Pb2+ inhibits the hyperpolarization-activated current in acutely isolated dorsal root ganglion neurons

Neuroscience 120 (2003) 57– 63

Pb2ⴙ INHIBITS THE HYPERPOLARIZATION-ACTIVATED CURRENT IN ACUTELY ISOLATED DORSAL ROOT GANGLION NEURONS al., 2000; Wang et al., 1997) and Ih was inhibited by Cs⫹ and ZD7288 (Satoh and Yamada, 2000). Modulation of Ih by neurotransmitters and intracellular second messengers in various cells has been reported, but the mechanisms responsible for such a modulation vary (Satoh and Yamada, 2001). However, little is known about the effects of Pb2⫹ on Ih. Therefore, the principal purpose of the present study was to determine the effects of acute application of Pb2⫹ on Ih and discuss possible mechanisms underlying the Pb2⫹ action.

X.-Q. DAI,a,b* E. KARPINSKIa AND X.-Z. CHENa* a

Membrane Protein Research Group, Department of Physiology, University of Alberta, Edmonton, Alberta, T6G 2H7, Canada b School of Life Science, University of Science and Technology of China, Hefei, Anhui 230027, China

Abstract—The hyperpolarization-activated h channel current (Ih) reported to be present in acutely isolated rat dorsal root ganglion (DRG) neurons is inhibited by Csⴙ and ZD7288. It was recently reported that lead (Pb2ⴙ) inhibits voltage-gated Ca2ⴙ and Kⴙ channels in DRG neurons but the effect of Pb2ⴙ on Ih has so far not been reported. Using whole-cell patch clamp technique we show that Pb2ⴙ specifically inhibited Ih. External application of 0.1, 1 and 10 ␮M Pb2ⴙ reversibly reduced the amplitude of Ih in a dose-dependent manner, with an IC50 value of 3.7 ␮M and a Hill coefficient of 1.1. Pb2ⴙ shifted the activation curve of Ih by 9.3 mV but had no effect on the slope factor. Pb2ⴙ inhibited Ih in a voltage-dependent manner and slowed down the activation process, indicating an action of Pb2ⴙ on the activation kinetics of h channels. Our studies thus demonstrated that Pb2ⴙ is a dose-dependent, voltage-dependent and reversible blocker of Ih in DRG neurons. © 2003 IBRO. Published by Elsevier Science Ltd. All rights reserved.

EXPERIMENTAL PROCEDURES All experiments conformed to the Health Sciences Animal Policy and Welfare Committee of The University of Alberta and other international guidelines. During all the course of the experiments, the number of animals used and their suffering were minimized as much as possible. Adult Wistar rats (15–21 days) were anesthetized deeply with pentobarbital sodium (80 mg/kg) injected intraperitoneally. Dorsal root ganglion (DRG) neurons were isolated according to the method described in our previous study (Dai et al., 2001). Medium size DRG neurons (32– 40 ␮m in diameter) were used. External solution contained (in mM): 140 NaCl, 4 KCl, 1.8 CaCl2, 1 MgCl2, 1 BaCl2, 10 HEPES, 10 D-glucose, and 0.5 ␮M tetrodotoxin, pH 7.3. Electrodes were filled with (in mM): 140 KCl, 4 NaCl, 2 Na2-ATP, 1 MgCl2, 10 EGTA, and 10 HEPES, pH 7.2. All drugs were purchased from Sigma. Pb2⫹-acetate of extra pure grade (extra pure, Merck Inc., Whitehouse Station, NJ, USA) was dissolved in distilled water at a concentration of 20 mM and added it to the external solution just before each experiment to avoid precipitation. The Pb2⫹ concentrations in the final external solution were calculated with the aid of the “Chelator” program (Schoenmakers et al., 1992) using appropriate stability constants reported by Martell and Smith (1982). All experiments were carried out at room temperature (22–24 °C). Patch electrodes were made by a PP-830 micropipette puller (Narishige, Japan) with resistance of 2–5 M⍀. Membrane currents were digitized at 0.5 ms/sample, filtered at 2 kHz and measured in conjunction with an EPC-9 amplifier operating Pulse/Pulsefit 7.89 software (HEKA Elektronik, Lambrecht, Germany). When wholecell configuration was established, the capacitance and series resistance were compensated. Leakage was subtracted on line. Neurons with an inadequate seal or a resting potential more depolarized than ⫺50 mV were excluded from data analysis. Data analysis was performed using Origin 6.0 (Microcal Software, Inc. Northampton, MA, USA). To examine dose dependence of the Pb2⫹-inhibited current, experimental data were fitted with the Logistic equation:

Key words: lead, Ih, DRG, rat.

Lead (Pb2⫹) is one of the most common neurotoxic metals present in our environment. Previous studies on the interaction of Pb2⫹ with voltage-gated calcium channels and transient outward currents have reported a reduction in current amplitude following acute exposure of various neurons to Pb2⫹ (Bu¨sselberg et al., 1994; Dai et al., 2001). Hyperpolarization-activated currents (Ih) refer to a mixed Na⫹ and K⫹ currents activated by membrane hyperpolarization and deactivated by membrane depolarization. Ih is an important contributor to the rhythmic (or pacemaker) activity in neurons and heart cells. The presence of Ih has been described in different neurons (Bal and Oertel, 2000; Brown et al., 1979; Ghamari-Langroudi and Bourque, 2000; McCormick and Pape, 1990; Santoro et *Correspondence to: X.-Q. Dai or X.-Z. Chen, Department of Physiology, University of Alberta, Edmonton, Alberta, T6G 2H7, Canada. Tel: ⫹1-780-492-2307; fax ⫹1-780-492-8915. E-mail addresses: [email protected] (X-Q. Dai) or xzchen@ ualberta.ca (X.-Z. Chen). Abbreviations: DRG, dorsal root ganglion; HEPES, HCO(3-)-free N-2hydroxyethylpiperazine-N'-2-ethanesulfonic acid; ICtrl, amplitude of Ih in control solution; Ih, hyperpolarization-activated current; Iins, instantaneous current; Iir, inward rectifier K⫹ current; Imax, maximal amplitude of Ih; IPb, amplitude of Ih in Pb2⫹ solution; Iss, steady-state current; k, slope factor of the Boltzmann equation; ␶, activation time constant; V1/2; half-activation voltage of the Boltzmann equation.

I⫽(I0⫺I1)/[1⫹(x/x0)p]⫹I1

(I) 2⫹

where I0 and I1 are is the minimal ([Pb ]⫽0) and maximal ([Pb2⫹]⬁) current values, respectively. x0 corresponds to the IC50 value and p is the Hill coefficient. The rate of the Ih channel activation was determined by fitting with a mono-exponential function of the form:

0306-4522/03$30.00⫹0.00 © 2003 IBRO. Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0306-4522(03)00279-3

57

58

X.-Q. Dai et al. / Neuroscience 120 (2003) 57– 63

Ih(t)⫽A·e(⫺t/␶)⫹C

(II)

where ␶ represents the activation time constant. The steady-state Iss was plotted as a function of the membrane potential and fitted with the following Boltzmann equation: I⫽Imax/兵1⫹exp关共V⫺V1/2)/k]}

(III)

Where Imax is the fitted maximal current amplitude. V1/2 and k represent the half activating voltage and slope factor of the I–V curve around the point V1/2, respectively. Mean values were expressed as means⫾S.E. and statistical comparisons conducted using one-way analyses of variance (Origin 6.0). Difference between two values was considered significant or very significant when P⬍0.05 or 0.01.

RESULTS From a holding potential of ⫺60 mV, in response to a series of hyperpolarizing voltage pulses ranging from ⫺60 to ⫺150 mV, in 15 mV increments, a hyperpolarizationactivated current (Ih) was observed, which was comprised of an initial instantaneous inward current (instantaneous current, Iins) and a subsequent slow activating inward current (steady-state current, Iss; Fig. 1A) in 84.3% of medium size DRG neuron bodies. Ih displayed reproducible timeand voltage-dependent activation patterns similar to those described by Mayer and Westbrook (1983). In 10 DRG neurons tested, bath application of 3 mM Cs⫹ resulted in a complete and reversible block of Iss (Fig. 1A, B), indicating that Iss is mainly due to K⫹ channels that have been shown to be selectively blocked by Cs⫹. When 3 mM Cs⫹ was replaced by 10 ␮M Pb2⫹ similar inhibition pattern was observed (Fig. 1C, D). In 14 DRG neurons tested, the Pb2⫹-inhibited current exhibited similar voltage dependence as the Cs⫹-inhibited current (Fig. 1E). Note that the Pb2⫹-inhibited current was slightly smaller in amplitude than the Cs⫹-inhibited current. Also, inhibition by Pb2⫹ and Cs⫹ differed in the half-maximal activation voltage (V1/2). The V1/2 values for the Pb2⫹-inhibited curve was significantly shifted in the hyperpolarizing direction (by 10.6 mV at 10 ␮M Pb2⫹, P⬍0.05), compared with the Cs⫹-inhibited I–V curve (Fig. 1E). In contrast, the k values for Cs⫹- and Pb2⫹-inhibited currents showed no significant difference (14.5⫾1.4 and 14.3⫾4.0, respectively). Furthermore, similar to Cs⫹, Pb2⫹ had a greater inhibitory effect on Iss than on Iins at voltages between ⫺60 and ⫺150 mV. These results together indicate that Pb2⫹ and Cs⫹ inhibited the same type of currents, i.e. the steady-state component of Ih. To further document that the two inhibitors block the same current, we examined effects of 10 ␮M Pb2⫹ in the presence of 3 mM Cs⫹. Indeed, no inhibition by Pb2⫹ was observed when Pb2⫹ was applied in the presence of Cs⫹ (Fig. 2A–C), demonstrating that both metals inhibit the Iss current. In addition to Ih, many neurons also possess a fast inward rectifier K⫹ current (Iir). It has been shown that in many cells, application of Ba2⫹ can pharmacologically dissect Ih from Iir (Scroggs et al., 1994). Ba2⫹ blocked Iir but, in large part, had no effect on Ih. In agreement with this, in all DRG

neurons tested (n⫽10), bath application of 1 mM Ba2⫹ had no significant effect on the steady-state Ih (Fig. 2D, E). Therefore, in the subsequent experiments on effects of Pb2⫹ on Ih, 1 mM Ba2⫹ was added to all external solutions. In 42 neurons tested, Ih was reduced following the application of Pb2⫹ for 0.5–2 min (Fig. 3A). Application of 0.1, 1 and 10 ␮M Pb2⫹ decreased Ih at ⫺150 mV by 2.1⫾0.4% (P⬎0.05), 22.3⫾4.6% (P⬍0.05) and 82.1⫾18.4% (P⬍0.01), respectively, indicating that the inhibition by Pb2⫹ was dose-dependent. Complete recovery was observed after the 0.1 or 1 ␮M Pb2⫹-containing external solution was replaced by the Pb2⫹-free solution for approximately 2 min, while limited recovery was seen after a 10 min washout following the application of 10 ␮M Pb2⫹. The IC50 value for Pb2⫹ was approximately 3.7 ␮M, and the Hill coefficient was 1.1 (equation I and Fig. 3B). The V1/2 and k values of the activation curve were ⫺95.5⫾1.2 mV and 15.1⫾1.1, respectively, in the absence of Pb2⫹, and ⫺104.8⫾3.2 mV, and 14.9⫾3.7, respectively, at 10 ␮M Pb2⫹ (Fig. 4A). Thus, Pb2⫹ produced a 9.3 mV hyperpolarizing shift in the activation curve of Ih without a significant change in the slope factor. The ratio of the Ih current at 10 ␮M Pb2⫹ (IPb) to that in the absence of Pb2⫹ (ICtrl), IPb/ICtrl, was used to assess the voltage dependence of the Pb2⫹ inhibition and plotted as a function of the membrane potential (Fig. 4B). With 10 ␮M Pb2⫹, the ratio decreased in a voltage-dependent manner from 64.1⫾4.2% at ⫺60 mV to 22.3⫾1.3 at ⫺150 mV (n⫽10), indicating that the potency of Pb2⫹ increased with hyperpolarization. From ⫺60 to ⫺90 mV, the inhibition potency increased steeply but from ⫺90 to ⫺150 mV, the inhibition was less voltage-dependent and more or less reached its maximum. To determine whether the Pb2⫹ blockade on Ih could be relieved at greater hyperpolarizing voltages, 5 ␮M Pb2⫹ was used. At ⫺150 mV, 5 ␮M Pb2⫹ inhibited the current by 47.1⫾6.3%; at ⫺165 and ⫺180 mV, it was 49.9⫾8.6% and 45.6⫾7.9%, respectively. This result rather supports that there is no obvious inhibition relief at hyperpolarization. The effects of 1 ␮M Pb2⫹ on the Ih activation time course in DRG neurons were also investigated. The inward sag associated with Ih activation was well fitted by a monoexponential function in the interval of ⫺75 and ⫺150 mV (equation II). The activation time constant (␶) of Ih was clearly voltage-dependent (Fig. 4C). In the absence of Pb2⫹, the ␶ value was 204⫾15, 156⫾13, 122⫾10, 100⫾8, 80⫾7 and 65⫾4 ms (n⫽12) at ⫺75, ⫺90, ⫺105, ⫺120, ⫺135 and ⫺150 mV, respectively. In comparison, In the presence of 1 ␮M Pb2⫹, ␶ was equal to 200⫾18, 168⫾14, 140⫾14, 122⫾10, 105⫾9 and 93⫾7 ms (n⫽12) at the same final voltages (Fig. 4C). Thus, 1 ␮M Pb2⫹ significantly slowed down the Ih activation process at hyperpolarizing voltages ranging from ⫺90 to ⫺150 mV (P⬍0.05), consistent with the voltage-dependent Pb2⫹ inhibition.

DISCUSSION In this present study, we have shown that Pb2⫹ reversibly inhibited a K⫹ current in rat DRG neurons, and the inhibi-

X.-Q. Dai et al. / Neuroscience 120 (2003) 57– 63

59

Fig. 1. Cs⫹ and Pb2⫹ inhibition of the hyperpolarization-activated current (Ih) in rat DRG neurons. (A) Representative recording of Ih in a DRG neuron in the absence of Cs⫹. Whole-cell currents were recorded in response to 800-ms voltage steps from a holding potential of ⫺60 mV in 15 mV increments (top). Arrows indicate the positions where Iins and Iss were measured. The dashed horizontal bars represent the zero-current level. (B) Ih in the presence 3 mM Cs⫹ from the same cell as panel A. (C) Representative Ih in the absence of Pb2⫹. (D) Ih in the presence of 10 ␮M Pb2⫹ from the same neuron as panel C. (E) The Cs⫹-inhibited and Pb2⫹-inhibited steady-state currents (Iss) versus the membrane voltage, each averaged from 14 neurons. Curves were fitted using the Boltzmann equation (equation III), and then all data were normalized by dividing by the fitted Imax obtained from the Boltzmann fit to the Cs⫹-inhibited currents. V1/2⫽⫺91.1⫾1.5 mV and k⫽14.5⫾1.4 for the Cs⫹-sensitive curve; V1/2⫽⫺101.7⫾3.6 mV and k⫽14.3⫾4.0 for the Pb2⫹-inhibited curve.

60

X.-Q. Dai et al. / Neuroscience 120 (2003) 57– 63

Fig. 2. Specific effects of Pb2⫹ on Ih and effects of Ba2⫹. (A) Representative recording of Ih in the absence of inhibitors. (B) Ih recording in the presence of 3 mM Cs⫹ from the same cell as panel A. (C) Ih in the presence of 3 mM Cs⫹ and 10 ␮M Pb2⫹ from the same cell as panels A and B. (D) Representative recording of Ih in the absence of Ba2⫹. (E) Ih recording in the presence of 1 mM Ba2⫹ from the same neuron as panel D.

tion was dose-dependent with an IC50 value of 3.7 ␮M. By comparing with the Cs⫹-inhibited current, we have demonstrated that the Pb2⫹-inhibited current was indeed mediated by the h channel. We have also shown that the Pb2⫹ inhibition was voltage-dependent, with greater inhibition at hyperpolarization, and increased the time constants of Ih. It was reported that Ih plays an important role in the control of electrical activity in DRG neurons during hyperpolarization (Mayer and Westbrook, 1983; Pape, 1996; Scroggs et al., 1994). The major functional roles of Ih are: (1) contribution to the resting membrane potential; (2) control of the rhythmic-oscillatory activity; (3) maintenance of the membrane potential toward depolarization; and (4) contribution to the production of afterhyperpolarization. Pb2⫹ may alter neurophysiological processes by exerting its neurotoxic effects. Suppression of Ih

by Pb2⫹ causes depressant effects on neuronal activity. Two analyses shed light on the consequences of the Pb2⫹ action on DRG neurons. First, the role of Ih in the determination of the resting potential makes it a particularly useful mechanism for providing pacemaker depolarization during generation of rhythmic-oscillatory activity (Pape, 1996). The normal resting potential is distinctively positive to the presumed K⫹ equilibrium potential in many types of neurons, indicating a contribution of depolarizing currents. Because part of the Ih conductance is active at the resting potential, the recruitment of the fraction of channels contributes to the total resting conductance and participates in depolarizing the membrane potential to the K⫹ equilibrium potential. The depolarizing action of Ih could be naturally exploited by neurons as an excitability-promoting factor. Whether the depolarization evokes spike activity or re-

X.-Q. Dai et al. / Neuroscience 120 (2003) 57– 63

61

Fig. 3. Dose dependence of the Pb2⫹ inhibition of Ih. (A) Representative trace recordings in the presence of 0, 1 and 10 ␮M Pb2⫹ obtained following a voltage step from ⫺60 to ⫺150 mV. (B) Dose-response curve of the Pb2⫹-inhibited currents. Shown data were averages of those obtained from 14 DRG neurons. The curve was a fit to a logistic equation. The IC50 value was 3.7 ␮M and the Hill coefficient 1.1.

mains sub-threshold is largely dependent upon the position of the Ih activation curve with respect to the threshold for the generation of Na⫹/K⫹-mediated action potentials. It is possible that the blockade of Ih by Pb2⫹ leads to a reduction of the depolarizing influence after the initial phase of afterhyperpolarization. The negative shift in Ih activation by Pb2⫹ will result in a slight hyperpolarization of the membrane. Therefore, Pb2⫹ could decrease the contribution made by Ih to the resting membrane conductance. The membrane potential would be expected to shift away from the level at which action potential is triggered. The inhibition of Ih by Pb2⫹ may help to maintain decreased excitability by preventing Ih from returning the membrane potential to an action potential threshold. Second, in DRG neurons showing spontaneous action potential activity, if the afterhyperpolarization of the spontaneous action potential is negative and lasts long enough to activate Ih, then Ih would contribute to facilitation of the firing discharges (Yagi and Sumino, 1998). Inhibition of Ih by Pb2⫹ may reduce the discharge frequency of spontaneously active DRG neurons, and delays initiation of the next action potential, resulting in the reduction of the membrane conductance and frequency of the repetitive action potentials evoked by a depolarizing current. Because Ih opposes membrane depolarization, the

inhibitory effects produced by Pb2⫹ would affect neuronal excitability. There are two possible mechanisms by which this might occur: a general binding of cations to negative charges of the surface (charge screening), and a specific, negatively charged binding site for Pb2⫹ formed within the channel protein. Pb2⫹ action through general charge screening would apply to many different ion channels, while binding of Pb2⫹ to an intramembrane site requires physical and charge fits between a Pb2⫹ metal ion and a binding site. The facts that the divalent cation Ba2⫹ exhibited no inhibitory effect on the Iss and that the Pb2⫹mediated inhibition was voltage-dependent favor the presence of a binding site buried within the transmembrane domain of the channel protein. Ba2⫹ and Pb2⫹ both are divalent cations and thus should exhibit similar effects if inhibition is via charge screening. Changes in the membrane voltage occur mostly within the membrane domain but not in the extra- and intracellular solutions. An inhibition binding site buried within the membrane is consistent with the observed voltage-dependent inhibition by Pb2⫹. The reversible Pb2⫹ inhibition suggests that Pb2⫹ can dissociate from the binding site relatively easily. Hyperpolarization increased the association of Pb2⫹ to the binding site and decreased the dissociation of Pb2⫹ from the site, resulting

62

X.-Q. Dai et al. / Neuroscience 120 (2003) 57– 63

Fig. 4. Voltage-dependent effects of the Pb2⫹ inhibition of Ih. (A) Activation curves in the absence and presence of Pb2⫹ and fitted with the Boltzmann equation. Shown data were normalized by dividing by Imax obtained from the Boltzmann fit to the control data (in the absence of Pb2⫹). V1/2⫽⫺95.5⫾1.2 mV and k⫽15.1⫾1.1 for the control curve; V1/2⫽⫺104.8⫾3.2 mV and k⫽14.9⫾3.7 for the curve in the presence of Pb2⫹. Vertical bars indicate S.E. (B) Voltage-dependence of inhibitory effects of Pb2⫹. Relative amplitude was assessed by the ratio IPb/ICtrl where IPb and ICtrl are the Ih values at 10 and 0 ␮M Pb2⫹, respectively. (C) Voltage dependence of the activation time constant (␶) obtained in the absence of Pb2⫹ (control, E) and in the presence of 1 ␮M Pb2⫹ (F; n⫽12).

in increased inhibition affinity. However, because Pb2⫹ inhibits non-specifically many membrane protein activities, we should not exclude that part of the Pb2⫹ inhibition may be mediated through the charge screening effect. For example, the conceptual model developed by Woodhull (1973) associated changes in activation time constants with a general charge screening effect. These screening actions of Pb2⫹ might also contribute to a reduction in signal transfer between neurons and a decreased neuronal excitability

In conclusion, Pb2⫹ is a dose- and voltage-dependent, reversible blocker of Ih in DRG neurons.

REFERENCES Bal R, Oertel D (2000) Hyperpolarization-activated, mixed-cation current (Ih) in octopus cells of the mammalian cochlear nucleus. J Neurophysiol 84:806 –817. Brown HF, DiFrancesco D, Noble SJ (1979) How does adrenaline accelerate the heart? Nature 280:235–236.

X.-Q. Dai et al. / Neuroscience 120 (2003) 57– 63 Bu¨sselberg D, Platt B, Michael D, Carpenter DO, Haas HL (1994) Mammalian voltage-activated calcium channel currents are blocked by Pb2⫹, Zn2⫹, and Al3⫹. J Neurophysiol 71:1491–1497. Dai XQ, Ruan DY, Chen JT (2001) The effects of lead on transient outwards potassium channels in acutely isolated dorsal root ganglion neurons. Brain Res 904:327–340. Ghamari-Langroudi M, Bourque CW (2000) Excitatory role of the hyperpolarization-activated inward current in phasic and tonic firing of rat supraoptic neurons. J Neurosci 20:4855–4863. Martell AE, Smith GL (1982) Critical stability constants. Plenum: New York. Mayer ML, Westbrook GL (1983) A voltage-clamp analysis of inward (anomalous) rectification in mouse spinal sensory ganglion neurons. J Physiol 340:19 –45. McCormick DA, Pape HC (1990) Properties of a hyperpolarizationactivated cation current and its role in rhythmic oscillation in thalamic relay neurons. J Physiol 431:291–318. Pape HC (1996) Queer current and pacemaker: the hyperpolarizationactivated cation current in neurons. Annu Rev Physiol 58:299 –327. Santoro B, Chen S, Lu¨thi A, Pavlidis P, Shumyatsky GP, Tibbs GR, Siegelbaum AA (2000) Molecular and functional heterogeneity of

63

hyperpolarization-activated pacemaker channels in the mouse CNS. J Neurosci 20:5264 –5275. Satoh TO, Yamada M (2000) A bradycardiac agent ZD7288 blocks the hyperpolarization-activated current (I(h)) in retinal rod photoreceptors. Neuropharmacology 39:1284 –1291. Satoh TO, Yamada M (2001) Niflumic acid reduces the hyperpolarization-activated current (Ih) in rod photoreceptor cells. Neurosci Res 40:375–381. Schoenmakers TJM, Visser GJ, Flik G, Theuvenet APR (1992) Chelator: an improved method for computing metal ion concentrations in physiological solutions. Biotechniques 12:870 –897. Scroggs RS, Todorovic SM, Anderson EG, Fox AP (1994) Variation in Ih, Iir, and Ileak between acutely isolated adult rat dorsal root ganglion neurons of different size J. Neurophysiol 71:271–279. Wang Z, Van Den Berg DB, Ypey DL (1997) Hyperpolarization-activated currents in the growth cone and soma of neonatal rat dorsal root ganglion neurons in culture. J Neurophysiol 78:177–186. Woodhull AM (1973) Ionic blockage of sodium channels in nerve. J Gen Physiol 61:687–708. Yagi J, Sumino R (1998) Inhibition of a hyperpolarization-activated current by clonidine in rat dorsal root ganglion neurons. J Neurophysiol 80:1094 –1104.

(Accepted 25 March 2003)