British Journal of Anaesthesia, 2002, Vol. 89, No. 1 62-78
© 2002 The Board of Management and Trustees of the British Journal of Anaesthesia
Molecular and Basic Mechanisms of Anaesthesia |
Anaesthetic actions on other targets:protein kinase C and guanine nucleotide-binding proteins
Departments of Anesthesiology and Physiology & Biophysics, School of Medicine,State University of New York, Stony Brook, NY 11794-8480, USA*Corresponding author
Abstract
Br J Anaesth 2002; 89: 62–78
Keywords: anaesthetics, molecular targets; enzymes, protein kinase C; proteins, G-proteins
Introduction
Current research into the mechanism of anaesthesia has been focused on a few ligand-gated ion channels that are sensitive and appear to be relevant targets for general anaesthetic agents.138 This is, however, no reason to rule out the hundreds of other proteins likely to bind these drugs.35 It is reasonable to suppose that among these, more than a handful are subject to anaesthetic perturbation, and a few could be relevant to the anaesthetic state. There is no reason a priori to believe that metabotropic receptors, which modulate synaptic transmission and bind the same ligands, should not be as sensitive or important to anaesthetic action. Similarly the downstream elements (guanine nucleotide (G)-proteins) that couple these receptors to various effector pathways (cAMP or inositol trisphosphate [InsP3]/Ca2+) or the feedback and response elements (protein kinase C [PKC]) may well be sensitive and relevant, yet few were examined until recently. One of the major reasons that attention has moved away from these second messenger generating pathways is the perception that they and their components are insensitive to anaesthetic agents. While this is undoubtedly true for some, this does not eliminate others – many ion channels are insensitive to general anaesthetics. Measurements of second messengers and downstream responses also do not necessarily eliminate upstream components, since their sensitivity will depend on the levels of receptor reserve and feedback. Here the bias towards ion channels is reflected in the nature of the experimental measures of anaesthetic action.
Sensitivity to any perturbation depends on what steps are measured and the concurrent feedback. Ligand-gated and other ion channels behave as self-amplifying units. Opening a channel produces a measurable change in current under the appropriate conditions. Although other channels may respond rapidly, and could thereby dampen or reinforce voltage changes, when measured under voltage clamp conditions (the typical experiment), feedback (generally) requires some mechanism intrinsic to the channel itself; hence, at this level of measurement, one examines directly the effects on the ion channel itself. Concepts such as receptor reserve and feedback are not meaningful here.
By contrast, G-protein coupled signals rely on many components operating in series. In tracing the cascade of events towards the response elements, the further away one gets from agonist binding, the more likely that some step has already been saturated with respect to the preceding signal. Such systems often demonstrate significant receptor reserve, which must be depleted before a perturbant, such as an anaesthetic, will have a measurable effect. If one measures the most proximal events, for example receptor/G-protein coupling, and works forwards to the response elements, performing measurements well below the reserve threshold, it should be possible to rule out certain pathways as being sensitive to anaesthetics. For PKC-modulated, G-protein coupled pathways this has rarely been done. Moreover, in systems where feedback is strong, changes in signal output caused by anaesthetics are likely to be suppressed. Therefore, the ability to measure responsiveness to anaesthetic depends very much on the system under study, its receptor reserve and the strength of its feedback elements, all of which will vary among different synapses. This problem is compounded when examining an ensemble of neurones, for several reasons. The most critical is related to measurement of second messengers (other than calcium ions), which normally entails sampling of large tissue volumes containing thousands of neurones. Even with receptor subtype specific agonists, spatial and time averaging of the signals across the many synapses present in this volume may well preclude observing any anaesthetic effects.
In this article we review the results obtained for two alternative anaesthetic targets – PKC and G-proteins – that are integral to metabotropic receptor signalling. A significant body of evidence has accumulated suggesting that certain forms of these proteins are sensitive to general anaesthetic agents in concentration ranges relevant to anaesthetic action.
Protein kinase C
Protein kinases catalyse phosphoryl transfers from ATP to serine, threonine, tyrosine and occasionally histidine residues of numerous regulatory proteins, many of which control neuronal excitability. Located pre- and post-synaptically, protein kinases are allosterically controlled, often directly driven by the levels of intracellular second messengers. Their critical role in the release and response to neurotransmitters, and their own regulation, suggest a potential site for the action of general anaesthetic agents. Nonetheless, the sensitivities of only a few protein kinases have been examined. The best studied is PKC, a group of serine/threonine kinases activated by lipid and calcium signals generated by receptor-stimulated phospholipase C (PLC).50
The PKC family targets a wide range of substrates, including ligand-,129 G-protein-29 and voltage-gated ion channels,134 and metabotropic receptors.96 Mammals express eleven different PKC subtypes, all multi-domain proteins containing a common catalytic core, composed of an ATP-binding C3 region and a substrate binding C4 region. The various subtypes can be broadly classified on the basis of sensitivity to calcium, and diacylglycerol (DAG) or the xenobiotic, phorbol ester. The conventional subtypes 
, ßI, ßII, and 
contain a C1 region consisting of two tandem C1 domains that bind DAG or phorbol esters, and a C2 domain that binds calcium and acidic phospholipids (usually phosphatidylserine [PS]); crystallographic structures of these domains have been published.75 The calcium-independent (novel) forms 
, 
I, II, 
and 
, have a ‘novel’ C2 domain, but are activated by DAG or phorbol esters through the C1 region. So called atypical forms 
and 
are missing the C2 region and have only a single C1 domain that does not bind lipid; thus, their activity is independent of DAG and calcium. All PKC subtypes contain a 17-residue pseudosubstrate sequence, located near the amino terminus, that constitutively blocks catalytic activity and figures prominently in their activation (Fig. 1).
|
Regulation
A general model of conventional PKC regulation involves tethering of the otherwise soluble enzyme to the cytoplasmic aspect of the membrane surface.31 126 The initial event involves binding of the C2 domain to calcium complexed with acidic phospholipids. This alters domain–domain interactions, aligning the C1 region with the membrane surface where it binds DAG or phorbol esters (Fig. 2). Subsequent conformational changes pull the pseudosubstrate sequence from the mouth of the catalytic core, thereby facilitating binding and phosphorylation of a subtype-specific array of substrates. Control of PKC activity, however, also involves upstream protein kinases and phosphatases in addition to calcium and lipid activators. Three sites, located in the catalytic core of conventional PKCs, are constitutively phosphorylated in living cells; autophosphorylation (by PKC itself) accounts for two, while the third, part of the activation loop of the catalytic core, is subject to phosphorylation by a newly discovered 3-phosphoinositide-dependent protein kinase, PDK-1.130 Newly synthesized PKC is constitutively phosphorylated by PDK-1, triggering autophosphorylation.126 In this state, the active site of the now cytoplasmic PKC is constitutively blocked by the pseudosubstrate sequence. Subsequent recruitment to the membrane by DAG, PS or other acidic lipids and calcium (conventional forms) leads to expulsion of the pseudosubstrate peptide, and activation. Thus PKCs are tightly controlled by coordinated, and in some cases, opposing pathways that include PDK-1, protein phosphatases, generation of DAG from polyphosphoinositides, and elevation of cytoplasmic calcium from both external and internal sources. Perturbation of any one of these pathways is predicted to alter conventional PKC activity.
|
Lateral organization in the membrane
Several PKC subtypes have protein binding partners known as receptors for activated C kinases (RACKs) that serve to organize the kinases, their substrates and associated regulatory proteins (Fig. 3).116 RACK1 is selective for PKC-ßII108 whereas RACK2 is selective for PKC-
.28 In addition to PKC and proteins associated with these pathways, RACKs are also capable of binding components of other signalling pathways and could therefore serve as coincidence detectors and sites of signal integration. Other proteins bind PKC, as well, laterally organizing kinases with their substrates. For example, inactivation-no-afterpotential D (INAD), a scaffold protein containing multiple postsynaptic density-95/discs large/zona occludens (PDZ) domains, organizes eye-specific PKC, fly PLC-ß isozyme and the transient-receptor-potential calcium channel, a PKC substrate that regulates photoreceptor sensitivity.131 How PKC and its substrates are organized in mammalian synapses is only now being unravelled.
|
Tissue and subcellular distribution
Immunoblotting (to measure protein), northern blotting and in situ hybridization (to measure mRNA levels) and indirect immunofluoresence techniques have been used to map distribution of the various PKC subtypes in the central nervous system (CNS). Neuronal PKC isoforms (
, ßI, ßII, , , , , , ) are differentially distributed among the major brain and spinal cord structures89 101 and all have distinct distributions within each area.111 Unfortunately, there are no definitive rules concerning distribution among different neurones classified by transmitter, and multiple different forms of each group are commonly expressed in the same cell. At a subcellular level, many PKC isoforms have their own distinct distributions, implying unique neuromodulatory roles for each subtype/class. For example, PKC-, a conventional form expressed primarily in the CNS, is often located presynaptically,111 whereas the form is often found throughout the neuronal cell body, and PKC- is associated with actin fibers.54
Anaesthetics and their effects on purified PKC subtypes
The effects of n-alkanols and volatile agents on the enzymatic activities of partly purified brain PKC containing conventional subtypes, as well as individual recombinantly expressed isoforms, have been examined, and both inhibition and stimulation have been reported.50 Slater and colleagues121 found that n-alkanols and volatile anaesthetics inhibit PKC activity when protamine sulphate is used as the activating surface and histone III-S as substrate. (Activation by protamine sulphate appears to be different from that induced by membrane cofactors;95 moreover, this compound is a substrate for PKC.) Under these conditions, halothane and enflurane were also inhibitory, albeit at high concentrations (IC50 0.8 and 0.95 mM, respectively). The n-alkanol potency followed relative hydrophobicity, which parallels anaesthetic potency. Inhibition was also observed using phosphatidylcholine (PC)/PS membranes containing phorbol ester, indicating that inhibition was not due to the absence of lipid. Although obtained using a highly artificial system, these observations suggested a direct lipid-independent effect of these compounds on PKC. Allosteric modulation of the enzyme was supported by the insensitivity of the constitutively active, trypsinized form of PKC to n-alkanols. Comparable results were also obtained with high anaesthetic concentrations of ethanol, halothane, thiopental, etomidate and diethylether, with modest inhibition observed using various substrates and PS/DAG as the activating lipid dispersion.37
Differences between short- and long-chain n-alkanols were observed in PC/PS membranes containing either phorbol ester or DAG, and myelin basic protein peptide (MBP) as substrate.123 Short-chained n-alkanols (C2–C6) inhibit DAG-stimulated PKC-, whereas the 12-O-tetradecanoylphorbol 13-acetate (TPA)-activated enzyme was insensitive. By contrast, long-chain n-alkanols (C7–C10) potentiated both DAG- and TPA-stimulated activities, and were more effective in potentiating the latter; long-chain alcohols also enhanced binding of fluorescent phorbol ester. These results point to distinct actions of long- and short-chain n-alkanols.
F-actin binds and activates PKC, and alcohols affect this interaction.124 125 In complex in vitro assays containing calcium, phorbol ester, and MBP, n-alkanols from C2 to C6 inhibited stimulation by F-actin, with potencies that paralleled hydrophobicity. By contrast, n-alkanols of more than six carbons were less potent. Ethanol (EC50 
100 mM) inhibited calcium/phorbol-ester-stimulated PKC- and ßI, but had no effect on or novel or atypical isoforms. The effects of n-alkanols on F-actin binding and activation were complex, and depended on the subtype and the presence and levels of activating cofactors. Whether and how anaesthetics in living cells modulate PKC isoforms associated with non-membranous structures remains to be determined.
Activation of PKC by anaesthetics and n-alkanols has also been reported. Comparing three different substrate/lipid activating systems, Hemmings and Adamo52 observed that partially purified PKC from brain is stimulated by halothane (EC50 2.2 vol%) and propofol (EC50 240 µM). In this case the stimulatory effects were dependent on substrate and the type of activating surface, yielding twofold stimulation with histone-H1 (substrate), and PC/PS/DAG vesicles, a more physiological activating model membrane. Lipid dispersions that maximally activated PKC showed no further stimulation by anaesthetic.50 52 A constitutively active, trypsin-cleaved form of PKC was insensitive. The stimulatory effect of halothane appeared to increase Vmax rather than decrease Km for substrates and reduced the amount of PS required for activation.49 Unfortunately the use of Michaelis–Menton kinetics can lead to erroneous conclusions when the enzyme and substrate bind to, or are part of, a membrane surface. Nonetheless, anaesthetics appear to activate, contradicting earlier results.122 Whether the differences are due to assay conditions, substrate or the source of enzyme is still unsettled.
Extending the observations of Hemmings and Adamo, Shen and colleagues119 showed that n-alkanols stimulated PKC- catalysed phosphorylation of histone substrate when PC/PS/DAG membranes were used as the activating surface. In this case, n-alkanols like octanol increased activity up to fourfold (EC50 250 µM). The EC50 was 10-fold higher when DAG was omitted. Similar results were obtained for C4 through C7, although the maximum activation declined with decreasing chain length. Importantly, binding to the PC/PS membrane was greatly enhanced, suggesting that n-alkanol could substitute for DAG. The degree of activation was dependent on the acyl chain composition of the phospholipid. Greater activation occurred when saturated PC and PS were combined with unsaturated DAG. Whilst PKC activity was measured at temperatures where the saturated lipid was fluid, increasing the mol fraction of octanol or pentanol caused a lowering and broadening of the transition temperature range. Maximal activation by pentanol was observed under conditions where both gel and liquid crystalline phases coexist, implying that n-alkanols activate PKC by both altering lipid structure and enhancing membrane binding. Indeed, the PKC enzyme and its constitutive domains (especially C1) are exquisitely sensitive to subtle changes in the membrane surfaces,55 supporting a modified lipid hypothesis. One problem with this idea is the lack of evidence for such transitions in natural membranes, although indirect evidence for the existence of liquid ordered phases (rafts) enriched in saturated lipids, cholesterol and sphingomyelin has been obtained.21 The possibility that these are potential sites for PKC activation, much less modulation by anaesthetics, is an open question.
Possible molecular explanations
The in vitro sensitivity of conventional PKCs suggests that anaesthetics disturb or enhance the normal allosteric modulation by DAG and calcium. Thus far no three-dimensional structure of the whole enzyme has been obtained, but several recent structures of the C1 and C2 domains26 suggest how anaesthetics might operate at the protein/lipid interface. Since the C1 domain must penetrate the bilayer to access DAG or phorbol ester,82 it would seem that increased lateral pressures near the interface, caused by the partitioning of anaesthetic into the interface,22 should inhibit PKC activation by raising the energetic costs of surface penetration, yet activation is generally reported when artificial membranes are used52 49 119 and inhibition is observed when the conventional PKCs are tested in the absence of membranes.123–125 Alternatively, anaesthetics may disrupt the normal allosteric transitions at domain–domain interfaces. The realignment of the C1 region, thought to occur upon Ca:PS binding to the C2 domain, would suggest a focus on this interface. Here, either activation or inhibition could be proposed, depending on which enzyme conformers are stabilized. In the former case, little further activation should be observed at maximal concentrations of calcium, PS or DAG. Activation of PKC by halothane is observed at saturating levels of calcium and DAG; however, sensitivity to halothane appears to depend on the mol fraction of PS in the membrane surface,49 suggesting enhanced C2 domain function. Significant activation by halothane is also observed in the absence of DAG. Similar DAG independence was noted for long-chain n-alkanols,119 consistent with alterations in how the C1 domain interacts with the membrane surface. How anaesthetics alter the membrane and C1/C2 interface remains to be investigated.
Biological membranes and synaptosomes
Results obtained in synaptosomes (resealed membrane vesicles pinched off from nerve terminals) are generally consistent with activation by long-chain n-alkanols and halothane. Halothane induced a twofold increase in PKC-catalysed phosphorylation of endogenous myristoylated alanine-rich C kinase (MARCK), an 83 kDa PKC substrate, as well as several other proteins (EC50 1.8 vol%) in whole and lysed synaptosomes.51 A selective peptide inhibitor of PKC blocked halothane-induced phosphorylation whereas a protein phosphatase inhibitor failed to affect the increase, consistent with stimulation of PKC. By contrast, basal phosphorylation of synapsin 1, a key snare complex protein that is under the control of multiple other kinases, was unaffected, although halothane did enhance synapsin 1 phosphorylation induced by calcium ionophore or membrane depolarization. It should be noted that these results were obtained in the absence of any overt physiological stimulus.
Halothane also enhances the effectiveness of phorbol esters, which induce rapid translocation of conventional PKC isoforms from the cytosol to the membrane, coincident with activation and down-regulation through a proteolytic pathway.92 When halothane alone is applied to synaptosomes, a small decrease in cytosolic and membrane PKC- immunoreactivity is observed, indicative of translocation followed by rapid degradation.51 Phorbol 12-myristate-13-acetate (PMA) alone causes a substantial translocation but halothane enhances this effect, clearing the entire cytosolic fraction of any remaining PKC-. Similar results were obtained for PKC-. Unfortunately, the relationship between phorbol ester concentration and anaesthetic dose was not explored and many other questions remain. What happens when a physiological stimulus is applied? To which membrane compartments do the PKC isoforms localize? Which isoform is responsible for phosphorylation of MARCKs and other relevant substrates? It is clear that further work is needed to explore the relationship between anaesthetics and the events leading to PKC activation/translocation.
Oocytes
Xenopus oocytes are a convenient system to express recombinant CNS proteins, especially metabotropic receptors coupled to PLC and ion channels. The read-out for coupling to PLC is a calcium-activated chloride conductance, endogenous to the oocyte, that is stimulated by calcium, mobilized from internal stores by the PLC product, InsP3. Anaesthetic agents (isoflurane, halothane or enflurane) and n-alkanols reversibly suppress a calcium-stimulated chloride conductance induced by agonist activation of M1, M3 muscarinic receptors, 5-HT2A, 5-HT1C serotoninergic and mGlu5 glutamate receptors.72 84–87 114 Importantly, the chloride conductance itself is not directly affected by these agents. Oocytes injected with cerebrocortical mRNA express M1 muscarinic and 5HT-1C serotoninergic receptors. Enflurane (1.8 mM) applied just before a low dose of acetylcholine or serotonin completely and reversibly blocked the increase in the calcium-dependent current.73 The lack of any effect on InsP3-induced calcium release (reflected in calcium-activated current) pointed to enflurane blockade upstream of InsP3, at the level of the receptor, G-protein or PLC. Enflurane also blocked the GTPS stimulated current, further suggesting that the site(s) of inhibition involved receptor/G-protein/PLC coupling. Similar results were obtained with halothane (0.34 mM) in oocytes injected with M1 receptor mRNA, although the AT1A angiotensin II receptor was resistant.34
Ethanol also reversibly inhibits M1 and 5HT-1C receptors in oocytes,114 although the GTPS-stimulated current was not affected. This n-alkanol was most potent at low agonist concentrations, increasing the agonist EC50 fourfold and decreasing the maximum response. Suppression was blocked by a PKC inhibitor peptide or staurosporine (a non-specific kinase inhibitor), suggesting that PKC, known to down-regulate these receptors, was essential to the ethanol response. In addition, ethanol, halothane, F3, an anaesthetic cyclobutane, and the related non-immobilizer, F6, inhibit 5HT-2A receptor-induced currents.87 n-alkanols inhibited best at low doses, with increasing potency up to C8; however, C10–C12 n-alkanols were comparatively ineffective. (Note that long-chain alcohols are most effective in stimulating conventional PKC isoforms in vitro; see above.) The PKC inhibitor bisindolylmalemide I (BIS) increases 5HT-enhanced current by 2.5-fold, while pretreatment with BIS prevents nearly all the effects of halothane, F3, ethanol and octanol, consistent with a requirement for PKC. Extending these experiments, when types 1 and 5 mGlu receptors were compared, it was found that ethanol (200 mM), halothane (0.25 mM) and F3 substantially inhibited mGlu5- but not mGlu1-induced chloride current.84 BIS prevented the anaesthetic effects on mGlu5-induced currents, again suggesting that PKC is required for the action of these anaesthetics. Results of two other experiments further support this interpretation. First, pretreatment of oocytes with calyculin A, a protein phosphatase inhibitor, prolonged the inhibitory effect of these drugs after wash out but did not affect control mGlu5-induced current. Secondly, S890G substitution in mGlu5 (an important PKC phosphorylation site) resulted in a leftwards shift in the glutamate activation curve, and resistance to anaesthetic inhibition. These results support the conclusion that anaesthetics somehow stimulate PKC activity and thereby reduce mGlu5 receptor generated signals through an enhanced negative feedback at the level of receptor phosphorylation. However, this fails to explain why the closely related mGlu1 receptor is anaesthetic resistant, since it is subject to PKC down-regulation via multiple phosphorylation sites.1
In comparable experiments, the receptor for substance P, an important transmitter and modulator of nociception, is similarly inhibited by halothane, isoflurane, enflurane, ether and ethanol.86 These effects are also blocked by the PKC inhibitor, BIS. Likewise, isoflurane (1.25–2.5 vol%) reversibly suppressed M3 receptor induced current.86 Chelerythrine (another PKC inhibitor) blocked the effects of isoflurane, whereas protein phosphatase inhibitors, such as okadiac acid, prolonged isoflurane action, similar to the mGlu5 receptor,33 except that chelerythrine did not affect the agonist response in the absence of anaesthetic, whereas BIS increased responses to M1 stimulation. Interestingly, isoflurane is much more effective at suppressing M3 than M1 receptor responses,33 which is curious because enflurane,73 halothane,34 87 and F385 are quite effective at suppressing the latter. Thus, anaesthetics do not have equivalent effects on different receptor subtypes, implying that stimulation of a single PKC isoform is unlikely to account for these results.
Muscarinic and other G-protein/PLC coupled receptors also modulate high-voltage-operated calcium channels through G proteins, second messengers and PKC.64 Although anaesthetics can directly inhibit some high voltage activated calcium channels, they also perturb receptor/PKC-mediated regulatory pathways.20 M1 receptors, and phorbol esters or DAG, which activate R-type channels through PKC, are differentially affected by volatile agents.65 When R-type channel subunit combinations are expressed in oocytes, pre-application of the volatile agents halothane (0.59 mM) or isoflurane (0.7 mM) blocks the effect of direct PKC activators,64 consistent with inhibition of PKC. Effects on receptor modulation, however, are more complex and depend on agonist dose and timing. (It is also worth noting that a relationship has been established between PKC and the anaesthetic sensitivity of voltage-regulated sodium channels from skeletal muscle and brain.88 98 These channels are insensitive to halothane when expressed alone in Xenopus oocytes but acquire sensitivity when co-expressed with PKC-.)
In summary, it is difficult to reconcile these results with a simple model wherein one predominant species of PKC is activated by anaesthetic. Whether an explanation lies in the multiplicity of PKC isoforms in oocytes is unclear.
Experiments in whole animals
Early experiments in tadpoles suggested an important role for protein kinases, possibly PKC, in anaesthesia. Staurosporine, a relatively non-specific kinase inhibitor, caused behavioural effects, immobility and loss of righting reflex, similar to that of halothane.38 More recently the anaesthetic sensitivity of mice lacking PKC- or isoforms has been tested. (Null mice have been generated for PKC , ß, and 
, but neurological and anaesthetic testing has been limited to and isoforms.91 128) Of particular interest to anaesthetic mechanisms, Harris and colleagues have shown that null mice have a reduced sensitivity to the sedative48 and anxiolytic17 effects of ethanol. The changes observed, however, are, to a large extent, dependent on the genetic background of the particular mouse strain,18 suggesting that sedative-related phenotypes are produced through the actions of multiple genes, in addition to PKC-. Interestingly, mice lacking PKC- have a reduced GABAA receptor response to ethanol,48 which normally enhances GABAA activity. The mechanism, however, is unclear. PKC generally down-regulates the GABAA receptor through direct phosphorylation of the channel19 and receptor internalization.27 36 Thus the absence of PKC- does not provide a simple explanation.
PKC- null mice also have complex phenotypes. These animals have diminished pain receptor sensitization, suggesting that PKC- plays a role in hyperalgesic states.67 Markedly enhanced sensitivity to ethanol, benzodiazapines and barbiturates is also observed,56 and this appears to be related to GABAA receptor sensitivity, suggesting that phosphorylation by this PKC subtype leads to changes in receptor function and responses to these allosteric modulators.
The contrary sensitivities of the and null mice suggest that PKC- facilitates while PKC- diminishes the effects of ethanol, altering neuronal responses that affect GABAergic transmission. Whether the phenotypes of these animals reflect any direct action of ethanol or other drugs on PKC is unclear. Other complications cloud even these broad interpretations. Unfortunately the behaviour of these mice indicates some generalized developmental changes that influence global patterns of CNS function, consequently affecting anaesthetic sensitivity. Similar difficulties have been encountered in mice lacking GABAA receptor subunits.57 To eliminate developmental differences, an inducible knock-out or knock-down (antisense RNA or interfering protein) design would be required. Alternatively, more subtle changes in anaesthetic-sensitive PKC subtypes could be engineered by gene knock-ins (replacing the natural exon with a mutated sequence) but this requires knowledge of the anaesthetic binding site(s). Even this elegant approach may provide equivocal results, since the actions of anaesthetics are so closely linked to binding of the endogenous ligand. Thus, it is unlikely that an anaesthetic-related mutation that did not also affect binding and action of the natural activator would be found.
Future directions
Whilst experiments in oocytes implicate PKC (or closely related kinases) as essential for the actions of anaesthetic compounds, the mechanistic implications are less clear. The simple interpretation that anaesthetics activate PKC, resulting in receptor and channel phosphorylation and thereby dampening subsequent agonist response, fails to account for all results, especially when one considers that ethanol has only been shown to inhibit PKC in vitro and that there are mechanistic differences among the anaesthetics, especially enflurane, which also blocks GTPS-stimulated chloride currents (calcium mobilization).73 115 Moreover, the differential sensitivity of receptor subtypes does not correlate with their susceptibility to down-regulation by PKC. Whether this inconsistency is the result of differential sensitivity and action of the various PKC subtypes or other kinase requirements is unclear. Also uncertain is whether some effects of anaesthetic on PKC activity involve changes in the cellular levels of activators such as DAG and calcium rather than a direct effect on PKC itself. (Although AT1A and mGlu1 receptor subtypes were insensitive to anaesthetic compounds, at least in the oocyte expression system, these receptors contain multiple PKC phosphorylation sites,104 and they do desensitize via PKC-catalysed phosphorylation, especially at low doses of agonist.11 39)
Do the effects of anaesthetics on metabotropic receptor coupling reflect a direct or permissive role for PKC? In altering the receptor, has the mutation simply erased a phosphorylation site or has the allosteric behaviour of the receptor changed? Could anaesthetics bound to the receptor, PLC or G-protein alter the receptor conformation so as to make it a better PKC substrate? These three proteins form an activation complex wherein the agonist-occupied receptor maintains an active heterodimer of G-protein and PLC, both of which are subject to PKC phosphorylation. It is possible that anaesthetics bound to this complex favour an agonist-occupied state of the receptor, one that is more readily phosphorylated, thereby accelerating the natural desensitization process and requiring PKC.
Despite the wealth of anaesthetic data, some important information is lacking, especially the phospho-states of the receptors and identities of the PKC isoforms involved. Identification of the PKC subtypes should be possible with isoform-specific peptide inhibitors117 – these results would help focus the attention of in vitro experiments. With current expression technologies it should be possible to test directly whether n-alkanols or volatile anaesthetics truly enhance PKC-catalysed phosphorylation of the receptors in a reconstituted system – which has yet to be demonstrated directly. Procedures for reconstituting muscarinic receptors with Gq and PLC in artificial membranes have been worked out.110 Using this experimental model and truncated forms of PKC that are catalytically active but insensitive to anaesthetics, it should be possible to test whether the receptor or other components become better substrates for the enzyme when anaesthetics are present. Given that localization of PKC is key to its action, studies of whether and how anaesthetic compounds affect subcellular distribution of specific PKC subtypes are also needed. This approach has already provided some insight into how ethanol and other short-chain n-alkanols may modulate PKC- and PKC- in cultured cells.43 44
G-proteins
Heterotrimeric guanine nucleotide-binding proteins (G-proteins) are bound to the inner leaflet of the plasma membrane, where they couple metabotropic and other receptors to an array of ion channels and second messenger generating effector enzymes.107 Widely expressed in the CNS, G-proteins directly and indirectly modulate neurotransmitter release from the presynaptic terminal and excitability of the post-synaptic membrane. Thus, any actions of anaesthetics that interfere with the function of G-proteins should have a profound effect on neurotransmission, in addition to processes outside direct CNS control. Recent evidence supports the view that early steps in metabotropic signalling are sensitive to haloalkanes, ethers and n-alkanols. Whether the molecular targets include G-proteins, receptors, or the protein kinases that modulate their coupling to receptor and effector proteins, is presently unclear.
G-proteins consist of a guanine nucleotide binding subunit and a ß heterodimer. G subunits are comprised of an -helical and a ras-like domain.127 Nucleotide occupies the cleft between the domains engaging flexible loops. The loops, part of the so-called switch region, recognize distinctive determinants on GDP and Mg2+-GTP, giving rise to three conformations: GDP, Mg2+-GTP, and empty, which are coupled to N and C terminal structures that form the binding/activation surfaces for ß subunits, receptors, effectors and other proteins. The N or C terminus or both are lipid modified, having subunit-specific combinations of myristoyl, palmitoyl, farnesyl and geranylgeranyl groups.
Gß subunits have propeller-like projections (WD-40 motifs) that form the binding surfaces for subunit, activated receptor and down-stream effector proteins.86 ß and subunits are coordinately expressed whereupon they form an irreversible complex associated through a coiled-coil structure. Each subunit is also lipid modified, linking the heterodimer to the membrane surface.
In the current view of G-protein activation, agonist favours a state of the receptor that binds the G-protein heterotrimer, catalysing exchange of GDP for GTP on the subunit. The GTP-charged subunit and free Gß heterodimer diffuse in the plane of the membrane, engaging downstream signalling components. Two modes of desensitization are normally triggered by activation: one involves receptor/G-protein phosphorylation by various protein kinases, including PKC, and the other involves regulators of G-protein signalling (RGS) proteins30 that down-regulate a cognate set of G-protein subtypes (Fig. 4). By stimulating GTP hydrolysis, RGS proteins return the subunit to an inactive GDP state from which it can retrieve the released ß heterodimer. In many cases the effector enzyme itself serves an RGS-like function, such as PLC.106
|
Combinatorial regulation
There are 23 different
subunits, divided into four major classes: s (olf), i (o, t, g, z), q (11, 14, 15)and 12 (13), each linked to a specific set of downstream effectors. In addition there are six different ß and twelve unique subunits. G-protein heterotrimers are generally classified according to which subunit they contain (Table 1). Particular combinations of , ß and appear to dictate the specificity for coupling each receptor subtype to a characteristic set of signalling pathways.137
|
Most
, ß and subunit combinations are possible, giving rise to more than 1600 possible combinations. When coupled to more than 2000 different receptor genes encoded in mammals,78 and considering that some receptors can interact with more than one G-protein subtype,41 the number of potential combinations becomes truly enormous. Adding further layers of combinatorial control and complexity, these receptors form homo- and heterodimers,41 suggesting mechanisms for modulating signal output, coincidence detection and signal integration. Finally, subunits are subject to phosphorylation, giving rise to functionally distinct phospho-states of each G-protein.24
Tissue distribution
G-protein tissue distribution is generally widespread, with a few exceptions such as 15 and 16, confined to hematopoietic cells, gust, expressed primarily in taste organs, and t1 and t2 (transducins), expressed in retinal rods and cones. Specific enrichments of the other subtypes in differing brain and spinal cord regions have been observed but no clear rules governing patterns have been uncovered.9 10 61
Work on the central nervous system
Studies of the CNS support the view that volatile anaesthetics disrupt receptor/G-protein coupling, perhaps at the level of the G-protein itself. Halothane prevents the shift to the low-affinity state of the muscarinic receptor, which is observed in the presence of the non-hydrolysable GTP derivative, guanyl-5'-yl imido diphosphate (GMPPNP).5 Similarly, halothane affects oxotremorine binding in rat brainstem membranes.32 Persistence of this high-affinity receptor state is also found with other volatile agents like chloroform, enflurane, isoflurane and diethylether.3 Volatile anaesthetics also cause a rightwards shift in the dose of clonidine, a partial 2-agonist, to inhibit adenylyl cyclase (AC). This same dose of halothane antagonizes the decrease in clonidine binding affinity observed in the presence of GMPPNP. These results suggest that volatile agents prevent agonist-induced dissociation of receptor from the heterotrimeric G-protein or decrease the binding of GTP analogue or both, thereby preserving the high-affinity state and uncoupling metabotropic receptors from their downstream effector proteins. Not all metabotropic receptors are affected by halothane in membrane preparations. Serotoninergic 5HT-1A and adensoine A1 receptors were insensitive to halothane (0.4–1.9 mM).5 81 Although a modest decrease in high-affinity binding was observed, the change in affinity due to GMPPNP was unaffected. (There is a disparity between the work of Aronstam and colleagues on the CNS and that of Bazil and Minneman,13 who found no effect of 1.25% halothane on binding of muscarinic agonist when a comparable experiment was performed at 37°C. They suggested that the effects reported by Aronstam and colleagues5 were obtained at tissue concentrations that were too high to be relevant, because they did not correct tissue concentrations for the lower [20°C] temperature, which was chosen because it produced the optimal coupling between receptors and G proteins in these isolated membranes.7 However, later studies by Aronstam and colleagues performed at 37°C2 and at 30°C90 confirmed earlier observations.)
Ethanol also affects the affinity of muscarinic receptors in rat brainstem membranes.2 Muscarinic-stimulated GTPase activity and GTP binding to membranes were also suppressed, consistent with an affect of this n-alkanol on receptor/G-protein coupling.120 Ethanol at low concentration also blocks Gi/o mediated inhibition of AC activity in rat brain cortex and in neuronal cell lines.12 23 These results may, however, be explained by activation of Gs rather than inhibition of the opposing Gi/o dependent pathway; indeed most evidence supports the view that ethanol promotes activation of AC by Gs.140 Although other n-alkanols operate similarly, agents like chloroform and halothane do not. In general, the volatile anaesthetics have a much less consistent effect on Gs coupling to AC, perhaps because of the differences in AC isoforms and their differential distribution among and within tissues. For example, halothane inhibits isoprenaline activation of lipolysis, but this appears to result from suppression of hormone-sensitive lipase downstream of cAMP production rather than an effect on coupling of the ß-adrenoceptor to AC.102
Acute ethanol treatment (50–200 mM) has been shown to decrease the binding affinity of agonists but not antagonists to the -opioid receptor expressed in neuroblastoma cells.42 The magnitude of the effect was temperature dependent: at 37°C, 200 mM ethanol caused a 20-fold decrease in apparent affinity and 2.5-fold increase in Bmax. This increase in maximum binding was accounted for by ethanol blockade of agonist-induced receptor internalization. Agonist-stimulated binding of GTPS to membranes was also suppressed (50% by 50 mM ethanol), although the G-proteins involved were not identified. By contast, n-alkanols (C4, C6, C8) inhibit binding of agonist and antagonist to 5HT1A receptors in hippocampal membranes in the presence or absence of GTPS. Here, ethanol (10–100 mM) had no effect on agonist but inhibited antagonist binding. In the presence of GTPS, inhibition by ethanol was diminished.47 These results highlight differences among the n-alkanols and suggest that the effects of ethanol may be linked to receptor-specific G-protein interactions.
Work on myocardium
Many halogenated anaesthetics have a negative ionotropic and chronotropic effect on the heart in isolation and in patients.99 Paradoxically, these agents, especially halothane, sensitize the myocardium to the effects of exogenous catecholamines, through a direct action on the heart. For example, halothane enhances contraction induced by coadministration of the ß-adrenoceptor agonist isoprenaline in electrically driven human ventricular preparations.15 In myocardial membranes, halothane increases stimulation of AC by isoprenaline, NaF, cholera toxin and GMPPNP, but not the stimulation produced by forskolin, which acts directly on this enzyme.15 90 118 Treatment with pertussis toxin (PTX) increases isoprenaline-, NaF- and GMPPNP-stimulated AC activity, but halothane is unable to increase this response further. Halothane completely suppresses muscarinic receptor GTPase activity in rat atrial membranes, with an EC50 of 0.3 mM, further evidence implicating receptor/G-protein coupling as a target.4 These observations imply that Gi is uncoupled from downstream effectors by PTX and halothane, operating through analogous mechanisms. However, many of these effects are observed in failing but not healthy myocardium,15 suggesting that the levels and activities of other signal transduction components (receptors, kinases, RGS proteins, etc.) may be critical for the effects of anaesthetics to rise above the natural feedback mechanisms and relative receptor reserves.
In ventricular myocytes, muscarinic receptor activation of Ik, a Gß regulated ion channel, is slowed by halothane and isoflurane, although the magnitude of Ik conductance is increased by the former.77 The slowing of Ik activation is consistent with an uncoupling of receptor/G-protein from the channel. One plausible mechanism for slowed activation could entail the sequestration of released ß subunits by receptor-associated G bound to anaesthetic. However, the increased current, which was attributed to an increased opening probability, is most consistent with an additional direct effect on the channel. This work has recently been extended in oocytes (see below).
General anaesthetics also effect Na+ channel activity – through a mechanism that is modulated by G-protein, albeit at high anaesthetic concentrations. Depolarization-induced Na+ channel conductivity in guinea-pig ventricular myocytes is substantially reduced by halothane (1.2 mM) and isoflurane (1 mM).136 Effects of both were reduced by GDPßS, suggesting the involvement of one or more G-proteins, but here the effect of halothane was enhanced by pretreatment with PTX whereas isoflurane was resistant. Further differences among anaesthetics were suggested by the finding that the PKC inhibitor BIS enhanced the effects of isoflurane but not halothane. The 1-adrenergic suppression of Na+ channels was enhanced by pre-application of isoflurane or halothane. Co-application of GDPßS, PTX or BIS markedly reduced the effects of agonist and anaesthetic, although as a percentage of the agonist response the effects of halothane were still observed, leading the authors to conclude that the interactions between this anaesthetic and agonist were independent of G-proteins and PKC. The residual effects of isoflurane on agonist inhibition were blocked however, leading to the converse conclusion with respect to this drug. In both cases, anaesthetic effects were already attenuated by these broadly acting reagents; thus, definitive mechanistic conclusions seem unjustified. Rather, the effects of PTX and GDPßS in modulating halothane sensitivity suggests a complex cross-talk among G-protein-coupled pathways that somehow governs anaesthetic sensitivity. How this influences 1-adrenergic regulation of the Na+ channel is unclear.
Work in smooth muscle
Volatile agents like halothane cause relaxation of certain smooth muscles and appear to disrupt the coupling of muscarinic receptors to G-protein dependent pathways.63 In permeabilized canine tracheal smooth muscle, halothane (0.8–0.9 mM) suppresses acetylcholine-induced potentiation of regulatory myosin light chain (MLC) phosphorylation,60 which normally enhances contraction at submaximal calcium levels. This potentiation is dependent on the presence of GTP. AlF-induced potentiation, a mimic of the GTPP, is also blocked, but much more effectively when halothane is added before AlF,63 suggesting that halothane inhibits dissociation of the AlF-GDP charged monomer from the ß heterodimer. Once these subunits are allowed to dissociate, the anaesthetic is less effective. Potentiation is also dependent on a monomeric rho-related GTPase downstream of the muscarinic receptor, the G-protein heterotrimer and MLC phosphatase, which is normally inhibited by this low molecular weight G-protein.63 These results are consistent with halothane causing a net increase in MLC phosphatase activity without changing MLC kinase activity.46 It is worth noting that earlier studies showed that pretreatment of guinea-pig ileum myenteric plexus longitudinal muscle with PTX greatly reduced the sensitivity to halothane, shifting the IC50 from 2% to 9% (vol).103 Clearly, at least some of the actions of halothane are intimately linked to Gi/o activity in smooth muscle.
In addition, butanol, hexanol and octanol are able to completely relax airway smooth muscle.113 When a muscarinic agonist is applied, these n-alkanols suppress the enhanced calcium sensitivity in a manner similar to halothane, although the exact mechanism has yet to be described. Ethanol also relaxes smooth muscle and inhibits rMLC phosphorylation induced by muscarinic stimulation,45 but here calcium sensitivity is increased, presumably via an independent mechanism.
Work on oocytes
As discussed in the section on PKC, Xenopus oocytes have been used to study the coupling between metabotropic receptors, G-proteins and PLC. From these studies it appears that receptors which couple strongly to PLC through Gq are suppressed by general anaesthetic agents via PKC-catalysed phosphorylation of the receptor. Nonetheless, anaesthetic can suppress GTPS-stimulated calcium mobilization,73 suggesting that some of these agents also block at the level of GTP binding proteins or downstream effectors, such as PLC. Here, an action on Gi or Go is also plausible, since stimulation of PLC in oocytes is PTX-sensitive.105
The effects of anaesthetic agents on receptor/G-protein regulated ion channels have also been examined. G-protein-coupled inwardly rectifying K+ channels (GIRKs), which play an important role in neurotransmission,79 80 are clearly affected when expressed in oocytes. Both brain (GIRK1/2) and cardiac (GIRK1/4) types are activated by ethanol,70 71 whereas other inwardly rectifying K+ channels are not. A homologous series of n-alkanols from C1 to C3 activate GIRK1/2 and GIRK1/4 with decreasing efficacy, whereas pentanol is inhibitory.70 71 Heterodimers containing the GIRK2 subunit and GIRK2 homodimers are most sensitive.71 PTX had no significant effect on ethanol activation of either channel heterodimer,70 suggesting a direct action on the channel itself. GABAB- and µ-opioid-receptor coupling to the channel was also enhanced by ethanol,70 71 with the degree of stimulation being independent of protein kinase A and PKC.71 By contrast with these results, clinical concentrations of halothane, isoflurane and enflurane and F3 inhibited basal GIRK1/2-, GIRK2- and µ-opioid-receptor-activated GIRK2 channel currents, whereas GIRK1/4 was resistant to all except F3.138 This demonstrates clear differences between the anaesthetics and their effects on specific subunit combinations. In a similar study, 0.1–0.3 mM halothane suppressed M2 muscarinic receptor activated GIRK1/4 and GIRK1 channels.135 Halothane also blocked activation by GTPS, suggesting that this anaesthetic disrupted the coupling of ß subunits with the channel. The apparent differential halothane sensitivity of M2135 and µ opioid138 receptors and stimulation of the GIRK1/4 channel is consistent with receptor-specific actions of this anaesthetic at the level of G-protein coupling.
Other systems
In turkey erythrocytes, low concentrations of halothane (0.1–0.3 mM) significantly inhibit stimulation of PLC by isoprenaline and P2Y purinergic agonists, whose receptors are coupled to PLC through Gi.109 In the case of isoprenaline, PLC activity was reduced nearly 50%; however, the effects were biphasic, with stimulation of PLC observed at high anaesthetic concentrations, obscuring any further inhibition. In endothelial cells cultured from aorta, ethanol (10–100 mM) enhances nitric oxide synthase activity in a PTX-sensitive manner.53 Interestingly, ethanol also enhances PTX-catalysed ADP-ribosylation of Gi, suggesting that this n-alkanol alters subunit conformation or ADP-ribosyl transferase activity.
Work in Caenorhabditis elegans
The nematode species C. elegans has been exploited to identify genes that modulate sensitivity to volatile anaesthetics. Among the most notable are syntaxin and the Go subunit. Mutations in syntaxin, part of the presynaptic complex of proteins regulating transmitter release, modulate anaesthetic sensitivity over a 33-fold range, measured as a loss of coordinated locomotion.132 133
Go and related G-proteins are negative modulators of transmitter release, suppressing, in a voltage-dependent manner, calcium channel opening and vesicle fusion in C. elegans and man.66 Four different loss-of-function mutations in the subunit resulted in decreased sensitivities to isoflurane (2-fold), but only one, an amino acid substitution near the C-terminus (sy192), was associated with both isoflurane and halothane resistance. Genetic analysis of sy192 suggested an interference with other components of pathways under Go control.132 Crossing a hypersensitive syntaxin strain with an isoflurane-resistant Go mutant leads to the latter phenotype, showing that syntaxin-related hypersensitivity depends on normal Go function. Consistent with these results, over-expression of the RGS protein, EGL-10, a negative modulator of Go, or a reduced function mutation in eat-16, an RGS that operates on Gq, produced similar resistance to both anaesthetics.132
Whatever the mechanism, it seems unlikely that Go is the sole/principal anaesthetic target of halothane in C. elegans. Rather, another component associated with transmitter release machinery (SNARE complex) of syntaxin, SNAP-25, VAMP and N/P type calcium channel is more likely. The interfering effects of the RGS, egl-10, and the sy192 mutation, which are located near the ß binding interface of Go, suggest a shift in the normal equilibrium between and ß subunits, resulting in a higher proportion of Go-GDP and ß heterotrimer complex and a reduction in the amounts of free ß heterodimer. Since free ß appears to inhibit the SNARE complex,141 these mutations should increase neurotransmission, which appears to be the case. Nonetheless, enhanced transmitter release opposing the actions of anaesthetics operating elsewhere cannot be the sole explanation for resistance to these drugs.132 Likewise, developmental defects cannot account for these observations, since inducing wild-type Go expression after development recovers anaesthetic sensitivity.
Studies in transgenic animals
Most of the genes encoding G subunits have been inactivated in mice, but targeted mutations of Gß or G have not been reported.93 Interestingly, Go-deficient mice appear to be hyperalgesic,59 Gz-deficient mice exhibit altered responses to a variety of psychoactive drugs,139 and Golf-deficient mice exhibit dramatically reduced electrophysiological responses to odorants.14 Mice lacking Gq develop ataxia, with clear signs of motor coordination deficits, although, like PKC- null mice, this appears to be a developmental problem.68 94 Despite the availability of G transgenic mice, the sensitivity of these mice to anaesthetic agents has yet to be tested.
Purified GTP-binding proteins
To test the idea that volatile anaesthetics directly affect G subunits, we examined exchange of bound GDP for GTP,100 a reaction normally stimulated by activated receptors.16 At clinical doses halothane substantially reduced the extent of nucleotide exchange for i1, i2, i3 and s, but not o. Of all the subunits tested, Gi2 was most sensitive. Comparisons of halothane and other volatile anaesthetics, isoflurane, enflurane and sevoflurane, showed that all drugs suppress exchange on i2 at clinical concentrations, with isoflurane being the most effective. Similar results were obtained when recombinant Gi1 lacking fatty acyl chain modification was tested in the absence of detergents and in membranes prepared from Sf9 cells overexpressing Gi1 and ß subunits.100 By contrast to the sensitive subunits, the low molecular weight G-protein, ras, was resistant to high concentrations of each drug. A recent study of transducin showed that this Gi/Go family member is insensitive to halothane, as measured by photocrosslinking of radiolabelled drug.58 Thus, the anaesthetic agents display measurable degrees of G-protein specificity in vitro.
Since binding and exchange of the guanine nucleotide are coupled to the conformations of the switch regions and the receptor/ß subunit interface, we tested whether anaesthetics could affect the binding of the subunit to the ß heterodimer. When charged with Mg2+-GTPS, the soluble subunit has a low affinity for the ß vesicles, but in the presence of anaesthetic, formation of ß complex is greatly enhanced (Fig. 4). A similar, although less dramatic, enhancement of ß complex formation is observed when the subunit is charged with GDP. These effects are reversible and dose dependent (EC50 0.25 mM).
These results obtained with the purified protein suggest an uncoupling of the normal conformational changes between the switch regions and the N/C-terminal regions of the G subunits that form the ß binding interface. (An effect of anaesthetics on ß seems unlikely given the stability of the heterodimer, whose WD-40 motifs form a rigid platform for the docking of other proteins; thus, the conformation of the ß binding interface in Mg2+-GTPS-charged subunits must resemble that of the GDP state.) In the GDP state, the disorder in the switch region permits the folding of the N-terminal and C-terminal residues into a stable micro-domain. Conversely, these regions become highly disordered when GTP and Mg2+are bound (Fig. 5). Thus, G subunits use the favorable energy of binding GTP and Mg2+ to select a stable protein conformer, flicking open a switch that triggers the release of ß subunits, thereby modulating downstream effector proteins.69 127 In the presence of anaesthetic, the binding interface is altered, enhancing subunit affinity for ß. Hence, anaesthetics are predicted to lock the metabotropic receptor into a persistent, high-affinity complex with the G-protein, which would explain how volatile anaesthetics prevent the decrease in receptor affinity for agonist that is normally induced by non-hydrolysable GTP analogues, as reported by Aronstam and others.5–8 42 112 This could have profound implications for the operation of this molecular switch, since it would effectively prevent GTP-charged and ß subunits from coupling to their downstream effectors.
|
Future directions
Many studies that measure responses downstream of G-protein coupling use high concentrations of anaesthetics, bringing into question the relevance of these pathways to the anaesthetic state. Nonetheless, as more reductionist approaches have simplified these systems, it is clear that receptor/G-protein coupling is sufficiently sensitive to a wide range of anaesthetics. Once the most sensitive metabotropic receptor/G-protein pairs are identified, clear predictions can be generated that should be testable in neuronal circuits of the brain and spinal cord.
More work is needed on a molecular level to explore the anaesthetic sensitivities of the G-proteins themselves, particularly Gq, which couples to many of the receptors examined in oocyte experiments. Thus far it appears that coupling to Gq is sensitive at the level of the receptor (through PKC phosphorylation) whereas coupling to Gi/o seems sensitive at the level of G-protein or receptor/G-protein complex. Definitive answers will only be provided by studying purified receptors and G-proteins reconstituted into artificial bilayers.
Conclusion and perspective
This review has focused exclusively on PKC and G-proteins because they are integral to metabotropic receptor signalling, which is clearly affected by general anaesthetics. Because of space limitations we did not discuss other important targets including calcium release channels such as the ryanodine receptor,25 97 and pumps such as the plasma membrane calcium ATPase,40 74 or actin.62 Other targets are also undoubtedly important, as shown in studies in C. elegans. Thus, the data do not support a single site theory unless one argues that all these other effects are irrelevant. A more realistic theory would be similar to the multi-site agent-specific hypothesis proposed by MacIver,76 wherein each drug is associated with its own array of protein targets (some common and some agent specific), with distinct subsets for each aspect of anaesthesia (sedation, amnesia and immobility).
Acknowledgement
The work in the authors’ laboratory is supported by NIH Grant GM-60376.
References
1 Alagarsamy S, Sorensen SD, Conn PJ. Coordinate regulation of metabotropic glutamate receptors. Curr Opin Neurobiol 2001; 11: 357–62[ISI][Medline]
2 Anthony BL, Dennison RL, Jr., Narayanan TK, Aronstam RS. Diethyl ether effects on muscarinic acetylcholine receptor complexes in rat brainstem. Biochem Pharmacol 1988; 37: 4041–6[ISI][Medline]
3 Anthony BL, Dennison RL, Aronstam RS. Disruption of muscarinic receptor-G protein coupling is a general property of liquid volatile anesthetics. Neurosci Lett 1989; 99: 191–6[ISI][Medline]
4 Aronstam RS, Anthony BL, Dennison RL, Jr. Halothane effects on muscarinic acetylcholine receptor complexes in rat brain. Biochem Pharmacol 1986; 35: 667–72[ISI][Medline]
5 Aronstam RS, Narayanan TK. Temperature effect on the detection of muscarinic receptor-G protein interactions in ligand binding assays. Biochem Pharmacol 1988; 37: 1045–9[ISI][Medline]
6 Aronstam RS. G-Protein systems. In: Yaksh TL, Lynch C, Zapol WM, Maze M, Biebuyck JF, Saidman LJ, eds. Anesthesia—Biological Foundations. Philadelphia: Lippincott-Raven, 1997; 45–53
7 Aronstam RS, Dennison RL, Jr. Anesthetic effects on muscarinic signal transduction. Int Anesthesiol Clin 1989; 27: 265–72[ISI][Medline]
8 Aronstam RS, Triggle DJ, Eldefrawi ME. Structural and stereochemical requirements for muscarinic receptor binding. Mol Pharmacol 1979; 15: 227–34
9 Asano T, Morishita R, Kato K. Two forms of G(o) type G proteins: identification and distribution in various rat tissues and cloned cells. J Neurochem 1992; 58: 2176–81[ISI][Medline]
10 Asano T, Nagahama M, Kato K. Subcellular distribution of GTP-binding proteins, Go and Gi2, in rat cerebral cortex. J Biochem 1990; 107: 694–8
11 Balmforth AJ, Shepherd FH, Warburton P, Ball SG. Evidence of an important and direct role for protein kinase C in agonist-induced phosphorylation leading to desensitization of the angiotensin AT1A receptor. Br J Pharmacol 1997; 122: 1469–77[ISI][Medline]
12 Bauche F, Bourdeaux-Jaubert AM, Giudicelli Y, Nordmann R. Ethanol alters the adenosine receptor-Ni-mediated adenylate cyclase inhibitory response in rat brain cortex in vitro. FEBS Lett 1987; 219: 296–300[ISI][Medline]
13 Bazil CW, Minneman KP. Effects of clinically effective concentrations of halothane on adrenergic and cholinergic synapses in rat brain in vitro. J Pharmacol Exp Ther 1989; 248: 143–8
14 Belluscio L, Gold GH, Nemes A, Axel R. Mice deficient in G(olf) are anosmic. Neuron 1998; 20: 69–81[ISI][Medline]
15 Bohm M, Schmidt U, Gierschik P, Schwinger RH, Bohm S, Erdmann E. Sensitization of adenylate cyclase by halothane in human myocardium and S49 lymphoma wild-type and cyc cells: evidence for inactivation of the inhibitory G protein Gi alpha. Mol Pharmacol 1994; 45: 380–9[Abstract]
16 Bourne HR. How receptors talk to trimeric G proteins. Curr Opin Cell Biol 1997; 9: 134–42[ISI][Medline]
17 Bowers BJ, Elliott KJ, Wehner JM. Differential sensitivity to the anxiolytic effects of ethanol and flunitrazepam in PKCgamma null mutant mice. Pharmacol Biochem Behav 2001; 69: 99–110[ISI][Medline]
18 Bowers BJ, Owen EH, Collins AC, Abeliovich A, Tonegawa S, Wehner JM. Decreased ethanol sensitivity and tolerance development in gamma-protein kinase C null mutant mice is dependent on genetic background. Alcohol Clin Exp Res 1999; 23: 387–97