA GxTX Kv2.1 bead +∆V bead Pra13 bead cells cells Discussion C D * * The functional roles of ion channels in biological circuits are B 50% actively debated. We optically detected activity of the Kv2.1

sll +4 channel subtype using VST-conjugate probes. Kv2.1 is ubiqui-

e

ct

h

t

iw tously expressed in the soma of neurons throughout the brain (52)

n

e

r sda and is the dominant delayed rectifier of central neurons (53, 54).

e 0 ’ h The channel s activity shapes the repetitive firing properties of

.

.2

da

eb

TxG

2

vK+ vK neurons (13), and Kv2.1 is homeostatically regulated in response to electrical activity (55). The up-regulation of Kv2.1 activity is

o

o

X

nX

n1

20 μm 100 μm T proposed to relieve excitotoxic stress by suppressing neuronal

x

G1 hyperexcitability and seizure activity (32). However, under path- E F * G ** 1.0 ological conditions such as ischemia from stroke, up-regulation of 60% 0.1 Kv2.1 activity leads to apoptosis (56). To disentangle the con-

sl

+

l

4h trasting neuroprotective and apoptotic roles of Kv2.1, it is es-

ect

F/F -1

t

i 0.5 s sential to know when and to what degree the channel is activated.

/

n

wsd ∆ 0.01 – ere Activity-sensing probes such as the GxTX fluorophores de- off

k

ae h scribed here can detect channel response to sustained de-

d

b a polarization or epochs of action-potential-like stimuli (Fig. 4), 0.0 0.001 0 and could provide a means to determine the extent to which Kv2 -100 0 100 control high control high time / s [K+] [K+] channels are involved with specific neural signaling pathways. The full potential of ion channel activity probes would be re- Fig. 5. Tarantula toxin-channel binding mediates cell adhesion to surfaces. alized with probes customized for particular channel types and (A) Depiction of GxTX Pra13 conjugation to azide beads and adhesion to specific questions related to their function. Developing activity cells. (B) Kv2.1 CHO cells adhering to a GxTX-coated bead. (C) Kv2.1 CHO cells probes for VGIC types requires careful consideration of their selectively adhere to GxTX beads. Beads functionalized with either GxTX (gray) or fluorophore (blue) were incubated with CHO-K1 cells expressing two components: a reporter amenable to sensitive detection and Kv2.1. Note the propensity of cells to accumulate on GxTX beads. (D) a conformation-specific VGIC ligand. Quantification of bead adhesion to cells as in A from experiments on sep- The ability of these probes to report channel activity without arate days. Percentage of beads with more than four cells visibly adhered disrupting electrical signals is limited by the signal-to-noise scored with GxTX on beads and Kv2.1 expressed on cells (gray, n = 6), with characteristics of the reporter. An important caveat to consider Kv2.1 cells but no GxTX on beads (green, n = 4), or with GxTX on beads but when using any activity-dependent probe is the physics of mi- no Kv2.1 in cells (white, n = 3).*P < 0.05. (E) Fluorescence change of Kv2.1 croscopic reversibility, which dictates that a state-dependent li- CHO cells after washout of 100 nM GxTX. Blue bar indicates application of 7 + gand will act allosterically to modulate channel function. Similar volumes of high [K ] external. Lines are fits of Eq. S4; gray, kcontrol = 0.0069 ± −1 −1 limitations affect all other affinity-based molecular probes. For 0.0002 s ; blue, khigh[K+] = 0.0335 ± 0.0003 s .(F) Summary of dissociation rates of GxTX-dy550. *P = 0.008 (G) GxTX-coated beads lose adhesion to instance, calcium dyes buffer calcium and voltage-dependent Kv2.1-expressing CHO cells in a depolarizing high [K+] solution. Each con- dyes increase capacitance. To minimize perturbation of the nected set of points represents time-matched samples from an experiment process studied, probes are ideally used at a trace concentration. conducted on a different day. Experimenter was blinded to solutions used. To avoid altering electrical signaling, VST probe can be applied **P < 0.05 Wilcoxon signed-rank directional test. at a concentration substantially below the Kd of its target VGIC

6of8 | www.pnas.org/cgi/doi/10.1073/pnas.1406876111 Tilley et al. Downloaded by guest on September 24, 2021 as in Fig. 4G. At these electrophysiologically inefficacious con- brain and body. Using these probes in intact tissue could enable PNAS PLUS centrations, only a small fraction of channels is bound by the measurement of which channel subtypes respond to different probe, yet optical signals from these probed channels can be stimuli within biological circuits. Our laboratories are currently detected. As the fraction of channel bound by a probe drops, the developing probes with improved optical and kinetic properties reporter signal diminishes, and hence a sensitive signal detection to report activation of ion channels in vivo. technique such as fluorescence microscopy is required. Report- ers with increased signal-to-noise ratios would enable use of even Conclusions lower trace concentrations of probe. Even brighter optical This study describes optical reporting of ion channel activation in reporters such as quantum dots or up-converting nanoparticles live cells. Our study demonstrates that: (i) voltage sensor toxins can (57) could enhance the VGIC activity-reporting capabilities of retain bioactivity after chemoselective conjugation to reporters, (ii) VST probes. fluorescent toxins colocalize with channels in live cells, (iii)toxin With a library of channel-specific optical probes, many secrets fluorescence reports channel activation, and (iv) toxins mediate of ion channel activity could be rapidly revealed. This library activity-dependent binding of channels to microbeads. The activa- could be engineered from the terrific variety of potent, VGIC tion-state-dependent binding of these reporters to live cells sug- subtype-selective VSTs that are increasingly being discovered. gests potential for dissecting specific contributions of ion channels Examples of VSTs from spiders include ω-Agatoxin-IVA, which + to cellular electrical activity and enabling discovery of novel ion binds P-type Ca2 channels (58) and the ProTx-I and -II peptides, + channel modulators. which bind Na channel subtypes (59–62), and heteropodatoxins, which bind Kv4 channels (63). The library of known VSTs is Materials and Methods growing rapidly as venom peptides and mRNAs of more preda- Details are in SI Materials and Methods. Peptides were synthesized by Fmoc tory animals are sequenced (64). Probes for different channel peptide synthesis, folded by air oxidation, and purified by reverse-phase activities could be developed from toxins that stabilize different HPLC. Identity of peptides and conjugates was verified by matrix-assisted conformations of ion channels. For example, the Magi5 peptide + laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry, stabilizes activated Na channels, whereas ProTx-II stabilizes + HPLC, and SDS/PAGE. Whole-cell voltage-clamp recording was used to mea- resting conformations (65, 66). Similarly, GxTX inhibits K sure currents from CHO-K1 cells stably expressing rat Kv2.1 channels. Live cell channels that are activated by the related VST, hanatoxin (35). imaging and patch-clamp fluorometry were performed on an inverted mi- All of these tarantula VSTs share a common fold, suggesting that croscope using light-emitting diode or laser excitation. Structural modeling each of them could be derivatized with chemoselective side was performed using the Rosetta-Membrane method (36–38). PHYSIOLOGY chains, as was GxTX. Fluorescent labeling of different VSTs could create a palette of probes to report activation and deac- ACKNOWLEDGMENTS. We thank Yi Liu of the Molecular Foundry for the 2+ + + BTTAA reagent. We are very grateful to many at the University of California, tivation of Ca ,Na ,andK channel subtypes. Davis: Oscar Cerda for his talented culturing of hippocampal neurons; James In addition, our results suggest a novel method of discovering Trimmer for ion channel plasmids and helpful discussion; Jie Zheng and Peter activity-dependent ligands for VGICs. A clear difference was Cala for feedback on the manuscript; and Sebastian Ayala, Christina Berry, seen in binding between Kv2.1 cells and GxTX beads when the and Yuanpei Li for providing technical assistance. This work was supported by US NIH Grants 5P30GM092328-02, R01NS042225-09S1, R25NS063307, and channels were voltage-activated (Fig. 5). This suggests that T32HL086350; American Association Grant 10SDG4220047; and conformation-selective ligands could be found by screening Milton L. Shifman Endowed Scholarship for the Neurobiology Course at combinatorial libraries using the OBOC method. In principle, an Woods Hole. Work at the Molecular Foundry was supported by the Office OBOC approach could be used to develop optical probes to of Science, Office of Basic Energy Sciences, of the US Department of Energy under Contract DE-AC02-05CH11231. This work is dedicated to the report activity of different VGIC subtypes. memory of coauthor Kenneth S. Eum (1987–2014). Ken was a talented Specific VGIC activity-sensing probes also have the potential PhD student, a driven and caring soul who brought joy to the lives of to map ion channel activation throughout large regions of the those who knew him.

1. Kass RS (2005) The channelopathies: Novel insights into molecular and genetic 14. Sack JT, Aldrich RW (2006) Binding of a gating modifier toxin induces intersubunit mechanisms of human disease. J Clin Invest 115(8):1986–1989. cooperativity early in the Shaker K channel’s activation pathway. J Gen Physiol 128(1): 2. Waxman SG (2013) Painful Na-channelopathies: An expanding universe. Trends Mol 119–132. Med 19(7):406–409. 15. Sack JT, Aldrich RW, Gilly WF (2004) A gastropod toxin selectively slows early tran- 3. Vacher H, Mohapatra DP, Trimmer JS (2008) Localization and targeting of voltage- sitions in the Shaker K channel’s activation pathway. J Gen Physiol 123(6):685–696. dependent ion channels in mammalian central neurons. Physiol Rev 88(4):1407–1447. 16. Lee HC, Wang JM, Swartz KJ (2003) Interaction between extracellular Hanatoxin and 4. Kaczorowski GJ, McManus OB, Priest BT, Garcia ML (2008) Ion channels as drug tar- the resting conformation of the voltage-sensor paddle in Kv channels. Neuron 40(3): gets: The next GPCRs. J Gen Physiol 131(5):399–405. 527–536. 5. Yu FH, Yarov-Yarovoy V, Gutman GA, Catterall WA (2005) Overview of molecular 17. Catterall WA, et al. (2007) Voltage-gated ion channels and gating modifier toxins. relationships in the voltage-gated ion channel superfamily. Pharmacol Rev 57(4): Toxicon 49(2):124–141. 387–395. 18. Catterall WA (1979) Binding of scorpion toxin to receptor sites associated with sodium 6. Norris AJ, Foeger NC, Nerbonne JM (2010) Neuronal voltage-gated K+ (Kv) channels channels in frog muscle. Correlation of voltage-dependent binding with activation. function in macromolecular complexes. Neurosci Lett 486(2):73–77. J Gen Physiol 74(3):375–391. 7. Norris AJ, Foeger NC, Nerbonne JM (2010) Interdependent roles for accessory KChIP2, 19. Catterall WA (1977) -dependent binding of scorpion toxin to the KChIP3, and KChIP4 subunits in the generation of Kv4-encoded IA channels in cortical Na+ ionophore. Studies with a toxin derivative prepared by lacto- pyramidal neurons. J Neurosci 30(41):13644–13655. peroxidase-catalyzed iodination. J Biol Chem 252(23):8660–8668. 8. Chen X, Johnston D (2006) Voltage-gated ion channels in dendrites of hippocampal 20. McDonough SI, Lampe RA, Keith RA, Bean BP (1997) Voltage-dependent inhibition of pyramidal neurons. Pflugers Arch 453(3):397–401. N- and P-type calcium channels by the peptide toxin omega-grammotoxin-SIA. Mol 9. Burkhalter A, Gonchar Y, Mellor RL, Nerbonne JM (2006) Differential expression of Pharmacol 52(6):1095–1104. I(A) channel subunits Kv4.2 and Kv4.3 in mouse visual cortical neurons and synapses. 21. Phillips LR, et al. (2005) Voltage-sensor activation with a tarantula toxin as cargo. J Neurosci 26(47):12274–12282. Nature 436(7052):857–860. 10. Johnston D (2010) The Na+ channel conundrum: Axon structure versus function. Nat 22. Swartz KJ (2007) Tarantula toxins interacting with voltage sensors in potassium Neurosci 13(7):784–785. channels. Toxicon 49(2):213–230. 11. O’Connell KM, Loftus R, Tamkun MM (2010) Localization-dependent activity of the 23. Swartz KJ, MacKinnon R (1997) Mapping the receptor site for hanatoxin, a gating Kv2.1 delayed-rectifier K+ channel. Proc Natl Acad Sci USA 107(27):12351–12356. modifier of voltage-dependent K+ channels. Neuron 18(4):675–682. 12. Fox PD, Loftus RJ, Tamkun MM (2013) Regulation of Kv2.1 K(+) conductance by cell 24. Li-Smerin Y, Swartz KJ (2000) Localization and molecular determinants of the Ha- surface channel density. J Neurosci 33(3):1259–1270. natoxin receptors on the voltage-sensing domains of a K(+) channel. J Gen Physiol 13. Liu PW, Bean BP (2014) Kv2 channel regulation of action potential and 115(6):673–684. firing patterns in superior cervical ganglion neurons and hippocampal CA1 pyramidal 25. Jung HJ, et al. (2005) Solution structure and lipid membrane partitioning of VSTx1, an neurons. J Neurosci 34(14):4991–5002. inhibitor of the KvAP potassium channel. Biochemistry 44(16):6015–6023.

Tilley et al. PNAS Early Edition | 7of8 Downloaded by guest on September 24, 2021 26. Milescu M, et al. (2007) Tarantula toxins interact with voltage sensors within lipid 47. Suchyna TM, et al. (2004) Bilayer-dependent inhibition of mechanosensitive channels membranes. J Gen Physiol 130(5):497–511. by neuroactive peptide enantiomers. Nature 430(6996):235–240. 27. Milescu M, et al. (2009) Interactions between lipids and voltage sensor paddles de- 48. Swartz KJ, MacKinnon R (1997) Hanatoxin modifies the gating of a voltage-dependent tected with tarantula toxins. Nat Struct Mol Biol 16(10):1080–1085. K+ channel through multiple binding sites. Neuron 18(4):665–673. 28. Wang JM, et al. (2004) Molecular surface of tarantula toxins interacting with voltage 49. Liu R, Marik J, Lam KS (2002) A novel peptide-based encoding system for “one-bead sensors in K(v) channels. J Gen Physiol 123(4):455–467. one-compound” peptidomimetic and small molecule combinatorial libraries. JAm 29. Herrington J, et al. (2006) Blockers of the delayed-rectifier potassium current in Chem Soc 124(26):7678–7680. pancreatic beta-cells enhance glucose-dependent insulin secretion. Diabetes 55(4): 50. Baek HG, Liu R, Lam KS (2009) Development of hydrogel TentaGel shell-core beads for 1034–1042. ultrahigh throughput solution-phase screening of encoded OBOC combinatorial small 30. Schmalhofer WA, et al. (2009) A KV2.1 gating modifier binding assay suitable for high molecule libraries. J Comb Chem 11(1):91–102. throughput screening. Channels (Austin) 3(6):437–447. 51. Yao N, et al. (2009) Discovery of targeting ligands for breast cancer cells using the 31. Li XN, et al. (2013) The role of voltage-gated potassium channels Kv2.1 and Kv2.2 in one-bead one-compound combinatorial method. J Med Chem 52(1):126–133. the regulation of insulin and somatostatin release from pancreatic islets. J Pharmacol 52. Trimmer JS (1991) Immunological identification and characterization of a delayed – Exp Ther 344(2):407 416. rectifier K+ channel polypeptide in rat brain. Proc Natl Acad Sci USA 88(23): 32. Speca DJ, et al. (2014) Deletion of the Kv2.1 delayed rectifier potassium channel leads 10764–10768. – to neuronal and behavioral hyperexcitability. Genes Brain Behav 13(4):394 408. 53. Du J, Haak LL, Phillips-Tansey E, Russell JT, McBain CJ (2000) Frequency-dependent 33. Lee S, et al. (2010) Solution structure of GxTX-1E, a high-affinity tarantula toxin in- regulation of rat hippocampal somato-dendritic excitability by the K+ channel sub- teracting with voltage sensors in Kv2.1 potassium channels. Biochemistry 49(25): unit Kv2.1. J Physiol 522(Pt 1):19–31. 5134–5142. 54. Murakoshi H, Trimmer JS (1999) Identification of the Kv2.1 K+ channel as a major 34. Smith JJ, Cummins TR, Alphy S, Blumenthal KM (2007) Molecular interactions of the component of the delayed rectifier K+ current in rat hippocampal neurons. J Neurosci gating modifier toxin ProTx-II with NaV 1.5: Implied existence of a novel toxin binding 19(5):1728–1735. site coupled to activation. J Biol Chem 282(17):12687–12697. 55. Misonou H, et al. (2006) Bidirectional activity-dependent regulation of neuronal ion 35. Milescu M, Lee HC, Bae CH, Kim JI, Swartz KJ (2013) Opening the shaker K+ channel channel phosphorylation. J Neurosci 26(52):13505–13514. with hanatoxin. J Gen Physiol 141(2):203–216. 56. Pal S, Hartnett KA, Nerbonne JM, Levitan ES, Aizenman E (2003) Mediation of neu- 36. Yarov-Yarovoy V, et al. (2012) Structural basis for gating charge movement in the ronal apoptosis by Kv2.1-encoded potassium channels. J Neurosci 23(12):4798–4802. voltage sensor of a . Proc Natl Acad Sci USA 109(2):E93–E102. 57. Gargas DJ, et al. (2014) Engineering bright sub-10-nm upconverting nanocrystals for 37. Yarov-Yarovoy V, Baker D, Catterall WA (2006) Voltage sensor conformations in the single-molecule imaging. Nature Nanotechnol 9(4):300–305. open and closed states in ROSETTA structural models of K(+) channels. Proc Natl Acad 58. Mintz IM, et al. (1992) P-type calcium channels blocked by the spider toxin omega- Sci USA 103(19):7292–7297. Aga-IVA. Nature 355(6363):827–829. 38. Yarov-Yarovoy V, Schonbrun J, Baker D (2006) Multipass membrane protein structure 59. Middleton RE, et al. (2002) Two tarantula peptides inhibit activation of multiple so- prediction using Rosetta. Proteins 62(4):1010–1025. dium channels. Biochemistry 41(50):14734–14747. 39. Gray JJ, et al. (2003) Protein-protein docking with simultaneous optimization of rigid- 60. Schmalhofer WA, et al. (2008) ProTx-II, a selective inhibitor of NaV1.7 sodium chan- body displacement and side-chain conformations. J Mol Biol 331(1):281–299. 40. Long SB, Tao X, Campbell EB, MacKinnon R (2007) Atomic structure of a voltage- nels, blocks action potential propagation in nociceptors. Mol Pharmacol 74(5): – dependent K+ channel in a lipid membrane-like environment. Nature 450(7168): 1476 1484. 376–382. 61. Sokolov S, Kraus RL, Scheuer T, Catterall WA (2008) Inhibition of sodium channel 41. Choi J, Yoon NM (1995) Synthesis of disulfides by copper-catalyzed disproportionation gating by trapping the domain II voltage sensor with protoxin II. Mol Pharmacol – of thiols. JOrgChem60(11):3266–3267. 73(3):1020 1028. 42. Besanceney-Webler C, et al. (2011) Increasing the efficacy of bioorthogonal click re- 62. Bosmans F, Martin-Eauclaire MF, Swartz KJ (2008) Deconstructing voltage sensor actions for bioconjugation: A comparative study. Angew Chem Int Ed Engl 50(35): function and pharmacology in sodium channels. Nature 456(7219):202–208. 8051–8056. 63. Sanguinetti MC, et al. (1997) Heteropodatoxins: Peptides isolated from spider venom 43. Shimony E, Sun T, Kolmakova-Partensky L, Miller C (1994) Engineering a uniquely that block Kv4.2 potassium channels. Mol Pharmacol 51(3):491–498. reactive thiol into a cysteine-rich peptide. Protein Eng 7(4):503–507. 64. Jungo F, Bougueleret L, Xenarios I, Poux S (2012) The UniProtKB/Swiss-Prot Tox-Prot 44. Li S, et al. (2009) Synthesis and characterization of a high-affinity alphavbeta6-specific program: A central hub of integrated venom protein data. Toxicon 60(4):551–557. ligand for in vitro and in vivo applications. Mol Cancer Ther 8(5):1239–1249. 65. Corzo G, et al. (2007) Solution structure and alanine scan of a spider toxin that affects 45. Hunter MJ, Komives EA (1995) Deprotection of S-acetamidomethyl cysteine-containing the activation of mammalian voltage-gated sodium channels. J Biol Chem 282(7): peptides by silver trifluoromethanesulfonate avoids the oxidation of methionines. Anal 4643–4652. Biochem 228(1):173–177. 66. Bosmans F, Puopolo M, Martin-Eauclaire MF, Bean BP, Swartz KJ (2011) Functional 46. Ruta V, MacKinnon R (2004) Localization of the voltage-sensor toxin receptor on properties and toxin pharmacology of a dorsal root ganglion sodium channel viewed KvAP. Biochemistry 43(31):10071–10079. through its voltage sensors. J Gen Physiol 138(1):59–72.

8of8 | www.pnas.org/cgi/doi/10.1073/pnas.1406876111 Tilley et al. Downloaded by guest on September 24, 2021