Analgesic α-Conotoxins Vc1.1 and Rg1A Inhibit N-Type Calcium Channels in Rat Sensory Neurons via GABAB Receptor Activation

α-Conotoxins Vc1.1 and Rg1A are peptides from the venom of marine Conus snails that are currently in development as a treatment for neuropathic pain. Here we report that the α9α10 nicotinic acetylcholine receptor-selective conotoxins Vc1.1 and Rg1A potently and selectively inhibit high-voltage-activated (HVA) calcium channel currents in dissociated DRG neurons in a concentration-dependent manner. The post-translationally modified peptides vc1a and [P6O]Vc1.1 were inactive, as were all other α-conotoxins tested. Vc1.1 inhibited the ω-conotoxin-sensitive HVA currents in DRG neurons but not those recorded from Xenopus oocytes expressing CaV2.2, CaV2.1, CaV2.3, or CaV1.2 channels. Inhibition of HVA currents by Vc1.1 was not reversed by depolarizing prepulses but was abolished by pertussis toxin (PTX), intracellular GDPβS, or a selective inhibitor of pp60c-src tyrosine kinase. These data indicate that Vc1.1 does not interact with N-type calcium channels directly but inhibits them via a voltage-independent mechanism involving a PTX-sensitive, G-protein-coupled receptor. Preincubation with a variety of selective receptor antagonists demonstrated that only the GABAB receptor antagonists, [S-(R*,R*)][-3-[[1-(3,4-dichlorophenyl)ethyl]amino]-2-hydroxy propyl]([3,4]-cyclohexylmethyl) phosphinic acid hydrochloride (2S)-3[[(1S)-1-(3,4-dichlorophenyl)-ethyl]amino-2-hydroxypropyl](phenylmethyl) phosphinic acid and phaclofen, blocked the effect of Vc1.1 and Rg1A on Ca2+ channel currents. Together, the results identify CaV2.2 as a target of Vc1.1 and Rg1A, potentially mediating their analgesic actions. We propose a novel mechanism by which α-conotoxins Vc1.1 and Rg1A modulate native N-type (CaV2.2) Ca2+ channel currents, namely acting as agonists via G-protein-coupled GABAB receptors.


Introduction
Conotoxins are small peptides derived from the venom of the Conus genus of predatory marine snails (Terlau and Olivera, 2004). Many conotoxins are highly selective antagonists of a diverse range of mammalian ion channels and receptors associated with pain pathways. Thus, conotoxins are attractive as drug leads for treating neuropathic pain. Previous studies have demonstrated the analgesic potential of several different classes of conotoxins Livett et al., 2004) and, most recently, antagonists of neuronal nicotinic acetylcholine receptors (nAChRs) have begun to show promise for the treatment of neuropathic pain (Sandall et al., 2003;Lang et al., 2005;Satkunanathan et al., 2005;Vincler et al., 2006). The ␣-conotoxins competitively inhibit neuronal-type nAChRs (Dutton and Craik, 2001) with varying degrees of selectivity for different nAChR subtypes (McIntosh et al., 1999).
␣-Conotoxin Vc1.1, which was first discovered using a PCR screen of cDNAs from the venom ducts of Conus victoriae (Sandall et al., 2003), contains two residues, Pro6 and Glu14, that are post-translationally modified to hydroxyproline and ␥-carboxyglutamate, respectively, in the native peptide, designated vc1a (Jakubowski et al., 2004). Synthetic Vc1.1 lacking these two modifications is a competitive antagonist of neuronal nAChRs in bovine adrenal chromaffin cells (Clark et al., 2006) and is potent at recombinant ␣9␣10 nAChRs expressed in Xenopus oocytes (Vincler et al., 2006;Nevin et al., 2007). Furthermore, Vc1.1 and its post-translationally modified analogs vc1a, [P6O]Vc1.1, and [E14␥]Vc1.1 have been reported to be equally potent at inhibiting ACh-evoked currents mediated by ␣9␣10 nAChRs . Vc1.1 also antagonizes the nicotineinduced increase in axonal excitability of unmyelinated C-fiber axons in isolated segments of peripheral human nerves (Lang et al., 2005). Blockade of nAChRs on unmyelinated peripheral nerve fibers may have an analgesic effect on unmyelinated sympathetic and/or sensory axons. Interestingly, synthetic vc1a, which incorporates the modified residues, has been reported to inhibit ␣9␣10  but not ␣3-containing (Clark et al., 2006) nAChRs expressed in Xenopus oocytes and is inactive in two rat neuropathic pain assays (Livett et al., 2006). In contrast, Vc1.1 alleviates neuropathic pain in three rat models of human neuropathic pain and accelerates the functional recovery of injured neurons (Satkunanathan et al., 2005). ␣-Conotoxin Rg1A from Conus regius, which also selectively inhibits ␣9␣10 nAChRs , has similarly been shown to significantly reduce chronic constriction nerve injury-induced hyperalgesia in rats (Vincler et al., 2006;Vincler and McIntosh, 2007). Thus, although Vc1.1 is currently under development as a treatment for neuropathic pain (Livett et al., , 2006, its specific receptor target has not been unequivocally identified. We show here that Vc1.1 and Rg1A act by an unexpected mechanism not reported previously for any conotoxin.

Materials and Methods
DRG neuron preparation. DRG neurons were enzymatically dissociated from ganglia of 2-to 30-d-old Wistar rats according to standard protocols. Rats were killed by cervical dislocation as approved by the University of Queensland Animal Ethics Committee, the spinal column was hemi-segmented, and the spinal cord was removed. Ganglia were removed and rinsed in cold HBSS (MultiCell). They were minced and incubated in 1 mg/ml collagenase (type 2; 405 U/mg) (Worthington Biochemical) in HBSS at 37°C for ϳ30 min. After incubation, ganglia were rinsed three times with warm (37°C) DMEM (Invitrogen) supplemented with 10% fetal calf serum and 1% penicillin/streptomycin, and gently triturated with a fire-polished Pasteur pipette. Cells were plated on glass coverslips, incubated at 37°C in 95% O 2 /5% CO 2 , and used within 4 -48 h.
Electrophysiological methods. The external recording solution for rat DRG neurons contained the following (in mM): 150 tetraethylammonium (TEA)-Cl, 2 BaCl 2 , 10 D-glucose, and 10 HEPES, pH 7.3-7.4. Recording electrodes were filled with an internal solution containing the following (in mM): 140 CsCl, 1 MgCl 2 , 5 MgATP, 0.1 Na-GTP, 5 bis(2aminophenoxy)ethane-N, N,NЈ,NЈ-tetra-acetic acid (BAPTA)-Cs 4 , and 10 HEPES, pH 7.3-7.4, with CsOH, and had resistances of 1-2.5 M⍀. Membrane currents were recorded using the whole-cell configuration of the patch-clamp technique with an Axopatch 200B amplifier (Molecular Devices). Neurons were voltage clamped at a holding potential of Ϫ90 mV and Ba 2ϩ currents were elicited by step depolarizations, first to Ϫ40 mV for 200 ms [for low-voltage-activated (LVA) currents] and then to Ϫ10 mV, from a holding potential of Ϫ50 mV, for 200 ms [for highvoltage-activated (HVA) currents] at 10 s intervals. In a series of experiments, BaCl 2 was replaced by CaCl 2 . A voltage protocol using step depolarizations from Ϫ70 to 0 mV was used when examining HVA currents alone. Test potentials 150 ms in duration were applied every 20 s. To construct current-voltage ( I-V) relationships, neurons were voltage clamped at Ϫ100 mV and voltage stepped from Ϫ80 mV to ϩ50 mV in 10 mV increments. Voltage-dependent inhibition was assessed by applying a prepulse to ϩ80 mV for 20 ms in duration before the test pulse. Leak and capacitative currents were subtracted using a ϪP/4 pulse protocol. Currents were generated by a computer using pClamp 9.2 software (Molecular Devices), filtered at 2 kHz, and sampled at 8 kHz by the Digidata 1322A (Molecular Devices). Sampled data were stored digitally on a computer for further analysis.
Before recording, oocytes were injected with 25 nl of 50 mM BAPTA to eliminate the endogenous Ca 2ϩ -activated chloride conductance. Depolarization-activated Ca 2ϩ channel currents were recorded using Ba 2ϩ as a charge carrier and a two-electrode virtual ground voltageclamp circuit with a GeneClamp 500B amplifier (Molecular Devices). The external bath solution contained the following (in mM): 85 TEA-OH, 5 BaCl 2 , 5 KCl, and 10 HEPES, pH 7.4, with methanesulfonic acid. During recording, oocytes were constantly perfused with bath solution using a gravity-fed perfusion system at a flow rate of ϳ1.5 ml/min. Electrodes were filled with a solution containing 3 M KCl (0.5-1.0 M⍀). Oocytes were voltage clamped at a holding potential of Ϫ80 mV and membrane currents were elicited by 200 ms step depolarization to 0 mV (for Ca V 2.2) or ϩ20 mV (for Ca V 2.1) applied every 10 s. To construct I-V relationships, cells were voltage-clamped at Ϫ80 mV and then stepped from Ϫ80 mV to ϩ50 mV in 10 mV increments. Data were low-pass filtered at 2 kHz, digitized at 10 kHz, leak subtracted online using a ϪP/4 protocol, and analyzed off-line.
Data analysis. Data were analyzed off-line using pClamp 9.2 software (Molecular Devices). Peak Ba 2ϩ current amplitude in response to a depolarizing pulse was calculated and plotted against time to monitor stability of the recording. Generally, Ba 2ϩ current reached steady state within 2-5 min of initial recording, and any recordings that did not reach steady state were discarded. Drugs were applied via perfusion in the bath solution. Current amplitudes obtained in the presence of the drug were normalized by dividing by the current amplitude obtained under control conditions. Concentration-response relationships were obtained using GraphPad Prism software by plotting normalized current amplitude as a function of drug concentration and were fitted using a logistic equation. Numerical data are presented as mean Ϯ SEM. , and pp60c-src peptide (521-533) corresponding to its C-terminal regulatory domain were purchased from Tocris Bioscience. The peptide binds pp60c-src at the Src homology 2 domain suppressing its tyrosine kinase activity (Roussel et al., 1991). Phaclofen (3-amino-2-(4-chlorophenyl) propylphosphonic acid was from BIOMOL. All other drugs were obtained from Sigma-Aldrich and were prepared as concentrated stock solutions which were then diluted in bath solution and applied via perfusion. The -conotoxins CVID and MVIIA were prepared as described previously (Lewis et al., 2000). ␣-Conotoxins Vc1.1, [P6O]Vc1.1, [E14␥]Vc1.1, vc1a, RgIA, ImI, MII, BuIA and [A10L]PnIA, were synthesized as reported previously (Gehrmann et al., 1999;Hogg et al., 1999;Blanchfield et al., 2003;Clark et al., 2006Clark et al., , 2008Jin et al., 2007) and provided as a stock concentration in H 2 O of ϳ1 mM and used at a final concentration of 0.1 nM to 1 M. Reduced Vc1.1 was prepared by incubation with dithiothreitol (2 mM; 1 h; room temperature) or alkylated with iodoacetamide in 0.1 M ammonium acetate buffer.

␣9␣10 nAChR-selective conotoxins, Vc1.1 and Rg1A, inhibit voltage-gated Ca 2؉ channel currents in rat DRG neurons
In an attempt to account for ␣-conotoxin Vc1.1-induced analgesia and to elucidate the receptor target and mechanism of action, the effects of the conotoxins, Vc1.1 and RgIA, were investigated on voltage-gated Ca 2ϩ channel currents in rat DRG neurons. Whole-cell Ca 2ϩ channel currents of rat DRG neurons are mediated by both LVA and HVA calcium channels (Scroggs and Fox, 1992). The effect of Vc1.1 on Ca 2ϩ channel currents recorded from neonatal (2-to 10-d-old) rat DRG neurons was initially assessed using a two-step protocol (see Materials and Methods). Vc1.1 (1 M) had no effect on LVA Ca 2ϩ channel current amplitude elicited by voltage step from Ϫ90 to Ϫ40 mV (97 Ϯ 12% of control, n ϭ 10). In the same cells, Vc1.1 inhibited HVA Ca 2ϩ channel currents elicited by a 150 ms voltage step to Ϫ10 mV to 56.5 Ϯ 4.4% of control (n ϭ 10) ( Fig. 1 A, B). Vc1.1 only inhibited VGCC currents evoked by step depolarization positive of Ϫ40 mV, demonstrating the lack of effect of Vc1.1 on LVA Ca 2ϩ channel currents (Fig. 1C). The effect of Vc1.1 on HVA Ca 2ϩ channel currents was therefore assessed using a 150 ms step depolarization from a holding potential of either Ϫ70 or Ϫ90 mV to the test potential of either Ϫ10 or 0 mV at a frequency of 0.05 Hz, with 2 mM Ba 2ϩ as the charge carrier. Vc1.1 inhibition of HVA Ca 2ϩ channel currents developed with a delay of ϳ2 min and maximum inhibition occurred ϳ7 min after the beginning of the response, independent of the concentration of Vc1.1 (n ϭ 74). Inhibition of HVA Ca 2ϩ channel currents by Vc1.1 was concentration dependent and the mean concentration-response data gave an IC 50 of 1.7 nM, with maximum inhibition to 57.8 Ϯ 3.8% of control occurring at 1 M (n ϭ 25) ( Fig. 1 D).
Vc1.1 inhibition of VGCC currents was not reversed during washout. However, reapplication of Vc1.1 produced an equivalent inhibition of HVA Ca 2ϩ channel currents in other DRG neurons on the same coverslip (n ϭ 5 sets), suggesting that depolarization-induced Ca 2ϩ channel activation is required for inhibition by Vc1.1.
To further assess the dependence of Vc1.1 inhibition on activation of HVA Ca 2ϩ channel currents, we varied the frequency and duration of the depolarizing pulse in the presence of Vc1.1 (Fig. 2). When the depolarization pulse was shortened to 15 ms with the frequency maintained at 0.05 Hz, 100 nM Vc1.1 inhibited the current to only 88.4 Ϯ 3.3% of control in all cells tested (n ϭ 24). However, when the frequency of the 15 ms pulse was increased to 0.1 Hz, the current was reduced to 51.1 Ϯ 5.6% of control in all cells tested (n ϭ 7) ( Fig. 2 A, B). Figure 2 B also shows the relative inhibition of Ba 2ϩ currents by Vc1.1, using the standard depolarizing pulse duration (150 ms) at both 0.05 Hz and 0.1 Hz. The action of Vc1.1 does not depend on a rise in intracellular divalent cation, because the inhibition was also observed with 2 mM Ca 2ϩ as the charge carrier. In 9 of 12 cells, Vc1.1 (100 nM) inhibited the HVA Ca 2ϩ current amplitude to 61.4 Ϯ 5.2% of control and was not reversible on washout (Ͼ30 min) (Fig. 2C). Furthermore, in fura-2-loaded DRG neurons, bath application of 1 M Vc1.1 alone failed to elicit any increase in intracellular Ca 2ϩ concentration (n ϭ 14, data not shown).

Post-translationally modified analogs of Vc1.1 and other ␣conotoxins do not inhibit HVA Ca 2؉ channel currents in DRG neurons
The effects of post-translationally modified analogs of Vc1.1 were examined on HVA Ca 2ϩ channel currents from rat DRG neurons under the same conditions as Vc1.1. The native peptide, vc1a, did not inhibit HVA Ca 2ϩ channel currents (100.0 Ϯ 4.8% of control, n ϭ 9); however, the same cells responded to Vc1.1 after washout of vc1a (Fig. 3 B, C).

Vc1.1 inhibits the N-type component of HVA Ca 2؉ channel currents in DRG neurons
The selective N-type Ca 2ϩ channel blockers, -conotoxins CVID and MVIIA, were applied to determine which component of the whole-cell HVA Ca 2ϩ channel currents Vc1.1 inhibited in DRG neurons. Bath application of 200 nM CVID reduced peak Ca 2ϩ channel current amplitude to 54.4 Ϯ 4.8% of control (n ϭ 6). Application of 1 M Vc1.1 in the presence of CVID (200 nM) produced no further reduction of the HVA Ca 2ϩ channel current amplitude (51.7 Ϯ 5.1% of control, n ϭ 6) ( Fig. 4 A-C). Figure 4C also shows the result of the converse experiment, that is, the average response of cells to CVID (200 nM) after Vc1.1 (Ն100 nM) application. Similar results were obtained with bath application of 100 nM -conotoxin MVIIA, with 51.6 Ϯ 5.6% of the current remaining after application of MVIIA and 50.4 Ϯ 0.1% remaining in the presence of MVIIA and 100 nM Vc1.1 (n ϭ 5) (data not shown).These data suggest that Vc1.1 inhibits the N-type Ca 2ϩ channel in DRG neurons.

Action of Vc1.1 on expressed Ca V 2.1 and Ca V 2.2 channels in oocytes
To assess whether Vc1.1 directly inhibits HVA Ca 2ϩ channels, the ␣ subunits of Ca V 2.1 (P/Q-type) and Ca V 2.2 (N-type) channels were expressed in Xenopus oocytes together with the auxiliary subunits ␤ 3 /␣ 2 ␦ 1 in the ratio 1:1:1. Bath application of 1 M Vc1.1 did not significantly inhibit the Ba 2ϩ current amplitude for oocytes expressing Ca V 2.2 channels (91.3 Ϯ 2.2% of control, n ϭ 8) (Fig. 5 A, C). Analogous to Ca V 2.2-expressing oocytes, 1 M Vc1.1 did not inhibit Ba 2ϩ currents mediated by Ca V 2.1 channels Figure 2. Vc1.1 inhibition of HVA calcium channel currents is use dependent. A, Representative time course of peak Ba 2ϩ current amplitude recorded using a 15 ms step depolarization from a holding potential of Ϫ80 mV to a test potential of Ϫ10 mV at a frequency of either 0.05 or 0.1 Hz, as indicated, in the absence and presence of 100 nM Vc1.1. B, Bar graph of the relative inhibition of HVA calcium channel currents by 100 nM Vc1.1 during a 15 ms depolarization at a frequency of either 0.05 Hz (n ϭ 24) or 0.1 Hz (n ϭ 7 responsive cells of 7 tested), and a pulse duration of 150 ms and frequency of either 0.05 Hz (n ϭ 36 responsive cells of 50 tested) or 0.1 Hz (n ϭ 9 responsive cells of 12 tested).***p Ͻ 0.001 unpaired t test. C, Superimposed traces of depolarization-activated whole-cell Ca 2ϩ channel currents recorded using 2 mM Ca 2ϩ as the charge carrier, elicited by a voltage step from a holding potential of Ϫ80 to 0 mV in the absence (a) and presence (b) of 100 nM Vc1.1. (88.1 Ϯ 8.2% of control, n ϭ 4) (Fig. 5 B, C). Raising the Vc1.1 concentration to 10 M similarly failed to inhibit the Ba 2ϩ currents mediated by either Ca V 2.1 or Ca V 2.2 channels expressed in oocytes. A similar lack of effect of Vc1.1 on Ca V 2.3 (R-type) and Ca V 1.2 (L-type) channels expressed in Xenopus oocytes was also observed (data not shown). To investigate whether the auxiliary subunits influence the sensitivity of recombinant Ca 2ϩ channels to Vc1.1 (Yasuda et al., 2004), oocytes were injected with the subunit combination Ca V 2.2/␤ 3 in a 1:1 ratio. Although cells expressing Ca V 2.2/␤ 3 elicited smaller Ca 2ϩ channel currents than those expressed by Ca V 2.2/␤ 3 /␣ 2 ␦ 1 channels, the Ba 2ϩ current amplitude was not appreciably changed (88 Ϯ 6.5% of control, n ϭ 3) in the presence of 1 M Vc1.1 (Fig. 5C). Together, these data indicate that Vc1.1 does not interact with either Ca V 2.1 (P/ Q-type) or Ca V 2.2 (N-type) channels directly, and that Vc1.1 modulation of HVA Ca 2ϩ channel currents occurs via an indirect mechanism.

Vc1.1 inhibits HVA Ca 2؉ channel currents via a G-proteincoupled receptor mechanism and src tyrosine kinase
The role of G-proteins in mediating Vc1.1 inhibition of HVA Ca 2ϩ channel currents in DRG neurons was examined by replacing GTP with GDP␤S, a nonhydrolyzable GDP analog, in the intracellular recording solution. GDP␤S antagonized the effect of Vc1.1 on Ca 2ϩ channel currents, with 96.8 Ϯ 4.5% of current  remaining after addition of 100 nM Vc1.1 in the presence of GDP␤S (n ϭ 15) (Fig. 6C).
These results suggest that Vc1.1 inhibits N-type Ca 2ϩ channel currents in DRG neurons via a G-protein-coupled receptor (GPCR) mechanism. Vc1.1 inhibition also persisted after application of depolarizing prepulses to ϩ80 mV, suggesting a voltageindependent inhibitory pathway (n ϭ 7) (Fig. 6C). Incubation of cells in medium containing 3 g/ml pertussis toxin (PTX) for 24 h antagonized the inhibition of VGCC currents by Vc1.1 (99.9 Ϯ 4.5% of control, n ϭ 11) (Fig. 6 A, C). Together, these results suggest that the mechanism by which Vc1.1 inhibits HVA Ca 2ϩ channel currents involves a PTX-sensitive GPCR of the G␣ i/o subfamily.

Vc1.1 inhibits HVA Ca 2؉ channel currents in DRG neurons via GABA B receptor
To investigate further the identity of the primary target for Vc1.1, a number of different receptor antagonists were tested for their ability to antagonize the effect of Vc1.1 ( Table 1). Antagonists of neuronal nAChRs, mecamylamine and hexamethonium, and desensitizing concentrations of ACh were bath applied ϳ5 min before application of Vc1.1 and in all cases failed to antagonize the inhibition of N-type VGCC currents in DRG neurons by Vc1.1. Similarly, the muscarinic acetylcholine receptor antagonist, atropine, the ␣ 1 -and ␣ 2 -adrenergic receptor antagonists, prazosin and yohimbine, respectively, the -opioid receptor antagonist, naloxone, and the GABA A receptor antagonist, bicuculline, all failed to antagonize the effect of Vc1.1 in the majority of cells tested.
Numerous studies have shown that activation of GABA B receptors leads to inhibition of N-type Ca 2ϩ channel currents in central and peripheral neurons (Menon-Johansson et al., 1993;Guyon and Leresche, 1995;Fujikawa et al., 1997;Harayama et al., 1998;Richman et al., 2005;Castro et al., 2007;Raingo et al., 2007), but conotoxins have not previously been contemplated to act on this pathway. The inhibition of N-type Ca 2ϩ channel currents after activation of GABA B receptors has been shown to occur in both a voltage-dependent and a voltage-independent manner, to be PTX-sensitive, and to involve c-Src (Richman et al., 2005;Raingo et al., 2007). Several GABA B receptor antagonists were tested for their ability to antagonize Vc1.1 inhibition of HVA Ca 2ϩ channel currents in DRG neurons. Application of 1 M Vc1.1 in the presence of the GABA B receptor antagonists phaclofen (50 M) and CGP55485 (1 M) did not significantly reduce the HVA Ca 2ϩ channel current in DRG neurons [95.5 Ϯ 4.4% (n ϭ 9 and 93.2 Ϯ 4.5% (n ϭ 14) of control, respectively] ( Fig. 7 A, B). RgIA (100 nM) also failed to inhibit HVA Ca 2ϩ channel current in the presence of CGP55485 (92.5 Ϯ 3.5% of control, n ϭ 5) (data not shown). Together, these data indicate that Vc1.1 and Rg1A inhibit HVA Ca 2ϩ channel currents via activation of the GABA B receptor in DRG neurons.
The GABA B receptor agonist baclofen has previously been shown to inhibit HVA Ca 2ϩ channel currents in neonatal rat DRG neurons Scott, 1987, 1990;Menon-Johansson et al., 1993;Fujikawa et al., 1997). We examined whether the effects of Vc1.1 and baclofen were additive in inhibition of the HVA Ca 2ϩ channel current. The inhibition produced after addition of 10 -30 M baclofen (63.6 Ϯ 10.6% of , when a prepulse to ϩ80 mV (ϩPP) was applied 10 ms before the test pulse, when GTP was replaced with GDP␤S, after overnight incubation in PTX, or when pp60c-src inhibitory peptide was included in the pipette recording solution. ***p ϭ 0.0001 compared with control. Numbers in parentheses reflect the number of cells. ϩ, Reversal of Vc1.1 inhibition; Ϫ, no effect; mAChR, muscarinic acetylcholine receptor.
control, n ϭ 4) was not significantly greater than that produced by 100 nM Vc1.1 alone (61.2 Ϯ 15.9% of control, n ϭ 4) in the same cells. Similarly, application of Vc1.1 (100 nM) after addition of 30 M baclofen did not further inhibit HVA Ba 2ϩ current amplitude (63.6 Ϯ 7.2% of control, n ϭ 5) compared with baclofen alone (64.6 Ϯ 6.6% of control, n ϭ 5) (Fig. 7C). Under the present recording conditions (150 ms depolarizing pulse to Ϫ10 mV applied at a frequency of 0.05 Hz), the time to maximum inhibition (9.3 Ϯ 1.2 min) and reversal (Ͻ5% after 10 min washout) by 30 M baclofen (n ϭ 11) was significantly slower than when the depolarizing pulse was shortened to 15 ms or the frequency of depolarization was increased to 0.1 Hz (time to maximum inhibition occurred in 2.8 Ϯ 0.9 min and complete recovery within 5 min after washout of baclofen, n ϭ 5).
In subsequent experiments, we observed that in ϳ70% of oocytes expressing recombinant Ca V 2.2 (N-type) but not Ca V 2.1 (P/Q-type), Ca V 2.3 (R-type), or Ca V 1.2 (L-type) calcium channels, application of baclofen (100 M) inhibited Ba 2ϩ currents (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). We therefore tested the effect of ␣-conotoxins Vc1.1 and Rg1A in these oocytes. Both Vc1.1 and Rg1A inhibited Ba 2ϩ currents through recombinant Ca V 2.2 channels in baclofen-responding oocytes (n Ն 28). Preincubation of oocytes with the selective GABA B receptor antagonist CGP55845 (1 M) abolished the inhibition of Ba 2ϩ currents observed with 1 M Vc1.1 (n ϭ 4) (supplemental Fig. 2, available at www.jneurosci.org as supplemental material).

Discussion
To date, the ␣-conotoxins have been classified as competitive nAChR antagonists (Dutton and Craik, 2001;Livett et al., 2006). The present study is the first to identify the N-type calcium channel of mammalian DRG neurons as a target of the ␣-conotoxins Vc1.1 and RgIA. To the best of our knowledge, this study is also the first to describe conotoxin modulation of an ion channel via a G-protein receptor-activated pathway.
N-type calcium channels are an appealing target of Vc1.1 because they are well known to play a central role in the detection and transmission of nociceptive stimuli in DRG neurons (Altier and Zamponi, 2004). Several studies have highlighted the importance of N-type calcium channels in neuropathic pain: N-type ␣ 1B channel knock-out mice have a decreased response to neuropathic pain (Hatakeyama et al., 2001;Kim et al., 2001;Saegusa et al., 2001), there is an upregulation of N-type ␣ 1B as well as ␣ 2 ␦ subunits in rat nerve injury models (Abe et al., 2002;Cizkova et al., 2002;Yokoyama et al., 2003), and currently used treatments, or treatments being developed, for pain relief include direct (conotoxin MVIIA, also known as Prialt) and indirect (morphine) inhibitors of N-type channels (McGivern, 2006).
HVA Ca 2ϩ channel currents of rat DRG neurons include contributions from N-, P/Q-, L-, and R-type calcium channels (Scroggs and Fox, 1992). The present study indicates that Vc1.1 targets solely the N-type component of the whole-cell Ca 2ϩ current, because it had no effect in the presence of the N-type Ca 2ϩ channel blockers, -conotoxins CVID and MVIIA. Vc1.1 inhibition of N-type Ca 2ϩ channel current was irreversible after washout of up to 30 min, but subsequent cells patched from the same coverslip were as sensitive to Vc1.1 as the initial cell exposed to Vc1.1. This suggests that Vc1.1 inhibition of N-type current is dependent on membrane depolarization. Consistent with this hypothesis, Vc1.1 had no effect on HVA Ca 2ϩ currents in DRG neurons when the duration of the depolarization pulse was shortened to one-tenth of its initial duration with the frequency maintained at 0.05 Hz. Vc1.1 inhibition was only observed with this short depolarizing pulse when the frequency was increased to 0.1 Hz.
DRG neurons have been shown to express a splice variant Ca V 2.2 that differs at exon 37 (Bell et al., 2004) from the peripheral splice variant we expressed, but extensive studies reported here indicate that the reason Vc1.1 is ineffective on recombinant N-type Ca 2ϩ channels is that it is acting via a G-protein receptoractivated pathway. Our results suggest that the GPCR responsible is the GABA B receptor: first, Vc1.1 inhibition of N-type Ca 2ϩ current was antagonized by selective GABA B receptor antagonists, and second, similar to previous studies of the GABA B receptor agonist baclofen in rat or chick DRG neurons (Dolphin and Scott, 1987; Richman and Diversé-Pierluissi, 2004; Raingo et al., 2007), Vc1.1 inhibition of N-type Ca 2ϩ currents in rat DRG neurons was antagonized when GTP was replaced in the patch pipette with GDP␤S, when the cells were incubated overnight with PTX, and when c-Src tyrosine kinase activity was inhibited. Activation of the GABA B receptor has been shown to inhibit N-type Ca 2ϩ channel current by voltage-dependent (relieved by a strong depolarizing prepulse) and voltage-independent pathways, and to be readily reversible after washout of baclofen (Richman and Diversé-Pierluissi, 2004;Raingo et al., 2007). No relief of Vc1.1 inhibition was observed after depolarizing prepulses to ϩ80 mV. Although baclofen inhibition during prolonged depolarizations (150 ms, 0.05 Hz) was also irreversible, baclofen inhibition of N-type Ca 2ϩ current during short depolarization pulses (15 ms, 0.05 Hz) was rapidly reversible after washout of baclofen (our unpublished observations). These results suggest that Vc1.1 does not inhibit the N-type current via the fast, G␤␥-mediated, membrane-delimited, voltage-dependent pathway associated with GABA B receptor activation. Rather, it activates a pathway that remains to be fully elucidated but involves pp60c-src tyrosine kinase and leads to a maintained inhibition of the N-type Ca 2ϩ channel current, possibly as a result of persistent binding of Vc1.1 to the GABA B receptor. Alternatively Vc1.1 may interact differently with the GABA B receptor than baclofen, which binds to the venus fly trap motif of the GABA B1 subunit, leading to its closure and activation of the receptor (for review, see Emson, 2007).
Vc1.1 was without effect on recombinant Ca V 2.1 (P/Q-type), Ca V 1.2 (L-type), and Ca V 2.3 (R-type) channels expressed in Xenopus oocytes. However, in subsequent experiments we found that baclofen activation of endogenous GABA B receptors in batches of Xenopus oocytes mediates inhibition of Ba 2ϩ current through recombinant Ca V 2.2 (N-type) channels. In oocytes expressing endogenous GABA B receptors, ␣-conotoxins Vc1.1 and Rg1A also inhibit recombinant Ca V 2.2 channels via agonist activation of a GABA B receptor-signaling pathway in Xenopus oocytes (Yang et al., 2001).
The characteristics of the G-protein-induced inhibition of voltage-dependent N-type calcium channels by Vc1.1 and Rg1A activation of GABA B receptors is reminiscent of the agonistinduced internalization of N-type calcium channels observed in rat DRG neurons (Altier et al., 2006) and embryonic chick DRG neurons (Puckerin et al., 2006;Tombler et al., 2006). Prolonged exposure of opioid receptor-like receptor 1 (ORL1) receptors to the agonist nociceptin has been shown to trigger the internalization of the ORL1 receptors and Ca v 2.2 channels coexpressed in human embryonic kidney tsA-201 cells (Altier et al., 2006), which is accompanied by a tonic inhibition of calcium entry. Baclofen, a GABA B agonist, has also been shown to induce the rapid removal of -conotoxin GVIA-sensitive (N-type) calcium channels from the plasma membrane of DRG neurons (Tombler et al., 2006). A mechanism proposed for the internalization of voltagedependent calcium channels involves the selective binding of ␤-arrestin 1 to the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor)-binding region of the calcium channel whereby, on GABA B receptor activation, receptors are recruited to the arrestin-channel complex and internalized (Puckerin et al., 2006). Furthermore, ␤-arrestins have been shown to be required to target nonreceptor tyrosine kinase c-Src to constitutively active G-protein-coupled -opioid receptors in DRG neurons, resulting in their internalization (Walwyn et al., 2007). It is likely, therefore, that ␣-conotoxin activation of GABA B receptors and the inhibition of N-type calcium channels in DRG neurons is mediated by a mechanism similar to that reported previously for internalization of GPCR-channel complexes.
In conclusion, our results identify Ca v 2.2 as a target of the analgesic ␣-conotoxins Vc1.1 and Rg1A. We propose a novel mechanism by which these ␣-conotoxins modulate N-type calcium channel currents in DRG neurons, that is, acting as agonists via G-protein-coupled GABA B receptors. The activation of GABA B receptors by agonists such as baclofen is well established to produce antinociceptive and antiallodynic actions in chronic pain models (Schuler et al., 2001;Bettler et al., 2004). Therefore, the activation of GABA B receptors by Vc1.1 and Rg1A and subsequent inhibition of Ca v 2.2 (N-type) channels is most likely the mechanism mediating their analgesic properties, despite their recently reported modulation, at lower efficacy, of the ␣9␣10 nAChR (Vincler et al., 2006;Vincler and McIntosh, 2007).