Essential role of the C148-C227 disulphide bridge in the human 5-HT2A homodimeric receptor.

The 5-HT2A receptor is a homodimeric G protein-coupled receptor implied in multiple diseases, including schizophrenia. Recently, its co-crystallisation with the antipsychotic drugs zotepine and risperidone has revealed the importance of its extracellular domains in its pharmacology. Previous studies have shown that the non-specific disruption of extracellular disulphide bridges in the 5-HT2A receptor decreases ligand binding and receptor activation. There is enough evidence to hypothesize that this decrease may be due to a reduction of the disulphide bridge that links transmembrane domain 3 (TM-3) and extracellular loop 2 (ECL-2) of the 5-HT2A receptor via cysteine 148 (C148) and C227. Thus, to study the influence of the C148-C227 disulphide bridge on 5-HT2A receptor pharmacology, we substituted C148 and C227 in the human 5-HT2A receptor (WT) with alanines, to obtain two single mutants (C148A and C227A) and a double mutant (C148A/C227A), and the resultant DNA constructs were used to generate four stable cell lines. These substitutions reduced the binding of the 5-HT2A receptor to [3H]lysergic acid diethylamide ([3H]LSD) and impeded the 5-HT2A receptor-mediated activation of phospholipase C (PLC). Furthermore, bioluminescence resonance energy transfer (BRET) and western blotting analysis revealed that these mutations did not alter the homodimeric nature of the 5-HT2A receptor. However, fluorescence microscopy showed that these mutations hindered receptor trafficking to the cell membrane. These results illustrate the importance of the disulphide bridge between TM-3 and ECL-2 in maintaining the correct 5-HT2A receptor conformation to allow ligand binding and migration of the homodimeric receptor to the cell membrane.

The role of GPCR dimerisation/oligomerisation in receptor signalling.

A wide range of techniques have been employed to examine the quaternary structure of G-protein-coupled receptors (GPCRs). Although it is well established that homo-dimerisation is common, recent studies have sought to explore the physical basis of these interactions and the role of dimerisation in signal transduction. Growing evidence hints at the existence of higher-order organisation of individual GPCRs and the potential for hetero-dimerisation between pairs of co-expressed GPCRs. Here we consider how both homo-dimerisation/oligomerisation and hetero-dimerisation can regulate signal transduction through GPCRs and the potential consequences of this for function of therapeutic medicines that target GPCRs. Hetero-dimerisation is not the sole means by which co-expressed GPCRs may regulate the function of one another. Heterologous desensitisation may be at least as important and we also consider if this can be the basis for physiological antagonism between pairs of co-expressed GPCRs. Although there may be exceptions (Meyer et al. 2006), a great deal of recent evidence has indicated that most G-protein-coupled receptors (GPCRs) do not exist as monomers but rather as dimers or, potentially, within higher-order oligomers (Milligan 2004b; Park et al. 2004). Support for such models has been provided by a range of studies employing different approaches, including co-immunoprecipitation of differentially epitope-tagged but co-expressed forms of the same GPCR, co-operativity in ligand binding and a variety of resonance energy transfer techniques (Milligan and Bouvier 2005). Only for the photon receptor rhodopsin has the organisational structure of a GPCR been studied in situ. The application of atomic force microscopy to murine rod outer segment discs indicated that rhodopsin is organised in a series of parallel arrays of dimers (Liang et al. 2003) and based on this, molecular models were constructed to try to define and interpret regions of contact between the monomers (Fotiadis et al. 2004). Only for relatively few other GPCRs are details of the molecular basis of dimerisation available but within this limited data set, recent studies on the dopamine D2 receptor suggest a means by which information on the binding of an agonist can be transmitted between the two elements of the dimer via the dimer interface (Guo et al. 2005). Although the availability of cDNAs encoding molecularly defined GPCRs has allowed high-throughput screening for ligands that modulate GPCR function, this is performed almost exclusively in heterologous cell lines transfected to express only the specific GPCR of interest. Given that the human genome contains some 400-450 genes encoding non-chemosensory GPCRs, it is clear that any individual cell of the body may express a considerable number of GPCRs. Interactions between these, either via hetero-dimerisation, via heterologous desensitisation or via the integration of downstream signals can potentially alter the pharmacology, sensitivity and function of receptor agonists and hence produce varied responses. In this article, we will use specific examples to consider the role of homo-dimerisation/oligomerisation in GPCR function and whether either direct hetero-dimerisation or heterologous desensitisation between pairs of co-expressed GPCRs affects the function of the receptor pairs.

Dimerization of alpha1-adrenoceptors.

Three distinct genes encode alpha(1)-adrenoceptors. Although homodimers of each subtype have been reported, certain but not all combinations of heterodimers of the alpha(1)-adrenoceptors appear to form. Key studies in this field are reviewed and the approaches that have been applied to monitoring the selectivity and the basis of alpha(1)-adrenoceptor dimerization are discussed.

Regional distribution and cellular localization of 5-HT2C receptor mRNA in monkey brain: comparison with [3H]mesulergine binding sites and choline acetyltransferase mRNA.

The distribution of serotonin 5-HT(2C) receptor mRNA in monkey brain was studied by in situ hybridization and compared with the distribution of [3H]mesulergine binding sites as visualized by receptor autoradiography. 5-HT(2C) receptor transcripts showed a widespread and heterogeneous distribution. The strongest hybridization signal was detected in choroid plexus. In neocortex, 5-HT(2C) mRNA was detected in layer V of all cortical regions examined except in the calcarine sulcus, which was devoid of signal. Several structures within the striatum and basal forebrain were strongly labeled: nucleus accumbens, ventral aspects of anterior caudate and putamen, septal nuclei, diagonal band, ventral striatum, and extended amygdala. Several thalamic, midbrain, and brainstem nuclei also contained 5-HT(2C) mRNA. Comparison of the distributions of 5-HT(2C) mRNA and specific [3H]mesulergine binding sites showed a good agreement in the majority of brain regions, suggesting a predominant somatodendritic localization of 5-HT(2C) receptors. A possible localization to axon terminals of 5-HT(2C) receptors is suggested by the disagreement observed in some regions such as septal nuclei and horizontal limb of the diagonal band (presence of mRNA with apparent absence of binding sites) and interpeduncular nucleus (presence of binding sites with apparent absence of mRNA). Comparison of 5-HT(2C) receptor and choline acetyltransferase mRNA distributions indicate that some regions where cholinergic cells are located are also enriched in cells containing 5-HT(2C) mRNA. Although the present methodology does not allow strict colocalization of both mRNA species to the same cells, the codistribution observed in several regions provides a possible anatomical substrate for the described modulation of acetylcholine release by 5-HT(2C) receptors.

Multiple conformations of native and recombinant human 5-hydroxytryptamine(2a) receptors are labeled by agonists and discriminated by antagonists.

We have expanded previous studies with the 5-hydroxytryptamine (5-HT)(2) receptor agonist (+/-)-1-(2,5-dimethoxy-4-[(125)I]iodophenyl)-2-aminopropane [(+/-)-[(125)I]DOI] in human brain that had shown biphasic competition curves for several 5-HT(2A) receptor antagonists by using new selective antagonists of 5-HT(2A) (MDL100,907) and 5-HT(2C) (SB242084) receptors together with ketanserin and mesulergine. Autoradiographic competition experiments were performed with these antagonists in human brain regions where (+/-)-[(125)I]DOI labels almost exclusively 5-HT(2A) receptors (frontal cortex and striosomes). Furthermore, the effect of uncoupling receptor/G protein complexes on antagonist competition was studied with guanosine-5'-(beta,gamma-imido)triphosphate [Gpp(NH)p]. Competition experiments with (+/-)-[(3)H]1-(4-bromo-2,5-dimethoxyphenil)-2-aminopropane [(+/-)-[(3)H]DOB] were also performed in membranes from Chinese hamster ovary cells (CHOFA4) expressing cloned human 5-HT(2A) receptors. In both systems, ketanserin and MDL100,907 displayed biphasic competition profiles, whereas SB242084 and mesulergine competed monophasically. In absence of antagonist, 100 microM Gpp(NH)p decreased brain (+/-)-[(125)I]DOI specific binding by 40 to 50% and (+/-)-[(3)H]DOB specific binding to CHOFA4 cells by 30%. The remaining agonist-labeled uncoupled sites were still displaced biphasically by ketanserin and MDL100,907, with unaltered affinities. Saturation experiments were performed in CHOFA4 cells. (+/-)-[(3)H]DOB labeled two sites (K(d(h))= 0.8 nM, K(d(l)) = 31.22 nM). Addition of 100 microM Gpp(NH)p resulted in a single low-affinity (K(d) = 24.44 nM) site with unchanged B(max). [(3)H]5-HT showed no specific binding to 5-HT(2A) receptors. These results conform with the extended ternary complex model of receptor action that postulates the existence of partly activated receptor conformation(s) (R*) in equilibrium with the ground (R) and the activated G protein-coupled (R*G) conformations. Thus, both in human brain and CHOFA4 cells, the agonists possibly label all three conformations and ketanserin and MDL100,907 recognize with different affinities at least two of these conformations.

Mapping of 5-HT2A receptors and their mRNA in monkey brain: [3H]MDL100,907 autoradiography and in situ hybridization studies.

The anatomic distribution of serotonin 5-HT2A receptors visualized with [3H]MDL100,907 and of their mRNA detected by in situ hybridization were studied in monkey brain. Both autoradiographic patterns of signal showed heterogeneous distributions and were in general in good agreement in the majority of brain regions. In most neocortical areas, [3H]MDL100,907 presented a four-banded pattern with layers I and III-IV more intensely labeled and layers II and V-VI showing weaker labeling. 5-HT2A receptor mRNA was detected in layers III and IV, and in some cases also in layers II and V. In intra- and extra-calcarine areas of striate cortex a five-banded pattern was distinguished, with layers III and IVc-V showing the highest densities of [3H]MDL100,907 labeling. These two areas showed the highest neocortical hybridization signal. An unexpected finding was the presence of low densities of [3H]MDL100,907 labeling and 5-HT2A receptor mRNA in choroid plexus. Comparison of the distribution of [3H]MDL100,907 and [3H]ketanserin binding sites in monkey brain regions with high nonspecific [3H]ketanserin binding (caudate, putamen, substantia nigra, inferior olive) revealed specific binding of [3H]MDL100,907 with very low nonspecific binding. Some differences were noted between the distribution of [3H]MDL100,907-labeled 5-HT2A receptors in monkey brain and the previously reported distribution of these receptors in human brain: absence of striosome labeling in monkey striatum and different patterns of neocortical labeling. The present results provide the first detailed comparison of 5-HT2A receptor and mRNA distribution in primate brain. The observed species differences in 5-HT2A receptor distribution should be considered when extrapolating results among different species.

Human striosomes are enriched in 5-HT2A receptors: autoradiographical visualization with [3H]MDL100,907,[125I](+/-)DOI and [3H]ketanserin.

5-HT2A receptors have been visualized with [3H]MDL100,907 in selected human brain areas by autoradiography. These areas included caudate and putamen, nucleus dentatus of the cerebellum, substantia nigra, nucleus raphe dorsalis, locus coeruleus and inferior olive. In the striatum [3H]MDL100,907 labelling was compared with the pattern obtained with [125I](+/-)DOI and [3H]ketanserin. [3H]MDL100, 907 and [125I](+/-)DOI showed an identical patchy distribution which was hardly observed with [3H]ketanserin. In the remaining regions, [3H]MDL100,907 and [3H]ketanserin autoradiographical signals and percentage of specific binding were compared. Whereas the pattern of distribution was identical for both radioligands, [3H]MDL100,907 presented a much lower percentage of nonspecific binding compared with [3H]ketanserin. These results confirm the presence of 5-HT2A receptors in human striosomes and in those areas where [3H]ketanserin presented a high nonspecific binding, and they highlight the advantage of using [3H]MDL100,907 to visualize these receptors.

[3H]MDL 100,907 labels 5-HT2A serotonin receptors selectively in primate brain.

The selective antagonist for the 5-HT2A serotonin receptor MDL 100,907, recently characterized autoradiographically in rat brain, has been characterized as a radioligand for the visualization of this receptor in human and monkey brain. In both species [3H]MDL 100,907 binding to brain sections was saturable, had sub-nanomolar affinity (Kd = 0.14-0.19 nM in human brain; Kd= 0.16-0.19 nM in monkey brain) and presented a pharmacological profile consistent with its binding to 5-HT2A receptors (rank order of affinity for [3H]MDL 100,907-labeled receptors: MDL 100,907 > spiperone > ketanserin > mesulergine). The autoradiographical signal obtained with [3H]MDL 100,907 was compared to the signal obtained with [3H]ketanserin, [3H]RP62203 and [3H]mesulergine in both species, and to the distribution of 5-HT2A receptor mRNA as determined by in situ hybridization in monkey brain. At variance with the other radioligands, [3H]MDL 100,907 showed a single population of binding sites with extremely low levels of non-specific binding. As expected, mesulergine showed low affinity for [3H]MDL 100,907-labeled receptors and the autoradiographic pattern shown by [3H]mesulergine confirmed the lack of labeling of the 5-HT2A receptor by this radioligand in primate brain. The similarity of the distribution of [3H]MDL 100,907-labeled receptors and 5-HT2A mRNA in monkey brain, supports the selectivity of this radioligand for 5-HT2A receptors and suggests a somatodendritic localization of these receptors. The present results confirm [3H]MDL 100,907 as the radioligand of choice at present for the autoradiographic visualization of 5-HT2A receptors in mammalian brain including post-mortem human brain.

Early localization of mRNA coding for 5-HT1A receptors in human brain during development.

The distribution of 5-HT1A receptor mRNA in the human brain was studied in neonatal, children and adult cases by means of in situ hybridization histochemistry, using an oligonucleotide derived from the coding region of the human receptor. A prenatal pattern of development was observed. The hippocampus, raphe nuclei and neocortex presented high levels of hybridization already at the fetal/neonatal stage, fully comparable to the adult. A high and transient hybridization signal was found in cerebellum. These results support a role for 5-HT1A receptors in the regulation of neural development.

Selective visualization of rat brain 5-HT2A receptors by autoradiography with [3H]MDL 100,907.

The recently developed 5-HT2A receptor selective antagonist [3H]MDL100,907 ((+/-)2,3-dimethoxyphenyl-1-[2-(4-piperidine)-methanol]) has been characterized as a radioligand for the autoradiographic visualization of these receptors. [3H]MDL100,907 binding to rat brain tissue sections was saturable, had sub-nanomolar affinity (Kd = 0.2-0.3 nM), and presented a pharmacological profile consistent with its binding to 5-HT2A receptors (rank order of affinity for [3H]MDL100,907-labelled receptors: MDL100,907 > spiperone > ketanserin > mesulergine). The distribution of receptors labelled by [3H]MDL100,907 was compared to the autoradiographical patterns obtained with [3H]Ketanserin, [3H]Mesulergine, and [3H]RP62203 (N-[3-[4-(4-fluorophenyl)piperazin-1-y1]propyl]-1,8-naphtalenes ultam) and to the distribution of 5-HT2A receptor mRNA as determined by in situ hybridization. As opposed to the other radioligands, [3H]MDL100,907 labelled a single population of sites (5-HT2A receptors) and presented extremely low levels of non-specific binding. The close similarity of the distributions of [3H]MDL100,907-labelled receptors and 5-HT2A mRNA further supports the selectivity of this radioligand for 5-HT2A receptors and suggests a predominant somatodendritic localization of these receptors. The present results point to [3H]MDL100,907 as the ligand of choice for the autoradiographic visualization of 5-HT2A receptors.

5-HT receptors in mammalian brain: receptor autoradiography and in situ hybridization studies of new ligands and newly identified receptors.

In recent years the family of mammalian serotonin receptors has grown to 14 different subtypes, characterized by pharmacological or molecular biological techniques. In parallel, new ligand molecules have been developed for their study. However, selective ligands are not yet available to study every one of them. In addition the degree of selectivity of ligands, hitherto regarded as specific for a particular receptor subtype has been called in question by their affinities for newly discovered receptors. Consequently, a re-evaluation of past ligand receptor autoradiography work is necessary in view of the redefined receptor profiles of these ligands, and the introduction of newly developed ligands. A further difficulty for the characterization of these receptors is the absence of selective antagonist ligands which, for some of the subtypes, have become available only recently. In an attempt to overcome these difficulties we have combined in situ hybridization histochemistry and receptor ligand autoradiography to study the regional and cellular localization of several serotonin receptors in the rodent brain. In addition, for some receptors, we have expanded these studies to primates, including humans. We have found that the distribution of 5-HT1A receptors in monkey brain, labelled with the agonist 3H-8-OH-DPAT and the antagonist 3H-WAY 100635 was very similar at the levels examined, and corresponded well with that observed for the cells containing mRNA coding for this receptor, confirming the somatodendritic localization of 5-HT1A receptors in monkey brain. The labelling conditions to visualize 5-HT1F receptors in guinea pig brain, namely 3H-sumatriptan in the presence of 10(-8) M 5-CT to block 5-HT1D receptors, are suitable for visualizing this receptor, since the results agreed with those observed by in situ hybridization. By using 3H-ketanserin and 3H-mesulergine in parallel with in situ hybridization using the corresponding oligonucleotides, we were able to show that these ligands label respectively 5-HT2A and 5-HT2C binding sites in monkey brain. 5-HT4 receptors were localized in the brain of several species including humans by using 125I-SB 207710. In situ hybridization experiments performed in guinea pig confirmed that 5-HT4 receptors are localized on the terminals of the striatopallidal and striatonigral projections. 5-HT7 binding sites were labelled in rat and guinea pig brains by incubating with 3H-5-CT in the presence of 100 microM WAY 100135 and 250 microM GR 127935; the distribution obtained in both species agreed, in general, with that of the corresponding mRNA coding for them. These results are an illustration of the understanding of our current knowledge of the chemical neuroanatomy of the mammalian 5-HT system.