European Journal of Pharmacology 672 (2011) 135–142

Contents lists available at SciVerse ScienceDirect

European Journal of Pharmacology

j ourna l homepage: www.e lsev ie r .com/ locate /e jphar

Cardiovascular Pharmacology

R(+)-pulegone impairs Ca2+ homeostasis and causes negative inotropism inmammalian myocardium

Sandra Valeria Santos de Cerqueira a, Antonio Nei Santana Gondim b,c, Danilo Roman-Campos b,Jader Santos Cruz b, Amilton Gustavo da Silva Passos a, Sandra Lauton-Santos a, Aline Lara d,Silvia Guatimosim d, Eduardo Antonio Conde-Garcia a,Evaleide Diniz de Oliveira a, Carla Maria Lins de Vasconcelos a,⁎a Laboratório de Biofísica do Coração, Departamento de Fisiologia, Universidade Federal de Sergipe, Aracaju, SE, Brazilb Laboratório de Membranas Excitáveis, Departamento de Imunologia e Bioquímica, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazilc Laboratório de Biofísica e Farmacologia do Coração, Departamento de Educação-Campus XII, Universidade do Estado da Bahia, Guanambi, BA, Brazild Laboratório de Sinalização Intracelular em Cardiomiócito, Departamento de Fisiologia e Biofísica, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil

⁎ Corresponding author at: Cidade Universitária ProfMarechal Rondon, s/n, Jardim Rosa Elze, CEP 49100-000Tel.: +55 79 21056642.

E-mail addresses: carlamlv@ufs.br, carlamlv@hotma

0014-2999/$ – see front matter © 2011 Elsevier B.V. Alldoi:10.1016/j.ejphar.2011.09.186

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 March 2011Received in revised form 19 September 2011Accepted 24 September 2011Available online 10 October 2011

Keywords:R(+)-pulegoneContractilityL-type calcium currentCalcium transientCardiac action potentialPotassium current

The present study aimed to investigate the inotropic effects of R(+)-pulegone, a monoterpene found in plantspecies belonging to the genus Mentha, on the mammalian heart. In electrically stimulated guinea pig atria,R(+)-pulegone reduced the contractile force (~83%) and decreased the contraction time measured at 50% ofthe maximum force amplitude (CT50) from 45.8±6.2 ms to 36.9±6.2 ms, suggesting that R(+)-pulegonemay have an effect on Ca2+ homeostasis. Nifedipine (40 μM), taken as a positive control, showed a very similarprofile. To explore the hypothesis that R(+)-pulegone is somehow affecting Ca2+ handling, we determinedconcentration–response curves for both CaCl2 and BAY K8644. R(+)-pulegone shifted these curves rightward.Using isolated mouse ventricular cardiomyocytes, we measured whole-cell L-type Ca2+ current and observedan ICa,L peak reduction of 13.7±2.5% and 40.2±2.9% after a 3-min perfusion with 0.11 and 1.1 mM of R(+)-pulegone, respectively. In addition, the intracellular Ca2+ transient was decreased (72.9%) by 3.2 mM R(+)-pulegone, with no significant changes in [Ca2+]i transient decay kinetics. Moreover, R(+)-pulegone at1.1 mM prolonged the action potential duration at 10, 50, and 90% of repolarisation. The lengthening ofthe action potential duration may be attributed to the substantial blockade of the outward K+ currents causedby 1.1 mM of R(+)-pulegone (90.5% at 60 mV). These findings suggest that R(+)-pulegone exerts its negativeinotropic effect on mammalian heart mainly by decreasing the L-type Ca2+ current and the global intracellularCa2+ transient.

. “José Aloísio de Campos”, Av., São Cristóvão, Sergipe, Brazil.

il.com (C.M.L. de Vasconcelos).

rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

R(+)-pulegone ((R)-2-isopropylidine-5-methyl-cyclohexanone) isa ketone monoterpene found in several plant essential oils, includingthose from the Mentha genus, such as Mentha pulegium (also knownas pennyroyal). Pennyroyal oil, which contains between 62% and 95%R(+)-pulegone (Gordon et al., 1982; Grundschober, 1979), is tradition-ally used not only therapeutically but also for its toxic effects as aborti-facient (Bowen and Cubbin, 1992). A number of studies have associatedpennyroyal toxicity with the presence of R(+)-pulegone (Andersonet al., 1996; Bakerink et al., 1996; Gordon et al., 1982; Molck et al.,1998; Moorthy et al., 1989; Sullivan et al., 1979). Surprisingly, despite

its reported toxic effects, pennyroyal oil is largely used as flavouringagent in chewing gums, toothpaste, and candies (Petrakis et al., 2009).

The complete pharmacological characterisation of essential oilsand their constituents is a very active and important area of currentresearch. In this context, several plants that produce R(+)-pulegoneas a secondary metabolite were the focus of recent pharmacologicalinvestigation. Previously, the essential oil of Mentha piperita (pepper-mint) was reported to have antispasmodic effects on rat tracheal(Souza et al., 2010) and on the guinea pig gastrointestinal smoothmuscle (Hills and Aaronson, 1991). Spasmolytic action on rat ileumwas reported with Calamintha glandulosa essential oil (Brankovicet al., 2009). Similarly, relaxant effects were also reported on uterinesmooth muscle exposed to the essential oil of M. pulegium (Soareset al., 2005), on rat vascular smooth muscle (Guedes et al., 2004),and guinea-pig intestinal smooth muscle also exposed to the essentialoil of Mentha x villosa (Sousa et al., 1997). According to these reports,R(+)-pulegone likely inhibits spontaneous and potassium-inducedmuscle contractions, and this effect has been reported to be related

136 S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

to a significant reduction in calcium influx (Brankovic et al., 2009).However, it is important to mention that, so far, there is no directevidence to support this mechanism of action.

Cardiac contractility is dependent on intracellular Ca2+homeostasis.Such homeostasis contributes to control the action potential waveform.There is evidence that R(+)-pulegone may act through diminution ofCa2+ influx, which could evoke changes in cardiac contractility. Thisevidence led us to ask the following questions: 1) Does R(+)-pulegonedirectly inhibit L-type Ca2+ channels? 2) Does R(+)-pulegone changethe cardiac intracellular global Ca2+ transient? 3) Does R(+)-pulegonemodify cardiac action potential properties?

2. Materials and methods

2.1. Drugs

The following drugs were purchased from Sigma-Aldrich(St. Louis, Missouri, USA): R(+)-pulegone; nifedipine; reserpine;S-(−)-1,4-dihydro-2,6-dimethyl-5-nitro-4-[2-(trifluoromethyl)phenyl]-3-pyridinecarboxylic acid methyl ester (S-(−)-BAY K8644); proteasetype XXIII (Catalogue # P4032); porcine pancreas trypsin (Catalogue# T0303); insulin; bovine serum albumin (Catalogue # A9318);HEPES; EGTA; tetraethylammonium; CsCl; CsOH; Dulbecco's modifiedEagle's medium (DMEM); and penicillin/streptomycin (Catalogue #P4333). Fluo4-AM was purchased from Molecular Probes, Inc. (Eugene,Oregon, USA). Collagenase type II was purchased from WorthingtonBiochemical Co. (Freehold, NJ, USA). Dimethyl sulfoxide (DMSO),NaCl, KCl, MgCl2, NaHCO3, CaCl2, NaH2PO4, NaOH, glucose, and heparinwere acquired from Vetec (Rio de Janeiro, Brazil) or Merck (Darmstadt,Germany).

2.2. Animals

The experiments designed to evaluate changes in contractilityelicited by R(+)-pulegone were performed using guinea pigs of bothsexes (Cavia porcellus, 400 to 700 g). Experiments on the effects ofR(+)-pulegone on the L-typeCa2+ current, intracellular Ca2+ transient,action potential and potassium currents were performed on isolatedventricular cardiac myocytes from C57Bl/6J mice (both sexes). This in-vestigation was approved by the Animal Research Ethics Committee(Protocol No. 37/10) of the Federal University of Sergipe, Brazil.

2.3. Inotropic effect of R(+)-pulegone in guinea pig left atria

The animals were sacrificed by decapitation. Each experiment wascarried out on isolated guinea pig left atrium mounted in an organchamber. There, the atrium was bathed in Tyrode's solution modifiedaccording to Dorigo et al. (1990) (in mM: NaCl 120, KCl 2.7, MgCl2 0.9,NaHCO3 11.9, CaCl2 1.37, glucose 5.5, NaH2PO4 0.4, pH 7.2), main-tained at 29±0.1 °C, and oxygenated with carbogen (95% O2 and 5%CO2). The atria were stretched to 5 mN and submitted to field stimu-lation (1 Hz). The stimulation pulses (400 V, 0.5 ms) were providedby a Digitimer 3072 Stimulator controlled by a Digitimer D4030 Pro-grammer (Welwyn Garden City, Hertfordshire, AL7 3BE, England).Atrial contraction force was recorded with an isometric force trans-ducer (HP FTA 10-1 Sunborn, HP 8805B, Chicago, Illinois, USA). Thesignals were stored in a computer to be processed and analysed off-line (DATAQ DI400, DI 205, WINDAQ PRO Acquisition, Ohio, Akron,USA).

We then determined the contractile force and times required at50% of the maximal force amplitude for contraction (CT50) and relax-ation (RT50). These data were measured under control conditions andafter addition of R(+)-pulegone or nifedipine to the organ bath. Tomeasure these parameters, 50 consecutive contractions were selectedand processed by the software CONEXON, which was developed in

the Cardiobiophysics Research Laboratory by Dr. E.A. Conde-Garcia(Patent deposit #00051104).

2.4. Effects of R(+)-pulegone on Ca2+ influx in guinea pig left atria

The experiments were carried out on isolated atria obtained fromreserpinised guinea pigs (5 mg/kg, i.p., 24 h before the experiment).Concentration–response curves for CaCl2 and BAY K8644 wereobtained before and after incubating the preparation with 3.2 mMR(+)-pulegone for 15 min. The results are expressed as percentagesof the maximal atrial contractile response to CaCl2 or BAY K8644.

2.5. Effects of R(+)-pulegone on the L-type Ca2+ current, action potentialand outward potassium current

Ventricular cardiomyocytes from mice (C57Bl/6J) were enzymati-cally isolated as previously described (Shioya, 2007), with few modi-fications. Whole-cell voltage-clamp recordings were obtained atroom temperature (22–25 °C) using an EPC-9.2 patch-clamp amplifi-er (HEKA Electronics, Rheinland-Pfalz, Germany) as routinely used inour laboratory (Lara et al., 2010; Oliveira et al., 2007). After thewhole-cell configuration was complete, the pipette solution wasallowed to equilibrate with the intracellular environment for 3 to5 min. The recording electrodes had tip resistances of 0.5 to 1.5 MΩ.Current recordings were low-pass filtered (2.9 kHz) and digitised at10 kHz before being stored on a computer. Myocytes showing a seriesresistance (Rs) larger than 6 MΩ were not used in the analysis. Rscompensation was set at 40 to 50%. For measurements of L-typeCa2+ current (ICa,L), recording pipettes were filled with an internalsolution containing (in mM) 120 CsCl, 20 TEACl, 5 NaCl, 10 HEPES,and 5 EGTA, and the pH was set to 7.2 with CsOH. ICa,L was recordedin the presence of 1.8 mM extracellular Ca2+. To measure ICa,Lsteady-state inactivation, cardiac cells were perfused with an externalsolution containing (inmM) 135 TEACl, 10 HEPES, 20 glucose, 5.4 CsCl,1 MgCl2, and 1.8 CaCl2, and the pH was set to 7.4 with CsOH.

For the time course analysis of the effects of 0.11 mM and 1.1 mMR(+)-pulegone on the L-type Ca2+ channels, ICa,L was elicited every10 s by test pulses from a holding potential of −80 mV to −40 mVfor 50 ms to inactivate Na+ and T-type Ca2+ channels. Thenwe steppedthe membrane potential to 0 mV for 300 ms. The same protocol wasperformed before, during, and after washing off the R(+)-pulegone.For the current vs. voltage relationship analysis, ICa,L was elicited by300 ms test potentials ranging from −50 to 50 mV in 10 mV incre-ments. The steady-state activation curves were obtained with theBoltzmann equation in the form: G/Gmax=1/{1+exp[(V−V0.5)/k]},where V0.5 is the voltage at which 50% of the maximum conductancewas attained and k is the slope factor. The voltage dependence of ICa,Lsteady-state inactivation was investigated using a double-pulse proto-col with a 1 s conditioning voltage step to potentials between −110and 10 mV in 10mV increments. This was followed by a 100 mstest pulse to 0 mV to evaluate the steady-state inactivation of L-typeCa2+ channels. For steady-state inactivation, the peak current value atthe corresponding conditioning membrane potential was normalisedto the Imax measured at −110 mV conditioning pulse and was fittedusing the Boltzmann equation: I/Imax=1/{1+exp[(V−V0.5)/k]},where V0.5 is the voltage at which 50% of the maximum conductancewas attained and k is the slope factor.

To measure action potential (AP) parameters and K+ currents,the pipette solution was filled with (in mM) 130 K-aspartate, 20KCl, 10 HEPES, 2 MgCl2, 5 NaCl, and 5 EGTA, and the pH was setto 7.2 with KOH. We used Tyrode's solution as the bath solution,which contained (in mM) 140 NaCl, 5.4 KCl, 1 MgCl2, 1.8 CaCl2, 10HEPES, and 10 glucose (pH set at 7.4). For the time course analysisof 1.1 mM R(+)-pulegone effects, we recorded 20–30 APs beforeapplying the drug. We evaluated the overshoot amplitude, maximalrate of depolarisation, duration at 10, 50, and 90% of AP repolarisation,

137S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

and the resting membrane potential. The APs were elicited by injectinga 1 nA depolarising current (3 to 5 ms duration). To investigate theeffects of 1.1 mM R(+)-pulegone on the voltage dependence of K+

currents, every 10 s we applied voltage ramps (120 mV/s), in which thecell membrane was maintained at −80 mV, depolarised to +60 mVand then repolarised to −120 mV over 1.5 s before, during, and afterwashing off the R(+)-pulegone. In the experiments assembled to recordAPs and K+ currents, there was a junction potential of approximately−20 mV that was not corrected.

2.6. Effects of R(+)-pulegone on intracellular Ca2+ global transient

Isolated myocytes obtained from mice (4 animals) were loadedwith 5 μM Fluo4-AM (stock solution prepared in pure DMSO) for30 min at room temperature and then washed with an extracellularsolution containing 1.8 mM Ca2+ to remove the excess dye. Cardiaccells were exposed to different concentrations of R(+)-pulegone(0.11, 1.1, and 3.2 mM) for 3 min. After this time, Ca2+ transientswere elicited by electrical field stimulation applied to the cardiomyo-cytes through a pair of platinum electrodes. Supra-threshold squarevoltage pulses of 0.2 ms duration were applied at 1 Hz to producesteady-state conditions. A Meta LSM 510 scanning system (ZeissGmbH, Jena, Germany) with a 63× oil immersion objective was usedfor confocal fluorescence imaging. Fluo4-AM was excited at 488 nm(Argon laser), and the emission intensity was measured at 510 nm.For recording Ca2+ transients, myocytes were scanned with a 512-pixel line. The scan linewas positioned randomly along the longitudinalaxis of the cell, although care was taken to avoid crossing the nucleus.Cells were scanned every 1.54 ms and sequential scans were stackedto create two-dimensional images with time on the x-axis. Digitalimage processing was performed with a custom routine developedwith the IDL programming language (Research Systems, Boulder, CO).The intracellular Ca2+ levels were reported as F/F0, where F0 is theresting Ca2+ fluorescence.

Fig. 1. The effects of R(+)-pulegone on guinea pig left atrium contractility. (A) Experimental re(bottom panel). (B) Superimposed curves (top panel) illustrating the time course of the isometand an artefact of stimulation (d). The effect of nifedipine on the isometric contraction is repor(c), and an artefact of stimulation (d). (C) Concentration–response curves of R(+)-pulegoneR(+)-pulegone (3.2 mM) and nifedipine (40 μM) on contraction time at 50% of the maxim(light grey), and washout (dark grey).

2.7. Statistical analysis

All results are shown as the means±S.E.M. GraphPad Prism(GraphPad Software, CA, USA) and SigmaPlot (Systat Software, IL, USA)were used for statistical analysis. Data were compared by Student'st-tests. Pb0.05 was taken as the significance level.

3. Results

3.1. Negative inotropic effect of R(+)-pulegone on guinea pig left atria

The inotropic effect of R(+)-pulegone was studied in isolated atriaelectrically paced at 1 Hz to avoid frequency-dependent changes inthe force of contraction. R(+)-pulegone impaired the developmentof atrial contractile force in a concentration-dependent manner.Fig. 1A shows the effects of R(+)-pulegone (top panel) and nifedi-pine (bottom panel), a well-known L-type Ca2+ channel blockerused as a positive control for atrial contractility. R(+)-pulegone(3.2 mM) significantly reduced atrial contractile force by approxi-mately 83% (n=4). This effect was only partially removed by wash-ing out R(+)-pulegone. It is interesting to note that a very similarpattern was also observed when nifedipine was used.

Fig. 1B (top panel) shows superimposed isometric recordings,illustrating the time course and decrease in contractility elicited byR(+)-pulegone at 1.1 mM (trace b) and 3.2 mM (trace c). Fig. 1B(bottom panel) also depicts the effect of nifedipine at 1 μM (trace b)and 40 μM (trace c). Recordings obtained during control conditionsare also shown (trace a).

EC50 values of 777±420 μM and 1.1±1.0 μM (n=4, Pb0.05) werecalculated from the cumulative concentration–response curvesobtained for R(+)-pulegone (Fig. 1C, open squares) and for nifedipine(Fig. 1C, closed squares), respectively. The negative inotropic effectinduced by R(+)-pulegone (3.2 mM) was accompanied by a slightshortening of the CT50 from 45.8±6.2 ms to 36.9±6.2 ms (Fig. 1D,n=4, Pb0.05). Nifedipine (40 μM) presented a similar behaviour

cord of negative inotropic effects produced by R(+)-pulegone (top panel) and nifedipineric contraction in the control (a), 1.1 mMR(+)-pulegone (b), 3.2 mMR(+)-pulegone (c),ted for comparison in the bottom panel: control (a), 1 μMnifedipine (b), 40 μMnifedipine(EC50=777±420 μM, n=4) and nifedipine (EC50=1.1±1.0 μM, n=4). (D) Effects ofum force amplitude (CT50) (n=4). Bars: control (white), R(+)-pulegone or nifedipine

138 S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

when compared to R(+)-pulegone, and theCT50 decreased from45.4±4.9 ms to 39.4±1.1 ms (Fig. 1D, n=4, Pb0.05). The RT50 was alsodecreased by R(+)-pulegone from 51.2±12.6 ms to 41.0±3.3 ms(n=4, Pb0.05) and by nifedipine from 55.1±7.0 ms to 44.8±7.2 ms(n=4, Pb0.05). These effects were partially reversible during drugwashout. Thus, we can conclude that R(+)-pulegone has a negativeinotropic effect on atrial musculature similar to that caused by thedihydropyridine nifedipine.

3.2. R(+)-pulegone reduces L-type Ca2+ current

The amount of Ca2+ entering to the cell through L-type Ca2+

channels during every cardiac cycle is very important because itregulates the contractile force magnitude in both atrial and ventricularmyocytes (Bers, 2002). To evaluatewhether the negative inotropic effectelicited by R(+)-pulegone was related to a decrease in Ca2+ influx, weperformed a series of experiments to determine if it could affect the ino-tropic response to both CaCl2 and BAY K8644. Our data showed thatR(+)-pulegone (3.2 mM) caused a rightward shift of the concentration–response curve for Ca2+ (EC50 increased from 1.2±0.2 mM to 5.2±0.04 mM; Fig. 2A, inverted triangles, n=5, Pb0.05), indicating that theCa2+ influxwas significantly reduced in the presence of R(+)-pulegone.To further evaluate the involvement of L-type Ca2+ channels in thenegative inotropic effect evoked by R(+)-pulegone, we performed anew set of experiments using BAY K8644, a selective L-type Ca2+

channel agonist. As expected, BAY K8644 had a positive inotropiceffect, but pre-incubationwith R(+)-pulegone (3.2 mM)decreased thispositive inotropic effect. In these experiments, the EC50 increased from0.1±0.06 μM to 0.6±0.1 μM in the presence of R(+)-pulegone(Fig. 2B, closed triangles, n=5, Pb0.05), suggesting that L-type Ca2+

channels may be involved in the cardiodepressor effect elicited byR(+)-pulegone.

To provide direct evidence that R(+)-pulegone decreases L-type Ca2+

current (ICa,L), we designed electrophysiological experiments to measureICa,L from cardiac myocytes. We found that R(+)-pulegone shows aninhibitory effect on peak ICa,L after 5 min of continuous perfusion andthat this effect was partially reversible during washout. Fig. 3A showsrepresentative current traces measured during the control, duringdrug perfusion, and after washout. The time course of the effect ofR(+)-pulegone (1.1 mM) on ICa,L is shown in Fig. 3B. After 5 min ofcontinuous perfusion, R(+)-pulegone inhibited ICa,L by approximately39%. This effect was only partially removed after washout. Compositedata are plotted in Fig. 3C, which shows that R(+)-pulegone reducedthe peak ICa,L in a concentration-dependent manner, with a reductionof 13.7±2.5% (n=6, Pb0.001) and 40.2±2.9% (n=8, Pb0.001) after5-min perfusions of 0.11 and 1.1 mM R(+)-pulegone, respectively.To better understand the possible mechanism responsible for thedecrease in ICa,L elicited by R(+)-pulegone, we evaluated its effecton the L-type Ca2+ current–voltage relationship.

Fig. 4A shows typical recordings of ICa,L at voltage membrane stepsfrom−40 mV to 0 mV before (Fig. 4Ai) and after 5 min in the presenceof R(+)-pulegone at 1.1 mM (Fig. 4Aii). R(+)-pulegone (1.1 mM)

Fig. 2. The effects of R(+)-pulegone on the concentration–response curves for thepositive inotropic effects of CaCl2 (A) and BAY K8644 (B). At 3.2 mM, R(+)-pulegonesignificantly reduced the positive inotropic effects of both CaCl2 and BAY K8644.

significantly reduced the ICa,L at all membrane voltages ranging from−10 mV to +30 mV (n=3, Pb0.05), but it did not affect the shape ofcurrent–voltage relationship curve (Fig. 4B). Normalising the channelconductance at each membrane potential to its maximum conductancerevealed a V50 and a slope factor of −15.0±0.65 mV and 5.2±0.5for control myocytes (n=3) versus −13.6±0.7 mV and 5.5±0.6 formyocytes exposed to 1.1 mM R(+)-pulegone (Fig. 4C, n=3, PN0.05).Thus, R(+)-pulegone did not alter the voltage dependence of steady-state activation. However, 1.1 mM R(+)-pulegone changed steady-state inactivation. R(+)-pulegone shifted V0.5 from −32.7±0.6 mVto−48.5±1.2 mV. The slope factor (k) did not change in the presenceof R(+)-pulegone (5.1±0.1 vs. 3.7±0.1; Fig. 4D, n=4, PN0.05).

3.3. R(+)-pulegone decreases intracellular Ca2+ global transient

It is well established that intracellular Ca2+ handling is critical forcontrolling cardiac contractility (Bers, 2008). Therefore, we asked ifthe reduction in Ca2+ influx caused by R(+)-pulegone could lead toan alteration of the sarcoplasmic global Ca2+ transient. To address thisissue, we performed experiments using isolated cardiomyocytespre-loaded with the intracellular Ca2+ indicator Fluo4-AM. Fig. 5 showsrepresentative recordings of intracellular Ca2+ transients obtained byelectrically stimulating isolated cardiac cells. (SeeMaterials andmethodsfor more details.) Panels A, B, C, and D show representative pseudo-colour images recorded with a confocal microscope in line-scan mode(top) and Ca2+ transients (bottom) taken under control conditions andin the presence of 0.11, 1.1, and 3.2 mM R(+)-pulegone, respectively.As shown in Fig. 5, 0.11 mM (panel B) and 1.1 mM R(+)-pulegone(panel C) significantly reduced the global Ca2+ transient, and at3.2 mM R(+)-pulegone almost eliminated the intracellular Ca2+ tran-sient (panel D).

Fig. 5E shows fluorescence changes as the ratio F/F0, where Frepresents the maximum fluorescence intensity, which is normalisedby the basal fluorescence (F0). Our analysis showed that the globalCa2+ transient was reduced after pre-incubation with R(+)-pulegone.Compared with the control group, the Ca2+ transient amplitudedecreased from 4.8±0.1 (n=104) to 4.1±0.2 at 0.11 mM R(+)-pulegone (14.6%, Pb0.05, n=26), to 2.6±0.2 at 1.1 mM R(+)-pulegone (45.8%, Pb0.0001, n=32), and to 1.3±0.02 at 3.2 mMR(+)-pulegone (72.9%, Pb0.0001, n=13). Although the resultsrevealed that the R(+)-pulegone decreased the global Ca2+ transient,most of the cardiomyocytes exposed to R(+)-pulegone (1.1 and3.2 mM) for 5 min did not show calcium transients in response tofield stimulation.

The results also showed that R(+)-pulegone did not alter intracel-lular Ca2+ transient decay kinetics. To confirm that, the intracellularCa2+ transient decay time was measured at 90% of this decay (t90).There was no difference when comparing the control (852.6±15.2 ms) with 0.11 mM (868.7±16.2 ms) and with 1.1 mM ofR(+)-pulegone (860.5±23.2 ms).

3.4. R(+)-pulegone modulates action potential and potassium current

It is well known that AP properties are essential for maintainingheart function (Nerbonne and Kass, 2005). Therefore, we decided toevaluate if R(+)-pulegone is able to alter the AP properties. Fig. 6summarises the results, showing that R(+)-pulegone indeed alteredresting membrane potential and AP depolarisation and repolarisationphases. Fig. 6A shows representative traces of the effects of 1.1 mMR(+)-pulegone on cardiac APs. Fig. 6B shows the time course of theeffects of R(+)-pulegone on the cardiac action potential duration(APD) at 50% of repolarisation. R(+)-pulegone markedly prolongedthe APD50 from 5.6 ms to 34.7 ms (619%). It is important to mentionthat exposure to 1.1 mM R(+)-pulegone resulted in a failure to triggerAPs, as shown in red circles in Fig. 6B. Fig. 6C clearly shows that R(+)-pulegone prolonged the APD in a reversible manner at 10%, 50%, and

Fig. 3. The effect of R(+)-pulegone on the L-type calcium currents in ventricular cardiomyocytes. (A) Typical recordings elicited by 300 ms test pulses to 0 mV from a holding potentialof−80 mV in control conditions (top trace), during the perfusionwith 1.1 mMR(+)-pulegone (middle trace), and after washout (bottom trace). (B) R(+)-pulegone inhibition of ICa,L incardiac ventricular myocytes. Each symbol indicates the net amplitude of ICa,L measured every 10 s at 0 mVmembrane potential under control conditions (black circles), during exposureto R(+)-pulegone at 1.1 mM(grey circles), and afterwashout (open circles). (C) Summaryof the effects of R(+)-pulegoneon the ICa,L amplitude. The effects of R(+)-pulegone at 0.11 mM(light grey bar, n=6) and 1.1 mM (dark grey bar, n=8) are compared with the control (white bar, n=14). The bars indicate the means±S.E.M. *Pb0.05 versus control values.

139S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

90% of repolarisation. Fig. 6D depicts the time course of the effects ofR(+)-pulegone (1.1 mM) on the resting membrane potential (RMP).R(+)-pulegone reduced RMP from −81 mV to −99 mV. Fig. 6E sum-marises the changes in the RMP elicited by R(+)-pulegone. The averagevalues in the control, in the presence of 1.1 mMR(+)-pulegone, and afterwashout were −80.5±0.5 mV, −93.5±2.1 mV, and −80.9±1.4 mV,respectively (n=5, Pb0.05). Additionally, R(+)-pulegone did not reducethe AP amplitude (n=5, PN0.05); however, 1.1 mM R(+)-pulegonedecreased the maximal rate of depolarisation from 319.1±15.2 V/s to206.4±37.1 V/s (n=5, Pb0.05).

Fig. 4. The effects of R(+)-pulegone on L-type Ca2+ current vs. voltage and steady-state acpotential of −80 mV, the cells were depolarised to −40 mV for 50 ms to inactivate Na+ an−40 and 0 mV in 10 mV steps. (A) Representative L-type Ca2+ current recordings before (CCa2+ current–voltage relationships before (open circles, n=3) and after exposure to 1.1 mML-type Ca2+ channels. Open circles represent the control, and closed triangles represent 1.1 mOpen circles represent the control and closed circles represent 1.1 mM R(+)-pulegone (n=4)

In cardiac myocytes, repolarisation of the AP is mainly due to thepresence of an outward K+ current (Nerbonne and Kass, 2005).Fig. 7A shows that 1.1 mM R(+)-pulegone was able to reduce inwardand outward K+ current components. Fig. 7B shows the reduction ofthe outward K+ current (51.5±4.4 A/F vs. 4.9±0.4 A/F, at 60 mV,Pb0.05, n=3). R(+)-pulegone reduced the inward rectifier K+ currentfrom−20.6±1.1 A/F to−12.3±0.7 A/F at−120 mV (Pb0.05, n=3).Fig. 7 (C and D) shows the time course of the effects of R(+)-pulegoneon the outward and inward K+ currents, respectively. Overall, these re-sults allowed us to establish a putative sequence of sub-cellular events

tivation and inactivation relationships in cardiac ventricular myocytes. From a holdingd T-type Ca2+ channels and then further depolarised to membrane potentials betweenontrol, panel Ai) and after application of 1.1 mM R(+)-pulegone (panel Aii). (B) L-typeR(+)-pulegone (closed triangles, n=3). (C) Steady-state activation relationships for

M R(+)-pulegone. (D) Steady-state inactivation relationships for L-type Ca2+ channels.. *Pb0.05 versus control values.

Fig. 5. Representative line-scan images recorded from field-stimulated cardiomyocytesloaded with the Ca2+ indicator Fluo4-AM in control conditions (A) and exposed toR(+)-pulegone at 0.11 mM (B), 1.1 mM (C), and 3.2 mM (D). The Ca2+ signal is shownas the fluorescence ratio (F/F0), with the fluorescence intensity (F) normalised to theminimum intensity measured in the diastolic phase (F0). (E) Averaged peak Ca2+

transient in control (n=104), with 0.11 mM (n=26), 1.1 mM (n=32), and 3.2 mMR(+)-pulegone (n=13). Values are means±S.E.M. *Pb0.05, ***Pb0.0001.

Fig. 6. Modulation of the action potential and resting membrane potential by R(+)-pulegongone effects on cardiac action potentials. (B) Time course of the effects of R(+)-pulegone ospond to a failure in triggering APs (to see coloured symbols the reader is referred to the web10%, 50%, and 90% repolarisation (n=5, *Pb0.05). (D) Effects of R(+)-pulegone (1.1 mM) opulegone (1.1 mM) effects on the resting membrane potential (n=5, *Pb0.05).

140 S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

that may explain how R(+)-pulegone decreases cardiac contractility,changes the intracellular global calcium transient, andmodulates actionpotential properties.

4. Discussion

In this study, we found that the negative inotropic effect of R(+)-pulegone on guinea pig atria is due to a reduction of ICa,L and a conse-quent decrease in the global Ca2+ transient. The evidence that the re-duction of twitch amplitude by R(+)-pulegone is related to theinhibition of ICa,L is that the positive inotropism was significantly affect-ed in response to Ca2+ or BAY K8644. Moreover, the negative inotropiceffect of R(+)-pulegone turns out to be very similar to that of the L-typeCa2+ channel blocker nifedipine. These findings strongly support thehypothesis that the negative inotropic effect of R(+)-pulegone is mostlikely mediated by the inhibition of L-type Ca2+ channels, which leadsto a reduced Ca2+ influx and, consequently, to a diminished Ca2+ re-lease from the internal stores. Taken together, these two actions con-tribute to the decrease in twitch amplitude.

These observations are consistent with previous reports showingthat R(+)-pulegone exposure results in spasmolytic action. At thispoint, it is important to mention that the effects of R(+)-pulegoneon atrial contractility are nearly reversible, making the possibility ofmyocardial damage unlikely.

Calcium cycling is essential for contraction, and Ca2+ influx via theL-type Ca2+ channel is a critical determinant of contractility. It is wellknown that the blockade of Ca2+ channels in atrial muscle leads to anegative inotropic response (Opie et al., 2000). Another very impor-tant mechanism involved in controlling cardiac muscle contractionand relaxation is related to myofibrillar biochemical properties(Bers, 2001). However, there are few studies exploring changes incardiac contractility in the presence of compounds that modify eitherintracellular Ca2+ regulation or myofibrillar properties (Lew, 1995;Vittone et al., 1994).

e in mice ventricular cardiomyocytes. (A) Representative traces of 1.1 mM R(+)-pule-n the cardiac action potential duration at 50% repolarisation (APD50). Red circles corre-version of this article). (C) Effects of R(+)-pulegone on the action potential duration atn the resting membrane potential (RMP) as a function of time. (E) Summary of R(+)-

Fig. 7. Effects of R(+)-pulegone on inward and outward potassium currents in mice ventricular cardiomyocytes. (A) Representative traces of potassium currents evoked by a rampprotocol (see Materials and methods for additional details) in the control (black trace), in the presence of 1.1 mM R(+)-pulegone (light grey trace), and after washing off the R(+)-pulegone (dark grey trace). (B) Effects of 1.1 mM R(+)-pulegone on outward and inward potassium currents. White and grey bars represent the outward currents evoked by+60 mV pulses (holding potential:−80 mV) in the control and in the presence of R(+)-pulegone, respectively (n=3, *Pb0.05). White and grey dashed bars represent the inwardcurrents evoked by−120 mV pulses (holding potential:−80 mV) in the control and R(+)-pulegone conditions, respectively (n=3, *Pb0.05). Effects of 1.1 mM R(+)-pulegone onthe time courses of outward (C) and inward (D) potassium current evoked by +60 mV and −120 mV pulses from a holding potential of −80 mV every 10 s.

141S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

Kondo et al. (1990) have reported that disopyramide, at submilli-molar concentrations, caused a negative inotropic response and ashortening of time to peak tension. In the present study, we also ob-served a significant shortening in the time to peak tension (byabout 1.24 times) in response to R(+)-pulegone. Our results are con-sistent with Baudet and Noireaud (1999), who demonstrated that al-tering SR Ca2+ handling does not necessarily lead to changes incontraction–relaxation indices.

The results of this study provide thefirstmeasurement of the cardiaceffects of R(+)-pulegone at both functional and sub-cellular levels. Tofurther understand the possible mechanism by which R(+)-pulegoneaffects themammalian heart, we performed experiments using isolatedventricular cardiac myocytes to investigate whether R(+)-pulegoneaffects the intracellular global Ca2+ transient.

In the heart, Ca2+ influx through the sarcolemmal membrane isessential for triggering Ca2+ release from the sarcoplasmic reticulum(SR). The combination of Ca2+ influx and SR Ca2+ release promotesan increase in the free intracellular Ca2+ concentration, allowing themuscle to contract (Bers, 2002).

The L-type Ca2+ channel conductance was not altered by R(+)-pulegone, suggesting that it may not interfere with the activationpathway of voltage-dependent L-type Ca2+ channels. These data cor-roborate the idea that R(+)-pulegone provokes ICa,L reduction by di-rectly blocking the channel. Additionally, R(+)-pulegone changedthe L-type Ca2+ channel steady-state inactivation, shifting the V0.5

for inactivation to more negative values (approximately −16 mV).These findings may contribute to understanding how R(+)-pulegonedecreases the ICa,L.

It is reasonable to think that a decrease of the L-type Ca2+ currentwould profoundly affect the Ca2+ mobilisation from the SR (Richardet al., 2006). Besides the inhibitory effect of R(+)-pulegone on ICa,L,our results showed that R(+)-pulegone also affects the Ca2+ tran-sient amplitude. R(+)-pulegone was able to inhibit SR Ca2+ releaseand thus impair cellular contraction in isolated cardiomyocytes.In addition, the contractile force depends on the number of

cardiomyocytes synchronically activated and recruited at each cardi-ac cycle (Bers, 2002). Because R(+)-pulegone promotes AP failure,the number of cardiac cells available to contract may be reduced,which would contribute to the cardiodepressor effect of R(+)-pulegone.

Cardiac Ca2+ ATPase (SERCA2a) plays a dominant role in loweringcytoplasmic Ca2+ levels (Bers, 2001; Nakayama et al., 2003). In mouseventricular myocytes, SERCA2a removes approximately 92% of sarco-plasmic Ca2+ (Bassani et al., 1994; Li et al., 1998). In our experiments,we observed that R(+)-pulegone did not significantly change theglobal Ca2+ transient decay kinetics measured at 90% of this decay(t90). These results suggest that SERCA2a activity is not altered byR(+)-pulegone.

In the present study, we also demonstrated that R(+)-pulegonewas able to prolong APs and inhibit voltage-gated K+ currents.Therefore, it is expected that the prolongation of APs leads tomore Ca2+ influx (Bassani et al., 2004; Volk et al., 1999); however,R(+)-pulegone also blocks L-type Ca2+ channels. In fact, prolonga-tion of the APD may contribute to making the EC50 of R(+)-pule-gone on cardiac contractile force relatively larger. Moreover, R(+)-pulegone induced hyperpolarisation of the RMP. It is well knownthat membrane hyperpolarisation should recruit more Na+ channelsto the resting state (Hille, 2001) and consequently a larger over-shoot should be observed. However, our data show that R(+)-pule-gone does not change the action potential amplitude. Additionally,R(+)-pulegone also reduces the maximal depolarisation rate. Theseresults suggest that R(+)-pulegone may act on cardiac Na+ channels,but this hypothesis needs to be confirmed.

The toxicity of R(+)-pulegone observed by others (Anderson et al.,1996; Bakerink et al., 1996), which is correlated with sudden death,may be attributable to a cardiac failure due to a reduction of the sarco-plasmatic calcium and changes on the AP electrical properties. Theseevents could lead to arrhythmias, causing sudden death (Guo et al.,2000; Miake et al., 2003); however, additional studies should be doneto evaluate whether this is the case.

142 S.V.S. Cerqueira et al. / European Journal of Pharmacology 672 (2011) 135–142

The present study clearly demonstrates that R(+)-pulegone de-creased myocardial contractility and markedly reduced both theintracellular Ca2+ transient and L-type Ca2+ current. Additionally,R(+)-pulegone caused changes in the action potential waveformand reductions of the outward and inward potassium current. Althoughit is possible that R(+)-pulegone acts on othermechanisms responsiblefor the control of cardiac intracellular Ca2+ handling, for instance at thelevel ofmitochondria, or onNa+/Ca2+ exchange, the simple blockage ofL-type Ca2+ channels and modulation of action potentials are sufficientto explain the negative inotropic actions of R(+)-pulegone.

Acknowledgements

This work was supported by grants from Conselho Nacional deDesenvolvimento Científico e Tecnológico (CNPq), Fundação deAmparo à Pesquisa do Estado de Minas Gerais (FAPEMIG), Coordena-ção de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and Pro-grama de Apoio a Núcleos de Excelência (PRONEX/FAPEMIG/CNPq).We would like to thank CEMEL (Biological Sciences Institute, FederalUniversity of Minas Gerais, Confocal Microscopy Facility).

References

Anderson, I.B., Mullen, W.H., Meeker, J.E., Khojasteh-Bakht, S.C., Oishi, S., Nelson, S.D.,Blanc, P.D., 1996. Pennyroyal toxicity: measurement of toxic metabolite levels intwo cases and review of the literature. Ann. Intern. Med. 124, 726–734.

Bakerink, J.A., Gospe, S.M., Dimand, R.J., Eldridge, M.W., 1996. Multiple organ failureafter ingestion of pennyroyal oil from herbal tea in two infants. Pediatrics 98,944–947.

Bassani, J.W.M., Bassani, R.A., Bers, D.M., 1994. Relaxation in rabbit and rat cardiac cells:species dependent differences in cellular mechanisms. J. Physiol. 476, 279–293.

Bassani, R.A., Altamirano, J., Puglisi, J.L., Bers, D.M., 2004. Action potential durationdetermines sarcoplasmic reticulum Ca2+ reloading in mammalian ventricularmyocytes. J. Physiol. 559, 593–603.

Baudet, S., Noireaud, J., 1999. Pharmacologic evaluation of isometric contraction–relaxation coupling indexes in rabbit ventricular muscle. J. Pharmacol. Toxicol.42, 21–30.

Bers, D.M., 2001. Inotropic mechanisms in cardiac muscle, Heart Physiology andPathophysiology, Fourth Edition. , pp. 779–788.

Bers, D.M., 2002. Cardiac excitation–contraction coupling. Nature 415, 198–204.Bers, D.M., 2008. Calcium cycling and signaling in cardiac myocytes. Annu. Rev. Physiol.

70, 23–49.Bowen, I.H., Cubbin, I.J., 1992. Mentha piperita and Mentha spicata. In: De Smet,

P.A.G.M., Keller, K., Hansel, R., Chandler, R.F. (Eds.), Adverse Effects of HerbalDrugs 1. Springer-Verlag, Berlin, pp. 171–176.

Brankovic, S.V., Kitic, D.V., Radenkovic, M.M., Veljkovic, S.M., Golubovic, T.D., 2009.Calcium blocking activity as a mechanism of the spasmolytic effect of the essentialoil of Calamintha glandulosa Silic on the isolated rat ileum. Gen. Physiol. Biophys. 28,174–178.

Dorigo, P., Gaion, R.M., Bergamin, M., Giacometti, A., Valentini, E., Moragno, I., 1990.Comparison between the cardiac effects induced by muzolimine and furosemidein guinea-pig atria. Cardiovasc. Drugs Ther. 4, 1477–1486.

Gordon, W.P., Forte, A.J., McMurtry, R.J., Gal, J., Nelson, S.D., 1982. Hepatotoxicity andpulmonary toxicity of pennyroyal oil and its constituent terpenes in the mouse.Toxicol. Appl. Pharmacol. 65, 413–424.

Grundschober, F., 1979. Literature review of pulegone. Perfumer Flavorist 4, 15–17.Guedes, D.N., Silva, D.F., Barbosa-Filho, J.M., de Medeiros, I.A., 2004. Endothelium

dependent hypotensive and vasorelaxant effects of the essential oil from aerialparts of Mentha × villosa in rats. Phytomedicine 11, 490–497.

Guo,W., Li, H., London, B., Nerbonne, J.M., 2000. Functional consequences of eliminationof Ito, f and Ito, s: early after depolarizations, atrioventricular block, and ventricular

arrhythmias in mice lacking Kv 1.4 and expressing a dominant-negative Kv 4 alphasubunit. Circ. Res. 87, 73–79.

Hille, B., 2001. Ion Channels of Excitable Membrane, Third Edition. Sinauer Associates,Inc, p. 44.

Hills, J.M., Aaronson, P.I., 1991. The mechanism of action of peppermint oil on gastroin-testinal smooth muscle. Gastroenterology 101, 55–65.

Kondo, N., Mizukami, M., Shibata, S., 1990. Negative inotropic effects of disopyramideon guinea-pig papillary muscles. Br. J. Pharmacol. 101, 789–792.

Lara, A., Damasceno, D.D., Pires, R., Gros, R., Gomes, E.R., Gavioli, M., Lima, R.F., Guimaraes,D., Lima, P., Bueno, C.R., Vasconcelos, A., Roman-Campos, D., Menezes, C.A.S., Sirvente,R.A., Salemi, V.M., Mady, C., Caron, M.G., Ferreira, A.J., Brum, P.C., Resende, R.R., Cruz,J.S., Gomez, M.V., Prado, V.F., de Almeida, A.P., Prado, M.A.M., Guatimosim, S., 2010.Dysautonomia due to reduced cholinergic neurotransmission causes cardiac remo-deling and heart failure. Mol. Cell. Biol. 30, 1746–1756.

Lew, W.Y., 1995. Asynchrony and ryanodine modulate load-dependent relaxation inthe canine left ventricle. Am. J. Physiol. 268, H17–H24.

Li, L., Chu, G., Kranias, E.G., Bers, D.M., 1998. Cardiac myocyte calcium transport inphospholamban knockout mouse: relaxation and endogenous CaMKII effects.Am. J. Physiol. 274, H1335–H1347.

Miake, J., Marban, E., Nuss, B., 2003. Functional role of inward rectifier current in heartprobe by Kir 2.1 overexpression and dominant-negative suppression. J. Clin. Invest.111, 1529–1536.

Molck, A.-M., Morten, P., Soren, T.L., Preben, O., 1998. Lack of histological cerebellarchanges in Wistar rats given pulegone for 28 days. Comparison of immersion andperfusion tissue fixation. Toxicol. Lett. 95, 117–122.

Moorthy, B., Madyastha, P., Madyastha, K.M., 1989. Hepatotoxicity of pulegone in rats:its effects on microsomal enzymes, in vivo. Toxicology 55, 327–337.

Nakayama, H., Otsu, K., Yamaguchi, O., Nishida, K., Date, M., Hongo, K., Kusakari, Y.,Toyofuku, T., Hikoso, S., Kashiwase, K., Takeda, T., Matsumura, Y., Kurihara, S.,Hori, M., Tada, M., 2003. Cardiac-specific overexpression of a high Ca2+ affinitymutant of SERCA2a attenuates in vivo pressure overload cardiac hypertrophy.FASEB J. 17, 61–63.

Nerbonne, J.M., Kass, R.S., 2005. Molecular physiology of cardiac repolarization. Physiol.Rev. 85, 1205–1253.

Oliveira, F.A., Guatimosim, S., Castro, C.H., Galan, D.T., Lauton-Santos, S., Ribeiro, A.M.,Almeida, A.P., Cruz, J.S., 2007. Abolition of reperfusion-induced arrhythmias inhearts from thiamine-deficient rats. Am. J. Physiol. Heart Circ. Physiol. 293,H394–H401.

Opie, L.H., Yusuf, S., Kubler, W., 2000. Current status of safety and efficacy of calciumchannel blockers in cardiovascular diseases: a critical analysis based on 100 stud-ies. Orig. Res. Artic. Prog. Cardiovasc. Dis. 43, 171–196.

Petrakis, E.A., Kimbaris, A.C., Pappas, C.S., Tarantilis, P.A., Polissiou, M.G.J., 2009. Quantitativedetermination of pulegone in pennyroyal oil by FT-IR spectroscopy. J. Agric. Food Chem.57, 10044–10048.

Richard, S., Perrier, E., Fauconnier, J., Perrier, R., Pereira, L., Gõmez, A.M., Bénitah, J.P.,2006. ‘Ca2+-induced Ca2+ entry’ or how the L-type Ca2+ channel remodels itsown signalling pathway in cardiac cells. Prog. Biophys. Mol. Biol. 90, 118–135.

Shioya, T.A., 2007. A simple technique for isolating healthy heart cells from mousemodels. J. Physiol. Sci. 57, 327–335.

Soares, P.M., Assreuy, A.M., Souza, E.P., Lima, R.F., Silva, T.O., Fontenele, S.R., Criddle,D.N., 2005. Inhibitory effects of the essential oil of Mentha pulegium on the isolatedrat myometrium. Planta Med. 71, 214–218.

Sousa, P.J., Magalhães, P.J., Lima, C.C., Oliveira, V.S., Leal-Cardoso, J.H., 1997. Effects ofpiperitenone oxide on the intestinal smooth muscle of the guinea pig. Braz. J.Med. Biol. Res. 30, 787–791.

Souza, A.A.S., Soares, P.M.G., Almeida, A.N.S., Maia, A.R., Souza, E.P., Assreuy, A.M.S.,2010. Antispasmodic effect of Mentha piperita essential oil on tracheal smoothmuscle of rats. J. Ethnopharmacol. 130, 433–436.

Sullivan, J.B., Rumack, B.H., Thomas, H., Peterson, R.G., Bryson, P., 1979. Pennyroyal oilpoisoning and hepatotoxicity. J. Am. Med. Assoc. 242, 2873–2874.

Vittone, L., Mundina-Weilenmann, C., Mattiazzi, A., Cingolani, H., 1994. Physiologic andpharmacologic factors that affect myocardial relaxation. J. Pharm. Toxicol. Methods32, 7–18.

Volk, T., Nguyen, T.H.-D., Schultz, J.-H., Ehmke, H., 1999. Relationship between transientoutward K+ current and Ca2+ influx in rat cardiac myocytes of endo- and epicardialorigin. J. Physiol. 519, 841–850.