Abstract
Although activation of G protein-coupled inward rectifying K+ (GIRK) channels by Gi/Go-coupled receptors has been shown to be important in postsynaptic inhibition in the central nervous system, there is also evidence to suggest that inhibition of GIRK channels by Gq-coupled receptors is involved in postsynaptic excitation. In the present study we addressed whether the Gq-coupled receptors of the bombesin family can couple to GIRK channels and examined the mechanism by which this process occurs. Different combinations of GIRK channel subunits (Kir3.1, Kir3.2, and Kir3.4) and bombesin receptors (BB1and BB2) were expressed in Xenopus oocytes. In all combinations tested GIRK currents were reversibly inhibited upon application of the bombesin-related peptides, neuromedin B or gastrin-releasing peptide in a concentration-dependent manner. Incubation of oocytes in the phospholipase C inhibitor U73122 or the protein kinase C (PKC) inhibitors chelerythrine and staurosporine significantly reduced the inhibition of GIRK currents by neuromedin B, whereas the Ca2+ chelator, BAPTA-AM had no effect. The involvement of PKC was further demonstrated by direct inhibition of GIRK currents by the phorbol esters, phorbol-12,13-dibutyrate and phorbol-12-myristate-13-acetate. In contrast, the inactive phorbol ester 4α-phorbol and protein kinase A activators, forskolin and 8-bromo cAMP did not inhibit GIRK currents. At the single-channel level, direct activation of PKC using phorbol ester phorbol-12,13-dibutyrate caused a dramatic reduction in open probability of GIRK channels due to an increase in duration of the interburst interval.
G protein-coupled inwardly rectifying K+ (GIRK) channels are important in regulating cell excitability in both the heart and the central nervous system (CNS; Dascal, 1997). In the CNS, GIRK channels are coupled to many types of inhibitory neurotransmitter receptors via pertussis toxin (PTX)-sensitive G proteins, e.g., γ-aminobutyric acid type B (GABAB), 5-hydroxytryptamine type 1A (serotonin), δ and μ opioid, somatostatin, and adenosine A1 receptors (Andrade et al. 1986; Trussell and Jackson, 1987; Penington et al., 1993; Velimirovic et al., 1995). In agreement with this, Lüscher et al. (1997) recently showed that GIRK channels perform an important role in inhibition of hippocampal neurons using Kir3.2 knockout mice. Furthermore, immunocytochemical staining has revealed the presence of Kir3.1 on presynaptic nerve terminals in many areas of the brain, suggesting a role for GIRK channels in presynaptic inhibition (Morishige et al., 1996; Ponce et al. 1996).
The mechanism by which neurotransmitters interact to activate the GIRK channel is now well established. In the presence of agonist, Gi/Go-coupled receptors catalyze the turnover of trimeric G proteins releasing βγ subunits, which directly bind to the GIRK channel complex, stabilizing phosphatidyl inositol 4,5-bis-phosphate binding, with consequent channel activation (Huang et al., 1997, 1998).
In addition to their role in neuronal inhibition, there is increasing evidence to suggest that GIRK channels have an important function in neuronal excitation induced by Gq-coupled neurotransmitter receptors (Velimirovic et al., 1995; Farkas et al., 1997). However, in contrast to the wealth of knowledge concerning Gi/Go-mediated activation of this channel complex, little is known regarding how Gq-coupled receptors inhibit these channels. In view of this, in the present study we have used the Xenopusoocyte expression system to examine the mechanism by which Gq-coupled receptors of the bombesin family interact with and inhibit GIRK channel activity. Some of these findings have recently been reported to The Physiological Society (Stevens et al., 1998).
Materials and Methods
Molecular Biology.
Rat Kir3.1 and Kir3.2 cDNAs engineered into pBG7.2 were linearized using NdeI, whereas rat bombesin receptor (BB1 and BB2) cDNAs engineered into pBluescript SK (Invitrogen, Netherlands) were linearized using BamHI and Spe1, respectively. Capped cRNAs were transcribed in vitro from linearized cDNAs using T7 polymerases (Promega, Southampton, UK).
Isolation of Xenopus Oocytes.
Xenopus laevis were anesthetized by immersion in 0.3% (w/v) 3-amino benzoic acid (Sigma, Poole, U.K.) and ovarian lobes were removed. Oocytes were dissociated using 0.3% (w/v) collagenase (Sigma) in Ca2+-free solution (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl2, and 5 mM HEPES, pH 7.6). Prepared oocytes were microinjected with 50 nl of cRNAs dissolved in water. Oocytes were incubated at 18°C in ND96 (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES).
Two-Electrode Voltage Clamp Recording.
Two-electrode voltage clamp recordings were performed 3 to 6 days after microinjection of cRNAs using a GeneClamp 500 amplifier (Axon Instruments, Burlingame, CA) interfaced to a Digidata 1200 A/D board with Clampex software (version 6, Axon Instruments) and recorded on DAT (Sony, Toyoko, Japan). Records were replayed on a Gould TA240 chart recorder. Oocytes were continually perfused with standard recording solution (90 mM KCl, 1 mM MgCl2, 1 mM CaCl2, and 1 mM HEPES, pH 7.4). Microelectrodes filled with 3 M KCl had resistances between 0.5 and 2 MΩ. Ten-millivolt hyperpolarizing steps were applied every second from a holding potential of −80 mV. Currents were filtered at 1 kHz. Data were analyzed using Clampfit (version 6; Axon Instruments) and Prism (version 2; GraphPad Software, San Diego, CA). Oocytes were preincubated with chelerythrine, staurosporine, U-73122 ({1-[6-((17β-3-Methoxyestra-1,3,5(10)-trien-17-yl)amino)hexyl]-1H-pyrrole-2,5-dione}), U73343 ({1-[6-((17β-3-Metho-xyestra-1,3,5(10)-trien-17-yl)amino)hexyl]-2,5-pyrrolidinedione} or BAPTA-AM (1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid acetoxymethyl ester) for 2 h. Statistical analyses were performed using a Student’s unpaired t test. Data were assumed to be normally distributed and a significant difference was accepted when the two-tailed P value was less than 0.01. Averaged data are presented as means ± S.E.M. Dose-response data were fitted with the equation:
Single-Channel Recording.
Single-channel recordings were performed using an Axopatch 200B patch clamp amplifier (Axon Instruments). Currents were recorded at 10 kHz and filtered at 2 kHz. The electrode solution contained: 140 mM KCl, 2 mM MgCl2, 1.8 mM CaCl2, and 0.1 mM GdCl3, pH 7.2, whereas the bathing solution contained: 140 mM KCl, 3 mM MgCl2, 5 mM EGTA, and 5 mM HEPES, pH 7.2. Patch electrodes were coated with dental wax and had resistances between 1 to 2 MΩ. Data was acquired using Fetchex (versin 6; Axon Instruments) and analyzed using TAC and TACfit (version 5; Bruxton Corp., Seattle, WA). All experiments were performed at 22 to 24°C.
All events were detected using the “50% threshold” technique and were visually inspected before being accepted. Events shorter than 100 μs were excluded from the histograms. Open and closed duration histograms were binned logarithmically and the distributions were fitted by a sum of exponential probability density functions using the maximum likelihood method.
The critical closed time (tcrit) is the duration below which closed times are considered to occur within a burst rather than between bursts and was used to define burst lengths. tcrit was calculated by the method of Colquhoun and Sigworth (1995) using the equation:
Open probability (P o) was calculated over a 5-min duration before and after application of phorbol-12-myristate-13-acetate (PMA).
Results
Inhibition of Kir3 by Bombesin Receptors.
Oocytes coinjected with GIRK channel and bombesin receptor cRNA were recorded using two-electrode voltage clamp. Oocytes expressing heteromeric Kir3.1-Kir3.2 channels and the BB1 receptor gave rise to Ba2+-sensitive currents at −80 mV in standard recording solution. A 2-min application of the preferred BB1 agonist, NMB, inhibited the GIRK current in a concentration-dependent manner that was slowly reversible on washout (Fig. 1, a, b, and d). Voltage ramps applied from −120 mV to +40 mV before and after NMB application, suggested that NMB reduced current amplitude, but not rectification properties (Fig. 1b). In initial experiments, the time course of GIRK channel inhibition by NMB was obscured due to the activation of an endogenous Ca2+-dependent Cl− current (Barish, 1983; Fig. 1a) at NMB concentrations above 50 pM (Fig. 2a–c). However, preincubation of oocytes with the membrane-permeant Ca2+ chelator BAPTA-AM abolished this Cl− conductance at all tested NMB concentrations while having no effect on the inhibition of the GIRK current (Fig. 1c and Fig. 3). Under these conditions, inhibition of the GIRK current was preceded by a small activation, which could be due to either residual Ca2+-activated C− currents or promiscuous coupling of the bombesin receptor to G proteins of the Gi/Go family. Using this approach, it was possible to construct a concentration-inhibition curve for NMB. Fitting these data to eq. 1 gave a Hill coefficient of 1.09 ± 0.25 and an IC50 of 5.4 ± 0.04 pM (Fig. 2a).
The effects of BB1 receptor expression on the amplitude of the GIRK current in the absence of agonist was investigated by comparing oocytes expressing Kir3.1-Kir3.2 with and without the BB1 receptor. GIRK currents in oocytes coexpressing Kir3.1-Kir3.2 and BB1receptor were not significantly different from oocytes expressing Kir3.1-Kir3.2 alone (6.9 ± 0.5 μA; n = 7, and 6.8 ± 0.8 μA; n = 6; at −80 mV), suggesting that the unstimulated BB1 receptor cannot catalyze the turnover of G proteins in the absence of agonists.
Currents flowing through homomeric Kir3.2 and heteromeric Kir3.1-Kir3.4 channels were also reversibly inhibited by activation of the BB1 receptor by 10 nM NMB (Fig. 1d). In addition, the related bombesin receptor, BB2 also inhibited Kir3.1-Kir3.2 currents in the presence of the preferred BB2 agonist, gastrin-releasing peptide (GRP; 10 nM) (Fig. 1d).
The most common GIRK channel in the CNS is the Kir3.1-Kir3.2 heteromultimer (Lesage et al., 1995; Liao et al., 1996) whose distribution overlaps with BB1 receptor in many areas of the brain (Wada et al. 1991; Ladenheim et al., 1992; Karschin et al. 1996; Murer et al., 1997). In attempt to investigate the interaction between these bombesin receptors and GIRK channels with relevance to the CNS, the Kir3.1-Kir3.2 and BB1receptor combination was chosen for more extensive study.
Inhibition Is Mediated by Protein Kinase C (PKC).
Bombesin receptors have previously been shown to couple to phospholipase C (PLC) via Gq in Xenopus oocytes, which activates both the inositol 1,4,5-triphosphate (IP3) and PKC pathways (Shapira et al., 1994). To determine whether PLC performed a role in inhibition of GIRK currents via interaction with the BB1 receptor, oocytes were preincubated in the PLC inhibitor U73122. GIRK current inhibition was significantly reduced by U73122, whereas the inactive analog, U73343 had no effect (Fig. 3).
To examine whether IP3-mediated release of Ca2+ from intracellular stores was involved in the present sequence of events, oocytes were preincubated in BAPTA-AM. However, as shown previously, this process had no effect on the ability of NMB to inhibit the GIRK current (Fig. 3), suggesting that Ca2+ release stimulated by IP3 is not involved in this coupling.
Because Kir3 subunits possess multiple consensus PKC phosphorylation sites, the role of PKC in channel inhibition was tested using PKC activators and inhibitors. Preincubation of oocytes in the PKC inhibitors staurosporine or chelerythrine before electrophysiological recording significantly reduced the ability of NMB to inhibit GIRK currents compared with untreated oocytes (Fig. 3). To verify that PKC is capable of GIRK current inhibition, the effects of phorbol esters on GIRK current amplitude was examined. Continuous application of 1 μM phorbol-12,13-dibutyrate (PDBu) or 1 μM PMA inhibited the GIRK current, which was irreversible over a 30-min period (n= 5) (Fig. 4, a, b, and d). As noted for PKC-mediated modulation of ion channels in other native cells and heterologous expression systems (Walsh and Kass, 1988; Henry et al., 1996; Shuba et al., 1996), the inhibitory action of PMA was maximal at 100 nM, because an increase in PMA concentration to 1 μM caused no further increase in inhibition (Fig. 4d). In contrast, the inactive phorbol ester 4α-phorbol (1 μM) had no effect on the magnitude of this current (Fig. 4, c and d), suggesting that the effect of PMA is mediated via PKC and not via some direct interaction with the channel complex. Activation of PKC by PMA was further demonstrated by incubation of oocytes in 10 μM staurosporine, which significantly blocked the inhibitory action of 100 nM PMA on GIRK currents (relative inhibition was 0.11 ±0.05 at −80 mV, n = 6). The protein kinase A (PKA) activators forskolin (1 μM) and 8-bromo cAMP (1 μM) had no significant inhibitory effect on the GIRK current (Fig.4d).
Single-channel Properties in Presence of PMA.
Single-channel recordings were performed in the cell-attached recording configuration. Inhibition of the GIRK current by the BB1 agonist NMB was of insufficient duration (<10 min) for complete single-channel analysis due to desensitization of the receptor (Corjay et al., 1991). Therefore, PKC-mediated inhibition of the GIRK current with PMA was used to examine the effect of PKC activation on GIRK channel activity. The single-channel properties of the Kir3.1-Kir3.2 complex were compared before and after treatment with PMA. Under control conditions, GIRK currents were found to exhibit a unitary amplitude of 2.71 ± 0.08 pA at −80 mV (n = 3). As previously reported in both native and heterologous expression systems, GIRK channel activity was characterized by bursting behavior (Grigg et al., 1996; Luchian et al., 1997), i.e., clusters of openings separated by brief closures (Fig. 5, a and b).
Following bath application of 100 nM PMA, there was a rapid reduction in channel open probability, Po, from 2.7 ×10−2 ± 8.1 × 10−3 to 8.6 × 10−4± 9.4 × 10−5 (n = 3). This reduction in channel activity was not associated with any change in single-channel current amplitude (2.63 ± 0.05 pA at −80 mV). Further kinetic analysis of channel activity revealed that PMA had no significant effect on either the open times (control, τf = 0.28 ± 0.03 ms, τs = 1.39 ± 0.05 ms; PMA, τf = 0.32 ± 0.04 ms, τs = 1.27 ± 0.33 ms, n = 3) or burst lengths (control, τf = 0.32 ± 0.04 ms, τs = 2.20 ± 0.05 ms; PMA, τf = 0.33 ± 0.03 ms, τs = 1.83 ± 0.27 ms, n = 3) of the GIRK channel complex (Fig. 5, a, b, and c). In contrast, PMA was found to dramatically alter channel closed times (Fig. 5d). In the absence of PMA, closed times were described by two exponential components with time constants of 0.42 ± 0.01 ms and 57.43 ± 21 ms (n = 3). The first component represents channel closing within a burst, whereas the second component represents channel closing between bursts. In the presence of PMA, closed times were again described by two exponentials with time constants of 0.43 ± 0.11 ms and 1499 ± 464 ms (n = 3). Thus, PMA-induced inhibition of the channel results in an increase in duration of the burst interval.
Discussion
This study demonstrates that activation of Gq-coupled bombesin receptors using the bombesin-related peptides NMB or GRP reversibly inhibits GIRK currents in Xenopus oocytes. The BB1 receptor has previously been shown to activate phospholipase C via a Gq-mediated pathway in oocytes (Shapira et al., 1994; Woodruff et al., 1996). This results in production of diacylglycerol and IP3. IP3releases Ca2+ from intracellular organelles, whereas diacylglycerol activates PKC (Sternweis et al., 1992). The involvement of PLC in coupling between BB1receptor and GIRK currents was demonstrated using the specific PLC inhibitor U73122, which decreased the degree of inhibition of GIRK currents in response to NMB. The increase in intracellular Ca2+ after binding of NMB to BB1 receptor is clearly seen in the present study by the activation of the endogenous oocyte Ca2+-dependent Cl− channel (Barish, 1983). Treatment of oocytes with the Ca2+ chelator, BAPTA-AM, abolishes this Cl− current, but has no effect on inhibition of GIRK current, suggesting that an increase in intracellular Ca2+ concentration is not involved in modulation of GIRK channel activity. In contrast, we demonstrate that PKC performs an important role in the modulation of GIRK channel activity using both PKC inhibitors and activators. The PKC inhibitors staurosporine and chelerythrine reduce the amount of GIRK current inhibition in response to NMB, whereas the PKC activators, the phorbol esters PMA and PDBu, induce inhibition of GIRK channel activity in the absence of receptor activation. The specificity of PDBu and PMA for PKC has been demonstrated using the inactive phorbol ester 4α-phorbol, which does not inhibit the GIRK current.
Previous studies on interactions between GIRK channels and the Gq/G11-coupled metabotropic glutamate receptors have given conflicting results. Saugstad et al. (1996) reported that type 1a metabotropic glutamate receptors (mGluR1a) activate Kir3.1-Kir3.4 heteromeric channels by coupling to PTX-sensitive G proteins, whereas Sharon et al. (1997) reported that mGluR1a receptors inhibit Kir3.1-Kir3.2 heteromeric channels by coupling to PTX-insensitive G proteins. Sharon et al. (1997) also showed that Kir3.1-Kir3.2 channels display an initial weak activation before the strong inhibition due to promiscuous coupling of the mGluR1a receptor to PTX-sensitive G proteins. In agreement with the present study, Sharon et al. (1997) showed that inhibition of the GIRK channel complex occurs through a PKC-dependent mechanism.
In the present study, single-channel recordings in the presence of PMA provide an insight into the mechanism of channel inhibition. Single-channel conductance, open-time distributions, and burst lengths remain unaffected, whereas the interburst interval is altered.Ivanova-Nikolova and Breitweiser (1997) reported that βγ binding changes the proportion of channels within three bursting modes, but does not change the burst duration. Termination of bursting occurs independently of βγ binding and has been shown to be caused by a C-terminal inactivation particle blocking the pore (Luchian et al., 1997). Because intraburst kinetics were unaffected by PKC activation in the present study, inhibition is unlikely to occur through a change in βγ binding. However, an increased interburst duration suggests that PKC activation stabilizes the inactivated state.
Although there is no direct evidence that phosphorylation by PKC acts directly on the GIRK channel rather than an accessory protein, Kir3.1, Kir3.2, and Kir3.4 contain consensus PKC sites on both the N and C termini of the channel subunits. Another inward rectifier, Kir2.3, has been shown to couple to the Gq-linked muscarinic M1 receptor and is inhibited by the phorbol esters PMA and PDBu (Henry et al., 1996), suggesting the involvement of PKC. Because Kir3 subunits and Kir2.3 possess a consensus PKC site at identical positions in the N terminus, it is possible to speculate that phosphorylation of this site may be involved in the channel inhibition process.
In order for GIRK channels to be a target for inhibition by coupling to Gq-linked receptors, a proportion of the channels must be activated. GIRK channels have been shown to display tonic activity in hippocampal neurons. Using patch clamp recording from brain slices, the membrane potentials of Kir3.2 knockout mouse CA1 neurons is 8 mV more depolarized than wild-type mouse CA1 neurons (Lüscher et al., 1997). Sodickson and Bean (1996) showed that dissociated rat hippocampal CA3 neurons give rise to GABAB-activated GIRK currents (EC50 = 1.6 μM) at extracellular GABA concentrations present basally in the brain (0.2–0.8 μM), suggesting that tonic activity of GIRK channels could arise from tonic activation of receptors. However, because GABAB and A1 antagonists do not affect resting potentials of CA1 neurons in slices, Lüscher et al. (1997) suggest that tonic activity of GIRK channels is independent of receptor activation. Tonic activity of GIRK channels in the absence of agonist has been shown to operate through constitutive turnover of G proteins, releasing βγ, in Xenopus oocytes (Stevens et al., 1997) and through a Na+-gating mechanism in both myocytes and oocytes (Lesage et al. 1995; Sui et al. 1996).
Negative coupling of GIRK channels to Gq-linked receptors has been demonstrated in several neuronal preparations. For example, in cultured rat locus coereleus neurons a GIRK channel activated by somatostatin and [Met]enkephalin has been identified by its single-channel conductance (30 picosiemen) and membrane-delimited activation (Grigg et al., 1996). The same GIRK channel is also inhibited by Substance P via a PTX-insensitive G protein (Velimirovic et al., 1995). Similarly, in GTPγS-loaded locus ceruleus neurons application of somatostatin induced a persistent increase in K+ conductance that was reversed by application of Substance P (Velimirovic et al., 1995). In cultured rat dopaminergic neurons of the midbrain ventral tegmental area a putative GIRK current is activated by dopamine (D2receptors) via PTX-sensitive G proteins and inhibited by neurotensin via PTX insensitive G proteins (Farkas et al., 1997)
The distribution of GIRK channels shares a high degree of overlap with BB1 and BB2 receptors throughout the CNS (Wada et al. 1991; Ladenheim et al., 1992; Murer et al., 1997; Karschin et al. 1996). Furthermore, inhibition of a resting K+ conductance by bombesin and related peptides has been reported in neurons of the rat dorsal raphe and suprachiasmatic nucleus (Pinnock and Woodruff, 1991; Reynolds and Pinnock, 1997). Because GIRK channels have been reported to be coupled to 5-hydroxytryptamine type 1A receptors in dorsal raphe neurons (Penington et al., 1993), it is possible that GIRK channels are the target for inhibition induced by bombesin.
In summary, in the present study we demonstrated unequivocally that Gq-coupled neuropeptides are able to inhibit GIRK channels in a nonmembrane-delimited manner involving PKC. Furthermore, we showed at the single-channel level this inhibition by PKC is mediated by a change in interburst interval of the channel complex. In future studies it will be important to examine the interaction between GIRK channels and Gq-coupled receptors in native neurons to test whether a similar mechanism is involved in postsynaptic excitation.
Acknowledgments
We are grateful to Dr. R. Murrell-Lagnado (Department of Pharmacology, University of Cambridge, United Kingdom) for generously supplying the vectors pBG7.2Kir3.1, pBG7.2Kir3.2, and pBG7.2Kir3.4. We are also grateful to J. F. Battey (Laboratory of Molecular Biology, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland) for providing us with clones of rat BB1 receptors, and to M. D. Hall for subcloning the receptor cDNA into Bluescript.
Footnotes
- Received November 16, 1998.
- Accepted March 9, 1999.
-
Send reprint requests to: Dr. Kevin Lee, Parke Davis Neuroscience Research Centre, Cambridge University Forvie Site, Robinson Way, Cambridge CB2 2QB, United Kingdom. E-mail:Kevin.Lee{at}wl.com
Abbreviations
- GIRK
- G protein-coupled inwardly rectifying K+ channel
- GRP
- gastrin-releasing peptide
- IP3
- inositol 1,4,5-triphosphate
- NMB
- neuromedin B
- PDBu
- phorbol-12,13-dibutyrate
- PKC
- protein kinase C
- PLC
- phospholipase C
- PMA
- phorbol-12-myristate-13-acetate
- PTX
- pertussis toxin
- The American Society for Pharmacology and Experimental Therapeutics