G Proteins and Opioid Receptor-Mediated Signalling

Abstract

Most opioid receptor-mediated functions appear to be mediated through G protein interactions, therefore an understanding of opioid signalling requires knowledge of those interactions. This review chronicles the studies examining these interactions for all the opioid receptor subtypes, both in vivo and in vitro.

Introduction

For millennia, people have found pain relief from opium, a natural product from the poppy containing a wide variety of alkaloids such as morphine and codeine. From these drugs came thousands of analogues which have proved invaluable in furthering our understanding of opioid pharmacology. Opiates act through specific receptors located on nervous tissue, first identified in 1973 1, 2, 3. Their discovery quickly led to identification within the brain of a wide variety of endogenous peptides with opioid activity, including enkephalins 4, 5, 6, endorphins 7, 8, and dynorphins 9, 10, 11. Extensive studies of the pharmacology of these agents revealed that a series of opioid receptor subtypes exist, each with a distinct selectivity profile. Our understanding of these receptors increased enormously with their cloning 12, 13, 14, 15, 16, 17, 18, 19and the confirmation that they were members of the G-protein coupled receptor superfamily. Although the cloned opioid receptor subtypes are highly homologous at the amino acid level, their interactions with G-protein subunits vary. We will now review the evidence linking the various opioid receptor subtypes to individual G-proteins.

At the present time, pharmacological and biochemical evidence supports the existence of three major classes of opioid receptors (Table 1 and Table 2): μ, δ, and κ 20, 21. Each of these major classes contains putative receptor subtypes. Although all these receptor subtypes are able to mediate analgesia, their individual binding profiles and other pharmacological activities clearly distinguish each from the others. Four separate receptors within the opioid family have been cloned. DOR-1 encodes a δ receptor while MOR-1 and KOR-1 encode μ and κ₁ receptors, respectively. When expressed, each of the receptors reveals the anticipated binding selectivity based upon brain homogenate binding studies.

A fourth member of the opioid receptor family, KOR-3 or ORL-l, has also been cloned 22, 23, 24, 25, 26, 27, 28, 29. While this clone possesses very high homology to opioid receptors, it binds most traditional opioid ligands with low affinity. Several pieces of evidence support the opioid receptor nature of this clone. The κ agonist Dynorphin A activates potassium channels in oocytes expressing ORL-1 [30]and antisense studies have revealed a close association between the KOR-3 clone and the κ₃ receptor 22, 23. This association is strengthened further by the ability of a κ₃-selective monoclonal antibody (mAb8D8) to recognize the KOR-3 in vitro translation product on immunoblots [23]. However, the two receptors are not the same. Although six antisense oligodeoxynucleotides directed against sequences downstream from a splice site in the first transmembrane region of the mouse clone (KOR-3) block κ₃-mediated analgesia, the inactivity of five additional antisense probes targeting sequences upstream of the splice site raise the possibility that the κ₃ opioid receptor may be a splice variant of KOR-3 22, 23. The differences between the two are illustrated further by the recent discovery of an endogenous peptide, orphanin FQ or nociceptin (OFQ/N) 31, 32. OFQ/N labels the expressed orphan clone with subnanomolar affinity in binding studies and is functionally active in cyclase studies. Pharmacologically, it is quite unique in that it can produce hyperalgesia 31, 32. More detailed studies reveal a far more complex pharmacology in mice, where OFQ/N also can elicit a naloxone-sensitive analgesia 33, 34. However, OFQ/N does not compete for binding to any of the traditional opioid receptor subtypes, including κ₃. Although KOR-3 is closely related to the κ₃ receptor, they are clearly distinct.

Although four members of the opioid receptor family have been cloned 12, 13, 14, 15, 16, 17, 18, 19, 22, 23, 24, 25, 26, 27, 28, 29, they do not appear to account for all the subtypes identified in binding and pharmacological studies, raising the possibility of additional genes or extensive alternative splicing. Already there is strong evidence for functionally significant splice variants of the MOR-1 clone [35]. Antisense studies reveal that probes based upon the sequence of exons 1 and 4 all block morphine analgesia without affecting the analgesic actions of morphine-6β-glucuronide (M6G). In contrast, antisense probes from exons 2 and 3, which are inactive against morphine analgesia, potently block M6G analgesia. Similar results are seen with antisense mapping of KOR-3, as discussed above.

Despite the strong evidence implying functionally significant alternative splicing, few splice variants have been described and they do not appear to represent functionally distinct subtypes. Two alternatively spliced versions of rat and human MOR-1, differing only in five and eight amino acids (respectively) at the carboxy terminus, have been described 36, 37. Although the rat isoforms have different distributions within the caudate-putamen [38]and desensitize to opioid agonist treatment at different rates [36], they do not appear to correspond to the pharmacologically defined μ₁ and μ₂ receptor subtypes 20, 21. Splice variants of the orphan clone have also been described 23, 37, but their significance remains unclear. Obviously this continues to be a major area of investigation.

Early on, binding studies suggested that opioid receptors were members of the G-protein-coupled receptor superfamily. Indeed, many of the binding characteristics of this receptor superfamily were initially described in the opioid receptor literature. High affinity opioid agonist binding is increased by divalent cations 39, 40, 41, 42, 43, 44, 45, 46and decreased by sodium ions 40, 41, 42, 43, 44, 45, 47, 48, 49, 50, 51, guanine nucleotide triphosphate analogues 40, 41, 42, 44, 45, 49, 50or pertussis toxin 52, 53, 54, consistent with their coupling to Gα_i and/or Gα_o.

At the molecular level, opioids are linked functionally to inhibition of cAMP accumulation through μ 50, 55, 56, δ 57, 58, κ₁ 59, 60, and κ₃ receptors 45, 61. Opioids also modulate calcium 62, 63, 64, 65, 66and potassium channel conductance 67, 68, 69, 70, 71, and regulate neurotransmitter release 55, 72, 73, 74. Their actions are mediated through pertussis toxin (PTX)-sensitive 52, 75, 76and -insensitive mechanisms 71, 77.

Recent studies report opioid-mediated interactions with phospholipase C (PLC) 78, 79, 80, 81, 82or protein kinase C (PKC) 83, 84, and opioid-mediated increases in intracellular calcium levels corresponding to inositol (1, 4, 5)tris-phosphate-mediated calcium release [85]. Opioids also inhibit phosphoinositide hydrolysis 37, 78in foetal rat brain aggregates maintained in culture for seven days. After 21 days in culture, opioids stimulated phosphoinositide formation 78, 79. Opioid-mediated stimulation of inositol phosphate production in rat primary cultures and SH-SY5Y cells was PTX-sensitive 78, 79, 81.

Guanine nucleotide binding proteins (G proteins) are heterotrimers, composed of three distinct subunits (α, β, γ), which couple a large number of receptors to a variety of effectors (see Table 3; [86]). There are many possible αβγ combinations, each with unique functional properties. The identity of a distinct G-protein oligomer is currently ascribed to the α subunit, an enzyme possessing intrinsic GTPase activity. Agonist binding to a receptor protein sets in motion a series of events including: 1) dissociation of GDP from the α-subunit which is complexed to βγ, 2) binding of GTP to the α-subunit and its activation, 3) dissociation of the G-protein from the receptor, 4) dissociation of the α-GTP from βγ, and 5) the subsequent effector regulation through α, βγ or both. The β- and γ-subunits function as a closely associated complex. Originally considered to be shared among different α-subunits, it now appears that the βγ complex can modulate the activity of various effectors (not necessarily the effectors modulated by the α-subunit, Table 3; [87]).

Adenylyl cyclases are a family of enzymes (Table 4) which convert ATP to cyclic adenosine monophosphate (cAMP), an intracellular signalling molecule. Cyclic AMP activates several cAMP-dependent protein kinases which control numerous events from metabolism to gene transcription through phosphorylation [88]. The eight known types of adenylyl cyclases 89, 90, 91, 92vary in their: 1) sensitivity to calcium/calmodulin, 2) regulation by β- and γ-subunits of G-proteins, and 3) regulation by PKC 89, 90. Gβγ inhibits type I adenylyl cyclase in the presence of Gα_s or calcium-calmodulin and potentiates the stimulatory effect of Gα_s on adenylyl cyclase types II and IV.

The PLC enzyme family is diverse and complex; members are subclassified as β, γ, and δ isoforms, with at least three subtypes of each [86]. They hydrolyze phosphatidylinositol bisphosphate (PIP₂) to produce two second messengers, inositol 1,4,5-trisphosphate (Ins(1,4,5)P₃) and diacylglycerol (DAG). Ins(1,4,5)P₃ mobilizes intracellular calcium stores while DAG activates a number of isoforms of PKC, which is involved in a host of other regulatory cellular functions [93]. Only the PLC-β isoforms are sensitive to G-proteins. PLC-β1 is activated by members of the Gα_q family while PLC-β2 is stimulated by Gβγ subunits. The two isoforms also differ in sensitivity to pertussis toxin. PLC-β1 is PTX-insensitive because its modulator, Gα_q, cannot be modified by PTX [86]. In contrast, PTX blocks the actions of PLC-β2 [94], suggesting that the βγ subunits involved with its action were previously associated with PTX-sensitive G protein α subunits such as Gα_i and Gα_o.