K⁺ channels form the largest family of mammalian membrane ion channels and play a pivotal role in function or dysfunction of excitable cells (Hille, 2001). Thus, there is substantial interest in identifying specific modulators of these ion channels. Tertiapin, a 21-amino acid-containing peptide isolated from the venom of the European honey bee (Gauldie et al., 1976; Xu and Nelson, 1993), has been characterized as a specific and potent blocker of recombinant GIRK1+GIRK4 heteromultimers and ROMK1 at nanomolar affinity (Jin and Lu, 1998) and ROMK2 at micromolar range (Sackin et al., 2003). On the other hand, tertiapin has been reported to be less effective on some Kir channels; for example, inhibition of Kir2.1 channels by 1 μM tertiapin is less than 10% (Jin and Lu, 1998). Tertiapin-Q is a stable (nonoxidizable) and functionally similar derivative of tertiapin whose methionine residue 13 is replaced by a glutamine (Jin and Lu, 1999). The mechanism of tertiapin-Q inhibition of Kir channels is not well understood, but it is believed to block the K⁺ channel pore by occluding its α helix into the channel vestibule (Jin et al., 1999). Kirs are highly K⁺-selective ion channels that carry K⁺ from the extracellular to the intracellular compartment more efficiently than the opposite direction (Yamada et al., 1998). This inwardly rectifying property allows the channel to set the resting membrane potential toward the K⁺ equilibrium potential without preventing action potential generation (Yamada et al., 1998; Hille, 2001).

The action potential in mammalian ganglion neurons is terminated by fast inactivation of an inward Na⁺ current and also by activation of outward K⁺ currents, including transient outward (I_A), delayed rectifier (I_Kv), and Ca²⁺-dependent (I_KCa) K⁺ currents (Adams and Harper, 1995; Hille, 2001). These outward K⁺ currents also form the basis of the action potential afterhyperpolarization (AHP) and play important roles in regulating neuronal excitability, spike frequency adaptation, and neurotransmission (Adams and Harper, 1995; Hille, 2001; Faber and Sah, 2003). The K⁺ channels underlying the AHP are inhibited by compounds such as 4-aminopyridine and tetraethylammonium, which depolarize the neurons and increase action potential duration (Adams and Harper, 1995; Hille, 2001). Ca²⁺-dependent K⁺ currents are generated by different K⁺ channels including small (SK), intermediate (IK), and large (BK) conductance channels (Hille, 2001; Faber and Sah, 2003). Due to their steep voltage dependence, BK channels contribute to the onset of the AHP with their high single-channel conductance (>100 pS in symmetrical K⁺ solutions), resulting in substantial K⁺ efflux upon depolarization during the upstroke of the action potential and then close rapidly following return of the membrane potential to negative values (Hille, 2001; Faber and Sah, 2003). BK channels are formed by a tetramer of the principle pore-forming α subunit that may also interact with an auxiliary β subunit (Adelman et al., 1992; Cui et al., 1997; Meera et al., 2000; Hille, 2001). The α subunit is encoded by the slo or KCNMA1 gene, and four β subunits are encoded by KCNMB1-4 genes (Adelman et al., 1992; Jan and Jan, 1997; Brenner et al., 2000; Hille, 2001; Ha et al., 2004). The pore-forming region of the α subunit shares significant homology with the pore regions of other K⁺ channels (Jan and Jan, 1997). Extensive splice variation of α (over 100) and some β (e.g., β3a-d) subunit RNAs, species differences, and developmental regulation further diversifies BK channel function and determines its sensitivity to voltage and intracellular Ca²⁺ (Adelman et al., 1992; Ramanathan et al., 1999; Brenner et al., 2000; Hu et al., 2003; Ha et al., 2004). BK channel α subunit homomultimers and α + β heteromultimers are mostly blocked by tetraethylammonium, charybdotoxin, iberiotoxin, paxillin, and penitrem A, with exceptions such as α and β4 heteromultimers being insensitive to charybdotoxin and iberiotoxin (Meera et al., 2000; Faber and Sah, 2003). Only iberiotoxin and paxillin appear to be selective for BK channels tested (Meera et al., 2000; Faber and Sah, 2003).

The aim of the present study was to investigate the effects of tertiapin-Q on recombinant and endogenous BK channels in dorsal root ganglion (DRG) neurons, which are known to play a key role in the reception, transmission, and modulation of nociceptive information between the periphery and the central nervous system (Kanjhan, 1995). Here, we report that, in addition to its high-affinity inhibition of some Kir channels, prolonged application of tertiapin-Q at nanomolar concentration also results in use-dependent block of recombinant BK channels expressed in oocytes, as well as native K⁺ channels underlying the AHP in DRG neurons.

Materials and Methods

All animal experiments were performed in accordance with the guidelines of the University of Queensland Animal Ethics Committee.

Preparation of Xenopus Oocytes and Two-Electrode Voltage-Clamp. The procedures for preparation and recording from oocytes were modified from previous studies (Kanjhan et al., 2003). Adult female Xenopus laevis frogs were anesthetized by immersion in dechlorinated water containing 0.13% (w/v) methane sulfonate salt of 3-aminobenzoic acid ethyl ester (MS-222; MP Biomedicals, Irvine, CA). Ovarian lobes were excised and placed in frog Ringer's solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl₂, 1.0 mM MgCl₂, and 10 mM HEPES, pH 7.2 adjusted with NaOH; 285 mOsm). The oocytes were transferred to a 50-ml sterile falcon tube containing 15 ml of collagenase type 1A (2 mg/ml; Sigma-Aldrich, St. Louis, MO) in Ca²⁺-free Ringer's medium (82.5 mM NaCl, 2.0 mM KCl, 1.0 mM MgCl₂, and 5.0 mM HEPES, pH 7.5 with NaOH). The tube was placed on gentle rocker for 2 to 3 h until single oocytes became free. Separated oocytes were washed initially in Ca²⁺-free Ringer's and subsequently in ND96 (96 mM NaCl, 2.0 mM KCl, 1.0 mM MgCl₂, 1.8 mM CaCl₂, and 5.0 mM HEPES, pH 7.5 with NaOH) with 5 mM pyruvic acid (Sigma-Aldrich) and gentamicin (0.05 mg/ml; Serva, Heidelberg, Germany) five times each to remove collagenase. Mature oocytes (stages V and VI) free of any follicular remains were selected under a stereo microscope using a plastic Pasteur pipette. Selected oocytes were transferred to sterile dishes with Ringer's solution containing gentamicin sulfate and stored at 19°C for a minimum of 2 h prior to capped RNA (cRNA) injection.

Plasmid for the human BK channel α subunit (hSlo1) was obtained from Merck (Whitehouse Station, NJ). Plasmids for the mouse GIRK1 and GIRK2 channels were a kind gift from Dr. Michel Lazdunski (Institut de Pharmacologie Moleculaire et Cellulaire, Centre National de la Recherche Scientifique, Valbonne, France). cRNAs were transcribed using the mMessage mMachine T7 Ultra Kit (Ambion, Austin, TX) according to the manufacturer's instructions following plasmid template linearization by restriction digest. Synthesized message was purified using the MEGAclear (Ambion) kit. Quality and quantity of cRNA was determined by gel electrophoresis and spectrophotometry. Oocytes were injected with a 25-ng (25 μl) BK channel α subunit or GIRK1 and GIRK2 cRNAs. Control oocytes were injected with the same volume of sterile milliQ water (Millipore Corporation, Billerica, MA). Injected oocytes were incubated in frog Ringer's at 19°C for a minimum of 48 h prior to recording.

Depolarization-activated membrane currents were recorded from oocytes using two-electrode voltage-clamp virtual ground at a holding potential (V_H = -40 or -80 mV) controlled by a Gene Clamp 500B amplifier (Axon Instruments Inc., Union City, CA). The voltage and current recording microelectrodes were pulled with a horizontal puller (P-87; Sutter Instrument Company, Novato, CA) from filamented borosilicate glass capillaries (GC150TF-10; Harvard Apparatus Inc., Holliston, MA) and filled with 3 M KCl (r = 0.1-0.6 MΩ). Oocytes were placed in a gravity-fed continuous-flow chamber (0.2-ml volume) and superfused at a rate of 2 ml/min with frog Ringer's solution. In Mg²⁺ replacement experiments, extracellular Ca²⁺ was increased from 1.8 to 2.8 mM, reducing Mg²⁺ nominally to 0 mM. In Ca²⁺ replacement experiments, extracellular Mg²⁺ was increased from 1 to 2.8 mM, reducing Ca²⁺ nominally to 0 mM. All recordings were carried out at room temperature (21-23°C). Some oocytes were stimulated by voltage steps (100-ms duration; from -80 to +80 mV at 0.1 Hz) to show use-dependent inhibitory effect of tertiapin-Q. Electrophysiological data were displayed on a real time digital oscilloscope (Tektronix TDS 210), acquired at a sampling rate of 5 kHz, low-pass filtered at 2 kHz, and stored on a PC computer using a digitizer interface and software (Digidata 1322A and PClamp 8.2; Axon Instruments Inc.). The data were analyzed off-line using Clampfit 9.0 software (Axon Instruments Inc.).

Preparation and Culturing of Mouse DRG Neurons. DRG ganglia located in cervical to sacral levels were dissected from newborn [postnatal day 0 (P0)] C57BL/6 mice following decapitation, and cells were plated at low density (∼5000 cells/cm²), as previously described (Coulson et al., 1999). Cells were grown in Monomed II medium (CSL, Melbourne, Australia) with 1% fetal bovine serum in nerve growth factor (2.5S NGF; 50 ng/ml; Alomone Labs, Jerusalem, Israel) and allowed to adhere overnight onto poly-dl-ornithine (500 μg/ml)- and laminin (10 ng/ml)-coated glass coverslips at 37°C.

Whole-Cell Patch-Clamp of Mouse DRG Neurons. Coverslips containing cultured DRG neurons were placed in an external solution containing 150 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM d-glucose (pH 7.4 adjusted with NaOH; 315 ± 5 mOsm). A gravity-fed system was used for the exchange of extracellular solution (1 ml/min) in the bath at room temperature (21-23°C). Patch pipettes (1-3 MΩ; borosilicate glass capillaries; Vitrex Modulohm, Herlev, Denmark) were pulled with a two-stage vertical puller (Narashige PC-10) and filled with an internal solution containing 130 mM K⁺ gluconate or KCl, 10 mM HEPES, 11 mM EGTA, 1 mM CaCl₂, 3 mM ATP-Mg²⁺, and 0.3 mM GTP-Tris salt (pH 7.25 adjusted with KOH; 305 ± 5 mOsm). The cells were voltage-clamped at a holding potential (V_H) of -50 mV controlled by a patch-clamp amplifier (Axopatch 1D; Axon Instruments Inc.). Some DRG neurons were stimulated by voltage steps (100-ms duration; from -50 to +40 mV at 0.1 Hz) to show use-dependent inhibitory effect of tertiapin-Q. Action potentials were evoked by applying 5- to 10-ms-long depolarizing current steps at threshold under current clamp at -50 mV. The data were acquired at a sampling rate of 5 kHz, low-pass filtered at 2 kHz, and stored on a PC computer using pClamp 8.2 software (Axon Instruments Inc.) and an analog to digital interface (Digidata 1320A; Axon Instruments Inc.). Statistical analysis, concentration-response curves were fitted by nonlinear regression analysis using Prism 3.0 software (GraphPad Software Inc., San Diego, CA). The data are presented as the mean ± S.D., and significance was accepted at the two-tailed P < 0.05 using paired Student's t test.

Tertiapin-Q. A stable and functionally similar derivative of tertiapin peptide [tertiapin-Q; amino acid sequence ALCNCNRIIIPHQCWKKCGKK where glutamine (Q) replacing methionine (M) at position 13] was commercially synthesized (Auspep, Melbourne, Australia) as described previously (Jin and Lu, 1999). Tertiapin-Q was dissolved to required concentration in the extracellular solution prior to bath application onto oocytes or cultured DRG neurons.

Results

Recombinant BK Channels Are Inhibited by Tertiapin-Q in a Use-Dependent Manner.Xenopus oocytes expressing the α subunit of BK channels displayed large outwardly rectifying currents in response to voltage steps between -50 and +80 mV at 10-mV intervals from a V_H of -80 mV (Fig. 1, A and B; control). These outwardly rectifying currents were inhibited by bath-applied tertiapin-Q in a concentration-dependent manner (Fig. 1), with half-maximal inhibition (IC₅₀) obtained at 5.8 ± 1.0 nM (n = 3) and complete block at 100 nM [1.98 ± 0.13 μA in the absence (control) and 0.10 ± 0.06 μA in the presence of tertiapin-Q; n = 3; P < 0.001]. The concentration-response curve for tertiapin-Q inhibition of BK-mediated currents, evoked by a voltage step to +80 mV, is shown in Fig. 1C. The Hill coefficient was 0.8 ± 0.1 (n = 3). The concentration-dependent block by tertiapin-Q was use-dependent because it required a minimum of 15 min of continuous stimulation by voltage steps (100-ms duration; from -80 to +80 mV at 0.1 Hz) to obtain maximum inhibition. Oocytes incubated in 100 nM tertiapin-Q for 30 min without depolarizing voltage steps did not exhibit any significant reduction in outward current amplitude (2.87 ± 0.50 μA before and 2.67 ± 0.47 μA after tertiapin-Q, n = 3; P > 0.05; Fig. 2, A and C). A subsequent 15- to 20-min stimulation of the same oocytes produced complete block. The complete block of BK channels with high concentrations of tertiapin-Q (> 100 nM) did not exhibit any signs of recovery following up to 30 min of washout (n = 4); however, a small recovery (<15%) was observed after 90 min of washout (n = 2).

In water-injected control oocytes, tertiapin-Q (1 μM) had no significant effect on the endogenous depolarization-activated outward currents under identical conditions for >20 min (Fig. 2, B and D). The endogenous outward current amplitude in response to a depolarizing voltage step to +80 mV was 0.42 ± 0.06 μA before and 0.40 ± 0.06 μA after tertiapin-Q application (n = 5; P > 0.05). Therefore, tertiapin-Q exhibited no significant effect on native defolliculated oocyte depolarization-activated currents, including Ca²⁺-activated chloride currents (Kuruma et al., 2000).

Recombinant GIRK1+GIRK2 Channels Are Blocked by Tertiapin-Q. Recombinant GIRK1 and GIRK2 channels coexpressed in Xenopus oocytes were also inhibited by tertiapin-Q (100 nM), under identical experimental conditions used above (Fig. 3). The maximum currents evoked by a voltage step to -160 mV from a V_H of -40 mV was -1.90 ± 0.26 μA in the absence (control) and -0.61 ± 0.09 μA in the presence of 100 nM tertiapin-Q for 15 min (n = 4; P < 0.001). The remaining inward current after exposure to tertiapin-Q is likely to be due to activation of endogenous hyperpolarization-activated nonselective cation and Ca²⁺-activated chloride currents described previously in defolliculated Xenopus oocytes (Kuruma et al., 2000). In control (noninjected) oocytes, 1 μM tertiapin-Q did not significantly affect currents activated by a hyperpolarizing voltage step to -200 mV from a holding potential of -40 mV (-0.59 ± 0.07 μA before and -0.58 ± 0.11 μA after tertiapin-Q, n = 5; P > 0.5) (Fig. 4). The endogenous hyperpolarization-activated currents in oocytes were sensitive to extracellular divalent cations, significantly blocked by 10 μM Zn²⁺ (66% block; n = 5; P < 0.001; Fig. 4) and 1 mM Ba²⁺ (45% block, n = 5; P < 0.001), and significantly potentiated by increasing extracellular Ca²⁺ or Mg²⁺ to 2.8 mM by replacing each other (920% increase by Ca²⁺ and 800% increase by Mg²⁺; n = 5 each; P < 0.001 for both) (Fig. 4B).

Tertiapin-Q Has No Effect on Hyperpolarization-Activated Inward Currents in Newborn Mouse DRG Neurons. Whole-cell recordings were made from a total of 43 cultured DRG neurons, including small-, medium-, and large-diameter cells (capacitance, mean ± S.D., 13 ± 5 pF; range 6-25 pF). These neurons had a resting membrane potential of -54 ± 5 mV (mean ± S.D.; range, -47 to -66 mV). Under voltage-clamp, the majority of DRG neurons (35 of 43, ∼81%) with small to medium cell bodies (12 ± 3 pF; 6-19 pF) exhibited small inward currents, without any inward rectification, in response to hyperpolarizing voltage steps to -120 mV at 20-mV intervals from a V_H of -50 mV (Fig. 5, Ai and Bi). The amplitude of these currents measured at their maximum at -120 mV was -0.12 ± 0.04 nA (range from -0.06 to -0.19 nA) and was not significantly altered by intracellular gluconate added to the internal pipette solution (Fig. 5, Aii and Bii), by 5- to 20-min bath application of 100 nM tertiapin-Q [-0.13 ± 0.04 nA in the absence (control) and -0.13 ± 0.04 nA in the presence of tertiapin-Q, n = 11; P > 0.05], 10 μM baclofen (-0.11 ± 0.04 nA before and -0.11 ± 0.04 nA after baclofen, n = 9; P > 0.5), or 1 mM Ba²⁺ (-0.10 ± 0.03 nA before and -0.11 ± 0.03 nA after Ba²⁺, n = 6; P > 0.05) (Figs. 5, C and D, and 6). Hyperpolarization-activated currents were not significantly affected by 100 μM Zn²⁺ (-0.15 ± 0.01 nA before and -0.15 ± 0.03 nA after Zn²⁺, n = 5; P > 0.05) (Fig. 6C). Bath application of carbachol (10 μM) also did not affect the hyperpolarization-activated currents in DRG neurons (n = 2; not shown). A minority of DRG cells (8 of 43 cells; ∼19%), which had a large diameter (capacitance 20 ± 3 pF; range 17-25 pF), exhibited characteristic signs of the hyperpolarization-activated cation (I_H) current, which activated slowly with strong inward rectification (Fig. 5, Aiii, Biii, Ci, and Di) and was sensitive to intracellular gluconate (130 mM; Figs. 5, Aiv and Biv, and 6C) or extracellular Cs⁺ (1 mM; Figs. 5, Ciii and Diii, and 6C), but were insensitive to 5- to 20-min bath application of 10 μM baclofen, 100 nM tertiapin-Q, 1 mM Ba²⁺, or 100 μM Zn²⁺ (Figs. 5, Cii and Dii, and 6, A and C).

Tertiapin-Q Blocks Outward Currents and Action Potential AHP in Mouse DRG Neurons in a Use-Dependent Manner. In contrast, bath application of tertiapin-Q (10-100 nM), combined with continuous stimulation by voltage steps from -50 to +40 mV at 0.1 Hz, inhibited outward currents in mouse DRG neurons in a time-dependent manner, attaining maximum inhibition by ∼20 min (Fig. 7, A, B, and D). The relative inhibition compared with maximum inhibition of outward currents evoked by a voltage step to +80 mV was 13, 40, 84, and 98% at 5, 10, 15, and 20 min, respectively (Fig. 7B). The relative inhibition at these time intervals was statistically significant compared with control, with most significant inhibition observed at 15 min compared with 10 min (P < 0.005; n = 4; Fig. 7B).

Tertiapin-Q inhibition of outward currents was voltage-dependent, with half-maximal inhibition obtained at -12 ± 4 mV (n ≥ 4; Fig. 7C). Test depolarization to +80 mV from V_H -50 mV evoked outward currents that were inhibited by 66.8 ± 4.6% compared with controls following 20-min application of 100 nM tertiapin-Q (before 2.73 ± 0.22 nA and after 0.91 ± 0.20 nA at the peak amplitude, n = 4; P < 0.001). Peak outward currents evoked by a depolarizing step to +20 mV from V_H -50 mV were inhibited by 60.8 ± 6.8% (before 0.86 ± 0.16 nA and 0.34 ± 0.09 nA after 20 min tertiapin-Q), which did not differ significantly from the inhibition of outward currents evoked by a depolarizing step to +80 mV (P > 0.05; n = 7; Fig. 7C). The relative inhibition obtained at 0 mV was significantly reduced to 43.6 ± 5.6% (before 0.44 ± 0.10 nA and 0.25 ± 0.07 nA after 20 min tertiapin-Q; P < 0.001 compared with +20 mV; n = 7; Fig. 7C).

The tertiapin-Q inhibition of outward currents was dependent on the stimulation voltage (Fig. 7D). When the neurons were stimulated with depolarizing pulses to 0 mV at 0.1 Hz, the time to maximum inhibition was significantly increased compared with more depolarized stimulation steps (P < 0.05 compared with +20 mV; n = 4). However, there was no significant difference in the time to maximum inhibition of outward currents by tertiapin-Q when the cells were stimulated with depolarizing steps to +20, +40, and +80 mV (P > 0.05; n = 4; Fig. 7D). The inhibition by tertiapin-Q was also dependent on the frequency of stimulation (Fig. 7D). Stimulation at 0.01 Hz with a voltage step from -50 to +40 mV significantly increased the time to reach maximum inhibition when compared with stimulation at 0.1 Hz (P < 0.001; n = 4). However, there was no significant difference between 0.1 and 1 Hz stimulation (P > 0.05; n = 4).

Under control conditions, DRG neurons current-clamped at -50 mV and stimulated with a brief depolarizing pulse exhibited action potentials of 76 ± 8 mV amplitude, followed by a small amplitude AHP (-5.8 ± 1.9 mV) with a half-width of 224 ± 101 ms (n = 8). Tertiapin-Q (100 nM) increased action potential duration and inhibited the AHP amplitude (Fig. 7, E and F). The half-width of action potential duration was 4.5 ± 1.1 ms in the absence (control) and significantly increased to 15.9 ± 6.2 ms in the presence (20 min) of 100 nM tertiapin-Q (n = 8; P < 0.001). The AHP was inhibited completely following 15- to 20-min bath application of 100 nM tertiapin-Q. AHP inhibition by low concentrations of tertiapin-Q (10 nM) exhibited minimal recovery following >60-min washout. However, extended treatment with tertiapin-Q for >25 min at high concentrations (>100 nM) was toxic to the DRG neurons continuously stimulated with depolarized voltage steps. The I-V relationship became linear due to a significant reduction in outward currents after >25-min exposure, and action potentials could no longer be evoked by membrane depolarization.

The inhibition by tertiapin-Q of the AHP in native DRG neurons was also use-dependent (Fig. 7, E and F). Tertiapin-Q block of action potential AHP currents was observed only in voltage-clamped neurons that were stimulated repeatedly by depolarizing voltage steps (to +40 mV at 0.1 Hz) and required a minimum of 15 min of exposure to tertiapin-Q to obtain maximum inhibition (compare Fig. 7 with Fig. 8). Other nonclamped DRG neurons in the recording chamber exposed to 100 nM tertiapin-Q for up to 120 min were not affected (Fig. 8, A-C), exhibiting an AHP of -6.2 ± 1.7-mV amplitude (n = 6; P > 0.5 unpaired Student's t test) and action potential half-width duration of 4.6 ± 1.5 ms (n = 6; P > 0.5, unpaired Student's t test), similar to control group of DRG neurons that were not exposed to tertiapin-Q (Fig. 8, A, D, and E). These results suggest that the blocking effects of tertiapin-Q are use-dependent and that the block of the channel occurs slowly when the pore of the channel is open.

Discussion

The present study demonstrates that tertiapin-Q inhibits recombinant and native BK-type K⁺ channels in a use-, concentration-, and voltage-dependent manner, in addition to inhibiting recombinant Kir channels made of GIRK1+GIRK2 heteromultimers. Therefore, the blocking actions of tertiapin-Q are not specific to recombinant GIRK1+ GIRK4 heteromultimeric and ROMK1 and ROMK2 homomultimeric channels as previously proposed (Jin and Lu, 1998, 1999; Kitamura et al., 2000; Sackin et al., 2003). Although the inhibition of GIRK and BK channels occur within a similar nanomolar range, the mechanisms of tertiapin-Q inhibition appear to be different among K⁺ channels. Inhibition of GIRK channels by tertiapin occurs within a minute and does not exhibit rapid recovery (Kitamura et al., 2000), whereas tertiapin-Q block of ROMK channels can take up to 10 min, and it is reversible with an overshoot of the outward conductance during washout (Sackin et al., 2003). In contrast, inhibition of BK channels is use- and voltage-dependent, requiring >15-min stimulation by depolarizing steps, and the recovery is slow. Therefore, it is likely that different and complex interactions, determining the relative permeability or block of the channel, occur during binding and dissociation of the tertiapin-Q peptide with GIRK, ROMK, and BK channels. A previous study has shown that tertiapin (10 μM) can inhibit Ca²⁺ binding to calmodulin (Dudkin et al., 1983). However, tertiapin-Q is a charged peptide and unlikely to cross the membrane to inhibit Ca²⁺ binding to BK channels. The present study demonstrates that a relatively slow block of BK channels by nM tertiapin-Q occurs only if the pore of the channel is open. Furthermore, tertiapin-Q block of outward currents in DRG neurons has been shown to be dependent on the stimulus voltage and frequency.

Tertiapin-Q inhibition of outward currents, AHP, and increase of action potential duration in DRG neurons suggests inhibition of BK channels shown to be present in DRG neurons using the inhibitors charybdotoxin and iberiotoxin (Scholz et al., 1998; Zhang et al., 2003). The main functions of BK channels in DRG neurons include shortening of action potential duration, enhancing speed of repolarization, and contributing to the fast AHP, leading to reduced repetitive firing (Scholz et al., 1998). However, complete blockade of the action potential AHP (Fig. 7F) by tertiapin-Q suggests inhibition of other K⁺ channels, such as Ca²⁺-activated voltage-insensitive SK and IK channels identified in DRG neurons contributing to AHP (Gold et al., 1996a,b; Mongan et al., 2005), given that iberiotoxin-sensitive BK channels contribute only to the fast component of the AHP (Scholz et al., 1998; Zhang et al., 2003), and apamin-sensitive SK channels constitute most of the remaining AHP in DRG neurons (Gold et al., 1996b).

Both tertiapin and apamin are short peptides isolated from honey bee venom with structural similarities such as disulfide bonds stabilizing their folding and an α-helical conformation adopted by their C termini (Xu and Nelson, 1993). Tertiapin also appears to act at least partially similarly to the scorpion venom toxin Lq2, which is known to inhibit ROMK1 as well as voltage- and Ca²⁺-dependent K⁺ channels (Lu and MacKinnon, 1997). Previous studies of tertiapin actions on native cells, often applied acutely, may have overlooked block of the BK channels by tertiapin due to its use dependence. For example, ATP-sensitive K⁺, voltage-dependent K⁺, and L-type Ca²⁺ channels in rabbit cardiac myocytes have been reported not to be effected by acute tertiapin at concentrations up to 1 μM (Kitamura et al., 2000). However, tertiapin potently blocks K_ACh formed by GIRK1+GIRK4 heteromultimers in myocytes, in a voltage-independent manner with an IC₅₀ of 8 nM (Kitamura et al., 2000).

There are over 100 isoforms of the α and β subunits producing numerous structurally and functionally distinct BK channels (e.g., variation in Ca²⁺ and voltage dependence), which are further diversified by species differences and developmental regulation (Adelman et al., 1992; Ramanathan et al., 1999; Brenner et al., 2000; Hu et al., 2003; Ha et al., 2004). However, we do not know which α and/or β isoforms form BK channels in DRG neurons, but the diversity seen in sensory cells involved in tuning of the ear to different sound frequencies (Ramanathan et al., 1999) suggests extensive heteromultimerization among DRG neurons with varying sensory functions. The local concentration of intracellular Ca²⁺ at microdomains close to BK channels in neurons is known to transiently increase up to tens of micromoles during action potentials (Regehr and Tank, 1992). Present results demonstrate that tertiapin-Q inhibition of DRG outward currents is voltage-dependent but similar between +20 and +80 mV (Fig. 7C). Therefore, DRG AHP currents inhibited by tertiapin-Q (Fig. 7F), potentially including BK and other Ca²⁺-activated K⁺ channels (i.e., SK and IK), are mostly active at +20 mV to allow physiological termination of action potential and concurrent activation of AHP currents to maintain and regulate neuronal excitability.

We demonstrate that tertiapin-Q (1 μM) has no effect on endogenous hyperpolarization- and depolarization-activated currents in defolliculated Xenopus oocytes, including Ca²⁺-activated chloride and hyperpolarization-activated nonselective cation currents (Kuruma et al., 2000). We also show that hyperpolarization-activated currents in oocytes are sensitive to extracellular divalent cations whereby they are inhibited by 10 μM Zn²⁺ and 1 mM Ba²⁺ but potentiated 8- to 9-fold by raising extracellular Ca²⁺ or Mg²⁺ by replacing each other, suggesting a strong negative interaction between Ca²⁺ and Mg²⁺ in control of opening of the channel. Therefore, tertiapin-Q can be safely used in oocytes, and Zn²⁺ blockade of endogenous hyperpolarization-activated currents may be useful in expression studies of recombinant Kir channels in Xenopus oocytes.

Inhibition of recombinant GIRK1+GIRK2 channels by tertiapin-Q reported here is consistent with previous studies comparing effects of tertiapin-Q between wild type and mice deficient in GIRK1 and GIRK2 (Marker et al., 2004). A functional and physical interaction between the GIRK1 and GIRK2 has been suggested and unlike GIRK1 and GIRK2 homomultimers, GIRK1 and GIRK2 coexpression results in large amplitude currents, indicating that heteromultimeric assembly is necessary for activity (Kofuji et al., 1995). Tertiapin-Q inhibition of GIRK1 and GIRK2 channels involving opioidergic signaling pathways in the dorsal horn of the spinal cord and in peripheral sensory endings suggests a potential modulatory role for tertiapin-Q in nociception (Marker et al., 2004).

The majority of small- and medium-sized adult rat DRG neurons exhibit Kir currents (Scroggs et al., 1994), and Kir2.2, Kir3.1, Kir3.2, and Kir3.3 mRNAs have been localized in rat DRG neurons at embryonic day 17 (Karschin and Karschin, 1997). Kir2.x currents in native cells are known to be rapidly activated and then slightly decreased by voltage steps to potentials more negative than -70 mV (Yamada et al., 1998). However, apart from a subpopulation of neurons exhibiting gluconate- and Cs⁺-sensitive I_H current, the present study found no evidence for inwardly rectifying currents in P0 mouse DRG neurons in response to hyperpolarizing voltage steps, either before or after treatment with GIRK agonists baclofen and carbachol (Yamada et al., 1998). Furthermore, GIRK inhibitors tertiapin-Q and Ba²⁺ had no effect on hyperpolarization-activated currents in DRG neurons. Consistent with our results, DRG precursors cultured from chicks at embryonic day 6 exhibit no detectable inward K⁺ channel activity, whereas outward K⁺ currents dominate (Gottmann et al., 1988). K_ACh channels (e.g., Kir3.1+Kir3.4 heteromultimers), with a slightly weaker inward rectification than Kir2.x, exhibit a low basal activity and are activated by acetylcholine (Yamada et al., 1998). However, carbachol (10 μM) had no effect on hyperpolarization-activated currents in DRG neurons tested. Therefore, it is unlikely that Kir2.x and Kir3.x (i.e., GIRK) channels play a significant role in modulating excitability of P0 DRG neurons. We hypothesize that functional expression of Kir channels may increase with exposure to sensory or painful stimuli during postnatal development, analogous to other sensory systems (see Geleoc et al., 2004). On the other hand, our identification of I_H current in a subpopulation of mouse DRG neurons with large diameter is consistent with data from the newborn rat, where it contributes to modulation of excitability as well as neurite outgrowth (Wang et al., 1997).

Although s.c. injection of bee venom has been reported to evoke tonic pain and hyperalgesia under normal conditions, beneficial effects of whole bee venom on chronic arthritis and inflammation have been long known and traditionally used in Oriental medicine (Lee et al., 2001; Kang et al., 2002). The bee venom peptides or associated substances responsible for these paradoxical effects have not been identified; the present data suggest tertiapin as a potential candidate. The hyperalgesic effects of bee venom under physiological conditions may be attributable to increased excitability as a result of depolarization (<10 mV) caused by partial block of action potential AHP currents in DRG neurons by tertiapin. In contrast, antinociceptive and anti-inflammatory effects of bee venom under pathologic conditions may be due to prolonged activity-dependent blockade of action potential AHP currents by tertiapin. The membrane depolarization that is sufficient to inactivate voltage-dependent Na⁺ channels in DRG neurons could produce so-called depolarization block of action potential firing.

In conclusion, inhibition of Kir and BK-type K⁺ channels and modulation of neuronal action potential firing by tertiapin-Q suggests a potential use in the treatment of inflammatory and persistent pain such as rheumatoid arthritis.