Wiley, John W., Hylan C. Moises, Robert A. Gross, and Robert L. Macdonald. Dynorphin A-mediated reduction in multiple calcium currents involves a Goα-subtype G protein in rat primary afferent neurons. J. Neurophysiol. 77: 1338–1348, 1997. We examined the effect of antisera directed at specific G-protein subtype(s) on dynorphin A (Dyn A)-mediated reduction of calcium currents in rat dorsal root ganglia (DRG) neurons. Whole cell patch-clamp recordings were performed on acutely dissociated neurons. Dyn A (1 μM)-mediated decrease in calcium currents was inhibited >90% by the preferential κ-receptor antagonist norbinaltorphimine. Dyn A (300–1,000 nM)-mediated reduction in calcium currents was examined during intracellular administration of antisera directed against specific regions of Goα, Gi1α/Gi2α, and Gi3α subunits. Intracellular dialysis with an antiserum specific for Goα for 20 min decreased calcium current inhibition by Dyn A (1 μM) in 13 of 15 neurons by an average of 75%. Dialysis with nonimmune serum did not affect Dyn A's action to reduce calcium currents. Intracellular dialysis with either anti-Gi1α/Gi2α or anti-Gi3α antisera did not affect Dyn A-induced changes in calcium currents. In the presence of the N-type calcium channel antagonist ω-conotoxin GVIA, the P-type calcium channel antagonist ω-Aga IVA, and ω-Aga MVIIC applied subsequent to the other toxins, the effect of Dyn A to reduce calcium currents was inhibited by 52, 28, and 16%, respectively. The L channel antagonist nifedipine did not affect the ability of Dyn A to inhibit calcium currents. These results suggest that in rat DRG neurons coupling of κ-opioid receptors to multiple transient, high-threshold calcium currents involves the Goα subclass of G proteins.
Activation of κ-opioid receptors inhibits membrane calcium currents in a variety of neurons (Gross et al. 1990; Macdonald and Werz 1986). In rat nodose ganglion neurons, reduction in calcium currents by the preferential κ-receptor agonist dynorphin A (Dyn A) was antagonized by pretreatment with pertussis toxin (PTX), which prevents activation of inhibitory Go/Gi-type G proteins (Gross et al. 1990). At least five PTX substrates (GoA, GoB, Gi1, Gi2, Gi3) now have been identified by molecular cloning and sequencing techniques (Simon et al. 1991). Moises et al. (1994) showed that μ-opioid receptors in rat dorsal root ganglia (DRG) neurons were coupled to multiple calcium channels via a Go-type G proteins. However, the subtype(s) of G proteins and the specific calcium currents regulated have not been elucidated for neuronal κ-opioid receptors.
Activation of μ- or κ-opioid receptors decrease high-threshold calcium current in rodent sensory neurons, but does not affect the low-threshold T current (Gross and Macdonald 1987; Rusin and Moises 1995). Opiates inhibit release of neurotransmitter from central endings of primary sensory afferent neurons and antagonize transmission of nociceptive information entering CNS presumably via their action to reduce calcium currents. The N-type channels are thought to provide the predominant target for inhibition by opiates, as well as by a variety of other inhibitory neurotransmitters and neuromodulators (Anwyl 1991). However, recent studies have demonstrated modulation of P-type currents by γ-aminobutyric acid (Mintz and Bean 1993) and adenosine (Mogul et al. 1993) and suppression of Q-type currents by 1S, 3R-ACPD, carbachol, 2-chloroadenosine, and baclofen (Wheeler et al. 1993). Thus it is possible that the depression of neurotransmitter release by opiates and other inhibitory neuromodulators might occur through suppression of activity of multiple types of high-threshold calcium channels.
μ-Opioid receptor activation decreases high-threshold, ω-conotoxin-GVIA-sensitive (N-type) calcium current in acutely dissociated rat DRG neurons (Rusin and Moises 1995). Furthermore μ-agonists reduce a fraction of high-threshold current that is resistant to blockade by saturating concentrations of ω-GVIA and the L channel blocker nifedipine (Rusin and Moises 1995). It is not known whether activation of κ-opioid receptors has a similar or distinct action on neuronal calcium currents. To characterize the calcium channel types that might contribute to κ-opioid-sensitive current, we examined the effects of application of the κ-opioid selective agonist Dyn A on high-threshold current components, isolated on the basis of their sensitivity to blockade by ω-GVIA (N type), nifedipine (L type), ω-MVIIC (preferential for P/Q type), and ω-Aga IVA (P type).
Using intracellular dialysis of antibodies specific for Gαi and Gαo proteins, we observed that κ-opioid receptors were coupled specifically to the Go-subtype G protein. Activation of κ-opioid receptors reduced three types of high-threshold calcium currents in rat DRG neurons, including a ω-GVIA-sensitive (N-type) channel, an ω-Aga IVA-sensitive (P-type) channel, and an MVIIC-sensitive, nifedipine/ω-GVIA/ω-Aga IVA-resistant (presumptive Q-type) channel.
Preparation of acutely dissociated neurons
The procedures used for the preparation of acutely dissociated DRG neurons are described in detail in a recent report (Moises et al. 1994). Briefly, DRG were dissected from the lumbar and thoracic regions of 7- to 14-day-old rats and treated with collagenase (Type II, 3 mg/ml, Sigma Chemical) for 50 min at 37°C for 50 min, followed by 10-min treatment with trypsin (Type I, 1 mg/ml, Sigma Chemical). A 5% solution of bovine serum albumin (20 mg/ml minimal essential medium, MEM, GIBCO Laboratories) then was added to the incubation medium to inhibit the enzymes, and the dissociated ganglia mechanically dispersed by triturating for four to five passages through a fire-polished Pasteur pipette. Cell suspensions (∼2 ml/dish) were plated onto collagen-coated culture dishes and incubated at 37°C for 1 h, after which an additional amount of MEM containing 10% horse serum (Sigma Chemical) and nerve growth factor (50 mg/ml, Boehringer Mannheim) was added to bring the total volume to 2 ml. Neurons were studied between 2 and 10 h after plating. Recordings were performed only on neurons without processes, and neurons were excluded from the analysis if inadequate space-clamp was obtained (e.g., delayed settling of capacitance transient with time constants >150 ms and broad tail currents).
G protein immunoblot analysis
DRGs were harvested as described above. Briefly, qualitative G protein immunoblot analysis was carried out patterned after the technique of Gierschik et al. (1985). Typically, 12 thoracic ganglia were used in these studies. Crude membrane homogenates (25 μg), prepared as described by Attali et al. (1989), were applied to a 4% stacking gel and electrophoresed on 7.5% polyacrylamide gels containing sodium dodecyl sulfate, blotted onto nitrocellulose, and incubated for 5 h with rabbit antibodies directed against the various G protein subunits. Blots then were incubated for 1 h with peroxidase-conjugated goat anti-rabbit IgG and stained for peroxidase activity. Stained bands were scanned with a Macintosh scanner (Silverscan 11) in conjunction with Image software (NIH, ver. 1.55). Staining, under these conditions, was proportional to the amount of protein applied to the gel within the range of 15–200 μg.
Whole cell patch-clamp recordings
Voltage-clamp recordings were obtained using the whole cell variation of the patch-clamp technique (Hamill et al. 1981). Glass recording patch pipettes, prepared from Fisher brand microhematocrit tubes and having resistances of 0.8–1.5 MΩ, were filled with recording solution of the following composition (in mM) 140 CsCl, 10 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonicacid (HEPES), 10 ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, 5 ATP (magnesium salt), and 0.1 GTP (lithium salt) (all reagents from Sigma Chemical). The pH was adjusted with 1 M CsOH to 7.3–7.35 after the addition of ATP and the osmolality (280–300 mOsm) adjusted to 10–15% below that of the bath solution. The neurons were bathed in a solution (pH 7.4, 310–330 mOsm) consisting of (in mM) 67 choline chloride, 100 tetraethylammonium chloride, 5.3 KCl, 5.6 glucose, 5.0 CaCl2, 0.8 MgCl2, and 10 HEPES. Under these conditions, sodium and potassium currents were suppressed.
Recordings were made at room temperature using an Axopatch 1-D patch-clamp amplifier (Axon Instruments, Foster City, CA). Pipette and whole cell capacitance and series resistance were corrected using compensation circuitry on the patch-clamp amplifier. Initial input resistances were in the range of 500 MΩ to 1.2 GΩ. Series resistance was estimated by cancellation of the capacitance-charging current transient after patch rupture. In most cases, series resistance compensation of 80–90% was obtained without inducing significant noise or oscillation, resulting in final series resistances ranging from 0.1 to 1.2 MΩ. No data were included in the analysis where series resistance resulted in a ≥5 mV error in voltage commands.
Voltage step commands of 100 ms duration were applied every 30 or 60 s, and the evoked currents were filtered with a 12-pole low-pass Bessel filter at 10 kHz (−3 dB). The filtered current records were digitized at 5 kHz, stored, and analyzed by a 386-based computer using the program pClamp (Axon Instruments, Foster City, CA). Leak current was estimated as the inverse of the current evoked with 100-ms hyperpolarizing commands of equal magnitude to the depolarizing commands used to evoke the inward currents. This current was subtracted digitally from the relevant inward current to obtain the calcium current. In some cases, Cd2+ (200–500 μM) was applied to the neuron at the end of an experiment, which always eliminated the inward calcium currents, including any that remained in the presence of the opioid agonist. In the presence of 200 μM Cd2+ to eliminate the calcium currents, no voltage-dependent outward currents were evoked at test potentials as positive as +30 mV.
Calculations of the magnitudes of Dyn A-mediated current inhibition and calcitonin gene-related peptide (CGRP)-mediated current increase were corrected for the decline in calcium current (rundown) that occurred in some neurons during the course of whole cell recording. To adjust for rundown, recovery currents were elicited after termination of agonist application until the amplitude of the postdrug current plateaued. The response to drug then was expressed as a percentage change in the averaged value of the predrug current and stabilized recovery current amplitudes. This approach relies on the fact that the establishment of agonist-induced inhibition in current occurs sufficiently fast (typically within several seconds) that the extent of current rundown during this time contributes marginally, if at all, to the absolute reduction in current measured.
Preparation and delivery of solutions
Dynorphin A and CGRP (Peninsula Laboratories, Belmont, CA) were stored frozen (at −20°C) in 10-ml aliquots of lyophilized peptide (dissolved in sterile water) and prepared as a fresh solution immediately before the experiment by dilution with normal extracellular bathing medium. The solution (pH 7.4, 310–330 mOsm) bathing the cells consisted of (in mM) 67 choline chloride, 100 tetraethylammonium chloride; 5.3 KCl, 5.6 glucose, 5.0 CaCl2, 0.8 MgCl2, and 10 HEPES. Dyn A or CGRP were applied either by using pressure ejection (2-s duration) from a blunt-tipped (10- to 20-mm tip diam) glass micropipette positioned ∼50 mm from the neuron. Depolarizing commands were initiated 2–3 s after application of the peptide. The effects of different G-protein antibodies were assessed by introducing them directly into the cell by intracellular dialysis from the recording pipette.
Specific antisera directed against the α-subunits of Go (GC/2) and Gi1/Gi2 (AS/7) were obtained from DuPont/NEN (Boston, MA) and Gi3 from Calbiochem (San Diego, CA). These are rabbit antisera that were raised against synthetic decapeptides corresponding to the amino and carboxyl termini of bovine brain Goα (GC/2) (Goldsmith et al. 1988; Spiegel 1991) and transducin-α (AS/7) (Goldsmith et al. 1987), respectively. The GC/2 antiserum specifically identifies Goα, but does not react with Giα-subunits; whereas, AS/7 recognizes both Gi1α and Gi2α, in addition to transducin-α, but does not react with Goα or Gi3α. Experiments were also performed using an affinity-purified rabbit antibody that only recognizes carboxyl-terminal epitopes specific to the α-subunit of Gi3. In some experiments, nonimmune rabbit serum was included in the recording pipette solution at the same dilutions used for antibody studies and served as a control. Recordings using nonimmune serum and anti-G protein antisera were performed during each experiment involving different batches of DRGs.
For all recordings in the presence of G-protein antibodies, the tip of recording pipette was filled to 1–2 mm with standard recording solution, and the pipette was back-filled with the experimental solution. In this way, the onset of action of antibody loading was delayed, which permitted the measurement of control responses to κ-opioid or CGRP application within the first few minutes of whole cell recording. Absence of protein-containing mixtures in the tip of the recording pipette facilitated the formation of giga-ohm seals and helped prevent clogging of the pipette tip after patch rupture.
DRUG SOLUTIONS AND APPLICATION.
A rapid-exchange U-tube system (Murase et al. 1989) was used to change the solution bathing the neuron under study. This consisted of a gravity fed U-shaped polyethylene tubing, which generated a laminar flow of bathing solution across the neuron. The delivery port of the tube was positioned within 200 μm of the neuron, whereas the distal end was interfaced to a vacuum line by means of a digitally controlled switching valve. Either control or drug solution flowed over the neuron throughout the recording, whereas the microenvironment around the cell could be exchanged rapidly (<1 s) by opening or closing the connection to the vacuum line.
The selective opioid receptor antagonists norbinaltorphimine (κ-opioid) and β-funaltrexamine (μ-opioid) were obtained from Research Biochemicals (Natick, MA). ω-GVIA (Peninsula Laboratories), ω-Aga IVA (Peptides International Pfizer), and MVIIC (Bachem, Torrance, CA) were prepared as 500, 100, and 1 μM stock solutions, respectively, and stored in the same manner. On the day of the experiment, aliquots of these channel blockers and freshly weighed amounts of nifedipine (Research Biochemicals) or opioid receptor antagonists were dissolved in the extracellular bathing solution to the desired concentrations. To minimize degradation caused by exposure to light, nifedipine was held in a light-proof container and applied to the recording chamber under restricted light conditions. Solutions with ω-GVIA, ω-Aga IVA, or MVIIC also contained cytochrome C (0.1%) to prevent peptide binding to containers.
Statistical comparisons between the effects of Dyn A or CGRP on calcium currents in control-, nonimmunoglobulin treated-, anti-G protein antisera-treated neurons were made using Student's two-tailed t-test. For experiments involving infusions of G-protein antibodies, a paired-sample t-test was first used to compare differences in the mean changes in the peak of the calcium currents produced by application of Dyn A or CGRP when tested within several minutes of patch rupture and again after prolonged dialysis of the neuron with an antiserum. Student's two-tailed t-test was used then to assess the statistical significance of any differences between the effects on Dyn A or CGRP responses produced by intracellular dialysis with a specific G-protein antiserum or nonimmune control serum. Values in the text are given as means ± SE.
Results obtained before and after administration of calcium channel blockers were compared statistically using the Student two-tailed t-test or repeated measures analysis of variance followed by Tukey-Kramer multiple comparisons test. Values in text and figures are given as means ± SE, unless otherwise indicated.
Immunoglobulins can be delivered intracellularly via patch pipettes
Pilot experiments were performed to determine whether cells could be loaded sufficiently with antibodies by intracellular dialysis via the patch pipette, as described in methods section. Briefly, DRG neurons were plated onto coverslips placed at the bottom of culture dishes and incubated for 1–2 h before recordings. Whole cell recordings were obtained with patch pipettes containing standard internal solution supplemented with either nonimmune serum (1:10–1:100 dilution) or a fluorescein-conjugated goat anti-rabbit immunoglobulin at a 1:10 dilution (DuPont/NEN) and examined during a 30-min period to document intracellular delivery of the immunoglobulin (n = 4). All the neurons demonstrated fluorescence under ultraviolet light within 5 min supporting the successful loading of neurons with the immunoglobulin. A representative neuron 20 min after patch rupture showing the general appearance of the cell (left) and the intensity of immunofluorescence with ultraviolet illumination (right) is shown in Fig. 1.
Rat DRGs contain immunoreactive proteins corresponding to Goα, Gi1α/Gi2α, and Gi3α
We next examined whether DRGs contained the immunoreactive substrates for the antisera employed in these studies. Representative Western blots demonstrating the presence of Goα, Gilα/Gi2α, and Gi3α in crude membrane homogenates of DRGs are depicted in Fig. 2 along with molecular weight markers corresponding to the 39-kDa immunoreactive protein recognized by the GC/2 antiserum (Goα) (A) and the 41-kDa immunoreactive protein recognized by the AS/7 antiserum (Gi1α/Gi2α) (B), or the protein recognized by specific anti-serum for Gi3α (C).
Dynorphin A inhibited and CGRP enhanced high-threshold calcium currents
In rat DRG neurons, Dyn A reduced transient high-threshold currents, but not the slowly inactivating high-threshold L-type or low-threshold T-type currents (Gross and Macdonald 1987). CGRP increased transient high-threshold calcium currents but slowly inactivating high-threshold or transient low-threshold calcium currents were not affected in primary afferent neurons (Ryu et al. 1988; Wiley et al. 1992). Therefore we restricted the present experiments to neurons that had little or no observable T-type current, which facilitated the analysis of agonist-induced effects after different experimental manipulations that were used to modify activity of G-protein-dependent processes. Unless otherwise indicated, responses to Dyn A and CGRP were quantified in terms of reductions or increases in peak current amplitude (I p), with the understanding that the opioid- and CGRP-sensitive currents may represent several types of high-threshold calcium channels, which exhibit either transient or more sustained inactivation kinetics. Thirty-eight of 48 (79%) neurons responded to Dyn A and 15 of 29 (51%) neurons responded to CGRP. The inhibition of calcium current by Dyn A (300 nM) was reversed by 88% (from 30.1 ± 3.8% to 3.5 ± 1.2%; n = 5) on application of the preferential κ-opioid receptor antagonist, norbinaltorphimine (100 nM) (Porthoghese et al. 1987) (Fig. 3). In contrast, responses to Dyn A in the same neurons largely were unaffected by administration of β-funaltrexamine (300 nM) an irreversible blocker of μ-opioid receptors (Takemori et al. 1981), with inhibition now averaging 28.5 ± 3.2% (n = 5).
Effect of intracellular delivery of G-protein antisera on Dyn A- and CGRP-mediated changes in DRG neuron calcium currents
Three known subtypes of Gi (Gi1, Gi2, and Gi3) and A and B subtypes of Go are among the family of G proteins that serve as substrates for ADP-ribosylation by PTX (Simon et al. 1991). We evaluated the effect of Dyn A during dialysis of neurons with G-protein antisera specific for a subunits of Go, Gi1/Gi2, and Gi3 to determine whether these subtype(s) of PTX-sensitive G protein(s) are involved in the coupling of κ-opioid receptors to calcium channels. To allow for comparisons between effects of the different antisera, responses to the peptides were examined at an early and late stage of recording so that each neuron could serve as its own control. In preliminary studies, we observed that administration of a protein load (e.g., bovine serum albumin) to neurons was often sufficient to affect calcium currents, producing stabilization or a slight increase in current amplitudes over the course of a typical 20-min recording. To control for such nonspecific effects that antiserum administration might have on the calcium currents or the responses to the peptides, control neurons were dialyzed with nonimmune rabbit serum at the same dilutions used for antibody solutions (1:10–1:100).
Dyn A (300 nM to 1 μM) produced concentration-dependent reduction in the peak of the high-threshold calcium currents evoked from a holding potential of −80 mV at a command potential of +10 mV. Calcium current reduction produced by application of Dyn A (1 μM) 20 min after patch rupture (38.4 ± 4.1%, n = 12) was not significantly different from that produced by application of the κ-agonist within the first 5 min of recording (42.5 ± 4.6%, P > 0.5). Intracellular administration of the anti-Goα (GC/2) antiserum attenuated the Dyn A (1 μM)-mediated reduction of calcium current in a concentration-dependent manner over the range evaluated (1:100–1:10). Infusions of anti-Goα antiserum at a 1:20 dilution during a 20-min period reduced the inhibitory effects of Dyn A (1 μM) in 13 of 15 neurons. In comparison, this antiserum was without significant effect on control currents or responses to Dyn A (1 μM) when included in the recording pipette at a 1:100 dilution (n = 4) and attenuated the response to the opioid receptor agonist in three of six neurons that were dialyzed with a 1:50 dilution. The pairs of current traces from the neuron shown in Fig. 4 compare the inhibitions in current produced by application of Dyn A (1 μM) after 20 min of recording in the presence of nonimmune antiserum (A) or the anti-Goα antiserum at a 1:20 dilution (B). Between 5 and 20 min after patch rupture, Dyn A-mediated inhibition in I p was reduced to <30% of its initial value (from 38 ± 3.7% to 10.4 ± 2.6% inhibition of I p) in the presence of the anti-Goα antiserum (P < 0.05, n = 15). The attenuation of the Dyn A-mediated inhibition in the presence of the Goα antiserum probably was unrelated to changes in control currents or to nonspecific effects associated with the infusion procedure, because administration of nonimmune rabbit serum did not significantly affect Dyn A-induced reductions in calcium current during the 20-min recording period from 41.7 ± 5.1% to 36.5 ± 4.2% (not significantly different; Fig. 4 A). The effect of the anti-Goα antiserum at a 1:10 dilution was not significantly different from that observed at a 1:20 dilution. Technical problems in forming stable giga-ohm seals were more prevalent at the 1:10 dilution (n = 4). In contrast, dialysis of neurons with the Gi1α/Gi2α antisera (1:20 dilution) had no significant effect on Dyn A (1 μM)-mediated inhibition of I p calcium current when compared with nonimmune rabbit serum (Fig. 5, n = 12). Similarly, dialysis of Dyn A-responsive neurons with Gi3α antiserum (1:20 dilution) did not significantly antagonize Dyn A-mediated inhibition in calcium currents, i.e., percent reduction in I p calcium current after 20 min dialysis with nonimmune serum or anti-Giα3 were 37 ± 4% and 34 ± 5%, respectively (not significantly different, n = 4).
CGRP increases calcium currents in rat primary afferent neurons via a pertussis toxin-sensitive mechanism (Wiley et al. 1992). Therefore it seemed of some interest to determinewhether dialysis of specific Goα and Giα antisera affected CGRP-mediated increases in calcium current in a manner similar to that found for responses to κ-opioid receptor activation. CGRP (1 μM) increased I p by 22.6 ± 2.8% at 5 min and 19.6 ± 2.1% at 20 min in neurons dialyzed with pipette solution containing nonimmune serum at a 1:20 dilution(n = 12, not significantly different). Dialysis (1:20 dilution) of neurons with either the Goα antiserum (Fig. 6, n = 8) or Gi1α/Gi2α antiserum (Fig. 6, n = 6) did not have a significant affect on CGRP-mediated increase in I p at 20 min when compared with the magnitude of enhancement observed in the presence of nonimmune serum. In addition, dialysis with the Gi3α antiserum did not have a significant affect on CGRP-induced enhancement in calcium currents at 20 min (19.2 ± 1.8%) when compared with the increase measured in the presence of nonimmune serum, 19.6 ± 2.1%, not significantly different, n = 4).
Dyn A reduces multiple high-threshold currents in rat DRG neurons
A principal goal of this study was to identify the subtypes of high-threshold calcium currents κ-opioid receptors modulate in rat DRG neurons. We reported previously that DRG neurons express at least five types of high-threshold calcium channels, differentiated by their sensitivity to blockade by selective calcium channel blockers (Rusin and Moises 1995). These channels include the DHP-sensitive (L-type) and ω-conotoxin GVIA (N-type) channels, P- and Q-type channels that can be differentiated by their sensitivity to blockade by low concentrations (30 nM, P type) and high concentrations (100–200 nM, Q type) of ω-Aga IVA, respectively, and an unidentified, toxin-resistant (R-type) channel.
The high-threshold calcium currents elicited in the five neurons examined in detail conformed closely with the profile of currents described previously in DRG neurons (Rusin and Moises 1995). The results from a representative neuron are depicted in Fig. 7. When the reductions in I p amplitude produced by administration of a particular calcium channel antagonist were averaged across the five neurons, we found that ω-GVIA (5 μM) irreversibly blocked 47.6 ± 4.3% (N type) and Aga IVA (100 nM) suppressed 13.2 ± 1.9% (P type) of the whole cell current. Application of MVIIC (1 μM) during blockade of N-type and P-type channels antagonized 17.3 ± 2.9% (Q type) and nifedipine (10 μM) blocked 14.5 ± 3.2% of the control whole cell calcium current. In addition, 7.4 ± 1.8% of the high-threshold calcium current was found to be cadmium-sensitive but resistant to blockade by all four calcium channel antagonists.
To identify the types of high-threshold calcium channels that are modulated by κ-opioid receptors, we examined the ability of Dyn A to reduce calcium current before and after selective blockade of N-, L-, and P-type and Q-like current components (n = 4). The traces in Fig. 7 illustrate the inhibitory effects of Dyn A (300 nM to 1 μM) on calcium currents evoked by 100-ms steps to 0 mV from a holding potential of −80 mV at the times and under the conditions indicated in the current versus time graph depicted in Fig. 7. Within the first few minutes after patch-rupture, the calcium current typically increased and then plateaued. The calcium current for the neuron depicted in Fig. 7 revealed maximum amplitude (4.8 nA) at 5.5 min after patch rupture. Application of Dyn A (1 μM for 30 s) at 2 min after patch rupture reduced the I p from 4.50 to 2.45 nA (a 45% decrease), and this effect was reversed completely on washout of the opioid. A second application of Dyn A (300 nM for 30 s) at 5.5 min after patch rupture reduced the I p from 4.8 to 3.2 nA (33% reduction). After recovery from the inhibition in current by Dyn A, ω-GVIA (3 μM) was administered for 30 s. ω-GVIA irreversibly suppressed the I p of the control current from 4.60 to 2.25 nA, and after ω-GVIA, the remaining current was decreased by Dyn A (300 nM) from 2.2 to 1.9 nA (representing 14% decrease). After establishment of an irreversible blockade in calcium currents, by ω-GVIA, application of ω-Aga IVA (100 nM) further reduced the I p from 2.1 to 1.3 nA (38% decrease), and this was associated with a diminution in the inhibitory response to Dyn A (from 1.30 to 1.15 nA). Administration of MVIIC (1 μM) in the presence of ω-Aga IVA reduced the I p from 1.4 to 0.6 nA and abolished any effect of Dyn A that remained after ω-AGA IVA. The application of nifedipine (10 μM) in the continued presence of ω-Aga IVA and MVIIC inhibited a fraction of the current that represented 6.3% (0.3 nA) of the original control current, whereas a cadmium-sensitive current that was resistant to blockade by the four antagonists contributed the remaining 4.2% (0.2 nA) of the current. When the reductions in peak current amplitude produced by a particular calcium channel antagonist were averaged across the four neurons examined in this manner, we found that ω-GVIA (3 μM) irreversibly blocked 48 ± 4.1%, ω- Aga IVA (100 nM) suppressed 20.5 ± 2.5%, MVIIC blocked 14.2 ± 2.2%, and nifedipine (10 μM) blocked 10.3 ± 1.7% of the control whole cell calcium current, with the remaining 7 ± 1.4% representing a pharmacologically undefined (R-type) component of calcium current. In three of the neurons, the inhibition in calcium current by Dyn A showed sensitivity to blockade by selective calcium channel blockers similar to that illustrated for the neuron illustrated in Fig. 7. In the remaining neuron, administration of Aga IVA (100 nM) had little effect on the calcium current or the response to Dyn A after irreversible blockade of N-type channels with ω-GVIA. In contrast, Dyn A-mediated inhibition of the Ip in this neuron was eliminated after application of MVIIC (1 μM). Figure 8 presents a graphic summary of these results and shows that the inhibitory effect of Dyn A was reduced significantly after selective blockade of N, P, and Q-type channels.
The principal finding of the present study was that activation of κ-opioid receptors inhibits multiple components of high-threshold Ca2+ current in rat DRG sensory neurons via a G-protein-dependent signaling pathway that involves activation of the alpha subunit of Go. Using antibodies raised against carboxyl-terminal portions of either Goα or Gi1α/Gi2α, McFadzean and coworkers (1989) showed that δ-opioid receptors in NG108-15 cells were coupled to calcium channels via Go but not Gi. However, results from biochemical reconstitution experiments suggest that purified μ-opioid receptors from rat brain membranes are coupled functionally to both Gi- and Go-type PTX substrates (Ueda et al. 1988). In addition, recent studies in clonal cells involving agonist-induced photoaffinity labeling of G protein α-subunits indicate that a single type of opioid receptor (e.g., native and cloned μ- or δ-opioid receptors) can interact with a variety of Gi or Go proteins (Chakrabarti et al. 1995; Prather et al. 1994). However, these kinds of data provide no information about the identity of Gi or Go proteins that transduce the inhibitory effect of κ-opioid receptor activation on neuronal calcium channels.
The results of our experiments in neurons dialyzed with specific G-protein antisera suggest that the reduction in calcium current by Dyn A was mediated by a Go-type protein. Thus κ-opioid-induced responses were attenuated in neurons dialyzed with the anti-Goα (GC/2) antiserum, whereas, intracellular administration of either the anti-Gi1α/Gi2α (AS/7) or anti-Gi3α antisera was without effect. The anti-Giα1α/Gi2α antiserum used in our studies previously has been shown to inhibit the reduction in adenylate cyclase activity by δ-opioids in NG108-15 cells (Ueda et al. 1988). Our Western blot studies indicated that the AS/7 and anti-Gi3α antisera do react with membrane proteins of the appropriate molecular weight corresponding to α-subunits of Gi1/Gi2 and Gi3, respectively. Therefore it seems unlikely that the lack of effectof the anti-Gi1α/Gi2α or anti-Gi3α antisera on Dyn A-induced reductions in calcium current was a consequence of the antisera having little affinity for native Gi proteins. It is possible that Dyn A could act nonselectively on opioid receptor subtypes other than the κ-receptor at the concentrations (300 nM to 1 μM) employed in our studies. However, we found that Dyn A-induced reduction in high-threshold calcium current was inhibited markedly (>90%) by the selective κ-opioid receptor antagonist norbinaltorphimine, but unaffected by administration of β-FNA, an irreversible blocker of μ-opioid receptors. Therefore we interpret the finding ofselective blockade of Dyn A-induced inhibition of calcium current by the GC/2 antiserum to indicate that κ-opioid receptors are coupled negatively to calcium channels in rat DRG neurons via Go-, but not Gi- type G proteins.
Molecular cloning techniques have identified two splice variants of Go (GoA, GoB) (Hus et al. 1990; Strathmann et al. 1990). Because the anti-Goα (GC/2) antiserum used here was raised against a peptide corresponding to an amino-terminal sequence of Goα, it should recognize both splice variants of Go. Therefore we cannot state whether or not adistinct isotope of Go was involved specifically in coupling κ-opioid receptors to calcium channels in rat DRG neurons. It is interesting to examine our results in view of recent reports supporting the important contribution of βγ subunits in mediating norepinephrine-induced decrease in calcium currents in sympathetic neurons (Herlitze et al. 1996; Ikeda 1996). Therefore both Gα and βγ subunits appear to participate in the regulation of high-threshold calcium currents. Additional studies will be required to assess the effect of either immunoneutralization or attenuating the expression of Gβγ subunits on κ-opioid receptor-mediated modulation of neuronal calcium channels.
We reported previously that CGRP-mediated enhancement in high-threshold calcium currents in rat nodose ganglion neurons was antagonized by GDP-β-S and PTX (Wiley et al. 1992). The present results suggest that the increase in calcium current by CGRP does not involve signaling through Goα, Gi1α, Gi1α/Gi2α, Gi2α, or Gi3α subunits. By contrast, findings by Gollash et al. (1993) indicate that Gi2 and protein kinase C may be required for thyrotrophin-releasing hormone-induced stimulation of voltage-activated calcium currents in GH3 cells. One possibility to consider is that CGRP receptors might be coupled to Gs and that the apparent PTX-sensitivity of CGRP-induced enhancement of calcium currents resulted from removal of “tonic” inhibitory G protein regulation of calcium channels. Tonic G protein-mediated inhibition of calcium channels is supported by double-pulse “facilitation” studies that suggest both PTX-sensitive and -insensitive G proteins tonically inhibit calcium channels (Kasai 1991). In addition, acute exposure to PTX was associated with an increase in the L-type calcium current in cardiac myocytes (Keung and Karlinger 1990).
κ-Opioid receptors are coupled to three types of high-threshold calcium channels
Previous work from this and other laboratories (Anwyl 1991; Rusin and Moises 1995) indicated that the profile of calcium current inhibition produced by κ-opioid activation in rat DRG sensory neurons cannot be attributed to the modulation of a single type of calcium channel. This conclusion was based on the fact that administration of ω-conotoxin GVIA in saturating concentrations always suppressed but was unable to completely eliminate the inhibition in calcium current by κ-opioid agonists in the vast majority of neurons. It is possible that κ-opioid receptors might regulate the activity of a subpopulation of N-type channels that are not blocked by ω-conotoxin GVIA. However, we consider this unlikely and favor the first hypothesis, because the fraction of Dyn A-sensitive current that persisted after ω-GVIA administration could be eliminated by combined application of other calcium channel antagonists. In a subset of DRG neurons (∼20%) administration of the selective P-type channel blocker ω-Aga IVA (100 nM), eliminated inhibitory responses to Dyn A that remained after irreversible block of N-type calcium channels by ω-GVIA was established. These data suggest that N- and P-type calcium channels contribute κ-opioid-sensitive current. Yet, in most of the sensory neurons examined (∼80%), κ-opioid receptors were coupled negatively to three types of high-threshold calcium channel, because ω-GVIA and ω-Aga IVA administered in saturating concentrations failed to eliminate completely the inhibition in calcium current by Dyn A. The pharmacological sensitivity of this residual component of κ-opioid-sensitive current to blockade by MVIIC suggested that it might be contributed by the so-called Q-type channels first described by Tsien and coworkers (1991). Overall, κ-opioid receptors appeared to be coupled differentially to three channels types with ω-GVIA-sensitive calcium current (N type) being most sensitive to modulation by Dyn A. Administration of Dyn A (1 μM) inhibited N-type current on average by ∼55%, whereas ω-Aga IVA-sensitive current (P type) was reduced by ∼20% and MVIIC-sensitive (Q like) current was inhibited by ∼30%. The finding in rat peripheral sensory neurons that κ-opioid receptors regulate activity of multiple high-threshold calcium current components is in keeping with a growing body of data that demonstrate modulation of other non-N-type high-threshold calcium currents in central neurons by a variety of neurotransmitters and neuromodulators (Mintz and Bean 1993; Mogul et al. 1993; Wheeler et al. 1993). In contrast, L-type channels do not appear to be regulated by either μ-opioid (Moises et al. 1994; Rusin and Moises 1995) or κ-opioid receptors (Gross and Macdonald 1987; Gross et al. 1990) in DRG sensory neurons.
In summary, we showed that the inhibition in calcium current by κ-opioids in rat DRG neurons occurred through activation of a Go type GTP-binding protein. Intracellular dialysis of neurons with an antisera specific for α-subunits of Go-type G-proteins greatly attenuated responses to κ-opioids, whereas antibodies specific for α-subunits of Gi1, Gi2, or Gi3 were without effect. These data support the notion that the Go subclass of GTP-binding proteins mediate inhibitory coupling of κ-opioid receptors calcium channels in rat DRG neurons. In addition, κ-opioid receptors are coupled negatively to three types of high-threshold calcium channels in rat DRG neurons, including ω-GVIA-sensitive (N-type), ω-Aga IVA-sensitive (P-type), and MVIIC-sensitive (presumptive Q-type) channels. Therefore, ω-GVIA-insensitive calcium channels, such as P- and/or Q-type channels, may participate together with ω-GVIA-sensitive N-type channels in the voltage-dependent influx of calcium that couples depolarization to neurotransmitter release (Wheeler et al. 1994).
We thank B. Dzwonek for providing expert technical assistance.
This work was supported in part by National Institutes of Health Grant DK-45820 and Veterans' Affairs Merit Awards to J. W. Wiley, DA-03365 to H. C. Moises, and DA-04122 to R. L. Macdonald.
Address for reprint requests: J. W. Wiley, University of Michigan Medical School, GRECC 11G, 2215 Fuller Rd., Ann Arbor, MI 48105.
Present address of R. A. Gross: Dept. of Neurology and Pharmacology, Box 711, University of Rochester, 601 Elmwood Ave., Rochester, NY 14642.