Activation of the Endothelin Receptor Inhibits the G Protein-Coupled Inwardly Rectifying Potassium Channel by a Phospholipase A2-Mediated Mechanism

Authors


  • Lippincott Williams & Wilkins, Inc., Philadelphia

  • The present address of Dr. C. Cyr is Maine Medical Center Research Institute, John Roberts Road, South Portland, ME 04106, U.S.A.

  • Ms. S. L. Rogalski and Dr. C. Cyr contributed equally to this work.

  • Abbreviations used: AACOCF3, arachidonyl trifluoromethyl ketone; DAMGO, [D-Ala2,N-MePhe4,Gly-ol5]-enkephalin; DIDS, 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid; GIRK or Kir 3, G protein-coupled inwardly rectifying potassium channel; HETA, human endothelin A; MOR, μ-opioid receptor; PKC, protein kinase C; PLA2, phospholipase A2; cPLA2, cytosolic PLA2; PLCβ, phospholipase C-β; PMA, phorbol 12-myristate 13-acetate.

Address correspondence and reprint requests to Dr. C. Chavkin at Department of Pharmacology, Box 357280, University of Washington, School of Medicine, Seattle, WA 98195-7280, U.S.A.

Abstract

Abstract: To develop a malleable system to model the well-described, physiological interactions between Gq/11-coupled receptor and Gi/o-coupled receptor signaling, we coexpressed the endothelin A receptor, the μ-opioid receptor, and the G protein-coupled inwardly rectifying potassium channel (Kir 3) heteromultimers in Xenopus laevis oocytes. Activation of the Gi/o-coupled μ-opioid receptor strongly increased Kir 3 channel current, whereas activation of the Gq/11-coupled endothelin A receptor inhibited the Kir 3 response evoked by μ-opioid receptor activation. The magnitude of the inhibition of Kir 3 was channel subtype specific; heteromultimers composed of Kir 3.1 and Kir 3.2 or Kir 3.1 and Kir 3.4 were significantly more sensitive to the effects of endothelin-1 than heteromultimers composed of Kir 3.1 and Kir 3.5. The difference in sensitivity of the heteromultimers suggests that the endothelin-induced inhibition of the opioid-activated current was caused by an effect at the channel rather than at the apioid receptor. The endothelin-1-mediated inhibition was mimicked by arachidonic acid and blocked by the phospholipase A2 inhibitor arachidonoyl trifluoromethyl ketone. Consistent with a possible phospholipase A2-mediated mechanism, the endothelin-1 effect was blocked by calcium chelation with BAPTA-AM and was not affected by kinase inhibition by either staurosporine or genistein. The data suggest the hypothesis that Gq/11-coupled receptor activation may interfere with Gi/o-coupled receptor signaling by the activation of phospholipase A2 and subsequent inhibition of effector function by a direct effect of an eicosanoid on the channel.

The conductance of the G protein-coupled inwardly rectifying potassium channel (GIRK or Kir 3) is regulated by signals generated by a large group of different G protein-coupled receptors. Activation of members of the pertussis toxin-sensitive family of Gi/Go-coupled receptors (e.g., opioids, GABAB, somatostatin, serotonin, dopamine, adenosine, and others; Jan and Jan, 1997) releases Gβγ that directly activates Kir 3 (Clapham and Neer, 1997). In contrast, the activation of members of the pertussis toxin-insensitive family of Gq/11-coupled receptors (e.g., substance P and endothelin) inhibit Kir 3 (Jan and Jan, 1997). The mechanism(s) responsible for the Gq/11-mediated inhibition of Kir 3 is less clear. Substance P activates calcium-dependent, arachidonic acid production (Garcia et al., 1994), and this production of arachidonic acid was suggested as the mechanism of Kir 3 inhibition in locus ceruleus (Koyano et al., 1994). In a similar manner, endothelin-1 activates a Gq/11-coupled receptor in glia, astrocytes, and neurons to stimulate phospholipase C-β (PLCβ), causes phosphoinositide hydrolysis, increases intracellular calcium, activates phospholipase A2 (PLA2), and produces arachidonic acid (Rubanyi and Polokoff, 1994). In contrast, others (Kurachi et al., 1992) have reported that arachidonic acid and its metabolites increase Kir 3 current. Recently, metabotropic glutamate receptors (mGluR1a and mGluR5) were found to inhibit Kir 3 expressed in Xenopus oocytes by a pertussis toxin-insensitive mechanism blocked by protein kinase C (PKC) inhibitors (Sharon et al., 1997). Thus, Gq/11-coupled receptors have been shown to affect Kir 3 channel current, but the underlying mechanism(s) remains uncertain.

To further explore how diverse G proteins signaling may interact to control Kir 3 current, we used Xenopus oocytes to coexpress the Gq/11-coupled endothelin A receptor, the Gi/o-coupled μ-opioid receptor, and Kir 3 to model the complex neuronal situation. Opioid receptors are coupled to pertussis toxin-sensitive Gi/o proteins, and receptor activation inhibits adenylyl cyclase, decreases calcium channel current, and increases both inwardly rectifying and delayed rectifying potassium channel currents (Kieffer, 1995). In this study, we found that endothelin receptor activation does reduce the response evoked by opioid receptor activation. The data obtained suggest that PLA2 antagonists blocked the endothelin-induced inhibition.

MATERIALS AND METHODS

Complementary DNA clones and mRNA synthesis

The rat μ-opioid receptor (MOR) clone was obtained from Dr. Lei Yu (GenBank accession no. L13069). cDNA for the human endothelin A (HETA) receptor (GenBank accession no. S67127) was obtained from Dr. Richard Kris. cDNAs for the Kir 3.1 (GIRK1) (GenBank accession no. U01071) and Kir 3.2 (GIRK2) (accession no. U11859) were obtained from Drs. Cesar Lebarca and Henry Lester. The Kir 3.4 (GIRK4) clone was provided by Dr. John Adelman (GenBank accession no. X83584). The Kir 3.5 (XIR) clone was provided by Dr. David Clapham (GenBank accession no. U42207). Plasmid templates for all constructs were linearized before in vitro mRNA synthesis as described (Dascal et al., 1993). cRNA was prepared by in vitro synthesis from cDNA templates using mMessage Machine (Ambion, TX, U.S.A.).

Oocyte maintenance and injection

Healthy stage V and VI oocytes were harvested from mature anesthetized Xenopus laevis (Nasco, Ft. Atkinson, WI, U.S.A.) and defolliculated enzymatically as described previously (Snutch, 1988). The oocytes were maintained at 18°C in standard oocyte buffer, ND96 (96 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES, pH 7.5), supplemented with 2.5 mM sodium pyruvate and 50 μg/ml gentamicin (Sigma Chemical, St. Louis, MO, U.S.A.). One day after harvest, each oocyte was injected with 50 nl of mRNA for the MOR, HETA, and GIRK heteromultimers Kir 3.1, Kir 3.1 and Kir 3.2, or Kir 3.1 and Kir 3.5 (XIR), into the vegetal pole. Recordings were made at least 72 h after cRNA injection.

Electrophysiological studies

An Axoclamp 2A amplifier was used for standard two-electrode voltage-clamp experiments. The pCLAMP program (Axon Instruments) was used for data acquisition and analysis. Oocytes were removed from incubation medium, placed in the recording chamber containing ND96 medium, and clamped at -80 mV. Recordings were made in hK buffer (2 mM NaCl, 16 mM KCl or 24 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES, pH 7.5). Microelectrodes were filled with 3 M KCl and had resistances of 0.5-1.0 MΩ. Currents were measured without leak subtraction. Individual comparisons of drug effects on Kir 3 were conducted using oocytes from the same harvest and injection batch. Pharmacological agents were perfused directly into the bath from freshly made stock solutions.

Materials

Stock solutions of arachidonic acid, BAPTA-AM [1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetraacetoxymethyl ester], and phorbol esters were dissolved in dimethyl sulfoxide; the final concentration of dimethyl sulfoxide applied to the oocytes was ≤0.02%. Arachidonyl trifluoromethyl ketone (AACOCF3) was dissolved in ethanol; the final concentration of ethanol applied to the oocytes was <0.01%. Nitrogen was bubbled through water before dissolving endothelin. All other drugs were dissolved in water. Alkaline phosphatase was obtained from New England Biolabs, Beverly, MA, U.S.A.; [D-Ala2,N-MePhe4,Gly-ol5]-enkephalin (DAMGO) was obtained from Peninsula Laboratories, Belmont, CA, U.S.A.; endothelin-1 (human and porcine) was obtained from Research Biochemicals International, Natick, MA, U.S.A.; arachidonic acid, AACOCF3, BAPTA-AM, and 4,4′-diisothiocyanatostil-bene-2,2′-disulfonic acid (DIDS) were from Calbiochem, La Jolla, CA, U.S.A. Quinacrine, genistein, staurosporine, phorbol 12-myristate 13-acetate (PMA), and dimethyl sulfoxide were from Sigma.

Statistical analysis

Data are presented as mean ± SEM values. The statistical significance of differences between results was calculated using a paired t test. A probability of p < 0.05 was considered statistically significant.

RESULTS

Effects of endothelin-1 on the DAMGO-elicited μ-opioid response mediated by the heteromultimer Kir 3.1 and Kir 3.2 expressed in Xenopus oocytes

Application of the μ-opioid agonist DAMGO (100 nM) in oocyte buffer containing 16 mM KCl increased the inward current (Fig. 1A). Previous in vitro expression studies in Xenopus oocytes showed that the opioid-induced current was caused by activation of the expressed GIRK (Dascal et al., 1993; Chen and Yu, 1994), and this was confirmed by current-voltage analysis of the opioid-activated current (data not shown). After the apioid-induced response reached steady state, oocytes were perfused with ND96 buffer for 4 min. Oocytes receiving a second application of 100 nM DAMGO showed a similar response to DAMGO (Fig. 1A); the amplitude of the second opioid response was 99 ± 2.9% (n = 41) of the first response (Fig. 1B). The opioid response was only observed in oocytes injected with both receptor and channel mRNA.

Figure 1.

Effects of endothelin-1 on the DAMGO-elicited MOR mediated by the heteromultimer Kir 3.1 and Kir 3.2 expressed in Xenopus oocytes. Oocytes were injected with a mixture of the following mRNAs: 0.8 ng MOR, 1.7 ng HETA receptor, 0.065 ng Kir 3.1, and 0.065 ng Kir 3.2, as summarized in the inset (circle). Recordings were performed 3-4 days after injection. Initially, oocytes were bathed in normal oocyte saline buffer containing 2 mM KCl and clamped at -80 mV. Oocytes were then perfused with a saline buffer containing 16 mM KCl. Elevation in potassium concentration facilitates current measurement by increasing the basal current through the inwardly rectifying potassium channels. Application of the μ-opioid agonist DAMGO (100 nM) in oocyte buffer containing 16 mM KCl increased the basal current. After the response reached steady state, normal oocyte saline buffer was washed on for 4 min. Buffer flow was decreased while an application of 25 μl of 100 nM endothelin-1 was applied directly to the oocyte. After the endothelin-1-induced calcium-activated chloride current returned to baseline, the bath flow was resumed for 4 min followed by a second treatment with 100 nM DAMGO in 16 mM KCl. Measurement of endothelin-induced inhibition of the second opioid response is expressed as a percentage of the first peak opioid response. A: Representative current traces from oocytes stimulated by perfusion with the μ-opioid agonist DAMGO (100 nM) in oocyte buffer containing 16 mM KCl in two applications. The first panel shows the response without endothelin-1. The second panel includes a 25-μl pulse of 100 nM endothelin-1 in ND96 applied directly to the oocyte in the recording chamber. Application of vehicle had no effect and the responses to endothelin were absent in oocytes not expressing the receptor. For these traces, oocytes were clamped at -80 mV in normal saline buffer. Horizontal bars indicate the duration of drug perfusion. B: The amplitude of the second opioid receptor response was expressed as a percentage of the first response. Columns represent mean ± SEM values of 22 oocytes from three batches.

FIG. 1.

Prior studies showed that the endothelin A receptor couples to PLC through G protein βγ subunits (Camps et al., 1992), thereby increasing intracellular calcium and thus activating an endogenous calcium-dependent chloride current (Shimada et al., 1991; Cyr et al., 1993) intrinsic to the oocyte (Barish, 1983) (Fig. 1A, right). Oocytes treated with endothelin-1 before the second DAMGO challenge showed a marked inhibition of the second opioid response. The amplitude of the second opioid response after endothelin-1 treatment was 33 ± 3.2% (n = 43) of the first opioid response (p < 0.01) (Fig. 1B). The endothelin-induced activation of the endogenous chloride current was only observed in oocytes injected with the endothelin A receptor mRNA.

Endothelin-1-induced inhibition of the opioid-activated potassium current was dose dependent. At 50 or 100 nM, endothelin-1 produced a large activation of the chloride current (4.1 ± 0.8 μA; n = 28). Moreover, the inhibition of the second opioid response produced by 100 nM endothelin-1 was the same as that produced by 1.0 μM endothelin-1 (n = 20). The mechanism of the endothelin-1 modulation of the opioid response was not clear and may involve either the action of the second messengers diacylglycerol or inositol trisphosphate (Shubeita et al., 1990), the elevation of intracellular calcium, or metabolites produced after the activation of PLA2 (Nishizuka, 1984). These possibilities were next explored.

Effects of anionic channel block

Because intracellular chloride has been reported to regulate G protein-gated channels in some cell types (Jentsch and Gunther, 1997), we tested if the chloride flux mediated by the calcium-activated chloride channel produced the observed inhibition of the opioid response. The chloride channel blocker DIDS (100 nM) (Tzounopoulos et al., 1995) was perfused into the recording chamber before endothelin-induced activation. Although the calcium-activated chloride current was significantly attenuated or blocked by DIDS, endothelin-1 still effectively inhibited the opioid response (Fig. 2). In oocytes pretreated with DIDS, the second opioid response was inhibited by 82 ± 10% (n = 8; p > 0.05) compared with the first DAMGO response, an effect not significantly different from that in the absence of DIDS.

Figure 2.

Effects of pretreatment with DIDS, staurosporine (Stauro), EGTA, BAPTA-AM, or AACOCF3 on endothelin-1 (Et-1) inhibition of the MOR response. The mechanism of the endothelin inhibition of the DAMGO response was determined using a series of pharmacological blockers of endothelin-activated second messenger systems. To normalize for daily differences in receptor expression, data are presented as percentages of the average maximal inhibition of the second opioid response produced by endothelin in control oocytes (100% equals maximum inhibition of the second DAMGO response for that batch of oocytes). Each group was compared with its own control in that day’s matched group. Aqueous EGTA was injected into the oocyte (46 nl of a 50 mM solution was injected to yield a final concentration of ∼2.5 μM) 4 min before recording. EGTA blocked the endothelin-activated chloride current, increased the basal Kir 3 current, but did not block the endothelin-induced inhibition of the second DAMGO response. The nonspecific anion channel blocker DIDS (100 nM) dissolved in ND96 was perfused for 4 min after the first peak response to DAMGO, as in Fig. 1. DIDS blocked the endothelin-activated increase in the chloride current but did not affect the endothelin-induced inhibition of the second DAMGO response. Oocytes were presoaked for 90-150 min with the membrane-permeant calcium chelator BAPTA-AM (50-100 μM) before recording. BAPTA-AM also blocked the endothelin-activated chloride current and blocked the endothelin-induced inhibition of the second DAMGO response. Oocytes were presoaked for 30 min with 15 μM AACOCF3 before recording. Data are mean ± SEM values. *p < 0.05; **p < 0.01.

FIG. 2.

Effects of calcium chelation

To determine if the observed inhibition of the opioid response was mediated by the endothelin-1-induced rise in cytoplasmic calcium, the effect of calcium chelators was next determined. In oocytes injected with EGTA (final intracellular concentration, ∼2.5 μM), the chloride channel response to endothelin was abolished (n = 6); however, the endothelin-1-induced inhibition of the opioid response was only slightly reduced. In oocytes pre-injected with EGTA, endothelin-1 caused an 84 ± 9% (n = 6) inhibition of the DAMGO response. This endothelin-1 effect was not significantly different from that observed using matched oocytes not injected with EGTA. To assess the result of a more effective calcium chelation, BAPTA-AM, the membrane-permeant analogue was used (Dieter et al., 1993). In oocytes soaked for 90-150 min in BAPTA-AM (50-100 μM) as previously described (Quamme, 1997), the chloride current was reduced by >90% (n = 12; p < 0.01). In BAPTA-AM-treated oocytes, the application of DAMGO after endothelin-1 produced only a 28 ± 6% (n = 12) response compared with the first DAMGO response (Fig. 2). Neither EGTA nor BAPTA-AM treatment altered the amplitude of the first opioid response to DAMGO, confirming that elevation of intracellular calcium was not required for the opioid-activated increase in potassium current. The greater effectiveness of BAPTA-AM was consistent with its more effective chelation of calcium in cytosolic and vesicular stores compared with EGTA (Tsien, 1980) and with the higher concentration of chelator achieved.

Effects of inhibition of kinase and phosphatase activity

To determine if the inhibition of the opioid response was the result of the action of an endothelin-activated kinase, oocytes were treated with the nonspecific kinase inhibitor staurosporine. Staurosporine (1 μM) treatment had no effect on resting membrane current, and the amount of endothelin-1 inhibition of the second opioid response was not significantly different in staurosporine-treated oocytes (p > 0.05) compared with control (108 ± 6%; n = 12) (Fig. 2). To validate the negative result obtained with staurosporine, control oocytes were perfused with 100 nM PMA for 15 min between DAMGO applications. PMA produced significant inhibition of both the basal and the opioid responses as previously reported (Dascal, 1997), and the PMA-induced inhibition was blocked by pretreatment with 1 μM staurosporine (n = 15). Treatment with the tyrosine kinase inhibitor genistein (Akiyama et al., 1987) (100 μM) also had no significant effect on the amount of endothelin-induced inhibition of the second opioid response (116 ± 8% of control; n = 5). Neither the magnitude of the endothelin-1 effect nor the second opioid response was different compared with control oocytes (p > 0.05). To determine if the endothelin-1 effects could be blocked by reduction in protein phosphorylation, alkaline phosphatase was injected into oocytes 30 min before recording (0.025 U/ml final concentration). Alkaline phosphatase did not effect the inhibition of the opioid response (99 ± 22% of control; n = 6). The data failed to support the hypothesis that a kinase or phosphatase directly mediated endothelin-induced channel inhibition. However, the negative data do not exclude the possibility of kinase or phosphatase regulation.

Effects of block of PLA2

To determine which calcium-dependent signal transduction pathway leads to endothelin-1-induced inhibition of the opioid response, oocytes were next treated with either quinacrine or AACOCF3 (Street et al., 1993), two inhibitors of PLA2. PLA2 activity leads to the direct release of arachidonic acid from phospholipids, which results in the formation of several biologically active arachidonate metabolites. In oocytes soaked in 15 μM AACOCF3 for 30 min, the inhibitory effects of 100 nM endothelin-1 were significantly reduced (Fig. 2); the endothelin-1 effect was only 57 ± 3.1% (n = 7) of the maximal effect in the absence of AACOCF3. As expected, AACOCF3 produced a more dramatic block of the effects of 25 nM than 100 nM endothelin-1 (Fig. 2). Pretreatment with AACOCF3 alone did not affect the opioid-activated Kir 3 response (n = 7). Quinacrine (50 μM) also blocked the endothelin-1 effect, but the interpretation of quinacrine’s actions was confounded by its direct inhibitory effects on Kir 3 current (data not shown). In addition, the endothelin-1 activation of the calcium-sensitive chloride channel was unchanged by the presence of AACOCF3. The chloride current evoked by 25 nM endothelin treatment was 930 ± 304 nA (n = 4), and was 940 ± 340 nA (n = 4) in the presence of 15 μM AACOCF3. These data suggest that AACOCF3 did not block PLC or inositol trisphosphate-mediated calcium release from cytoplasmic stores.

Arachidonic acid effects on GIRK heteromultimers

The partial reversal of the endothelin-1 effects by inhibition of PLA2 suggested that a product of PLA2 metabolism might be responsible for channel inhibition. Endothelin receptors have been shown to stimulate arachidonic acid synthesis (Rubanyi and Polokoff, 1994). Moreover, ion channels, including Cl- (Anderson and Welsh, 1990), delayed rectifier K+ (Villarroel and Schwarz, 1996), and the inward rectifier in cardiac myocytes (Kurachi et al., 1992), are targets of arachidonic acid action. To determine if a product of PLA2 metabolism reproduced the observed inhibition caused by endothelin-1, 20 μM arachidonic acid was perfused for 6 min before activation of the second opioid response. In oocytes expressing the channel heteromultimer composed of Kir 3.1 and Kir 3.2, arachidonic acid produced an inhibition of the second opioid response; the second DAMGO response was 61 ± 3.9% of the first response (p < 0.05) (Fig. 3). It is interesting that for oocytes expressing the channel heteromultimer composed of Kir 3.1 and Kir 3.5, arachidonic acid produced a significantly smaller inhibition. In matched oocytes treated with 20 μM arachidonic acid, the second opioid response was 88 ± 4.4% of first response (Fig. 3). Arachidonic acid did not affect the basal channel currents for either Kir 3.1-2 or Kir 3.1-5 combinations (Fig. 3).

Figure 3.

Arachidonic acid effects on GIRK heteromultimers. Oocytes were injected with a mixture of the following mRNAs: 0.8 ng MOR, 0.065 ng Kir 3.1, and 0.065 ng Kir 3.2, or 0.8 ng MOR, 0.13 ng Kir 3.1, and 0.13 ng Kir 3.5. Comparative RNA dose—response studies were used to define the minimum concentrations required to elicit similar currents with each combination. Initially, oocytes were bathed in normal oocyte saline buffer containing 2 mM KCl and clamped at -80 mV. Oocytes were then perfused with a saline buffer containing 16 mM KCl. Each column represents the amplitude of the second response (after arachidonic acid application). Arachidonic acid did not affect the basal responses of either Kir 3 channel. Application of the μ-opioid agonist DAMGO (100 nM) in oocyte buffer containing 16 mM KCl increased the current. The peak opioid response to 100 nM DAMGO was followed by perfusion with 20 μM arachidonic acid for 6 min at a flow rate of 4-5 ml/min and a second stimulation with DAMGO. Oocyte membrane stability was not maintained with concentrations higher than 20 μM arachidonic acid (data not shown); thus, higher doses of arachidonic acid were not evaluated. The second μ-opioid response is expressed as a percentage of the first response. Columns represent mean ± SEM values of 20 oocytes from four donors.

FIG. 3.

Endothelin effects on Kir 3 heteromultimers

As shown above, we found that endothelin-1 significantly inhibited the opioid response in oocytes expressing channel heteromultimers composed of Kir 3.1 and Kir 3.2. To determine if the magnitude of the endothelin effect also depended on the subunit composition of the Kir 3, oocytes from the same batch were injected with either the combination of (1) Kir 3.1 and Kir 3.2, (2) Kir 3.1 and Kir 3.4, or (3) Kir 3.1 and Kir 3.5. Channel expression was controlled so that the amplitudes of the both basal Kir 3 currents and the opioid-activated responses were similar among the groups by adjusting the RNA doses injected. For the combination of Kir 3.1 and Kir 3.2, the second opioid response after endothelin-1 treatment, expressed as a percentage of first response, was 11 ± 3.1% (n = 10) (Fig. 4). In a similar manner, for the heteromultimer consisting of Kir 3.1 and Kir 3.4, the second opioid response after endothelin-1 treatment, expressed as a percentage of first response, was 7.6 ± 2.8% (n = 12). For the heteromultimer consisting of Kir 3.1 and Kir 3.5, the second opioid response after endothelin-1 was 40.8 ± 7.4% (n = 9). Expression of Kir 3.1 alone in Xenopus oocytes showed a reduced response to endothelin-1 treatment, with the second opioid response 37 ± 6.3% (n = 7) of the first response (Fig. 4). The marked differences in endothelin-1 sensitivity among the channel heteromultimers indicated that endothelin-1 inhibition of Kir 3 was channel subtype specific. Furthermore, the observed differences in Kir 3 sensitivity to endothelin-1 suggest that endothelin effects may be directly on the channel. As endothelin-1 and arachidonic acid both produce an inhibition of the channel current that depends on the channel subtype expressed, the observed inhibition of the opioid-activated response may occur by a similar mechanism. These observations are consistent with the observation that PLA2 antagonists inhibited endothelin-1 effects.

Figure 4.

Endothelin effects on GIRK heteromultimers. The experiment was performed as in Fig. 1. All data are presented as comparisons between the second peak opioid response and the first peak opioid response. Oocytes were injected with 0.8 ng MOR, 1.7 ng HETA receptor, 0.065 ng Kir 3.1, and 0.065 ng Kir 3.2, or 0.8 ng MOR, 1.7 ng HETA receptor, 0.05 ng Kir 3.1, and 0.05 ng Kir 3.4, or 0.8 ng MOR, 1.7 ng HETA receptor, 0.13 ng Kir 3.1, and 0.13 ng Kir 3.5, or 2.6 ng MOR, 2.6 ng HETA receptor, and 8 ng Kir 3.1. All data are presented as comparisons between the second peak opioid response and the first peak opioid response. n = 15 from three batches.

FIG. 4.

DISCUSSION

We expressed the MOR, the HETA receptor, and Kir 3 heteromultimers in Xenopus oocytes, and used two-electrode voltage-clamp recording techniques to characterize common signals that allow diverse G protein-coupled receptors to regulate Kir 3. Treatment of oocytes with the μ-opioid agonist DAMGO increased Kir 3 current. Oocytes treated with endothelin-1 before a second DAMGO challenge showed a marked inhibition of the second opioid response. Although alternatives have not been excluded, our data suggest that the endothelin inhibition of Kir 3 was mediated by calcium-dependent PLA2. PLA2 antagonists blocked the endothelin effect. Endothelin-1 had different effects on Kir 3 heteromultimer combinations that paralleled the differences in channel sensitivity to arachidonic acid. The results obtained suggest that activation of the endothelin receptor activates PLC, thereby increasing cytosolic calcium, and thus increasing PLA2 activity. PLA2 may produce an arachidonic acid metabolite that directly blocks Kir 3 current.

As protein kinases and phosphatases also regulate neuronal excitability, we used pharmacological inhibitors to determine if a phosphatase or kinase mediated the endothelin-induced inhibition or Kir 3. The nonspecific kinase inhibitor staurosporine, the tyrosine kinase inhibitor genistein, and phosphatase inhibitor alkaline phosphatase were used to block intracellular kinase and phosphatase activity. The data suggested that a kinase or phosphatase did not directly mediate channel inhibition. We and others (Sharon et al., 1997) note that treatment of the oocytes with the phorbol ester PMA, to activate PKC, mimicked the endothelin-induced inhibition of Kir 3. In our study, PMA inhibition of Kir 3 was reversed by staurosporine, but pretreatment with staurosporine, but pretreatment with staurosporine did not block the endothelin-induced inhibition of the opioid response. We suggest that PKC is not directly involved in the endothelin-induced inhibition of the opioid response. However, the data do not exclude a possible role of kinase or phosphatase regulation of Kir 3.

Several studies using phorbol esters and inhibitors of PKC provide evidence that PKC may be involved in endothelin-induced nuclear signaling mechanisms in other cell types (Rubanyi and Polokoff, 1994). AT-1 cells treated with endothelin-1 were shown to be selective for the calcium-insensitive isoform PKCε. In this system, endothelin stimulation of phosphoinositide hydrolysis increased the membrane association of PKCε and increased cytosolic calcium in a subpopulation of cells (Jiang et al., 1996). These studies suggest that endothelin activation of PKC initiates multiple events that may indirectly contribute to the observed endothelin-induced inhibition of the opioid response in other cell types.

Endothelin-1 activation leads to the release of intracellular calcium as evident from the transient calcium-activated chloride current after agonist treatment. The BAPTA-AM data suggest that endothelin-1-induced increase in intracellular calcium is necessary for signals that lead to the inhibition of the opioid response. In addition to its role as a calcium chelator, BAPTA-AM was shown to down-regulate PKC in liver macrophages after prolonged exposure (Dieter et al., 1993). We presume that BAPTA-AM treatment does not down-regulate all of the isoforms of PKC expression in the oocyte, as the staurosporine data suggest that PKC is not directly involved in the endothelin-induced inhibition of Kir 3. However, alternative mechanisms of endothelin regulation of Kir 3 are not excluded by these data.

In C6 glioma, endothelin mobilization of intracellular calcium-activated arachidonic acid release was reversed by quinacrine, a nonspecific inhibitor of PLA2 (Dunican et al., 1996). In a culture of human pericardium smooth muscle cells, it was shown that endothelin-1-evoked arachidonic acid release was dependent on the elevation of intracellular calcium (Wu et al., 1996). Phospholipases A2 comprise a family of secretory and cytosolic enzymes that regulate arachidonic acid release through hydrolysis of membrane phospholipids (Bonventre, 1992). Increased cellular calcium may selectively activate the 85-kDa, calcium-sensitive, cytosolic PLA2 (cPLA2) (Clark et al., 1991). We determined that endothelin-induced inhibition of Kir 3 was calcium dependent and partially reversed by treatment with the AACOCF3, the trifluoromethyl ketone analogue of arachidonic acid in which the COOH group is replaced with COCF3 to form a selective, high-affinity inhibitor of cPLA2 (Street et al., 1993). Stimulation by calcium-mobilizing ligands may lead to cPLA2 translocation to the phospholipid membrane. cPLA2 may be the intermediate between agonist-activated G protein-linked receptors that mobilize calcium and the stimulation of arachidonic acid release (Clark et al., 1991). Because the specificity of AACOCF3 has not been previously tested in Xenopus oocytes, potential inhibitory effects on other phospholipases cannot be determined. Our results suggest that cPLA2 or a related enzyme may be the mediator of endothelin-induced inhibition of Kir 3. Although AACOCF3 did not inhibit the magnitude of the endothelin-induced chloride current, we do not exclude the possibility of nonspecific effects on isoforms of PLC.

Gβγ subunits may act as signaling molecules to downstream effectors, providing one potential mechanism to activate downstream signals. It has been reported that Gβγ subunits activate PLA2 in a reconstituted system of purified retina (Jelsema and Axelrod, 1987). Gβγ subunits positively regulate many effectors, including PLCβ (Clapham and Neer, 1997), and it has been suggested that Gβγ may directly activate PLA2 (Axelrod, 1990, 1995). Our data suggest that PLA2 activation may occur by an indirect mechanism, following an increase in intracellular calcium. However, as suggested by Jelsema and Axelrod (1997), Gβγ subunits may also directly activate PLA2.

These results obtained by using the Xenopus oocyte expression system suggest a possible means by which Gq/11 receptor signaling may influence Gi/o in brain. Examples of this type of receptor “cross-talk” have been noted previously in neuronal recording (Jan and Jan, 1997), although the underlying mechanisms have not been clarified. Furthermore, the suggestion that an arachidonic acid analogue may directly inhibit Kir current is important to explore in future studies. Understanding the signal transduction mechanisms that regulate the inwardly rectifying potassium channel helps to explain how disparate G proteins regulate Kir 3 in neurons and myocytes.

Acknowledgements

We thank Dr. Guy Chan for helpful discussion. This study was supported by USPHS grant DA04123 and DA07278 from NIDA.

Ancillary