Receptors for Dopamine and Somatostatin: Formation of Hetero-Oligomers with Enhanced Functional Activity

See allHide authors and affiliations

Science  07 Apr 2000:
Vol. 288, Issue 5463, pp. 154-157
DOI: 10.1126/science.288.5463.154


Somatostatin and dopamine are two major neurotransmitter systems that share a number of structural and functional characteristics. Somatostatin receptors and dopamine receptors are colocalized in neuronal subgroups, and somatostatin is involved in modulating dopamine-mediated control of motor activity. However, the molecular basis for such interaction between the two systems is unclear. Here, we show that dopamine receptor D2R and somatostatin receptor SSTR5 interact physically through hetero-oligomerization to create a novel receptor with enhanced functional activity. Our results provide evidence that receptors from different G protein (heterotrimeric guanine nucleotide binding protein)–coupled receptor families interact through oligomerization. Such direct intramembrane association defines a new level of molecular crosstalk between related G protein–coupled receptor subfamilies.

In the brain, somatostatin (SST) is found in interneurons as well as projection neurons in different regions, and is thought to be an important physiological regulator of numerous functions (1). The actions of SST are mediated by a family of G protein–coupled receptors (GPCRs) with five subtypes, SSTR1 to SSTR5, that are widely distributed with high concentrations in the deeper cortical layers, the striatum, and most regions of the limbic system (2). Dopamine, like SST, acts through its own family of five GPCRs, D1R to D5R, that also display rich expression in the cerebral cortex, striatum, and limbic structures (3, 4). The SSTR and DR families share ∼30% sequence homology and appear to be structurally related. Behavioral and clinical evidence indicates an interaction between the somatostatinergic and dopaminergic systems (5–8). Intracerebroventricular injections of SST produce dose-dependent neurobehavioral changes progressing from hyperkinesia to catatonia (6). The dual excitatory and inhibitory effects occur through differential activation of postsynaptic or presynaptic DRs, respectively (3). Central administration of dopamine likewise activates both SST and SSTRs in the striatum (7).

We characterized the interaction between the long form of the human D2R and human SSTR5, a subtype previously reported to undergo ligand-dependent homodimerization and to form heterodimers with other SSTRs (human SSTR1) (9). DRs have also been shown to exist as dimers on the plasma membrane (10). Both receptors signal through inhibition of adenylyl cyclase via Gi proteins (2, 3). By immunocytochemistry, D2R and SSTR5 were colocalized in distinct neuronal subsets with the morphological characteristics of medium aspiny neurons in the striatum and pyramidal neurons in the cerebral cortex (Fig. 1) (11). We next investigated hetero-oligomerization of the two subtypes in a heterologous system by functional rescue of a partially active C-tail deletion mutant of human SSTR5 (Δ318 SSTR5) (12). This mutant displays complete loss of adenylyl cyclase coupling while retaining full agonist binding affinity. Δ318 SSTR5 was stably expressed in CHO-K1 cells either alone or with D2R. The mutant showed complete loss of the ability to inhibit forskolin-stimulated adenosine 3′,5′-monophosphate (cAMP) levels upon somatostatin-14 (SST-14) treatment (12). When coexpressed with D2R, however, SST induced dose-dependent inhibition of cAMP levels by the cotransfectants to a maximum 26 ± 2% at 10–6M; this effect was completely abolished by pertussis toxin treatment (12, 13). Such functional complementation suggests that the Δ318 SSTR5 and D2R receptors associate as hetero-oligomers to constitute a functional Gi protein–linked effector complex. Treatment of the cotransfectants with dopamine produced 30 ± 2% reduction in cAMP. However, the addition of sulpiride (10−4 M), a dopamine receptor antagonist, completely abolished the ability of SST to inhibit forskolin-stimulated cAMP by Δ318 SSTR5-D2R putative hetero-oligomers. Such differential ability of agonist- or antagonist-bound D2R to rescue Δ318 SSTR5 could be explained by different conformational states of the agonist- or antagonist-occupied receptor and suggests a critical role of receptor conformation in promoting receptor association.

Figure 1

Immunohistochemical colocalization of D2R and SSTR5 in rat brain cortex and striatum. (A andB) Serial sections of rat striatum showing many neurons expressing D2R (A). A subset of these neurons, with the morphological characteristics of medium aspiny neurons, show coexpression of SSTR5 (B). (C to H) Confocal images of striatal (C to E) and cortical (F to H) regions double-labeled for D2R and SSTR5. D2R is localized by Cy3 imaged in red fluorescence in (C) and (F). SSTR5-positive neurons are localized by FITC imaged in green and identified in the same section in (D) and (G). Coexpression of D2R with SSTR5 can be seen by the yellow-orange color in the merged images in (E) and (H). Scale bar, 25 μm.

To investigate the interaction between wild-type receptors, we stably cotransfected D2R and HA-SSTR5 (human SSTR5 tagged at the NH2-terminus with a nonapeptide of the hemagglutinin protein) in CHO-K1 cells to achieve levels of expression (maximum binding capacity B max = 107 ± 29 fmol of protein per milligram for D2R, 163 ± 22 fmol/mg protein for HA-SSTR5) comparable to the density of endogenous receptor expression (3, 13, 14). Competition analysis of125I-labeled Leu8-d-Trp22-Tyr25-SST-28 (LTT-SST-28) showed a 3000% increase in the binding affinity of SST-14 upon addition of the D2R agonist quinpirole (10–4 M) (fromK i 1.5 ± 0.2 nM toK i 0.05 ± 0.01 nM, whereK i is the inhibition constant) (Fig. 2A). In contrast, addition of the DR antagonist sulpiride caused an 80% reduction in the affinity of SST-14 for binding to the putative HA-SSTR5–D2R oligomers (fromK i 1.5 ± 0.2 nM toK i 7.5 ± 1.2 nM) (Fig. 2A). Because neither the dopamine agonist nor antagonist is capable of binding to HA-SSTR5 directly, these results suggest that the binding affinity of HA-SSTR5 for SST-14 is modulated by different conformational states of the agonist- or antagonist-occupied D2R through putative HA-SSTR5–D2R hetero-oligomers (Fig. 2A). Displacement analysis of125I-labeled spiperone binding by sulpiride showed a doubling of the affinity of sulpiride for the D2R in the presence of low SST concentrations (from K i 17.2 ± 2.6 nM to K i 8.2 ± 1.4 nM) (Fig. 2B). This suggests a synergistic role of SST on D2R affinity at low but not high concentrations because of a more suitable conformation for125I-labeled spiperone binding by putative HA-SSTR5–D2R hetero-oligomers.

Figure 2

Functional interaction of D2R with SSTR5. (A) Displacement analysis using the SSTR-specific125I-labeled LTT-SST-28 radioligand (12,13). Membranes coexpressing HA-SSTR5 and D2R were treated with increasing amounts of SST-14 alone (open circles), SST-14 and quinpirole (10–4 M) (open triangles), or SST-14 and sulpiride (10–4 M) (solid triangles). DR ligand concentrations were selected to reach saturation binding, as well as maximal signaling (quinpirole) or inhibition of signaling (sulpiride). (B) Displacement analysis using the D2R-specific125I-labeled spiperone radioligand. Membranes coexpressing HA-SSTR5 and D2R were treated with increasing amounts of sulpiride alone (open circles) or sulpiride and SST-14 (10–6 M) (solid squares). The SSTR agonist concentration was selected to ensure maximal saturation binding and receptor signaling. (C) G protein coupling of the expressed receptors was assessed by investigating the effect on 125I-labeled LTT-SST-28 binding of incubating membranes from cotransfectants with GTPγS (10–4 M) for 30 min at 37°C. Maximum specific binding obtained in the absence of GTPγS (open bar) was ∼15%. SST-14 (10–6 M) was used to define nonspecific binding. GTPγS treatment significantly reduced specific binding of125I-labeled LTT-SST-28 to HA-SSTR5, reflecting G protein coupling. This response was enhanced by addition of quinpirole (Quin, 10–4 M) and dopamine (DA, 10–4 M), and was not affected by eticlopride (Et, 10–4 M) but was reduced by sulpiride (SUL, 10–4 M). Ligand concentrations used were chosen to reach saturation binding. (D) Receptor coupling to adenylyl cyclase (12) was assessed as inhibition of forskolin-stimulated cAMP after treatment of the cotransfectants with SST-14 (open circles), quinpirole (solid triangles), or SST-14 and quinpirole (open squares). Data are means ± SEM of three independent experiments. *P < 0.01 compared with control GTPγS treatment (Dunnett's post hoc one-way analysis of variance).

G protein coupling of HA-SSTR5 and D2R was assessed by monitoring the effect of guanosine 5′-O-(3′-thiotriphosphate) (GTPγS) treatment on membrane 125I-labeled LTT-SST-28 binding (Fig. 2C). GTPγS treatment of the cotransfectants led to a 41 ± 3% decrease in specific radioligand binding. Both dopamine and the dopamine agonist quinpirole increased inhibition by GTPγS. The dopamine antagonist eticlopride displayed no effect on SST-14–induced G protein coupling of HA-SSTR5, whereas sulpiride significantly reduced inhibition by GTPγS. Treatment of the cotransfectants with SST-14 or quinpirole induced maximum 36 ± 3% and 39 ± 3% inhibition of forskolin-stimulated cAMP, respectively (Fig. 2D). Simultaneous application of both agonists potentiated the cAMP inhibitory response to a maximum of 52 ± 4%. Sulpiride reduced SST-14–induced cAMP inhibitory response (maximum inhibition 25 ± 2%) and acted as a partial antagonist of SST-14 signaling.

To obtain direct evidence for association of SSTR5 and D2R in intact cells, we investigated receptor hetero-oligomerization by photobleaching fluorescence resonance energy transfer (pbFRET) microscopy (15, 16) (Fig. 3). In the basal state, we found a low effective FRET efficiency of 2 ± 2%, reflecting an insignificant amount of preformed hetero-oligomers (Fig. 4). Treatment with either SST-14 (10–6 M) or dopamine (10–4 M) resulted in a strong increase in FRET efficiency to 18 ± 2% and 16 ± 2%, respectively, suggesting that oligomerization of SSTR5 and D2R is induced by either agonist. Simultaneous treatment with both agonists (10–6 M SST-14 and 10–4 M dopamine) resulted in a similar FRET efficiency of 20 ± 2%. Treatment with the D2R-specific antagonists sulpiride (10–4 M) and eticlopride (10–4 M) led to a significantly lower FRET efficiency of 7 ± 2% and 3 ± 3%, respectively.

Figure 3

Representative photobleaching experiment for treatment with SST-14 (10–6 M). (A) Photobleaching of donor in absence of acceptor (selection of images shown). Only the high-intensity membrane region was considered in analysis; low-intensity background and intracellular regions were masked (black). Two leftmost panels: Unmasked images showing absence of rhodamine and initial donor fluorescence. Right panel: Histogram of time constants obtained from single-exponential fits to pixel-based photobleaching decay curves. The average time constant of 15.8 s (solid bar) was taken as τD–A. (B) Photobleaching of donor in presence of acceptor. The presence of acceptor led to larger donor photobleaching time constants, with an average, τD+A, of 22.2 s, reflecting FRET between FITC and rhodamine.

Figure 4

Effective FRET efficiency in the basal state as well as after treatment with SST-14 (10–6 M), dopamine (10–4 M), SST-14 (10−6 M) plus dopamine (10−4 M), sulpiride (10–4 M), and eticlopride (10–4 M) on plasma membrane of CHO-K1 cells coexpressing HA-SSTR5 and wild-type D2R. Ligand concentrations were selected to reach saturation binding, as well as maximum signaling (agonists) or maximum inhibition of signaling (antagonists). The number of cells analyzed for each condition was ∼45.

Our results demonstrate that D2R and SSTR5 associate on the plasma membrane and that receptors from different GPCR families can interact as functional oligomers. Although SSTR5 forms homodimers upon ligand activation as revealed by Western blots, and the D2R exists as preformed homodimers, it remains to be determined whether the two receptors assemble in vivo as simple heterodimers or larger oligomers. The D2R-SSTR5 oligomer is pharmacologically distinct from its receptor homodimers, as it is characterized by a much greater affinity for binding both dopamine and SST agonists, and is associated with enhanced G protein and effector coupling to adenylyl cyclase. Hetero-oligomerization led to synergy such that the binding affinity for the second radioligand and signaling were increased as a result of receptor occupancy by the first ligand. Given the endogenous coexpression of D2R and SSTR5 in striatal and cortical neurons, hetero-oligomerization of the two receptors may be one explanation for the reported enhancement of dopaminergic and somatostatinergic transmission induced by in vivo administration of SST or dopamine agonists. While we have characterized the pharmacological properties of one DR-SSTR hetero-oligomer pair, there may be similar oligomeric interactions between other members of these two receptor families, which could explain the full range of functional biological interactions between the two transmitter systems.

There is increasing biochemical and functional evidence that a number of GPCRs exist as dimers or larger oligomers, and several models have been proposed to explain their regulation (17). The use of high-density recombinant receptor expression systems for detection of dimers by Western blots in several of these studies, however, could account for a high level of basal dimerization as an artifact of receptor overexpression. We have thus used lower density expression systems to mimic endogenous receptor expression levels, coupled with FRET analysis to monitor receptor oligomerization. The high sensitivity of FRET is ideally suited for studying molecular interaction at low expression level. pbFRET has been previously used to demonstrate oligomerization of the EGF receptor (18). Here we show that hetero-oligomerization of SSTR5 and D2R is also induced by ligand binding, that ligand binding to either receptor can trigger hetero-oligomer formation, and that there are no preformed hetero-oligomers in the absence of ligand. The results of FRET analysis complement the pharmacological data in showing that hetero-oligomerization induced by either SST or dopamine agonists goes hand in hand with activated receptor function. There was, however, no strict correlation between the level of oligomerization and the activity state of the receptor. Whereas pharmacological data demonstrated enhanced functional activity of the heteromeric receptor complex when occupied by both SST and dopamine ligands, no such synergy was found by FRET, which showed the same level of hetero-oligomerization for either agonist alone or both agonists applied together. Such dissociation is to be expected, given that FRET monitors only the presence of oligomers, not their activity state. Unlike SST and dopamine agonists, which both promoted D2R-SSTR5 hetero-oligomer formation, antagonists produced modest or no receptor oligomerization. Because no selective SSTR5 antagonists are currently available, we could not test the effect of antagonism at the SSTR5 receptor on D2R-SSTR5 hetero-oligomerization. Our results suggest a model in which agonist induces oligomerization, with the heteromeric receptor complex simultaneously occupied by two ligands being the most active signaling form. Antagonists may act by preventing hetero-oligomer formation or by promoting the formation of inactive hetero-oligomers. Hetero-oligomerization defines a new level of functional diversity in endogenous GPCR signaling. There are likely to be many such heteromeric GPCRs between members of different GPCR families that exist endogenously and that should constitute novel and hitherto unrecognized drug targets for combinations of agonists or antagonists, as well as for heterodimer-selective single-ligand molecules.

  • * To whom correspondence should be addressed. E-mail: yogesh.patel{at}


View Abstract

Navigate This Article