1521-0111/84/5/763–773$25.00 http://dx.doi.org/10.1124/mol.113.088971MOLECULAR PHARMACOLOGY Mol Pharmacol 84:763–773, November 2013Copyright ª 2013 by The American Society for Pharmacology and Experimental Therapeutics

Subtype-Selective Activation of Kv7 Channels by AaTXKb(2–64),a Novel Toxin Variant from the Androctonus australisScorpion Venom

Zied Landoulsi, Francesco Miceli, Angelo Palmese, Angela Amoresano, Gennaro Marino,Mohamed El Ayeb, Maurizio Taglialatela, and Rym BenkhalifaLaboratoire des Venins et Molécules Thérapeutiques, Institut Pasteur de Tunis, Université Tunis-El Manar, Tunis-Belvédère,Tunisia (Z.L., M.E.A., R.B.); Division of Pharmacology, Department of Neuroscience, University of Naples Federico II, Naples, Italy(F.M., M.T.); Department of Chemical Sciences, University of Naples Federico II, Naples, Italy (A.P., A.A., G.M.); Department ofMedicine and Health Science, University of Molise, Campobasso, Italy (M.T.); and Unidad de Biofísica, Consejo Superior deInvestigaciones Cientificas, Universidad del Pais Vasco, Leioa, Spain (M.T.)

Received August 2, 2013; accepted September 9, 2013

ABSTRACTKv7.4 channel subunits are expressed in central auditory pathwaysand in inner ear sensory hair cells and skeletal and smooth musclecells. Openers of Kv7.4 channels have been suggested to improvehearing loss, systemic or pulmonary arterial hypertension, urinaryincontinence, gastrointestinal and neuropsychiatric diseases, andskeletal muscle disorders. Scorpion venoms are a large source ofpeptides active on K1 channels. Therefore, we have optimizeda combined purification/screening procedure to identify specificmodulator(s) of Kv7.4 channels from the venom of the NorthAfrican scorpion Androctonus australis (Aa). We report the isola-tion and functional characterization of AaTXKb(2–64), a novel var-iant of AaTXKb(1–64), in a high-performance liquid chromatography

fraction from Aa venom (named P8), which acts as the first peptideactivator of Kv7.4 channels. In particular, in both Xenopus oocytesand mammalian Chinese hamster ovary cells, AaTXKb(2–64),but not AaTXKb(1–64), hyperpolarized the threshold voltage ofcurrent activation and increased the maximal currents ofheterologously expressed Kv7.4 channels. AaTXKb(2–64) alsoactivated Kv7.3, Kv7.2/3, and Kv7.5/3 channels, whereashomomeric Kv1.1, Kv7.1, and Kv7.2 channels were unaffected.We anticipate that these results may prove useful in unravelingthe novel biologic roles of AaTXKb(2–64)-sensitive Kv7 chan-nels and developing novel pharmacologic tools that allowsubtype-selective targeting of Kv7 channels.

IntroductionThe Kv7 subfamily of voltage-gated potassium channels

consists of five members (Kv7.1–5) that each play distinctpathophysiologic roles in various cell types (Soldovieri et al.,2011). Kv7.1 subunits represent the molecular basis of IKs,a cardiac current involved in action potential repolarization,whereas Kv7.2 and Kv7.3 genes have a predominant neuronalexpression. Heteromeric assembly of Kv7.2 and Kv7.3 sub-units underlies the M-current (IKM), a subthreshold currentinhibiting repetitive action potential firing. Kv7.4 channelsare expressed in central auditory pathways and in inner ear

sensory hair cells (Kubisch et al., 1999), in skeletal musclecells (Iannotti et al., 2013), and in smooth muscle cells of thegastrointestinal tract (Jepps et al., 2009; Ipavec et al., 2011).In blood vessels, Kv7.4 channels contribute to vascular reactivityto vasopressors and vasodilators (Yeung et al., 2007; Mani andByron, 2011; Martelli et al., 2013), and down-regulation of Kv7.4contributes to primary and secondary hypertension (Jepps et al.,2011). Kv7.5 subunits are expressed in neurons, showinga distribution that largely overlaps those of Kv7.2 and Kv7.3,and are thought to contribute to IKM heterogeneity (Schroederet al., 2000); Kv7.5 transcripts and/or proteins have been alsoidentified in skeletal and smooth muscle cells (Soldovieriet al., 2011).Activation of K1 currents mediated by distinct Kv7 subunit

combinations is currently regarded as an effective pharmaco-logic strategy against several human diseases, with cardiacKv7.1 channels being potential targets for antiarrhythmics(Tamargo et al., 2004), and neuronal Kv7.2 and Kv7.3channels for antiepileptics and analgesics (Miceli et al.,

This work was supported by grants from the Pasteur Institute of Tunis,Ministry of Higher Education and Scientific Research [Tox11] (to R.B.);Telethon [GGP07125], the Fondazione San Paolo–IMI [Project Neuroscience],Regione Molise [Convenzione AIFA/Regione Molise], the Italian Ministry forEducation and Research [PRIN 2009], Provincia di Avellino [2012] (to M.T.);Regione Campania 2013 [PON01_01802 and PON01_00117] (to A.A.); anda traineeship grant from the International Division of the Pasteur Institute[RIIP/EC/MAM/N 117/11] (to Z.L.).

dx.doi.org/10.1124/mol.113.088971.

ABBREVIATIONS: Aa, Androctonus australis; ACN, acetonitrile; a-KTx, short chain potassium channel toxin; b-KTx, long chain potassium channeltoxin; CHO cells, Chinese hamster ovary cells; HPLC, high-performance liquid chromatography; LC-MS/MS, liquid chromatography–tandem massspectrometry; KTX, kaliotoxin; KTX2, kaliotoxin-2; MALDI-TOF, matrix-assisted laser desorption/ionization–time-of-flight; MarTx, martentoxin;MBS, modified Barth’s solution; retigabine, N-(2-amino-4-(4-fluorobenzylamino)-phenyl)-carbamic acid ethyl ester; SIM, single-ion monitoring; TFA,trifluoroacetic acid.

763

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

2008). Kv7.4-selective openers may improve hearing loss(Leitner et al., 2012), systemic or pulmonary arterialhypertension (Mani and Byron, 2011), urinary incontinence(Rode et al., 2010), gastrointestinal (Jepps et al., 2009; Ipavecet al., 2011) and neuropsychiatric (Hansen et al., 2008)diseases, and skeletal muscle disorders (Iannotti et al.,2010, 2013). However, most of available synthetic compoundsdiscriminated poorly among Kv7 channels formed by distinctcombinations of Kv7.2-5 subunits (Xiong et al., 2008).Therefore, compounds showing increased Kv7 subtype-selectivity, beside representing important probes to elucidatethe molecular pathophysiology of Kv7 channels at distinctsites, may offer additional therapeutic benefits (Wickendenet al., 2008).Scorpion venoms are a large source of peptides active on K1

channels (KTx) (Rodriguez de la Vega and Possani, 2004).According to their size, KTx can be classified into short chaintoxins (a-KTx), consisting of 23 to 42 amino acids peptidesstabilized by three or four disulfide bridges (Tytgat et al.,1999), and long chain toxins (b-KTx) of 45 to 75 residuesstabilized by three disulfide bridges (Diego-Garcia et al.,2007). Two structurally distinct domains can be recognized inb-KTx: a putatively a-helical N terminus and a cysteine-richC-terminal domain containing a common structural motifnamed Cs-a/b for cysteine-stabilized a-helix/b-sheet (Zhuet al., 2004). Remarkably, venom peptides mostly act as K1

channel blockers whereas naturally occurring K1 channel ac-tivators have been rarely found.Several KTx have been described in Androctonus scorpion

species. These include a-KTx such as kaliotoxin (KTX) (Crestet al., 1992) and kaliotoxin-2 (KTX2) (Laraba-Djebari et al.,1994), and among b-KTx AaTXKb (Legros et al., 1998).Therefore, we investigated Androctonus australis (Aa) scor-pion venom as a potential source of novel molecular entitiesactive on Kv7 channels. We report the biochemical andelectrophysiologic characterization of AaTXKb(2–64), a novelmolecular variant of AaTXKb, as the first peptide activator ofKv7.4 channels which also acts as a subtype-selectiveactivator of channels formed by Kv7.3, Kv7.2/3, and Kv7.5/3subunits.

Materials and MethodsTrypsin, a-cyano-4-hydroxycinnamic acid, dithiothreitol, and

NH4HCO3 were obtained from Sigma-Aldrich (St. Louis, MO). High-performance liquid chromatography (HPLC)-grade trifluoroaceticacid (TFA) was obtained from Carlo Erba (Milan, Italy). Retigabine[N-(2-amino-4-(4-fluorobenzylamino)-phenyl)-carbamic acid ethyl es-ter] was obtained from Valeant Pharmaceuticals (Aliso Viejo, CA). Allother reagents and solvents were of the highest purity available(J. T. Baker, Phillipsburg, NJ).

Scorpion Venom. Venom of Aa was collected from Beni Khedach(Tunisia) by the veterinarian service of the Pasteur Institute of Tunisand was kept frozen at 220°C in its crude form until use.

Purification Procedures. Crude venom was dissolved in waterand loaded on Sephadex G50 gel filtration chromatography columns(2 � K26/100; Pharmacia; GE HEalthCare, Velizy-Villacoublay,France). Columns were equilibrated and eluted with 0.1 M aceticacid buffer, pH 4.7. After freeze drying, the resolved fractions werestored at220°C until use. The elution profile of M2Aawas eventuallycollected in four subfractions (M2a, M2b, M2c, and M2d). HPLCpurification of M2c was performed using a C8 reversed-phase HPLCcolumn (5 mm, 4.6 � 250 mm; Beckman Coulter, Brea, CA) equipped

with a Beckman series 125 pump and Beckman diode array detectorset at 214 and 280 nm. Elution was controlled by means of the GOLDsoftware (32 Karat version 7.0; Beckman Coulter) at a rate of 0.8 ml/min using a linear gradient (30 minutes) from 20% to 50% of solutionB [0.1% TFA in acetonitrile (ACN)] in solution A (0.1% TFA in water).The protein concentration was measured by the Bradford proteinassay (Fermentas, Cornaredo, Italy) using bovine serum albumin asa standard.

cRNA Transcription. The cDNA encoding for human Kv7.4 wascloned in the Xenopus oocyte expression vector pTLN. After lineariza-tion with HpaI, capped cRNA was transcribed in vitro using the SP6mMessage mMachine kit (Ambion, Foster City, CA).

Expression in Xenopus laevis Oocytes. X. leavis females wereanesthetized with 0.17% solution of 3-aminobenzoic acid ethyl estermethanesulfonate salt (Sigma-Aldrich, Saint-Quentin Fallavier,France), and parts of ovaries were surgically removed from theabdominal cavity and bathed in sterile modified Barth’s solution (MBS)of the following composition (in mM): 88 NaCl, 1 KCl, 0.4 MgCl2,2.4 NaHCO3, 0.8 MgSO4, 10 HEPES, 2.4 CaCl2, and 0.3 Ca(NO3)2,pH 7.4. Xenopus oocytes were then defolliculated enzymaticallyby incubation for 2 hours in sterile MBS containing 2 mg/mlcollagenase (type A and type B; Roche, Indianapolis, IN), followed bythree to four washes in MBS. Stage V and VI oocytes were injectedwith 50 nl of mRNA (10 ng/oocyte) using an automatic microinjector(Nanoject; Drummond Scientific, Broomall, PA). Oocytes were kept at18°C in sterile MBS supplemented with 0.1 mM gentamycin untilelectrophysiologic experiments were performed.

Cell Culture, Transient Transfection, and Plasmids. Chan-nel subunits were expressed in Chinese hamster ovary (CHO) cells bytransient transfection, using plasmids containing cDNAs encodinghuman Kv7.1, Kv7.2, Kv7.3, Kv7.4, Kv7.5, and Kv1.1, all cloned inpcDNA3.1 (Life Technologies, Monza, Italy). Two additional plasmidsencoding for human Kv7.5 subunits were also tested: Kv7.5 inpcDNA3zeo, and Kv7.5 in a pIRES2-DsRed-Express vector from OpenBiosystems (Thermo Scientific, Lafayette, CO). According to theexperimental protocol, these plasmids were expressed individually orin combination, together with a plasmid-expressing enhanced greenfluorescent protein (Clontech, Palo Alto, CA) used as a transfectionmarker. Total cDNA in the transfection mixture was kept at 4 mg.CHO cells were grown in 100-mm plastic Petri dishes in Dulbecco’smodified Eagle’s medium containing 10% fetal bovine serum, non-essential amino acids (0.1 mM), penicillin (50 U/ml), and streptomycin(50 mg/ml) in a humidified atmosphere at 37°C with 5% CO2. At 24hours before transfection, the cells were plated on glass coverslips(Carolina Biologic Supply Company, Burlington, NC) coated withpoly-L-lysine, and were transfected on the next day with theappropriate cDNA using Lipofectamine 2000 (Life Technologies),according to the manufacturer’s protocol. Electrophysiologic experi-ments were performed 24 hours after transfection.

Two-Microelectrode Voltage Clamp. Four to six days aftermRNA injection, the two-microelectrode voltage-clamp measure-ments in Xenopus oocytes were performed at room temperature (22–24°C) using a Geneclamp 500B amplifier combined with Digidata1322A (Molecular Devices, Foster City, CA). Micropipettes werepulled from borosilicate glass capillaries on a Flaming/Brown typepipette puller (P-97; Sutter Instruments, Novato, CA) and had a tipresistance of 1–2 MV when filled with 3 M KCl. Data were filtered at1 KHz, and voltage step protocols and current analysis were performedwith pCLAMP 8.2 software (Molecular Devices). During the record-ings, oocytes were perfused with ND96 solution containing (in mM):96 NaCl, 4 KCl, 1 MgC12, 1.8 CaC12, 5 HEPES, at pH 7.6, in a smallchamber (2 ml volume). The perfusion system was controlled bya Manifold Solution Changer (MSC-200; Bio-Logic, Grenoble, France).From a holding potential of 280 mV, Kv7.4 currents were activatedwith voltage pulses from 2100 to 140 mV in 10-mV steps, followed bya constant pulse to 2120 mV of 300-millisecond duration.

Patch Clamp. Currents from CHO cells were recorded at roomtemperature (20–22°C) using the whole-cell configuration of the

764 Landoulsi et al.

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

patch-clamp technique, with glass micropipettes of 3–5 MV re-sistance. No compensation was performed for pipette resistance andcell capacitance. During the recording, constant perfusion of extra-cellular solution was maintained. The extracellular solution con-tained (in mM): 138 NaCl, 2 CaCl2, 5.4 KCl, 1 MgCl2, 10 glucose, and10 HEPES, at pH 7.4, with NaOH. The pipette (intracellular) solutioncontained (in mM): 140 KCl, 2 MgCl2, 10 EGTA, 10 HEPES, 5 Mg-ATP, at pH 7.3–7.4, with KOH. From a holding potential of 280 mV,Kv7.4 currents were activated either by 1.5-second depolarizingpulses from 280 mV to 140 mV in 110 mV increments, followed byan isopotential pulse at 0 mV of 500-millisecond duration, or by 3-second voltage ramps from 280 mV to 120 mV applied at 0.08 Hzfrequency. Current were recorded using an Axopatch-200A amplifier,filtered at 5 kHz, and digitized using a DigiData 1440A (MolecularDevices). The pCLAMP software (version 10.2) was used for dataacquisition and analysis (Molecular Devices).

Electrophysiologic Data Analysis. For Kv7.4 currents inXenopus oocytes, whole-cell conductance (G) was calculated accordingto the following equation: G 5 Imax/(V 2 EK), where I is the steady-state current measured at the end of each depolarizing step, V is thestep potential, and EK is the reversal potential for potassium, whichwas calculated to be 298 mV. Conductance values were expressed asa function of membrane potential and were fit to a single Boltzmanndistribution of the following form: y 5 max/[1 1 exp(V1/2 2 V)/k],where V is the test potential, V1/2 is the half-activation potential, andk is the slope factor (Wickenden et al., 2008).

For Kv7.4 currents in CHO cells, conductance-voltage curves (G/Vcurves; activation curves) were generated by normalizing to themaximal value of the instantaneous isopotential currents at 0 mV andexpressing the normalized values as a function of the precedingvoltages. Data were fit to a sum of two independent Boltzmanndistributions (B1 and B2) of the following form: y5max1/(11 exp(V12V)/k1) 1 max2/(1 1 exp(V2 2 V)/k2), where V is the test potential, V1

and V2 the half-activation potentials, k1 and k2 the slopes, and max1and max2 the relative amplitudes for each Boltzmann distribution(Castaldo et al., 2002). Ramp-evoked currents were analyzed bymeasuring the currents at 240 mV (a membrane potential value closeto the activation threshold) and at 0mV (amembrane potential value atwhich Kv7.4 conductance is largely saturated), and the effect of eachvenom fraction was expressed as the percent of current increase atthese two membrane voltages. Data analysis and fit were performedusing the Origin software (version 8.0; OriginLab, Northampton, MA).

Disulfide Bridges Reduction and Enzymatic Hydrolysis.HPLC-purified venom fractions were dissolved in 50 mM NH4HCO3

buffer at pH 8.0, and dithiothreitol was added to a final concentrationof 2 mM. Samples were allowed to react at 37°C for 2 hours. Enzy-matic digestion was then performed by adding trypsin with anenzyme/substrate ratio of 1/50 w/w at 37°C for 16 hours in thepresence of dithiothreitol.

MALDI Mass Spectrometry. Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS) experiments wereperformed on a Voyager-DE STR matrix-assisted laser desorption/ionization–time-of-flight (MALDI-TOF) mass spectrometer (AppliedBiosystems, Framingham, MA) equipped with a nitrogen laser(337 nm). We mixed 1 ml of the sample (1/1, v/v) with a 10 mg/ml solutionof a-cyano-4-hydroxycinnamic acid in ACN/50 mM citrate buffer,70/30 v/v. Spectra were acquired using a mass (m/z) range of 400–5000amu for the peptide mass fingerprint and 3000–8000 amu for themolecular weight determination.

Tandem Mass Spectrometry. The peptide mixtures werefiltered using 0.22 mm polyvinylidene fluoride membrane (Millipore,Billerica, MA) and were analyzed using a 6520 Accurate-Massquadrupole time-of-flight mass spectrometer liquid chromatography/mass spectrometer system equipped with a 1200 HPLC system anda chip cube (Agilent Technologies, Palo Alto, CA). After sampleloading, the peptide mixture was first concentrated and washed on40 nl enrichment column (Agilent Technologies chip), with 0.1% formicacid in 2% ACN as the eluent. The sample was then fractionated on

a C18 reverse-phase capillary column at flow rate of 400 nl/min, witha linear gradient of eluent B (0.1% formic acid in 95% ACN) in A (0.1%formic acid in 2% ACN) from 7 to 80% in 50 minutes. Peptide analysiswas performed using data-dependent acquisition of one MS scan (massrange from 300 to 1800 m/z) followed by tandem mass spectrometry(MS/MS) scans of the five most abundant ions in each MS scan.Dynamic exclusion was used to acquire a more complete survey of thepeptides by automatic recognition and temporary exclusion (0.5minutes) of ions from which definitive mass spectral data had beenpreviously acquired. Nitrogen at a flow rate of 3 l/min and heated to325°C was used as the dry gas for spray desolvation. MS/MS spectrawere measured automatically when the MS signal surpassed thethreshold of 50,000 counts. Double and triple charged ions werepreferably isolated and fragmented over single charged ions.

Protein Identification. The acquired MS/MS spectra weretransformed in mz.data format and used for protein identificationwith a licensed version of MASCOT 2.1 (Matrix Science, Boston, MA);the calculation of false discovery rate in this Mascot version is proneto several errors (http://www.matrixscience.com/pdf/2011WKSHP5.pdf), thus the false discovery rate values were not calculated forour peptides. Raw data from nano liquid chromatography–tandemmass spectrometry (LC-MS/MS) analyses were used to query theSwissProt database available at the Uniprot KB site (http://web.expasy.org/docs/relnotes/relstat.html); it contained a total numberof 468,851 sequences (166,149,756 residues). No taxonomic restrictionwas specified.

Mascot search parameters were: trypsin as enzyme, allowednumber of missed cleavage 2, oxidation of methionine (115.99492Da) and pyro-Glu N-termQ (217.02655 Da) as variablemodifications,10 ppm MS tolerance, 0.6 Da MS/MS tolerance, and peptide chargefrom 12 to 13. Spectra with a MASCOT score ,25 were low qualityand were rejected.

Single-Ion Monitoring Analysis. Peptide mixtures, obtained aspreviously described, were analyzed by LC-MS/MSusing a 4000QTrap(Applied Biosystems) coupled to a NanoLC-2D Plus system equippedwith a cHiPLC-Nanoflex System (Eksigent, Dublin, CA). The mixturewas loaded on an Eksigent Nano cHiPLC Trap column (200 mm� 0.5mmChromXP C18-CL 3 mm 300 Å) at 2 ml/min (A solvent 0.1% formicacid, loading time 10 minutes). Peptides were separated on anEksigent reverse-phase Nano cHiPLC column (75 mm � 15 cmChromXP C18-CL 3 mm 120 Å), at a flow rate of 0.3 ml/min with a 5%to 65% linear gradient in 60 minutes (A solvent 0.1% formic acid, 2%ACN in water; B solvent 0.1% formic acid, 2% water in ACN).Nanospray source was used at 2.8 kV with liquid coupling, witha declustering potential of 80 V, using an uncoated silica tip fromNew Objective (Woburn, MA) (O.D. 150 mm, I.D. 20 mm, T.D. 10 mm).Spectra acquisition was based on a survey enhanced resolution scanat 250 amu/s for the ion at m/z 722.45, 481.96, 658.40, and 439.27.Enhanced product ion scans (MS/MS) were performed at 4000 amu/s,and collision voltages were calculated automatically by rolling collisionenergy. Data were acquired and processed using Analyst software(Applied Biosystems).

NH2-Terminal Amino Acid Sequence Determination. TheNH2-terminal sequence of reduced/alkylated peptides (1 nmol) wasperformed for the first 10 cycles using an automatic liquid-phaseprotein sequencer (model 476A; Applied Biosystems, Foster City, CA)using Edman protein degradation, according to standard procedures.

Sequence Data Analysis and Statistics. Similarity searches onprotein sequences were performed after the BLAST analysis (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The theoretical molecular weightwas calculated by ProtParam (http://us.expasy.org/cgibin/protparam),and the trypsin potential cleavage sites were predicted usingPeptideCutter (http://web.expasy.org/peptide_cutter/).

Statistical differences between data groups, expressed as mean 6S.E.M., were tested using an unpaired two-tailed Student’s t test,assuming that the population follows a Gaussian distribution.Differences were considered statistically significantly different versusrespective controls when P , 0.05.

A Novel Toxin Activating Kv7 Channel Subtypes 765

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

ResultsEffects of Gel Filtration-Purified Aa Venom Frac-

tions on Kv7.4 Channels Expressed in XenopusOocytes. To search for naturally occurring peptides activeon Kv7.4 channels from the Aa scorpion, crude Aa venom wasseparated by gel filtration. The resulting profile (Fig. 1A) showsfive partially resolved fractions; among these, four correspond tonontoxic fractions (M1, M2, M3, and M4), and one (G50) wastoxic in vivo (Miranda et al., 1970). Fractions M1, M2, and G50,containing molecules having a molecular weight range .2 kDa,were selected for functional screening on Kv7.4 channels ex-pressed in Xenopus oocytes by two-microelectrode voltage-clamp; M3 andM4 fractions were not studied, as they containedlow-molecular-weight material in relatively low abundance.Kv7.4 currents were activated by membrane depolarizationabove 240 mV, and displayed slow activation kinetics and lackof inactivation (Fig. 1, B, left, and C) (Kubisch et al., 1999). Kv7.4currents were blocked by 76%6 5% (n5 3;140 mV) by 100 mMof the nonselective Kv7 channel blocker linopirdine. Oocyteperfusion for 5 minutes with M2 (50 mg/ml) caused a 38.6% 62.5% increase in the Kv7.4 peak currents at 140 mV, togetherwith a negative shift in the midpoint potential of activation(Fig. 1, B, right, and C).In fact, analysis of the G/V curves for Kv7.4 currents revealed

that the half activation voltage (V1/2) was 215.8 6 0.7 mV incontrols and 225.0 6 2.5 after M2 exposure (n 5 5; P , 0.05);no statistically significant change was instead observed in theslopes (k) of the G/V curve between controls (14.8 6 0.4 mV/e-fold) andM2-treated (14.26 0.7 mV/e-fold) G/V relationships(n 5 5; P . 0.05). Both these effects were largely reversibleafter 5 minutes washout of the M2 fraction. The effect of M2 onKv7.4 channels was dose-dependent between 1 and 100 mg/ml(Fig. 1D); higher concentrations could not be tested because ofthe limited material availability. In contrast to M2, perfusionwith G50 (50 mg/ml) did not modify Kv7.4 currents, whereas50 mg/ml M1 significantly decreased Kv7.4 current amplitude

by 51.4%6 9% (Fig. 1E). Given our interest in Kv7.4 activatorsrather than blockers, this latter effect was not pursued anyfurther.

Effect of M2 Subfractions on Kv7.4 Channels Ex-pressed in Mammalian Cells. The M2 fraction containsa heterogeneous mixture of molecules ranging from 7 to 12 kDa.To achieve a better size-separation, the M2 fraction elutedfrom the Sephadex G-50 column was collected into foursubfractions: M2a, M2b, M2c, and M2d (Fig. 2A). Given thelittle amount of purified fractions available, the effect on Kv7.4channels exerted by each M2 subfraction was then studiedusing the whole-cell configuration of the patch-clamp techniquein mammalian cells transiently transfected with Kv7.4 cDNA;this strategy allowed us to use much less material than thatrequired for Xenopus oocytes experiments.In addition, to decrease the length of each experiment and

further reduce the amount of material needed, we tested theeffects of the four M2 subfractions using a voltage protocol inwhich Kv7.4 currents were activated by 3-second voltage rampsfrom 280 to 120 mV applied at 0.08 Hz frequency. Perfusionwith 50 mg/ml of the M2c subfraction, corresponding to thedescending flank of the M2 fraction (Fig. 2A), enhanced ramp-evoked Kv7.4 currents (Fig. 2B).As shown in Fig. 2C, the extent of Kv7.4 current enhancement

by M2c was larger at240 mV (144.4%6 3.8%) than at 0 mV(124.5% 6 1.7%), thus largely reproducing the previouslydescribed effects exerted by the whole M2 fraction on Kv7.4channels expressed in Xenopus oocytes. M2c was the only M2subfraction active on Kv7.4 channels; in fact, perfusion withM2a, M2b, and M2d (each at 50 mg/ml) failed to affect thethreshold activation potential or the maximal currentamplitude of the expressed channels (Fig. 2C).

Effect of M2c Subfractions on Kv7.4 ChannelsExpressed in Mammalian Cells. The M2c gel filtrationsubfraction was resolved on a C8 analytic HPLC column,leading to the identification of nine peaks (P1–P9) (Fig. 3A).

Fig. 1. Effects of Aa venom purified gel-filtrationfractions on Kv7.4 channels expressed in Xenopusoocytes. (A) Gel filtration chromatogram of the Aavenom. (B) Effect of the M2 fraction on Kv7.4currents in a representative cells. Currents tracesfrom Kv7.4 channels were recorded before (leftpanel) and 5 minutes after the application of M2(50 mg/ml; right). As shown in the panel, currentswere activated by 2-second depolarizing pulsesfrom a holding potential of 280 mV to potentialsranging from 2100 mV to +40 mV in 10-mVincrements, followed by a 300-millisecond step to2120mV. Scale bar: vertical 0.5 mA; horizontal 0.5seconds. (C) Effect of the M2 fraction (50 mg/ml) onthe activation curve of Kv7.4 channels (n = 5). Thesolid lines represent the fits of the experimentaldata to a Boltzmann distribution (see Materialsand Methods). The dotted line is the M2 data fittrace normalized to controls. (D) Concentration-dependent effects of M2 on Kv7.4 currents. Thecalculated EC50 value is 58 mg/ml, and the Hillcoefficient is 1.6. (E) Effect of M1, M2, and G50 Aafractions (each at 50 mg/ml) on Kv7.4 currents at+10 mV (n = 5 for M1; n = 8 for M2; and n = 4 forG50; *P , 0.05 versus controls).

766 Landoulsi et al.

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

When each of these purified peaks were separately tested(each at 20 mg/ml) for their effect on Kv7.4 currents by patch-clamp experiments, it was found that only P8, eluted at19 minutes, enhanced Kv7.4 currents (Fig. 3B). Again, asalready observed for the whole M2 fraction and the M2csubfraction, P8 effectively increased Kv7.4 currents at everypotential tested, although the enhancement was more

significant around the membrane potential values closer tothe activation threshold (240 mV) than at 0 mV, a membranepotential value at which the K1 conductance is largelysaturated (Fig. 3C; see also Fig. 4B). None of the other peaksresulting from M2c purification modified Kv7.4 currents at240 or 0 mV (Fig. 3B). Figure 3C shows a representativeexample of Kv7.4 current activation by cell perfusion with

Fig. 2. Effect of M2 subfractions on Kv7.4 channelsexpressed in mammalian cells. (A) The gel filtrationchromatogram of fraction M2 was divided into foursubfractions (M2a, M2b, M2c, and M2d). (B) Kv7.4 currentsexpressed in CHO cells were activated by a 3-second voltageramp from 280 mV to +20. Recording were obtained beforeand after the application of M2c (50 mg/ml). The dotted lineindicates the zero-current level. Scale bar: vertical 0.2 nΑ;horizontal 500 milliseconds. (C) Effect of the 4 M2subfractions on Kv7.4 currents at 240 mV (filled bars) or0mV (empty bars; n = 3 forM2a andM2d; n = 5 forM2b; andn = 6 for M2c; *P , 0.05 relative to controls).

Fig. 3. Effect of M2c HPLC subfractions on Kv7.4 channelsexpressed in mammalian cells. (A) HPLC separation ofM2c. We loaded 50 mg of M2c per HPLC run on an analyticC8 reverse-phase column and separated using a gradient ofbuffer B (0.1% TFA in ACN) as described in the experi-mental procedures and represented in the figure by a dottedline. Collected fractions were labeled with numbers (from 1to 9). (B) Effect of the nine HPLC peaks (each at 20 mg/ml)purified from M2c and of retigabine (RTG; 10 mM) on Kv7.4current amplitude at 240 mV (filled bars) or 0 mV (emptybars) relative to controls (n = 3 for P1, P6, P9; n = 4 for P3;n = 5 for P2, P4, P5; n = 6 for P7; n = 19 for P8; and n = 10 forRTG). Currents were expressed as a percentage of thecontrol current amplitude at 240 mV and 0 mV for eachfraction; statistical comparison for each fraction wasperformed relative to their respective controls (*P , 0.05).(C) Ramp-evoked Kv7.4 currents recorded before and afterthe application of 20 mg/ml of P8. The dotted line indicatesthe zero-current level. Scale bar: vertical 0.2 nΑ, horizontal500 milliseconds. (D) Time course of Kv7.4 current poten-tiation by P8 (50 mg/ml) and RTG (10 mM). The steady-stateamplitudes of the peak currents at 240 mV were plotted asa function of time. The bars on top of the plot indicate theduration of drug application.

A Novel Toxin Activating Kv7 Channel Subtypes 767

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

20 mg/ml of P8 using the previously described voltage rampprotocol.Notably, the effects of P8 on Kv7.4 currents were quali-

tatively similar to those of the prototype Kv7 activatorretigabine (Fig. 3B). To compare the kinetics of onset andoffset of the Kv7.4-activating effects of P8 with those ofretigabine, time-course experiments were performed. Asshown in Fig. 3D, P8-induced Kv7.4 activation developedslowly, requiring about 1 minute to reach steady-state levels;similarly, basal current values were recovered after about 1minute upon drug removal. By contrast, the kinetics of onsetand offset of 10 mM retigabine appeared significantly faster,requiring less than 30 seconds to reach steady-state levels(Fig. 3D). During the 4 minutes of 50 mg/ml P8 perfusion, themaximal response decreased by an average of 14.2% 6 4.3%(n 5 8).The voltage ramp protocol used in the previously described

experiments did not allow us to precisely define the effect ofthe P8 fraction on the steady-state properties of the Kv7.4conductance; therefore, in subsequent experiments, Kv7.4channels were activated by depolarizing pulses at constantvoltages. Figure 4A shows a family of Kv7.4 currents recordedbefore (left panel) and after (right panel) P8 exposure.Activation curves revealed that P8 effectively increased bothmaximal conductance and caused a leftward shift in theactivation threshold of Kv7.4 currents. In particular, experi-mental data analysis revealed that a sum of two Boltzmannequations (B1 and B2) was required to properly fit the data(Castaldo et al., 2002). Under control conditions, the twoBoltzmann components had a 60/40 ratio in their relativeamplitude (max1 and max2 accounted for 59% 6 4% and42% 6 4% of the total conductance, respectively), and distincthalf-activation potentials (V1 5 238.1 6 1.0 mV; V2 5 27.3 6

2.1 mV) and slopes (k1 5 9.2 6 0.6 mV/e-fold; k2 5 15.0 61.3 mV/e-fold). After P8 exposure, no statistically significantchanges were observed in the relative amplitudes (max1 andmax2 accounted for 62% 6 1% and 44% 6 1% of the totalconductance, respectively), nor in V2 (25.1 6 2.3 mV) or k2(19.26 3.6 mV/e-fold); by contrast, both V1 (247.76 2.4 mV)and k1 (12.2 6 0.5 mV/e-fold) were statistically significantlydifferent from controls (n 5 5; P , 0.05). On the other hand,retigabine failed to affect the relative amplitudes (max1 andmax2 accounted for 66% 6 12% and 35% 6 12% of the totalconductance, respectively), or the slopes (k15 11.16 1.8 mV/e-fold ; k2 5 11.1 6 2.2 mV/e-fold) of B1 and B2; the midpointpotentials of both Boltzmann distribution were shifted in thehyperpolarizing direction by retigabine (V1 5 257.56 2.1 mV;V2 5 229.4 6 6.8 mV) (n 5 5; P , 0.05). The inset of Fig. 4Bcompares the normalized activation curves for control, P8, andretigabine.Finally, within the tested range (5–50 mg/ml), the effect of

P8 on Kv7.4 current enhancement at 240 mV was concentra-tion dependent (Fig. 4C).Mass Spectral Analysis of the P7 and P8 Fractions.

Given that P8 contained the biologically active molecule(s) ofinterest, this fraction was directly submitted to MALDI-TOFmass spectrometry analysis; the same analysis was alsoperformed on P7, which despite showing an HPLC elutionprofile very close to P8 was unable to modify Kv7.4 currents.MALDI-MS spectra revealed the occurrence of a major signalhaving a m/z value of 6991.42 (Fig. 5A) and 7119.15 (Fig. 5B)in P8 and P7, respectively. Both P8 and P7 fractions weresubjected to disulfide bridges reduction and trypsin di-gestion, and the obtained peptide mixture was analyzed byMALDI-MS and LC-MS/MS. LC-MS/MS data analysis led tothe identification in both fractions of AaTXKb from Aa

Fig. 4. Effect of P8 on the activation curve of Kv7.4channels expressed in CHO cells, and dose-de-pendent enhancement of the currents at 240 mVby P8. (A) Representative macroscopic currenttraces recorded from a CHO cell expressing Kv7.4homomeric channels in response to the voltageprotocol indicated before (left) and after (right)application of 20 mg/ml P8. Scale bar: vertical0.5 nΑ; horizontal 500milliseconds. (B) Steady-stateactivation curves for Kv7.4 currents before (emptysymbols) and after (filled symbols) application of20 mg/ml P8 (n = 6). Continuous lines represent fitsof the experimental data to a double Boltzmanndistribution. The inset shows a comparison amongthe normalized activation curves for P8 (20 mg/ml;dashed line) and retigabine (RTG; 10 mM; dottedline) and those of controls (Ctl; continuous line). (C)Concentration-dependent effects of P8 on Kv7.4currents calculated at 240 mV (n = 3 for 5 mg/ml;n = 4 for 10 mg/ml; n = 19 for 20 mg/ml; and n = 8 for50 mg/ml). Asterisks indicate values significantlydifferent (P , 0.05) versus controls.

768 Landoulsi et al.

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

(UniProtKB P69939), whose sequence was inferred from anAa cDNA library (Legros et al., 1998). Figure 5, C and D,reports the P8 and P7 fractions sequence coverage obtained byboth MALDI-MS and LC-MS/MS experiments, mapped ontothe AaTXKb mature sequence. In both P8 and P7 fractions,the N terminus peptide escaped mass spectral analysis.Because the mass difference detected in MALDI analysisbetween P8 and P7 major species (127.73 Da) could well bedue to the removal during the proprotein maturation of the N-terminal lysine residue (expected molecular mass 128.18 Da),P8 and P7 mixtures were submitted to LC-MS/MS analysis insingle-ion monitoring (SIM) mode (Fig. 6), searching for bothpeptides 1–13 (KLVKYAVPVGTLR) and2–13 (LVKYAVPVGTLR).In P7, the presence of the signal related to peptide 1–13 and thecontemporary absence of the signal related to the peptide 2–13(Fig. 6, B and D) demonstrated the occurrence of the full-lengthAaTXKb in this fraction. On the other hand, SIM analysisrevealed the presence of the signal relative to peptide 2–13 asthe major specie in P8 (Fig. 6, A and C). A good sequencecoverage for both 2–13 and 1–13 peptides was obtained by theoccurrence of intense y-ion series signals as shown in Fig. 6,C and D. However, no signals attributable to the fragmentpeaks that appear to extend from the amino terminus (“b ions”)for both P7 and P8 peptides could be detected in the masstandem spectra, probably due to the physicochemical charac-teristics of the peptides.Because the MS/MS data did not unambiguously identify

the N-terminal sequences, Edman degradation of S-pyridyl-ethylated P8 and P7 was performed, leading to the identi-fication of the N-terminal amino acid partial sequencesLVKYAV and KLVKYAV in P8 and P7, respectively.Altogether, results from the molecular weight determination,mass mapping analysis, SIM analysis, and Edman degrada-tion strongly suggested that the P8 peptide active on Kv7.4channels corresponds to a novel variant of AaTXKb that lacksthe N-terminal lysine; this peptide is thus called AaTXKb(2–64).By contrast, the same experimental procedures revealed thatthe inactive P7 fraction corresponded to the full-length sequenceof AaTXKb(1264).Only a monomeric form was observed in the mass spectra of

both AaTXKb(1264) and AaTXKb(2264) polypeptides, suggest-ing that no intermolecular disulphide bridges were formed.These results indicate that the six cysteines of AaTXK b(1264)

and AaTXK b(2264) are engaged in three intramoleculardisulphide bridges, as previously hypothesized (Legros et al.,1998).Effect of P8 on Kv7.1, Kv7.2, Kv7.3, Kv7.2/3, Kv7.5/3,

and Kv1.1 Currents. To investigate whether, in addition toKv7.4 channels, other K1 channels were also sensitive toAaTXKb(2264), we performed experiments in CHO cells ex-pressing homomeric Kv7.1, Kv7.2, Kv7.3, and Kv1.1 chan-nels as well as heteromeric Kv7.2/3 and Kv7.5/3 channels.Given that the results obtained by MALDI-TOF, LC-MS/MS,SIM, and Edman degradation experiments revealed thatAaTXKb(2264) was the major component of P8, in consider-ation of the little amount of P8 available for the experiments,the P8 fraction was used for these experiments, thus avoidinganother step of AaTXKb(2264) purification.When activated by the described voltage ramp protocol,

Kv7.3 and Kv7.2/3 channels showed similar sensitivity as Kv7.4to modulation by P8 (Fig. 7A). In fact, at potentials close tothreshold values (240 mV), P8 significantly increased Kv7.3,Kv7.2/3, and Kv7.4 currents by about 70% (Fig. 7B). It isnoteworthy that, at more depolarized potentials (0 mV), P8only increased the maximal currents in homomeric Kv7.4 andheteromeric Kv7.2/3 channels, whereas no effect was observedon homomeric Kv7.3 channels (Fig. 7, A and B). Moreover,20mg/ml P8 enhanced the currents carried byKv7.5/3 heteromericchannels at 0mV by 25.5%6 1.4% (n5 4). Currents at240mVin Kv7.5/3-transfected cells as well as currents at any potentialin cells transfected only with Kv7.5 cDNA from three differentsources (see Materials and Methods) were too small to assessthe effects of P8 (data not shown). Finally, P8 failed to modifycurrent amplitudes at depolarized potentials or activationthreshold in Kv7.1, Kv7.2, and Kv1.1 channels (Fig. 7, A and B).

DiscussionWe have found that the nontoxic fraction from the venom of

the North African scorpion Aa (called M2) dose-dependentlyenhanced Kv7.4 currents in Xenopus oocytes and caused a10-mV negative shift in the midpoint potential of the activationcurve. Biologic activities of peptides from nontoxic venomfractions have been previously described in Buthus occitanustunetanus scorpion venom; these include lipolytic properties(Soudani et al., 2005), activation of nicotinic receptors in

Fig. 5. Biochemical characterization of C8 HPLC-elutedpeptides in P8 and P7. (A and B) Molecular weights ofpurified peptides determined by MALDI-TOF in P8 (A),and P7 (B). (C and D) Amino acid sequences of P8 (C) andP7 (D) determined by LC-MS/MS are in bold and under-lined, and mapped onto the full-length sequence ofAaTXKb(1–64).

A Novel Toxin Activating Kv7 Channel Subtypes 769

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

skeletal muscle cells (Cheikh et al., 2007), and inhibition ofcardiac L-type calcium currents (Cheikh et al., 2006).Within the nine HPLC fractions of M2, only the subfraction

called P8 displayed Kv7.4-activating properties. Indeed, P8caused an increase in the maximal currents, and it hyper-polarized and decreased the steepness of the activation curveof Kv7.4 channels. In P8, mass mapping data identifiedAaTXKb, a long chain b-KTX whose sequence was inferred

upon screening of a cDNA library from Aa venom glands(Legros et al., 1998). Surprisingly, same analysis performedon the P7 HPLC peak, which did not enhance Kv7.4 currents,also led to the identification of AaTXKb. The data obtainedshow that two different forms of AaTXKb, having very distinctbiologic activities and differing by 127.73 Da in molecularmass, were present in P7 and P8 fractions. The combinedresults from the SIM analysis and Edman degradation

Fig. 6. LC-MS/MS analysis in SIMmode of the N-terminal region ofHPLC-eluted peptides in P8 and P7.(A and B) Total ion chromatogram ofMS/MS analysis in SIM mode of P8(A) and P7 (B) searching for peptides2–13 (LVKYAVPVGTLR) and 1–13(KLVKYAVPVGTLR) in both sam-ples. Spectra acquisition was basedon a survey enhanced resolution scanat 250 amu/s for the ion atm/z 722.45,481.96, 658.40, and 439.27. (C and D)The MS/MS spectra corresponding tothe major species detected in SIManalyses are reported and properlyannotated. Enhanced product ion scans(MS/MS) were performed at 4000 amu/s.Peak intensity is expressed as counts perseconds (cps).

Fig. 7. Differential effect of P8 on currents carriedby distinct potassium channels. (A) Representativewhole-cell current traces from Kv7.1, Kv7.2, Kv7.3,Kv7.2/3, Kv7.4, and Kv1.1 channels activated by thepreviously described ramp protocol (except for Kv1.1in which the duration of the ramp protocol from280to +20 mV was 250 milliseconds instead of 3seconds). Asterisks indicate currents recorded afterperfusion with P8 (20 mg/ml). In each panel, thedotted lines indicate the zero-current level. (B)Effect of P8 (20 mg/ml) on current amplitudes ofthe indicated channels at 240 mV (filled bars) or0 mV (empty bars) (n = 3 for Kv7.1; n = 6 for Kv7.2;n = 3 for Kv7.3; n = 5 for Kv7.2/3; n = 19 for Kv7.4; andn = 3 for Kv1.1; *P , 0.05).

770 Landoulsi et al.

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

experiments revealed that the most abundant peptide in theP7 fraction unable to activate Kv7.4 currents was the full-length AaTXKb(1–64), whereas the P8 AaTXKb componentactive on Kv7.4 currents lacked the first N-terminal lysine[AaTXKb(2–64)]. Thus, the removal of the N-terminal residuefrom the protein mature form, possibly as a further step in thematuration process, does infer this dramatic functionalchange.The biologic functions of b-KTx are poorly characterized

(Diego-Garcia et al., 2007). Selective interference with Kv

channels has been verified only for TsTXKb (inhibition of86Rb1 efflux carried by noninactivating K1 channels in ratbrain synaptosomes; Rogowski et al., 1994), BmTXKb (block-ade of ITO current in rabbit atrial myocytes; Cao et al., 2003a),MeuTXKb1 (inhibition of 125I-kaliotoxin binding to rat brainsynaptosomal membranes; Zhu et al., 2010), and a chemicallydigested C-terminal part of HgebKTX (reduction of Kv1channel currents expressed in Xenopus oocytes; Diego-Garciaet al., 2008). By contrast, no study has been performed toisolate the corresponding peptide from Aa venom, and theactivity of AaTXKb on K1 channels (or on any ion channel) hasnever been investigated. AaTXKb has a high sequencesimilarity (∼88% identity) with MeuTXKb1 from Mesobuthuseupeus (Fig. 8), a peptide showingweak affinity for K1 channelsbut exhibiting anti-Plasmodial and cytolytic activities, andwhose structural features have been resolved (Zhu et al., 2010).It has been suggested that, when compared with TsTXKb(Rogowski et al., 1994), the weak affinity of MeuTXKb1 for K1

channels may be explained by its extended N-terminal domainwith nine extra residues, highlighting a critical role of thisregion for target recognition. AaTXKb also lacks this N-terminal extension and, similarly to TsTXKb, could be a“truncated” version of MeuTXKb1. Notably, a critical in-fluence of the N-terminal region for scorpion toxins interactionwith K1 channels has been already described for KTX2(Legros et al., 1997) and for HgebKTX (Diego-Garcia et al.,2008). Similarly, chemical modifications and mutagenesisstudies have revealed that the N-terminal amino group or the

N-terminal basic residues of scorpion b-neurotoxins (or both)are critical for their interaction with voltage-gatedNa1 channels(Polikarpov et al., 1999).Altogether, these data suggest that b-KTx having the

classic Cs-a/b scaffold not only can produce inhibitory effectson K1 channels but can also activate ionic conductances.This ability has been previously described for martentoxin(MarTx), an a-KTx purified from the East-Asian scorpionMesobuthus martensi venom, which enhances the activities ofnative and cloned BK channels (Tao et al., 2011), and DkTx(double-knot toxin) purified from the venom of the Chinesebird tarantula Ornithoctonus huwena, which selectively andirreversibly activates transient receptor potential vanilloid-1channels (Bohlen et al., 2010). Notably, a MarTx variant(MarTx-1) lacking of the N-terminal phenylalanine displayedsimilar biologic activities when compared with the full-lengthMarTx (Cao et al., 2003b).The ability of AaTXKb(2-64) to cause an hyperpolarization

shift in Kv7.4 channels gating is reminiscent of the effectsexerted on voltage-gated Na1 channels by b-scorpion toxinsacting at site 4 (Dutertre and Lewis, 2010); these toxinsenhance voltage-dependent activation of Na1 channels bymodifying the movement of the S4 segments via a voltage-sensor trapping mechanism (Catterall et al., 2007). Toxin-induced sensor-trapping can also occur in voltage-gated Ca21

channels (agatoxins from spiders) (McDonough, 2007), and involtage-gated K1 channels of the Kv2 and Kv4 (hanatoxinsfrom tarantula) (Swartz and MacKinnon, 1997) or Kv3 andKv11 (sea anemone toxins) (Diochot and Lazdunski, 2009)subclasses. It has been reported that a conserved structuralmotif in voltage-gated K1 and Ca21 channels can accountfor the “promiscuity” of both hanatoxin and grammotoxin(another tarantula toxin) to interact with both channelclasses (Li-Smerin and Swartz, 1998) and, possibly, withnon–voltage-gated channels of the transient receptor potentialvanilloid-1 subfamily (Siemens et al., 2006).Distinct similarities and differences can be found between

AaTXKb(2–64) and previously described synthetic Kv7

Fig. 8. Amino acid sequence of the major components of P8 and P7, and comparison with other b-KTx. Amino acid sequence determined from MALDI-MS + LC/MS-MS, LC-SIM, and Edman degradation results of native P8 (AaTXKb(2–64)) and P7 (AaTXKb(2–64); Legros et al., 1998) were aligned withanalogous toxins: BmTXK-b-2 (Zhu et al., 1999) from Mesobuthus martensii, BuTXK-b from Buthus occitanus israelis, MeuTXK-b1 and MeuTXK-b2(Zhu et al., 2010) from Mesobuthus eupeus, Tdi-b-KTX and TdiKIK (Diego-Garcia et al., 2007) from Tityus discrepans, Tco-b-KTX and TcoKIK (Diego-Garcia et al., 2005) from Tityus costatus, Ttr-b-KTX and TtrKIK from Tityus trivittatus (Diego-Garcia et al., 2007), Hge-scorpine fromHadrurus gertschi(Diego-Garcia et al., 2008), and TsTXK-b from Tityus serrulatus (Rogowski et al., 1994). Gaps (–) have been introduced to optimize alignment. Signalpeptide and propeptide sequences are not reported. Residues are colored according to the following scheme: magenta, basic; blue, acid; red, nonpolar;green, polar. The six highly conserved C residues are highlighted in green. (|) indicates trypsin cleavage sites after marked amino acids for both nativeP8 [AaTXKb(2–64)] and P7 [AaTXKb(1–64)].

A Novel Toxin Activating Kv7 Channel Subtypes 771

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

channels openers. In fact, similarly to retigabine, the pro-totype Kv7 opener in clinical use as an anticonvulsant (Miceliet al., 2008), AaTXKb(2–64) hyperpolarizes the activationthreshold voltage and increases the maximal currents ofKv7.4 channels; however, at the concentrations tested, theefficacy of AaTXKb(2–64) on Kv7.4 currents seems lower thanthat of retigabine. The potency of retigabine in activatingchannels formed by distinct Kv7 subunits appears slightlydifferent, with Kv7.2 and Kv7.3 showing higher (EC50 , 1 mM),and Kv7.4 and Kv7.5 lower (EC50 . 1 mM) drug sensitivity(Tatulian et al., 2001; Dupuis et al., 2002). However, since thesedifferences are quantitatively rather small, retigabine iscurrently regarded as a poorly selective opener for channelscomposed of Kv7.2–5 subunits, being instead ineffective inKv7.1 channels.In our present experiments, in addition to homomeric Kv7.4,

AaTXKb(2–64) produced a hyperpolarizing shift in openingvoltage-dependence through homomeric Kv7.3 and hetero-meric Kv7.2/3 channels. At potential above 0mV, AaTXKb(2–64)

increased the maximal current in homomeric Kv7.4 andheteromeric Kv7.2/3 and Kv7.5/3 channels, but not in homo-meric Kv7.3 channels. Strikingly, both homomeric Kv7.1 andKv7.2, as well as the distantly related Kv1.1 channels wereunaffected by AaTXKb(2–64). In Kv7.4 channels, the effect ofAaTXKb(2–64) was relatively slow compared with retigabineand largely reversible, possibly suggesting an extracellular siteof action.Multiple mechanisms might cause the subtype-selective

macroscopic current increase at depolarized voltages byAaTXKb(2–64) in Kv7 channels; these include an increase inthe number of functional channels, of the single channelconductance, or of the maximal channel open probability (Po).The possibility that AaTXKb(2–64) recruits more active chan-nels at the plasma membrane seems unlikely, given that thetoxin actions were rather fast, requiring about 1 minute toachieve maximal effect, and were largely reversible. Consider-ing thatmost toxins acting as gatingmodulators of Kv channelsbind at a distance from the pore and do not affect single-channel conductance (Swartz and MacKinnon, 1997), we thusspeculate that AaTXKb(2–64) increases the Po of Kv7.4 channels.Although single-channel experiments were not performed toresolve this issue, the fact that AaTXKb(2–64) caused anhyperpolarization shift with no significant augmentation ofmaximal currents in Kv7.3 channels whose maximal Po is closeto unity (Li et al., 2004) seems in line with such speculation.Recently, acidic scorpion toxins belonging to new k-KTX2

and a-KTX28 subfamilies have been found to inhibit Kv7.1channels, although they did not affect other K1 channels thatare canonical targets for basic KTx (such as Kv1.3); un-fortunately, the pharmacologic profile of these toxins forchannels formed by other Kv7 subunit combination iscurrently unknown (Chen et al., 2012). Other compoundswith a certain degree of selectivity for Kv7 subtypes are: 1) theacrylamide derivative (S)-1 which preferentially activatesKv7.4 and Kv7.5 channels (Bentzen et al., 2006); 2) thesubstituted benzamide ICA-27243, which selectively in-creased the maximal conductance and negatively shifted thesteady-state activation curve in Kv7.2/Kv7.3 heterotetramers,with lower potency on Kv7.4 or Kv7.3/Kv7.5 channels and noactivity on Kv7.1 channels (Wickenden et al., 2008); 3) thediphenylamine carboxylate derivative NH29, which increasesthe current amplitude of Kv7.2, is weakly effective in Kv7.4,

and is ineffective on Kv7.3 and Kv7.1 channels (Peretz et al.,2010); 4) the cysteine-alkylating agent N-ethylmaleimide,which increases the currents carried by Kv7.2, Kv7.4, or Kv7.5but not by Kv7.1 and Kv7.3 channels (Li et al., 2004); and 5)zinc pyrithione, which potently activates Kv7.1, Kv7.2, Kv7.4,and Kv7.5 but not Kv7.3, channels (Xiong et al., 2007).Notably, at least two different binding regions, one in thepore and/or in the gating hinge and one in the voltage sensoritself, appear to contribute to the pharmacologic profileshown by synthetic openers for distinct Kv7 subunit com-binations (Miceli et al., 2011); whether these sites alsocontribute to the selectivity pattern of AaTXKb(2–64) remains tobe investigated.In conclusion, in the present study, we report the isolation

and functional characterization of AaTXKb(2–64), a novelvariant of AaTXKb from the nontoxic fraction of the Aascorpion acting as the first peptide activator of Kv7 channels.We anticipate that these results might prove useful to unravelnovel biologic roles of AaTXKb(2–64)-sensitive Kv7.4 channels,and to develop novel pharmacologic tools allowing subtype-selective targeting of Kv7 channels.

Acknowledgments

The authors thank Prof. Thomas J. Jentsch (Max-Delbrück-Center for Molecular Medicine, Berlin, Germany) for Kv7.2, Kv7.3,Kv7.4, and Kv7.5 cDNAs in pcDNA3.1; Prof. Mark Shapiro(University of Texas Health Science Center, San Antonio, TX) forKv7.5 cDNA in pcDNA3zeo; Prof. Anastasios Tzingounis (University ofConnecticut, Storrs, CT) for Kv7.5 cDNA in a pIRES2-DsRed-Expressvector; Prof. Enzo Wanke (University of Milan Bicocca, Milan, Italy)for Kv1.1 cDNA; and Dr. Eckhard Ficker (Rammelkamp Center,Cleveland, OH) for Kv7.1 cDNA. The authors also thank Prof.Hechmi Louzir (General Director of the Pasteur Institute of Tunis),and Dr. Najet Srairi-Abid (Venoms and Therapeutic MoleculesLaboratory, Pasteur Institute of Tunis) for help during the biochemicalpurification of P8; Dr. Zakaria Ben Lasfar and collaborators (Veteri-nary Laboratory, Pasteur Institute of Tunis) for providing Aa venomand Xenopus care; and Dr. Mejdoub Hafedh (University of Sciences,Sfax) for Edman degradation experiments. Further, the authors alsothank Dr. Kathy-Ann Koralek and the Dargut and Milena KemaliFoundation for organizing and sponsoring, respectively, the FirstKemali-IBRO Mediterranean School of Neuroscience held in Naples,on September 21–30, 2009, where this project was conceived, and Prof.Lucio Annunziato (Department of Neuroscience, University of NaplesFederico II, Naples, Italy) for generous support to the project.

Authorship Contributions

Participated in research design: Marino, El Ayeb, Taglialatela,Benkhalifa.

Conducted experiments: Landoulsi, Miceli, Palmese, Amoresano.Performed data analysis: Landoulsi, Miceli, Palmese, Amoresano,

Marino, Taglialatela, Benkhalifa.Wrote or contributed to the writing of the manuscript: Landoulsi,

Amoresano, Marino, Taglialatela, Benkhalifa.

References

Bentzen BH, Schmitt N, Calloe K, Dalby Brown W, Grunnet M, and Olesen SP(2006) The acrylamide (S)-1 differentially affects Kv7 (KCNQ) potassiumchannels. Neuropharmacology 51:1068–1077.

Bohlen CJ, Priel A, Zhou S, King D, Siemens J, and Julius D (2010) A bivalenttarantula toxin activates the capsaicin receptor, TRPV1, by targeting the outerpore domain. Cell 141:834–845.

Cao Z, Xiao F, Peng F, Jiang D, Mao X, Liu H, Li W, Hu D, and Wang T (2003a)Expression, purification and functional characterization of a recombinant scorpionvenom peptide BmTXKb. Peptides 24:187–192.

Cao ZY, Shen WQ, Pan YP, Xiao X, Liu XM, Wang XL, Liang XT, and Yu DQ (2003b)Purification, characterization of two peptides from Buthus martensi Karch. J PeptRes 62:252–259.

772 Landoulsi et al.

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from

Castaldo P, del Giudice EM, Coppola G, Pascotto A, Annunziato L, and TaglialatelaM (2002) Benign familial neonatal convulsions caused by altered gating of KCNQ2/KCNQ3 potassium channels. J Neurosci 22:RC199.

Catterall WA, Cestèle S, Yarov-Yarovoy V, Yu FH, Konoki K, and Scheuer T (2007)Voltage-gated ion channels and gating modifier toxins. Toxicon 49:124–141.

Cheikh A, Benkhalifa R, Bescond J, El Ayeb M, Raymond G, Cognard C, and PotreauD (2006) Depression of cardiac L-type calcium current by a scorpion venom fractionM1 following muscarinic receptors interaction involving adenylate cyclase path-way. Toxicon 48:373–387.

Cheikh A, Benkhalifa R, Potreau D, Raymond G, El Ayeb M, and Cognard C (2007) Anovel depolarizing activity of scorpion venom fraction M1 due to activation ofskeletal muscle nicotinic receptors. Toxicon 49:117–122.

Chen ZY, Zeng DY, Hu YT, He YW, Pan N, Ding JP, Cao ZJ, Liu ML, Li WX, and Yi Het al. (2012) Structural and functional diversity of acidic scorpion potassiumchannel toxins. PLoS ONE 7:e35154.

Crest M, Jacquet G, Gola M, Zerrouk H, Benslimane A, Rochat H, Mansuelle P,and Martin-Eauclaire MF (1992) Kaliotoxin, a novel peptidyl inhibitor of neuronalBK-type Ca21-activated K1 channels characterized from Androctonus mauretanicusmauretanicus venom. J Biol Chem 267:1640–1647.

Diego-García E, Batista CV, García-Gómez BI, Lucas S, Candido DM, Gómez-Lagunas F, and Possani LD (2005) The Brazilian scorpion Tityus costatus Karsch:genes, peptides and function. Toxicon 45:273–283.

Diego-García E, Schwartz EF, D’Suze G, González SA, Batista CV, García BI, de laVega RC, and Possani LD (2007) Wide phylogenetic distribution of Scorpine andlong-chain beta-KTx-like peptides in scorpion venoms: identification of “orphan”components. Peptides 28:31–37.

Diego-García E, Abdel-Mottaleb Y, Schwartz EF, de la Vega RC, Tytgat J,and Possani LD (2008) Cytolytic and K1 channel blocking activities of beta-KTxand scorpine-like peptides purified from scorpion venoms. Cell Mol Life Sci 65:187–200.

Diochot S and Lazdunski M (2009) Sea anemone toxins affecting potassium channels.Prog Mol Subcell Biol 46:99–122.

Dupuis DS, Schrøder RL, Jespersen T, Christensen JK, Christophersen P, JensenBS, and Olesen SP (2002) Activation of KCNQ5 channels stably expressed inHEK293 cells by BMS-204352. Eur J Pharmacol 437:129–137.

Dutertre S and Lewis RJ (2010) Use of venom peptides to probe ion channel structureand function. J Biol Chem 285:13315–13320.

Hansen HH, Waroux O, Seutin V, Jentsch TJ, Aznar S, and Mikkelsen JD (2008) Kv7channels: interaction with dopaminergic and serotonergic neurotransmission inthe CNS. J Physiol 586:1823–1832.

Iannotti FA, Barrese V, Formisano L, Miceli F, and Taglialatela M (2013) Specifi-cation of skeletal muscle differentiation by repressor element-1 silencing tran-scription factor (REST)-regulated Kv7.4 potassium channels. Mol Biol Cell 24:274–284.

Iannotti FA, Panza E, Barrese V, Viggiano D, Soldovieri MV, and Taglialatela M(2010) Expression, localization, and pharmacological role of Kv7 potassium chan-nels in skeletal muscle proliferation, differentiation, and survival after myotoxicinsults. J Pharmacol Exp Ther 332:811–820.

Ipavec V, Martire M, Barrese V, Taglialatela M, and Currò D (2011) KV7 channelsregulate muscle tone and nonadrenergic noncholinergic relaxation of the rat gas-tric fundus. Pharmacol Res 64:397–409.

Jepps TA, Chadha PS, Davis AJ, Harhun MI, Cockerill GW, Olesen SP, Hansen RS,and Greenwood IA (2011) Downregulation of Kv7.4 channel activity in primary andsecondary hypertension. Circulation 124:602–611.

Jepps TA, Greenwood IA, Moffatt JD, Sanders KM, and Ohya S (2009) Molecular andfunctional characterization of Kv7 K1 channel in murine gastrointestinal smoothmuscles. Am J Physiol Gastrointest Liver Physiol 297:G107–G115.

Kubisch C, Schroeder BC, Friedrich T, Lütjohann B, El-Amraoui A, Marlin S, Petit C,and Jentsch TJ (1999) KCNQ4, a novel potassium channel expressed in sensoryouter hair cells, is mutated in dominant deafness. Cell 96:437–446.

Laraba-Djebari F, Legros C, Crest M, Céard B, Romi R, Mansuelle P, Jacquet G, vanRietschoten J, Gola M, and Rochat H et al. (1994) The kaliotoxin family enlarged.Purification, characterization, and precursor nucleotide sequence of KTX2 fromAndroctonus australis venom. J Biol Chem 269:32835–32843.

Legros C, Céard B, Bougis PE, and Martin-Eauclaire MF (1998) Evidence for a newclass of scorpion toxins active against K1 channels. FEBS Lett 431:375–380.

Legros C, Feyfant E, Sampieri F, Rochat H, Bougis PE, and Martin-Eauclaire MF(1997) Influence of a NH2-terminal extension on the activity of KTX2, a K1

channel blocker purified from Androctonus australis scorpion venom. FEBS Lett417:123–129.

Leitner MG, Feuer A, Ebers O, Schreiber DN, Halaszovich CR, and Oliver D (2012)Restoration of ion channel function in deafness-causing KCNQ4 mutants by syn-thetic channel openers. Br J Pharmacol 165:2244–2259.

Li Y, Gamper N, and Shapiro MS (2004) Single-channel analysis of KCNQ K1

channels reveals the mechanism of augmentation by a cysteine-modifying reagent.J Neurosci 24:5079–5090.

Li-Smerin Y and Swartz KJ (1998) Gating modifier toxins reveal a conserved structuralmotif in voltage-gated Ca21 and K1 channels. Proc Natl Acad Sci USA 95:8585–8589.

Mani BK and Byron KL (2011) Vascular KCNQ channels in humans: the sub-threshold brake that regulates vascular tone? Br J Pharmacol 162:38–41.

Martelli A, Testai L, Breschi MC, Lawson K, McKay NG, Miceli F, Taglialatela M,and Calderone V (2013) Vasorelaxation by hydrogen sulphide involves activation ofKv7 potassium channels. Pharmacol Res 70:27–34.

McDonough SI (2007) Gating modifier toxins of voltage-gated calcium channels.Toxicon 49:202–212.

Miceli F, Soldovieri MV, Martire M, and Taglialatela M (2008) Molecular pharma-cology and therapeutic potential of neuronal Kv7-modulating drugs. Curr OpinPharmacol 8:65–74.

Miceli F, Soldovieri MV, Iannotti FA, Barrese V, Ambrosino P, Martire M, Cilio MR,and Taglialatela M (2011) The voltage-sensing domain of Kv7.2 channels as a mo-lecular target for epilepsy-causing mutations and anticonvulsants. Front Phar-macol 2:2.

Miranda F, Kupeyan C, Rochat H, Rochat C, and Lissitzky S (1970) Purification ofanimal neurotoxins. Isolation and characterization of eleven neurotoxins from thevenoms of the scorpions Androctonus australis hector, Buthus occitanus tunetanusand Leiurus quinquestriatus quinquestriatus. Eur J Biochem 16:514–523.

Peretz A, Pell L, Gofman Y, Haitin Y, Shamgar L, Patrich E, Kornilov P, Gourgy-Hacohen O, Ben-Tal N, and Attali B (2010) Targeting the voltage sensor of Kv7.2voltage-gated K1 channels with a new gating-modifier. Proc Natl Acad Sci USA107:15637–15642.

Polikarpov I, Junior MS, Marangoni S, Toyama MH, and Teplyakov A (1999) Crystalstructure of neurotoxin Ts1 from Tityus serrulatus provides insights into thespecificity and toxicity of scorpion toxins. J Mol Biol 290:175–184.

Rode F, Svalø J, Sheykhzade M, and Rønn LC (2010) Functional effects of the KCNQmodulators retigabine and XE991 in the rat urinary bladder. Eur J Pharmacol638:121–127.

Rodríguez de la Vega RC and Possani LD (2004) Current views on scorpion toxinsspecific for K1-channels. Toxicon 43:865–875.

Rogowski RS, Krueger BK, Collins JH, and Blaustein MP (1994) Tityustoxin K alphablocks voltage-gated noninactivating K1 channels and unblocks inactivating K1

channels blocked by alpha-dendrotoxin in synaptosomes. Proc Natl Acad Sci USA91:1475–1479.

Schroeder BC, Hechenberger M, Weinreich F, Kubisch C, and Jentsch TJ (2000)KCNQ5, a novel potassium channel broadly expressed in brain, mediates M-typecurrents. J Biol Chem 275:24089–24095.

Siemens J, Zhou S, Piskorowski R, Nikai T, Lumpkin EA, Basbaum AI, King D,and Julius D (2006) Spider toxins activate the capsaicin receptor to produce in-flammatory pain. Nature 444:208–212.

Soldovieri MV, Miceli F, and Taglialatela M (2011) Driving with no brakes: mo-lecular pathophysiology of Kv7 potassium channels. Physiology (Bethesda) 26:365–376.

Soudani N, Gharbi-Chihi J, Srairi-Abid N, Yazidi CM, Planells R, Margotat A,Torresani J, and El Ayeb M (2005) Isolation and molecular characterization ofLVP1 lipolysis activating peptide from scorpion Buthus occitanus tunetanus.Biochim Biophys Acta 1747:47–56.

Swartz KJ and MacKinnon R (1997) Hanatoxin modifies the gating of a voltage-dependent K1 channel through multiple binding sites. Neuron 18:665–673.

Tamargo J, Caballero R, Gómez R, Valenzuela C, and Delpón E (2004) Pharmacologyof cardiac potassium channels. Cardiovasc Res 62:9–33.

Tao J, Shi J, Yan L, Chen Y, Duan YH, Ye P, Feng Q, Zhang JW, Shu XQ, and Ji YH(2011) Enhancement effects of martentoxin on glioma BK channel and BK channel(a1b1) subtypes. PLoS ONE 6:e15896.

Tatulian L, Delmas P, Abogadie FC, and Brown DA (2001) Activation of expressedKCNQ potassium currents and native neuronal M-type potassium currents by theanti-convulsant drug retigabine. J Neurosci 21:5535–5545.

Tytgat J, Chandy KG, Garcia ML, Gutman GA, Martin-Eauclaire MF, van der WaltJJ, and Possani LD (1999) A unified nomenclature for short-chain peptides isolatedfrom scorpion venoms: alpha-KTx molecular subfamilies. Trends Pharmacol Sci 20:444–447.

Wickenden AD, Krajewski JL, London B, Wagoner PK, Wilson WA, Clark S, RoeloffsR, McNaughton-Smith G, and Rigdon GC (2008) N-(6-chloro-pyridin-3-yl)-3,4-difluoro-benzamide (ICA-27243): a novel, selective KCNQ2/Q3 potassium channelactivator. Mol Pharmacol 73:977–986.

Xiong Q, Gao Z, Wang W, and Li M (2008) Activation of Kv7 (KCNQ) voltage-gated potassium channels by synthetic compounds. Trends Pharmacol Sci 29:99–107.

Xiong Q, Sun H, and Li M (2007) Zinc pyrithione-mediated activation of voltage-gatedKCNQ potassium channels rescues epileptogenic mutants. Nat Chem Biol 3:287–296.

Yeung SY, Pucovský V, Moffatt JD, Saldanha L, Schwake M, Ohya S, and GreenwoodIA (2007) Molecular expression and pharmacological identification of a role for K(v)7 channels in murine vascular reactivity. Br J Pharmacol 151:758–770.

Zhu S, Li W, Zeng X, Jiang D, Mao X, and Liu H (1999) Molecular cloning andsequencing of two ‘short chain’ and two ‘long chain’ K1 channel-blocking peptidesfrom the Chinese scorpion Buthus martensii Karsch. FEBS Lett 457:509–514.

Zhu S, Huys I, Dyason K, Verdonck F, and Tytgat J (2004) Evolutionary traceanalysis of scorpion toxins specific for K-channels. Proteins 54:361–370.

Zhu S, Gao B, Aumelas A, del Carmen Rodríguez M, Lanz-Mendoza H, Peigneur S,Diego-Garcia E, Martin-Eauclaire MF, Tytgat J, and Possani LD (2010)MeuTXKbeta1, a scorpion venom-derived two-domain potassium channel toxin-like peptide with cytolytic activity. Biochim Biophys Acta 1804:872–883.

Address correspondence to: Maurizio Taglialatela, Department of Medicineand Health Sciences, University of Molise, Via De Sanctis, 86100, Campobasso,Italy. E-mail: m.taglialatela@unimol.it; and Rym Benkhalifa, Laboratoiredes Venins et Molécules Thérapeutiques, Group of Cellular Electrophysiology,Institut Pasteur de Tunis, BP74 - 1002 Tunis, Tunisia. E-mail: rym.benkhalifa@pasteur.rns.tn

A Novel Toxin Activating Kv7 Channel Subtypes 773

at ASPE

T Journals on N

ovember 30, 2018

molpharm

.aspetjournals.orgD

ownloaded from