Endogenous RGS Protein Action Modulates µ -Opioid Signaling Through G α o: Effects on Adenylyl Cyclase, Extracellular Signal-Regulated Kinases and Intracellular Ca 2+ Pathways*

Regulators of G protein signaling (RGS) proteins are GTPase activating proteins (GAPs) for the G α subunits of heterotrimeric G proteins and act to regulate signaling by rapidly cycling of G protein. RGS proteins may integrate receptors and signaling pathways by physical or kinetic scaffolding mechanisms. To ask whether this results in enhancement and/or selectivity of agonist signaling we have prepared C6 cells stably expressing the µ -opioid receptor and either pertussis toxin insensitive, or RGS and pertussis toxin insensitive G α o. We have compared the activation of G protein, inhibition of adenylyl cyclase, stimulation of intracellular calcium release and activation of the extracellular signal-regulated kinase (ERK1/2) mitogen-activated protein kinase (MAP kinase) pathway between cells expressing mutant G α o that is either RGS insensitive or RGS sensitive. The µ -receptor agonist DAMGO and partial agonist morphine were much more potent and/or had an increased maximal effect in inhibiting adenylyl cyclase and in activating MAP kinase in cells expressing the RGS insensitive G α o. In contrast µ -opioid agonist increases in intracellular calcium were less affected. The results are consistent with the hypothesis that the GAP activity of RGS proteins provides a control that limits agonist action through effector pathways and may contribute to selectivity of activation of intracellular signaling pathways.

4 increase in the antinociceptive potency of morphine (10). Although these changes are small, they are suggestive of a role for RGS proteins in opioid coupling efficiency.
An important question is whether RGS proteins alter the efficiency of all intracellular signaling pathways equally or whether there is a variable effect that would provide for selectivity. Selectivity for particular pathways may be obtained by several mechanisms. RGS containing proteins have a wide variety of non-RGS domains (11)(12)(13) that, when RGS binds to Gα, can link other proteins and signaling pathways to provide for diversity of signaling. In addition the interaction of RGS protein with receptors may contribute to selectivity, for example RGS12 binds with the carboxy terminus of the interleukin-8 receptor (14) and inhibition of Ca 2+ signaling in rat pancreatic acinar cells by RGS4 is selective for muscarinic receptors relative to bombesin and cholecystokinin receptors (15) possibly through interaction of the N-terminal domain of RGS4 with the receptors (16). Recently, Wang and colleagues (17) have demonstrated, using ribozyme technology, that RGS3 is a negative modulator of muscarinic m3 receptor signaling while RGS5 is a negative modulator of angiotensin AT1a receptor signaling through Gq/11. In addition to RGS proteins selectively modulating the coupling of different receptors to a single effector, it is possible that RGS proteins could selectively modulate the coupling of a single receptor to different effectors. Indeed, we have recently proposed a "kinetic scaffolding" model for G protein signaling in which RGS proteins confer selectivity for signaling pathways by their ability to shorten the lifetime of Gα-GTP (18). In this model, RGS accelerates hydrolysis of the Gα bound GTP, permitting recombination of Gα-GDP and Gßγ and recoupling of the heterotrimer and receptor, allowing rapid reactivation by agonist bound receptors. This maintains active Gα-GTP and Gβγ proteins in the close vicinity of the receptor, 5 Here we test the hypothesis that RGS proteins differentially regulate µ-opioid receptor coupling to signaling pathways thus contributing to selectivity of receptor activation of second messenger pathways. Since 30 mammalian proteins with RGS activity have been identified to date (12,13) the choice of which RGS to study is a difficult one. We have, therefore, made use of a point mutation in Gαo (G184S) that is known to block interaction with all members of the RGS family without affecting GTPase activity (RGSi) (19), together with a mutation (C351G) to confer pertussis toxin (PTX) insensitivity (PTXi) (20). In this way, when the RGS and pertussis toxin resistant Gαo (GαoRGS/PTXi) is expressed in a cellular system, coupling to endogenously expressed G proteins can be inactivated by PTX treatment and the system must then signal through the expressed mutated Gαo (21).
Our findings demonstrate that the µ-opioid agonists DAMGO and morphine show increased potency and/or efficacy of signaling to adenylyl cyclase in cells expressing RGS insensitive Gαo compared with RGS sensitive Gαo. Signaling through the MAP kinase pathway also shows an increased potency with the full agonist DAMGO, but not an increased maximal effect, though the maximal effect of the partial agonist morphine is significantly enhanced. In contrast, the ability of DAMGO or morphine to stimulate the release of calcium from intracellular stores is altered to a much lesser extent in cells expressing RGS insensitive Gαo compared with RGS sensitive Gαo. These results confirm that RGS proteins can modulate effector signaling by a single G protein and may play an important role in directing effector responses to µ-opioid receptor signaling.    ). GαoPTXi membranes by DAMGO at 2 min was concentration dependent (Fig. 2B). GST-RGS8

Stimulation of GTPase ActivityBasal
increased the maximal stimulation over basal by DAMGO from 60 + 7 % to 151 + 19 % (p < 0.05), with a shift in the EC 50 value from 34 + 12 nM to 92 + 31 nM, though this did not reach significance (p = 0.16). Maximal stimulation of GTP hydrolysis by 10 µM morphine at 2 min was 45 + 5 % in the C6µ GαoPTXi membranes and increased significantly (p < 0.05) to 94 + 17 % in the presence of 1 µM GST-RGS8, though relative to DAMGO, morphine was significantly less efficacious (p < 0.05) in the presence of GST-RGS8 (61 + 3 %) than in its absence (75 + 1 %). RGS8 was chosen for these studies because it is structurally a simpler RGS protein and is known to be a GAP for Gαo (22).   This effect of the GαoRGSi mutant on agonist-mediated inhibition of adenylyl cyclase was particularly marked for the partial µ agonist morphine that became almost (90%) as efficacious as the full agonist DAMGO in the C6µ GαoRGS/PTXi expressing cells. This indicates that RGS proteins may be more effective when the receptor-G protein-effector system is signaling at sub-maximal level. In agreement with this several authors have shown that at high agonist concentrations, RGS proteins are less effective (32,33). In addition, RGS5-mediated reduction in intracellular calcium release by the angiotensin 1a receptor (Gq linked) is less effective when receptors are expressed at high levels (34). Since the relative efficacy of an agonist is tissue specific, it may be possible that differential expression of RGS proteins in tissues is one factor in determining agonist efficacy. The present data show that endogenous RGS proteins may differentially affect signaling by a single G protein depending on the effector pathway to which the G protein couples. Several mechanisms could account for this specificity. One is that the GAP activity of endogenous RGS proteins controls signaling by a kinetic scaffolding mechanism (18). The kinetic scaffolding model predicts that RGS action reduces depletion of local Gα-GTP levels and so permits rapid recycling of G protein, rapid re-coupling of the receptor and maintains local G protein activation.

Inhibition of Adenylyl
The adenylyl cyclase and MAP kinase pathways are poorly signaled to in the presence of RGS activity, but when this activity is blocked, as in cells expressing the GαoRGSi mutant, then signaling can occur because spatial control is lost, allowing spill over of Gα-GTP and Gβγ subunits to more distant effectors. In contrast, coupling to intracellular calcium stores is the more similar in cells expressing GαoPTXi or GαoRGS/PTXi. Thus, for the kinetic scaffolding model to account for this effect the G proteins involved in coupling to this pathway must be organized closely with receptor and effector such that they show a reduced RGS-dependent effect.
The differential effect of RGS on the three pathways examined is consistent with, but does not provide direct proof for this theory. Other mechanisms may explain the findings. The increased opioid effect at adenylyl cyclase and MAP kinase may simply be due to an increased life-time of Gα-GTP in the absence of RGS GAP activity, but if so then the question arises as to why the intracellular calcium signal is not enhanced to a similar extent. There may be differential location or compartmentalization of effectors within the cell (36) (38).
In summary, we have shown in a transfected C6 cell that RGS proteins differentially regulate µ-opioid receptor mediated signaling to different effectors through Gαo, consistent with a kinetic scaffolding mechanism. Coupling to adenylyl cyclase and the MAP kinase pathway appears to be efficiently limited by endogenous RGS proteins, whereas coupling to intracellular calcium stores is less susceptible to RGS protein action. Since the potency and maximal effect of agonist is altered it is possible that differential expression of RGS proteins in tissues plays a role in tissue-specific differences in agonist selectivity and efficiency. Finally, since cyclic AMP (39) and MAP kinase (40,41) have been implicated in contributing to the tolerance associated with long-term opioid administration the effect of endogenous RGS proteins on these cellular adaptations merits further investigation.   Procedures. Shown is a representative blot of phosphorylated and total MAP kinase (A). Bands were quantified as sum intensity (pixels) and plotted as percent of basal (without ligand). The basal sum intensity was lower in C6µ GαoPTXi (157,000 + 17,000 pixels) than in C6µ GαoRGS/PTXi (67,000 + 12,000 pixels). Shown is the combined data from 3-4 assays.