ION CHANNELS, RECEPTORS AND TRANSPORTERS

Tetrodotoxin blocks L-type Ca2+ channels in canineventricular cardiomyocytes

Bence Hegyi & László Bárándi & István Komáromi &Ferenc Papp & Balázs Horváth & János Magyar &

Tamás Bányász & Zoltán Krasznai &Norbert Szentandrássy & Péter P. Nánási

Received: 3 April 2012 /Revised: 3 May 2012 /Accepted: 3 May 2012 /Published online: 22 May 2012# Springer-Verlag 2012

Abstract Tetrodotoxin (TTX) is believed to be the mostselective inhibitor of voltage-gated fast Na+ channels inexcitable tissues, including nerve, skeletal muscle, andheart, although TTX sensitivity of the latter is lower thanthe former by at least three orders of magnitude. In thepresent study, the TTX sensitivity of L-type Ca2+ current(ICa) was studied in isolated canine ventricular cells usingconventional voltage clamp and action potential voltageclamp techniques. TTX was found to block ICa in a revers-ible manner without altering inactivation kinetics of ICa.Fitting results to the Hill equation, an IC50 value of55±2 μM was obtained with a Hill coefficient of unity(1.0±s0.04). The current was fully abolished by 1 μMnisoldipine, indicating that it was really ICa. Under actionpotential voltage clamp conditions, the TTX-sensitive

current displayed the typical fingerprint of ICa, whichwas absent in the presence of nisoldipine. Stick-and-ballmodels for Cav1.2 and Nav1.5 channel proteins wereconstructed to explain the differences observed betweenaction of TTX on cardiac ICa and INa. This is the firstreport demonstrating TTX to interact with L-type calciumcurrent in the heart.

Keywords Calcium channels . Dog heart . Sodiumchannels . Tetrodotoxin . Voltage clamp

Introduction

It is generally believed that the marine guanidine toxintetrodotoxin (TTX), similar to saxitoxin (STX), is a highlyselective inhibitor of voltage-gated Na+ channels in mam-malian excitable tissues. This is really the case regardingskeletal muscle and neural tissues where INa is blocked byTTX in the nanomolar range. In cardiac tissues, however,micromolar concentrations of TTX are required to suppressNa+ current effectively [3, 5], except for Na+ channels in thefish heart which displayed TTX sensitivity in the nanomolarrange [12]. As a consequence, TTX concentrations as highas 30 μM had to be applied in order to achieve a fullsuppression of INa in voltage-clamped mammalian cardio-myocytes [5].

TTX-sensitive Ca2+ current components have beenidentified in cardiac tissues under pathological—butnever under physiological—conditions such as in hyper-trophied guinea pig [6] or infarcted rat hearts [2]. Fur-thermore, a TTX-sensitive fraction of a current, carriedby Ca2+ ions in the absence of external Na+, has beendescribed in a variety of cardiac tissues, including

B. Hegyi : L. Bárándi :B. Horváth : J. Magyar : T. Bányász :N. Szentandrássy : P. P. Nánási (*)Department of Physiology, University of Debrecen,P.O. Box 22, 4012 Debrecen, Hungarye-mail: nanasi.peter@med.unideb.hu

I. KomáromiClinical Research Center and Thrombosis Haemostasisand Vascular Biology Research Group of the Hungarian Academyof Sciences, University of Debrecen,P.O. Box 40, 4012 Debrecen, Hungary

F. PappDepartment of Biophysics and Cell Biology,Cell Biology and Signaling Research Group of the HungarianAcademy of Sciences, University of Debrecen,P.O. Box 39, 4012 Debrecen, Hungary

Z. KrasznaiDepartment of Biophysics and Cell Biology,University of Debrecen,P.O. Box 39, 4012 Debrecen, Hungary

Pflugers Arch - Eur J Physiol (2012) 464:167–174DOI 10.1007/s00424-012-1114-y

human [17], rat [1], and guinea pig [7, 14] myocardium.From these experiments, one may conclude that Ca2+ isthe charge carrier of this current likely flowing throughNa+ channels in the absence of external Na+ [7, 14, 23].Similarly to the studies mentioned above (i.e., underNa+-free conditions), the most important cardiac iso-forms of T-type Ca2+ current (Cav3.1 and Cav3.2) wereshown to be resistant to TTX [25]. On the other hand,little is known on the effect of TTX on the L-type Ca2+

channel, except for one study performed in mice whereno effect of TTX on ICa could be observed [24]. Weaimed to reexamine the selectivity of TTX on INa overICa by studying the effect of TTX on L-type Ca2+

current in isolated canine ventricular myocytes. Thispreparation was chosen because its electrophysiologicalparameters are believed to be most similar to those ofhuman regarding the distribution and kinetic propertiesof transmembrane ion currents [26, 27].

Methods

Isolation of single canine ventricular myocytes

Adult beagle dogs of either sex were anesthetized withintravenous injections of 10 mg/kg ketamine hydrochlo-ride (Calypsol, Richter Gedeon, Budapest, Hungary)+1 mg/kg xylazine hydrochloride (Sedaxylan, EurovetAnimal Health BV, Bladel, The Netherlands) accordingto a protocol approved by the local ethical committeeand conforming to the “Principles of laboratory animalcare” (NIH publication no. 85–23, revised 1985). Thehearts were quickly removed and placed in Tyrode’ssolution. Single myocytes were obtained by enzymaticdispersion using the segment perfusion technique [16].Briefly, a wedge-shaped section of the ventricular wallsupplied by the left anterior descending coronary arterywas dissected, cannulated, and perfused with oxygenizedTyrode’s solution which contained (in millimolar): NaCl,144; KCl, 5.6; CaCl2, 2.5; MgCl2, 1.2; HEPES, 5; anddextrose, 11; at pH07.4. Perfusion was maintained untilthe removal of blood from the coronary system, andthen it was switched to a nominally Ca2+-free Jokliksolution (Minimum Essential Medium Eagle, JoklikModification, Sigma-Aldrich Co., St. Louis, MO, USA)for 5 min. This was followed by 30-min perfusion withJoklik solution supplemented with 1 mg/ml collagenase(Type II., Worthington Biochemical Co., Lakewood, NJ,USA) and 0.2 % bovine serum albumin (Fraction V.,Sigma-Aldrich Co., St. Louis, MO, USA) containing50 μM Ca2+. Portions of the left ventricular wall werecut into small pieces, and the cell suspension, obtainedat the end of the procedure predominantly from the

midmyocardial region of the left ventricle, was washedwith Joklik solution. Finally, the Ca2+ concentration wasgradually restored to 2.5 mM. The cells were stored inMinimum Essential Medium until use.

Conventional voltage clamp

All electrophysiological measurements were performed at37 °C. The rod-shaped viable cells showing clear striationwere sedimented in a plexiglass chamber allowing con-tinuous superfusion with oxygenized Tyrode’s solution.Suction pipettes, fabricated from borosilicate glass, hadtip resistance of 2 MΩ after filling with pipette solutioncontaining (in millimolar): KCl, 110; KOH, 40; HEPES,10; EGTA, 10; TEACl, 20; K-ATP, 3, at pH07.2. ICawas recorded in Tyrode’s solution supplemented with3 mM 4-aminopyridine with Axopatch-2B amplifier(Axon Instruments Inc., Foster City, CA, USA) usingthe whole-cell configuration of the patch clamp tech-nique [10]. The current was activated at +5 mV using400-ms long depolarizations arising from the holdingpotential of −40 mV. Simultaneous recording of INaand ICa was performed by applying voltage ramps from−80 to −40 mV before ICa measurement (as shown inFig. 3a).

After establishing high (1–10 GΩ) resistance seal bygentle suction, the cell membrane beneath the tip of theelectrode was disrupted by further suction or by apply-ing 1.5 V electrical pulses for 1 ms. The series resis-tance was typically 4–8 MΩ before compensation(usually 50–80 %). Experiments were discarded whenthe series resistance was high or substantially increasingduring the measurement. Outputs from the clamp ampli-fier were digitized at 100 kHz under software control(pClamp 6.0, Axon Instruments Inc., Foster City, CA,USA). When ion currents were normalized to cell ca-pacitance, it was determined in each cell using shorthyperpolar iz ing pulses f rom −10 to −20 mV.Concentration-dependent effects of TTX were deter-mined in a cumulative manner by applying increasingconcentrations of the drug between 10 and 300 μM.Each concentration was superfused for 3 min. Whenstudying the reversibility of TTX action, a washoutperiod lasting usually for 10 min was introduced. Theseincubation and washout periods were sufficient to de-velop steady-state drug effects and practically fullreversion.

Action potential voltage clamp

Using the whole-cell configuration of the patch clamp tech-nique, action potentials were recorded in current clampmode from the myocytes superfused with Tyrode’s

168 Pflugers Arch - Eur J Physiol (2012) 464:167–174

solution. The cells were continuously paced through therecording electrode at steady stimulation frequency of1 Hz so that a 1–2-ms gap between the stimulus artifactand the upstroke of the action potential could occur. Tensubsequent action potentials were recorded from eachcell, which were analyzed online. One of these actionpotentials, having an APD90 value closest to the averageof the ten recorded action potentials, was delivered tothe same cell at the identical frequency as commandvoltage after switching the amplifier to voltage clampmode. The current trace obtained under these conditionsis a horizontal line positioned at the zero level exceptfor the very short segment corresponding to the actionpotential upstroke [8]. The profile of the TTX-sensitiveion current was determined by subtracting the predrugcurve from the postdrug one, and the difference currentwas displayed with an inverted polarity. In our graphs,therefore, the TTX-sensitive current appeared as an in-wardly directed current.

Simulation of TTX binding to Cav1.2 and Nav 1.5 channelproteins

The selectivity filter region of the human Cav1.2 andNav1.5 channels was built using the theoretical model ofthe selectivity filter region of Nav1.4 channel in a complexwith the TTX molecule [22] and the existence homologybetween the Nav1.4, Nav1.5, and Cav1.2 channels [21,22, 25]. The alignment between Nav1.4 and Cav1.2was given in the paper published by Tikhonov andZhorov [22], which was used without modification.Alignment for the Nav1.4 and Nav1.5 sequences hasbeen already published [25]. Proper position for thehighly conserved DEKA ring [21] in Na+ channelscorresponding to D372, E898, K1419, and A1711 forNav1.5 channel as well as the D406, E761, K1244, andA1536 residues for Nav1.4 channel confirmed thisalignment. The residue mutations corresponding toNav1.4 to Nav1.5 and Nav1.4 to Cav1.2 transforma-tions were then carried out by using the YASARAsoftware package. The energy of TTX-Nav1.5 andTTX-Cav1.2 complex were minimized using the samesoftware. The graphical representations of the complexgeometries were prepared by the Chimera suites ofsoftware [20].

Statistics

Results are expressed as mean±SEM values. Statistical sig-nificance of differences was evaluated using Student’s t testfor paired data. Differences were considered significantwhen p was less than 0.05. Drugs were obtained fromSigma-Aldrich Co. (St. Louis, MO, USA).

Results

As presented in Fig. 1, ICa was progressively suppressedby increasing concentrations of TTX. When dataobtained in four cells were fitted to the Hill equation,an IC50 value of 55±2 μM was obtained. The Hillcoefficient of unity (1.0±0.04) is congruent with a sin-gle binding site of TTX on the Ca2+ channel. It isimportant to know that the current recorded under ourexperimental conditions was, in fact, L-type Ca2+ cur-rent. To demonstrate this, the effect of 30 μM TTX wastested after pretreatment with nisoldipine, the potentblocker of ICa. Nisoldipine (1 μM) blocked 93 % ofICa, and TTX failed to modify the current significantlyin the presence of nisoldipine (Fig. 2a, b). In anotherseries of experiments, the current remaining after partial(30 μM) TTX blockade was successfully eliminated by1 μM nisoldipine (Fig. 2c, d). Although TTX furtherdiminished the small current remaining after nisoldipinetreatment, this change was not significant statisticallyand may probably be accounted for some contaminationwith INa. Inactivation of ICa followed biexponential ki-netics. No significant changes in the decay time con-stants were observed; however, the amplitudes of boththe fast and slow components were significantly de-creased by 30 μM TTX (Fig. 2e).

In order to record INa and ICa simultaneously, a rampof 100-ms duration arising from −80 to −40 mV wasapplied before starting the conventional ICa protocol(Fig. 3a). Of course, peak INa could not be properlyresolved under our experimental conditions, but thetiming of INa and ICa blockade was easy to follow. Asshown in Fig. 3b, c, INa fell to zero, while ICa was only

TTX ( M)

Blo

ck o

f cal

cium

cur

rent

(%

)

0

25

50

75

100

0 10 30 100 300

n = 4

IC50 = 55 ± 2 µM

Hill = 1.0 ± 0.04

ba

Time (ms)

Cal

cium

cur

rent

(nA

)

0 100 200-1.5

-1.0

-0.5

0

0.5

Control

TTX 10 µM

TTX 30 µM

TTX 100 µM

TTX 300 µM

μ

Fig. 1 Cumulative concentration-dependent effects of TTX on ICa. aSuperimposed ICa traces taken at +5 mV in Tyrode’s solution and in thepresence of increasing concentrations (from 10 to 300 μM) of TTX.Dashed line indicates zero current level. b Average results obtainedwith TTX on peak ICa in four myocytes. Solid line indicates the resultof the Hill plot. Symbols and bars are mean±SEM values, respectively

Pflugers Arch - Eur J Physiol (2012) 464:167–174 169

partially blocked by 30 μM TTX. The timing of theseevents was fully synchronized—similarly to recovery ofINa and ICa upon washout of TTX (Fig. 3c). Impor-tantly, the TTX-induced inhibition of ICa was almostfully reversible (Fig. 3d).

The profile of an ion current may be markedly dif-ferent when comparing under conventional voltageclamp and action potential clamp conditions [9, 11].An advantage of the action potential clamp techniqueis that it enables us to record true current profilesflowing during an actual cardiac action potential. Asdemonstrated in Fig. 4a–d, 30 μM TTX dissected aninward current, having a fingerprint similar to that

described for ICa under action potential voltage clampconditions [4, 18]. However, TTX failed to dissect cur-rent in the presence of 1 μM nisoldipine (Fig. 4e–h). Asalready seen in the conventional voltage clamp experi-ments, this effect of TTX was fully reversible uponwashout. These results strongly support the view thatthe TTX-sensitive current found in canine ventricularcells is really ICa.

In order to understand in detail how TTX blocksvoltage-gated Cav1.2 and Nav1.5 channels, structuralinformation on their selectivity filter would be desirable.Based on the theoretical model elaborated originally forthe Nav1.4 Na+ channel, a model describing the binding

Cal

cium

cur

rent

(nA

)

-1.5

-1.2

-0.9

-0.6

-0.3

0

Time (ms)0 50 100

Control

TTX 30 µM

TTX 30 µM + Nisoldipine 1 µM

25 75

a b

Cal

cium

cur

rent

(pA

/pF

)

*

Con

trol

TT

X 3

0 µM

TT

X +

Nis

oldi

pine

1 µ

M

*

-10

-8

-6

-4

-2

0

n = 8

e

0

10

20

30

50

40C

ontr

ol

TT

X

0

2

4

6

8

Con

trol

TT

X

0

2

4

6

8

*

Con

trol

TT

X

fast

(ms)

slow

(ms)

10

Am

plitu

defa

st(p

A/p

F)

0

1

2

3

Con

trol

TT

X

Am

plitu

desl

ow(p

A/p

F)

*

c d

Cal

cium

cur

rent

(pA

/pF

)

-10

-8

-6

-4

-2

0

Con

trol

Nis

oldi

pine

1 µ

M

Nis

oldi

pine

+ T

TX

30

µM

* *

Cal

cium

cur

rent

(nA

)

-1.5

-1.2

-0.9

-0.6

-0.3

0

Time (ms)0 50 10025 75

Control

Nisoldipine 1 µM

Nisoldipine + TTX 30 µM

τ

τ

Fig. 2 Blocking effect of 1 μMnisoldipine and 30 μM TTX onICa. Superimposed ICa records(a, c) and average data (b, d). Ina and b, TTX was applied afternisoldipine pretreatment, whilein c and d, TTX was firstapplied, and the remainingcurrent was eliminated bynisoldipine. e Effects of 30 μMTTX on inactivation kinetics ofICa. Decay of ICa was fitted as asum of a two exponentialcomponents. Columns and barsrepresent mean±SEM, asterisksindicate significant (p<0.05)changes from control values

170 Pflugers Arch - Eur J Physiol (2012) 464:167–174

of TTX to cardiac Nav1.5 and Cav1.2 channels hasbeen proposed. The geometry of the selectivity filterregion of the TTX-Nav1.5 and TTX-Cav1.2 complexesare shown in Fig. 5. In our model, similar to Nav1.4channels, the guanidinium group binds to D and E inrepeats I and II, respectively. A hydrophobic site ofTTX interacts with the side chain of W. Despite theeven more acidic highly conserved EEEE ring of theselectivity filter region of Cav1.2, only E can form asalt bridge with TTX when it is bound to Cav1.2.According to our model, it is also evident that thenumber of residues making direct contact with theTTX molecule is less in Cav1.2 than in Nav1.5. Theseresults predict a weaker interaction between the Cav1.2and TTX than the Nav1.5 and TTX complexes whichultimately results in a diminished blocking potency ofTTX against Cav1.2 channels.

Discussion

Although the marine toxins TTX and STX are generallybelieved to be selective inhibitors of fast Na+ current, datain literature are accumulating to indicate that this is notexactly the case. For instance, both toxins have been dem-onstrated to interact with native and cloned T-type Ca2+

channels [25], while STX was shown to inhibit L-typeCa2+ channels [24]. This is the first report to demonstratethe inhibitory action of TTX on cardiac ICa. Previously, theeffects of TTX and STX on ICa were studied in murinecardiac tissues [24]. In that study, STX, but not TTX,effectively blocked ICa. Since no striking differences be-tween the experimental designs of the two studies (i.e.,Ref. [24] and the present work) can be detected, this diver-sity can probably be accounted for interspecies differencesin murine and canine Ca2+ channels.

a

b

-2.0

-1.5

-1.0

-0.5

0

Control

TTX 30 µM

WashoutC

urre

nt (

nA)

TTX 30 µM

ControlWashout

0030020 100Time (ms)

c

Con

trol

TT

X 3

0 µM

Was

hout

Cal

cium

cur

rent

(pA

/pF

)

0

-8

-4

-10

-6

-2

*

n = 4

-80 mV

-40 mV

+5 mV

INa

ICa

Cur

rent

(nA

)

Time (min)0 5 10 15

-1.5

-1.2

-0.9

-0.6

-0.3

0

TTX

INa

ICa

d

Fig. 3 Simultaneous recordingof INa and ICa. a Avoltage rampfrom the holding potential of−80 to −40 mV preceded theconventional ICa protocol. Thisprotocol allowed simultaneousmonitoring of timing of TTXeffects on INa and ICa (b and c).Of course, peak INa was out ofscale when not suppressed. dAverage results demonstratingthe reversible inhibitory effectof TTX on ICa in four myocytes

Pflugers Arch - Eur J Physiol (2012) 464:167–174 171

Cardiac INa is relatively insensitive to TTX [3, 5]. Indeed,a concentration of 30 μM was required to fully suppress INain rat ventricular cells [5]. Since this high concentration ofTTX suppressed 35 % of ICa in our canine ventricularcardiomyocytes, serious errors can thus be introduced. Theproblem is especially accentuated when trying to dissectwindow Na+ current using TTX under action potential volt-age clamp conditions. The amplitude of this current is in therange of some tens of picoamperes in healthy myocyteswhich was shown to increase up to 200 pA in hypertrophiedhearts [6]. It is easy to imagine the magnitude of errorintroduced by blocking only a small fraction of ICa byTTX. Similarly, the reported hypertrophy-induced changesin the window Na+ current might seriously be contaminatedby the TTX-induced suppression of ICa.

It is not exceptional that a drug believed to interact withINa selectively binds to and thus modifies other transmem-brane ion channels such as L-type Ca2+ channels. The

“selective” Na+ channel activator alkaloid, veratridine, wasshown to block Ca2+ channels in frog skeletal muscle [19].These cross-reactions can be well explained with the knownsimilarities in the structures of Na+ and Ca2+ channels [15].The pore-forming subunit of both channel types containsfour repeats as a single linear sequence, each containing sixtransmembrane domains. The fourth transmembrane seg-ment, rich in positively charged amino acid residues, actsas a voltage sensor, while the loop between segments fiveand six forms the permeation pathway starting with theselectivity filter. A detailed analysis of the amino acid se-quence of Na+ and Ca2+ channels revealed further similari-ties, believed to be due to the common evolutionary originof these channel types [13].

In summary, ICa was shown to be TTX-sensitive in ca-nine ventricular myocytes. Similar results were observedusing STX in mice [24]. Since these toxins still appear tobe relatively selective to INa over ICa, they are widely

Cur

rent

(nA

)

-80

-40

0

40

Mem

bran

e po

tent

ial (

mV

)

0 50 100 150 200 250

Time (ms)

-1.0

-0.5

0

0.5

Cur

rent

(nA

)

-1.0

-0.5

0

0.5

-1.0

-0.5

0

0.5

Cur

rent

(nA

)

Control

TTX 30 µM

Washout

0 50 100 150 200 250

Time (ms)

Washout

TTX 30 µM

Control

-80

-40

0

40

Mem

bran

e po

tent

ial (

mV

)C

urre

nt (

nA)

-1.0

-0.5

0

0.5

Cur

rent

(nA

)

-1.0

-0.5

0

0.5

-1.0

-0.5

0

0.5

Cur

rent

(nA

)

No drug Nisoldipine 1 µM

b

a

d

c

f

e

h

g

Fig. 4 Effect of 30 μM TTXon net membrane current underaction potential voltage clampconditions in the absence (a–d)and presence (e–h) of 1 μMnisoldipine. Representativerecords of a command signal (a,e) and the underlying currenttraces obtained before (b, f), inthe presence of (c, g), and afterwashout of TTX (d, h). Sinceion currents were obtained fromthe same cell that provided thecommand action potential, thepredrug (control) trace was ahorizontal line at the zero level.The postdrug difference currentis clearly L-type Ca2+ current.For each trace, the dashed lineindicates zero current level ex-cept for the action potential re-cord where it represents zerovoltage. Fast INa, since it wasalso dissected by TTX, largelyoverlapped the capacitive tran-sient, which has been retouchedfrom the figures

172 Pflugers Arch - Eur J Physiol (2012) 464:167–174

applied in biomedical research. However, depending on theexperimental conditions, extra care must be taken whenusing these agents to block INa selectively.

Acknowledgments Financial support for the studies was providedby grants from the Hungarian Research Fund (OTKA-K100151,OTKA-PD101171, OTKA-K101196, and CNK-77855). The authorsthank Mrs. Vighné Katalin Horváth for excellent technical assistance.

Conflict of interest The authors declare that that they have noconflict of interest.

Author Note The experiments comply with the current laws ofHungary.

References

1. Aggarwal R, Shorofsky SR, Goldman L, Balke CW (1997)Tetrodotoxin-blockable calcium currents in rat ventricular

myocytes; a third type of cardiac cell sodium current. J Physiol(Lond) 505:353–369

2. Alvarez JL, Salinas-Stefanon E, Orta G, Ferrer T, Talavera K,Galán L, Vassort G (2004) Occurrence of a tetrodotoxin-sensitivecalcium current in rat ventricular myocytes after long-term myo-cardial infarction. Cardiovasc Res 63:653–661

3. Baer M, Best PM, Reuter H (1976) Voltage-dependent actionof tetrodotoxin in mammalian cardiac muscle. Nature263:344–345

4. Bányász T, Fülöp L, Magyar J, Szentandrássy N, Varró A, NánásiPP (2003) Endocardial versus epicardial differences in L-typecalcium current in canine ventricular myocytes studied by actionpotential voltage clamp. Cardiovasc Res 58:66–75

5. Brown AM, Lee KS, Powell T (1981) Sodium current in single ratheart muscle cells. J Physiol (Lond) 318:479–500

6. Chorvatova A, Snowdon R, Hart G, Hussain M (2004) Effects ofpressure overload-induced hypertrophy on TTX-sensitive inwardcurrents in guinea pig left ventricle. Mol Cell Biochem 261:217–226

7. Cole WC, Chartier D, Martin M, Leblanc N (1997) Ca2+ perme-ation through Na+ channels in guinea pig ventricular myocytes.Am J Physiol 273:H128–H137

8. Doerr T, Denger R, Trautwein W (1989) Calcium currents in singleSA nodal cells of the rabbit heart studied with action potentialclamp. Pflugers Arch 413:599–603

9. Doerr T, Denger R, Doerr A, Trautwein W (1990) Ionic currentscontributing to the action potential in single ventricular myocytesof the guinea pig studied with action potential clamp. PflugersArch 416:230–237

10. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ (1981)Improved patch-clamp techniques for high-resolution current re-cording from cells and cell-free membrane patches. Pflugers Arch391:85–100

11. Hancox JC, Levi AJ, Witchel HJ (1998) Time course and voltagedependence of expressed HERG current compared with the native“rapid” delayed rectifier K current during the cardiac ventricularaction potential. Pflugers Arch 436:843–853

12. Haverinen J, Hassinen M, Vornanen M (2007) Fish cardiac sodiumchannels are tetrodotoxin sensitive. Acta Physiol (Oxf) 191:197–204

13. Heinemann SH, Terlau H, Stühmer W, Imoto K, Numa S (1992)Calcium channel characteristics conferred on the sodium channelby single mutations. Nature 356:441–443

14. Heubach JF, Köhler A, Wettwer E, Ravens U (2000) T-type andtetrodotoxin-sensitive Ca2+ currents coexist in guinea pig ventric-ular myocytes and are both blocked by mibefradil. Circ Res86:628–635

15. Hille B (2001) Ion channels in excitable membranes. SinauerAssociates Inc, Sunderland

16. Horváth B, Magyar J, Szentandrássy N, Birinyi P, Nánási PP,Bányász T (2006) Contribution of IKs to ventricular repolarizationin canine myocytes. Pflugers Arch 452:698–706

17. Lemaire S, Piot C, Seguin J, Nargeot J, Richard S (1995)Tetrodotoxin-sensitive Ca2+ and Ba2+ currents in human atrialcells. Receptors Channels 3:71–81

18. Linz KW, Meyer R (2000) Profile and kinetics of L-typecalcium current during the cardiac ventricular action potentialcompared in guinea-pigs, rats and rabbits. Pflugers Arch439:588–599

19. Nánási PP, Varró A, Lathrop DA, Bryant SH (1994) Effects ofveratridine on Na and Ca currents in frog skeletal muscle. GenPharmacol 25:1661–1666

20. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM,Meng EC, Ferrin TE (2004) UCSF Chimera—a visualization sys-tem for exploratory research and analysis. J Comput Chem25:1605–1612

D372

E375

Y378

W904

D1714

W1421

TTX

Repeat I

Repeat II

Repeat IV

Repeat III

E736

N739

L399

M742

W1146T396

TTX

Repeat I

Repeat III

Repeat IV

Repeat II

a

b

E898

Fig. 5 Complex structure formed by TTX and the selectivity filter ofNav1.5 (a) and Cav1.2 (b) voltage-gated ion channels. Only the P-loopdomains (cartoon representation), the residues of P-loop domainsinvolved in TTX binding (stick representation), and the TTX molecule(ball-and-stick representation) are shown. Residue numbering,corresponding to the amino acid sequence number, starts with theinitiator methionin. For further details, see the relevant parts of “Meth-ods” and “Results”

Pflugers Arch - Eur J Physiol (2012) 464:167–174 173

21. Tikhonov DB, Zhorov BS (2011) Possible roles of exceptionallyconserved residues around the selectivity filters of sodium andcalcium channels. J Biol Chem 286:2998–3006

22. Tikhonov DB, Zhorov BS (2005) Modeling P-loops domain ofsodium channel: homology with potassium channels and interac-tion with ligands. Biophys J 88:184–197

23. Santana LF, Gómez AM, Lederer WJ (1998) Ca2+ flux throughpromiscuous cardiac Na+ channels: slip-mode conductance. Sci-ence 279:1027–1033

24. Su Z, Sheets M, Ishida H, Li F, Barry WH (2004) Saxitoxin blocksL-type ICa. J Pharmacol Exp Ther 308:324–329

25. Sun H, Varela D, Chartier D, Ruben PC, Nattel S, Zamponi GW,Leblanc N (2008) Differential interactions of Na+ channel toxinswith T-type Ca2+ channels. J Gen Physiol 132:101–113

26. Szabó G, Szentandrássy N, Bíró T, Tóth IB, Czifra G, Magyar J,Bányász T, Varró A, Kovács L, Nánási PP (2005) Asymmetricaldistribution of ion channels in canine and human left ventricular wall:epicardium versus midmyocardium. Pflugers Arch 450:307–316

27. Szentandrássy N, Bányász T, Bíró T, Szabó G, Tóth IB, Magyar J,Lázár J, Varró A, Kovács L, Nánási PP (2005) Apico-basal inho-mogeneity in distribution of ion channels in canine and humanventricular myocardium. Cardiovasc Res 65:851–860

174 Pflugers Arch - Eur J Physiol (2012) 464:167–174