Differential Regulation by CAMP-dependent Protein Kinase and Protein Kinase C of the p Opioid Receptor Coupling to a G Protein-activated K' Channel*

A p opioid receptor and a G protein-activated K' chan- nel were coexpressed inXenopus oocytes. Stimulation of the p opioid receptor induced an inwardly rectifying current that was blocked by opioid receptor antagonist naloxone, indicating that the p opioid receptor is functionally coupled to the K' channel. The coupling is me- diated by G proteins, since pertussis toxin treatment reduced the K' current and injection of GTPyS (guanosine 5'-O-(thiotriphosphate)) enhanced it. Re- peated stimulation of the p receptor leads to desensitization, as the K' current from the second stimulation was reduced to 70% of that from the first one. Both CAMP-dependent protein kinase (PKA) and protein kinase C (PKC) regulate this process, but in opposite di- rection. Activation of PKC by treatment of the oocyte with phorbol ester potentiated the desensitization of the p receptor-induced current. However, incubation of the cell with a membrane-permeable cAMP analog, 8-chlo-rophenylthio-cAMP, completely abolished the desensitization. The CAMP effect appears to be mediated by PKA, since injection of a PKA catalytic subunit showed the same effect as cAMP incubation. These results suggest that PKA and PKC differentially regulate the p opioid receptor coupling to the G protein-activated K'

Opioids, both endogenous peptides and exogenous alkaloids, affect the functioning of the central nervous system by interacting with membrane receptors (1,2). Pharmacological studies suggest the presence of three major types of opioid receptors in the brain and spinal cord: 1-1, 6, and K , of which p opioid receptor plays an important role in supraspinal analgesia and development of morphine tolerance and dependence ( 3 , 4 ) . Opioid receptor activation has been shown to mediate the inhibition of neuronal firing and neurotransmitter release in a variety of brain areas ( 5 , 6). Stimulation of the p opioid receptor results i n a membrane hyperpolarization caused by an increase Grant NS28190 (to L. Y.1. The costs of publication of this article were * This work was supported in part by National Institutes of Health defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. i: Recipient  in an inwardly rectifying K' conductance in both locus coeruleus and hippocampus neurons (7)(8)(9). The coupling of t h e p opioid receptor to the brain inwardly rectifying K' channel appears to be mediated by a pertussis toxin (PTX1'-sensitive G protein(s1 ( 5 , 10). However, the molecular identity of the K' channel that is coupled to the p opioid receptor is still unclear, and little is known about the functional regulation of the coupling.
Recently, a p opioid receptor and a G protein-activated K' channel have been molecularly cloned (11-13). The p opioid receptor was isolated from rat brain and is functionally coupled to G proteins and adenylyl cyclase (11). The G protein-activated K' channel was originally isolated from the atrial cells, and RNA blot analyses indicate that i t is also abundantly expressed i n the brain (12,13). In addition, Dascal et al. (13) isolated an almost identical clone from the brain. The brain expression of this clone suggests that i t m a y be a K' channel identical or similar to that activated by the p opioid receptor. To test this hypothesis, we expressed the p opioid receptor and the G protein-activated K' channel in X e n o p u s oocytes and investigated the modulation by PKA and PKC.
MATERIALS AND METHODS Chemicals-f"H1DAMGO (38 Ciimmol) was from the National Institute on Drug Abuse. DAMGO was from Bachem. Naloxone was from Research Biochemicals International. All other chemicals were from Sigma.
Complementary DNA Clones for the p Opioid receptor and the G Protein-actiuated K+ Channel-A cDNA clone, MOR-1, containing the protein coding region of a rat p opioid receptor has been described (11 I. Based on the cDNA sequence for the rat G protein-activated K+ channel (12,13). two oligonucleotide primers were synthesized corresponding to the 5'-and 3'-untranslated regions, respectively: CTCGGATCCGTAT-TATGTCTG and ATAGTCGACTAAAACTAAATC. PCR was performed in an air cycler (Idaho Technology) with 10 s a t 94 "C, 20 s at 56 "C, and 1 min at 75 T for 35 cycles, using the two primers and the purified lambda DNA from a rat brain cDNA library (14). A Deep Vent" DNA polymerase (New England Biolabs) was used to reduce PCR errors (151. The 1.7-kilobase pair polymerase chain reaction product was cloned into a TA-cloning vector (Invitrogen). Both cDNA clones were used to synthesize mRNA by in uitro transcription as described (161. Oocyte Injection and Binding Assay-Xenopus oocytes were prepared as described (16). In vitro transcribed RNA (1-2 ng/oocyte) was injected into oocytes with a Drummond microinjector. Oocytes were incubated in L-15 medium supplemented with 0.8 msl glutamine and 10 pg/ml of gentamycin at 20 "C for 3-4 days before analysis. Binding of the injected oocytes was carried out in regular ND96 (96 m41 NaC1, 2 mM KC1, 1 mhi MgC12, and 1.5 mhl CaC1,) solution at 20 "C for 90 min, using 1 ny ["HIDAMGO. Binding was terminated by vacuum filtration through a Whatman G F B filter pretreated with 1% polyethyleneimine. Three milliliters of ND96 was used to wash the oocytes, and nonspecific binding was determined using 1 PM naloxone. The radioactivity of the oocytes were determined in 6 ml of Scintiverse (Fisher) with a Beckman LS5801 scintillation counter.
Electroph.ysiology-Oocytes were voltage-clamped at -80 mV with two electrodes filled with 3 M potassium chloride and having a.resistance of 0.5-2 megohms, using an Axoclamp-2A and the pCLAMP software (both from Axon Instruments). Oocytes were superfused with either ND96 containing 6 mM CaCl, or a high potassium solution (96 mM KCl, 2 mM NaCl, 1 mM MgCl,, and 1.5 m M CaC1,l. Analysis of variance and Student's t test were used to determine the statistic significance among different grouus.

Kinase Regulation of p Opioid Receptor-K+ Channel Coupling
Coupling of the p opioid receptor to the G proteinactivated K' channel. Electrophysiologic analysis of oocytes injected with mRNAs for the rat p opioid receptor and the G protein-activated of -80 mV. Oocytes were exposed to 1 VM DAMGO (left trace) or 1 p~ K' channel. A, membrane current traces recorded a t a holding potential DAMGO plus 10 p~ naloxone (right trace) as indicated. Inward current is downward. B, membrane currents with voltage steps ranging from -160 mV to +40 mV were recorded before and 1 min after DAMGO superfusion. The DAMGO-induced net currents were derived by subtracting the currents before DAMGO application from those after, and are shown in the leftpanel. The rightpanel shows the I-V curve of these currents.

RESULTS
Coupling of the p Opioid Receptor to the G Protein-activated K+ Channel-To determine whether the p opioid receptor couples to the G protein-activated K' channel, we expressed both proteins inXenopus oocytes. Messenger RNAs of these two clones were generated by in vitro transcription, and oocytes were microinjected with each mRNA alone or both mRNAs.
[3H]DAMG0, a highly selective ligand for 1-1 opioid receptors, was used in whole-cell binding assay t o determine the expression of the p opioid receptor, and nonspecific binding was the residual binding not blocked by naloxone. Oocytes injected with mRNAs for both the p opioid receptor and the K+ channel displayed a specific binding of about 1 fmol/oocyte, whereas oocytes injected only with the K' channel mRNA did not show any appreciable binding to r3H]DAMG0 (data not shown).
Coupling of the p opioid receptor to the KC channel was studied by two-electrode voltage clamp. In the oocytes injected with either the p receptor mRNA or the K+ channel mRNA alone, no membrane current was observed with the p receptor agonist DAMGO (data not shown), indicating that there are no endogenous currents in oocytes that are activated by DAMGO, and that either the p receptor or the K' channel alone is not sufficient to generate DAMGO-induced currents. However, coexpression of both proteins gave rise to membrane currents upon DAMGO stimulation (Fig. 1). Exposure of the oocytes to 1 p~ DAMGO produced an inward membrane current that was completely blocked by the opioid receptor antagonist naloxone (Fig. lA). In agreement with the inwardly rectifying nature of this G protein-activated K+ channel, the current-voltage relationship of the DAMGO-induced membrane current showed a characteristic inward rectification (Fig. 1B) as the current magnitude increased with progressive membrane hyperpolarization, whereas there was little current when the membrane was depolarized above 0 mV. As expected for a K+ channel, membrane current was completely blocked by 100 p~ Ba2+ (data not shown). G Protein Involvement in the Coupling-Previous studies in neurons suggested that the coupling of opioid receptors to the membrane K' conductance involves a FTX-sensitive G protein ( 5 , 10). To test whether the coupling between the p opioid receptor and the K' channel in oocytes is affected by PTX, cells injected with both mRNAs were incubated with 0.5 pg/ml PTX for 24 h. PTX treatment reduced the DAMGO-induced membrane current by 60% (Fig. M ) , and this reduction was proportionally uniform across the voltage range (Fig. 2B). These data indicate that a PTX-sensitive G protein(s) is needed for the p receptor activation of the K' channel, accounting for a t least 60% of the coupling.
The involvement of heterotrimeric G proteins in the coupling was further studied using GTPyS, a nonhydrolyzable GTP analog that interacts with G protein and keeps it in an activated state. After DAMGO-induced current reached a plateau, intracellular injection of GTPyS elicited an additional increase of the current (Fig. 3A). When the time course of the normalized current was plotted using the peak current value before GTPyS injection as the standard, GTPyS injection resulted in a gradual rise of the current which, after reaching the maximum, decreased toward the base line following a similar time course as that of control oocytes (Fig. 3 B ) . However, injection of GTPyS itself without stimulation of the 1-1 receptor by DAMGO did not induce appreciable membrane current change (data not shown), indicating that the GTPyS-mediated enhancement of the K' conductance is dependent on the activation of the receptor.
Differential Regulation of the Coupling by PKA and PKC-To determine whether the coupling between the p opioid receptor and the K+ channel is regulated by PKA-and PKC-mediated phosphorylation, we used a protocol shown in Fig. 4A. The oocyte was superfused with high potassium solution (HK) while DAMGO-induced response was measured. Then the superfusate was switched to ND96 solution, and the cell was either treated with a chemical to stimulate a kinase or microinjected with the catalytic subunit of PKA. The cell was allowed to recover after the treatment, and DAMGO-induced response was measured again in HK solution. Comparison between the maximum responses before and after the treatment thus reveals how much desensitization has occurred after the first DAMGO stimulation. As shown in Fig. 4B, the DAMGO-induced membrane current recorded approximately 15 min after primary exposure was only about 70% of the first response, indicating that desensitization has taken place. Treatment of the oocytes with phorbol ester PMA, a PKC activator, further reduced the second response, suggesting a negative regulation by PKC of the p opioid receptor-activated K+ current. Membrane currents were recorded from oocytes injected with both the p receptor and the K+ channel mRNAs. A, a representative current trace recorded in a oocyte a t a holding potential of -80 mV illustrating the experimental protocol. The cell was bathed in HK solution, and l p~ DAMGO was applied by superfusion to elicit the K+ current. After the first DAMGO stimulation, the superfusate was switched to ND96 containing 6 mM CaCl,, and the oocyte was either untreated (as in this example) or subjected to drug treatment or enzyme injection (see below). The superfusate was then switched back to HK solution to record the second DAMGO-induced membrane current. B, relative response of the DAMGO-induced membrane currents from different treatment groups a t a membrane potential of -80 mV. Data are expressed as the percentage of the peak current induced by second DAMGO stimulation over that of the first stimulation, and are presented as mean 2 S. E. (n = 4). Treatment is labeled on the bottom of each bar. Result of variance analysis is shown a s ** with p < 0.01 as compared to the untreated group. Different treatments used in this experiment are as follows: 8-CPT-cAhfP, incubation with 1 mM 8-CPT-CAMP for 10 mi.n; PMA, incubation with 100 nM PMA for 10 min; PKA, injection of the catalytic subunit of PKA (50 fmol/cell).
Surprisingly, treatment of the oocyte with 8-CPT-CAMP, a membrane-permeable CAMP analog that can diffuse into the cell and stimulate PKA, completely abolished the desensitization observed in untreated oocytes (Fig. 4B). To determine whether the 8-CPT-CAMP effect on preventing desensitization is mediated by PKA, the catalytic subunit of PKA was injected into the oocytes after the first DAMGO stimulation. This resulted in the same effect as 8-CPT-CAMP incubation (Fig. 4B ).
The current-voltage relation was determined at the peak of

Receptor-K+ Channel Coupling 7841
both the first and the second DAMGO-induced response. Activation of PKC by PMA treatment enhanced desensitization over the entire voltage range, whereas either activation of PKA by CAMP or direct enzyme injection prevented desensitization across the voltage range (data not shown). These data suggest that the two kinases have opposite effect on the p opioid receptor-activated K' current, exerting differential regulation on this process.

DISCUSSION
Neurotransmitters modulate the excitability of neurons by affecting ion channels, K' channel being one of the primary targets of such modulation. In fact, many neurotransmitters have been shown to couple to a K' conductance in neurons (17)(18)(19)(20). The effect of neurotransmitters on K' channel involves a receptor-mediated mechanism, and opioids are no exception. In both locus coeruleus and hippocampus, p opioid receptors have been shown to regulate II K' conductance, leading to membrane hyperpolarization and a decrease in neuronal firing rate (8)(9)(10). The recent cloning of a p opioid receptor, as well as a G protein-activated K' channel, provided the opportunity to examine the molecular mechanism of this coupling. The K' channel was isolated from the heart atrial cells, where it is mainly involved in the heart beat regulation mediated by muscarinic receptors (21,22). However, both RNA blot analysis and cloning effort suggested that the same channel also exists in the brain (12,13); thus, it may mediate the neuronal effect of various neurotransmitters. In this report, we showed that the p opioid receptor and the G protein-coupled K' channel, when coexpressed in Xenopus oocytes, are functionally coupled. Although we cannot exclude the possibility that other K' channels may be involved in the coupling to the p opioid receptor, our data suggest that this G protein-coupled inward rectifier may be the long-sought K' channel that is linked to the p opioid receptor and other neurotransmitter receptors. Kubo et al. (12) have shown that this channel can be activated by injection of purified G protein GaiS, PI, and y2 subunits. Our experiments with PTX ( Fig. 2) and GTPyS (Fig. 3) suggested the involvement of a PTX-sensitive G-protein(s1 in the coupling. It has been shown that opioid receptors are associated with G proteins of the Gi and Go subtypes (23)(24)(25). Therefore, it is not surprising that similar PTX-sensitive G proteins of the Gi andor Go subtypes in Xenopus oocyte can mediate the coupling between the p opioid receptor and the K' channel. However, the fact that PTX treatment did not completely block the DAMGO-induced K' current ( Fig. 2) suggests that other G proteins not sensitive to PTX may also be involved in the p receptor coupling.
Phosphorylation by kinases is one of the most important mechanisms for functional regulation of many cellular proteins including neurotransmitter receptors and ion channels, and PKA and PKC are two of the most widely studied kinases (26,27). Phosphorylation of P2-adrenergic receptor by either PKAor PKC leads to its uncoupling from G proteins, resulting in desensitization to further agonist stimulation (28,29). In the case of voltage-dependent Ca2+ channels such as the endogenous oocyte Ca2' channel, PKA-and PKC-mediated phosphorylation is able to potentiate channel activity (30,31). Cystic fibrosis transmembrane conductance regulator, a C1-channel associated with cystic fibrosis, is also regulated by CAMP through PKA pathway (32,33). Regulation of the inwardly rectifying K' channels by either PKA or PKC, however, is not clear. Molecular cloning has shown that the p opioid receptor and the G protein-activated K' channel possess multiple putative sites for PKA and PKC phosphorylation (11)(12)(13). In this study, we found that the coupling of the p opioid receptor to the K' channel desensitizes upon repeated stimulation by the p receptor agonist DAMGO, as the peak current by the second DAMGO ap-

Kinase Regulation of y Opioid Receptor-K+ Channel Coupling
plication is reduced to 70% of that by the first one (Fig. 4).
Treatment of the cells with phorbol ester enhanced this desensitization (Fig. 4B ), suggesting PKC-mediated phosphorylation. Surprisingly, treatment with 8-CPT-CAMP or injection of the catalytic subunit of PKA completely abolished the desensitization (Fig. 4B). Thus, PKAand PKC appear to exert opposite effects on this p receptor-induced K' current. Our results, however, do not reveal the molecular entities of PKA-and PKCmediated phosphorylation. Further studies using mutagenesis are needed to determine the correlation between specific phosphorylation sites on these membrane proteins and the regulation by PKA and PKC.
of PKA and Mingting Tian for technical assistance.