Report

Dual Signaling Regulated by Calcyon, a D1 Dopamine Receptor Interacting Protein

See allHide authors and affiliations

Science  03 Mar 2000:
Vol. 287, Issue 5458, pp. 1660-1664
DOI: 10.1126/science.287.5458.1660

Abstract

The synergistic response of cells to the stimulation of multiple receptors has been ascribed to receptor cross talk; however, the specific molecules that mediate the resultant signal amplification have not been defined. Here a 24-kilodalton single transmembrane protein, designated calcyon, we functionally characterize that interacts with the D1 dopamine receptor. Calcyon localizes to dendritic spines of D1 receptor–expressing pyramidal cells in prefrontal cortex. These studies delineate a mechanism of Gq- and Gs-coupled heterotrimeric GTP–binding protein–coupled receptor cross talk by which D1 receptors can shift effector coupling to stimulate robust intracellular calcium (Ca2+ i) release as a result of interaction with calcyon. The role of calcyon in potentiating Ca2+-dependent signaling should provide insight into the D1 receptor–modulated cognitive functions of prefrontal cortex.

Dopamine (DA), acting through D1 receptors, modulates synaptic transmission in neural circuits, which mediate learning and memory (1, 2). In heterologous expression systems, D1 DA receptors stimulate formation of adenosine 3′,5′-monophosphate (cAMP) by coupling to Gs heterotrimeric GTP-binding (G) proteins (3). However, in brain and kidney, D1 receptor agonists also produce increases in inositol 1,4,5-trisphosphate (IP3) turnover and intracellular calcium (Ca2+ i) (4). It is not yet known whether production of these second messengers involves linkage of D1 receptors to multiple G proteins or activation of alternative D1-like receptor subtypes.

We conducted a yeast two-hybrid screen (5) with the COOH-terminal 81 residues (residues 365 to 446) of the human D1 (hD1) receptor used as bait (6). One interacting clone, designated calcyon (for “calcium on”), contained a cDNA encoding a 217-residue protein (7). Calcyon is a putative type II membrane protein with a predicted single transmembrane segment extending from residues 88 to 103 (8). BLAST searches of GenBank databases indicated that calcyon displays a high degree of sequence similarity with two previously identified proteins of unknown function, P19 and P21 (9). The three proteins exhibit extensive sequence similarity (7). Calcyon appears to be the most highly diverged member of the P19/21/calcyon family of proteins.

Antibodies to a 20-residue segment of calcyon bound to a strong band of about 34 kD and a weaker band of about 28 kD on immunoblots of microsomal protein fractions purified from rhesus monkey brain and spleen (7, 10). The bands were present in samples prepared from prefrontal cortex and caudate putamen but not spleen. Preincubation of antibodies with immunizing peptide prevented detection of the bands. The predicted calcyon protein includes a potential N-linked glycosylation site at residue 73 (7). After digestion of prefrontal cortex microsomal proteins with N-glycosidase F, calcyon antibodies reacted with an ∼24-kD protein, suggesting that the 28- and 34-kD bands corresponded to calcyon protein modified by N-linked oligosaccharides (7). Chaotropic salts failed to solubilize the immunoreactive protein, consistent with the notion that calcyon is an integral membrane protein (10). Together, these results suggest an NH2-terminal extracellular, COOH-terminal intracellular transmembrane orientation of calcyon.

Calcyon antibodies labeled cell bodies and processes of neurons throughout the cerebral cortex and hippocampus (11) (Fig. 1, B and F). The labeling was characterized by high density in the vicinity of the plasma membrane and variably in neuronal processes. D1 receptors exhibit a similar distribution in these brain regions (Fig. 1, A and E). Calcyon appears to be preferentially expressed in pyramidal neurons, but expression in interneurons cannot be discounted until a more detailed anatomic investigation is conducted. Extensive labeling was also observed in many other subcortical structures (Fig. 1D). Omission of calcyon antibodies, or preincubation of antibodies with the immunizing peptide, prevented neuronal labeling. Immunogold electron microscopy of prefrontal cortex further revealed calcyon protein in small and medium-sized dendrites and dendritic spines receiving asymmetric inputs (Fig. 2, A and B). Similar to D1 receptors (12), calcyon was localized at the periphery of postsynaptic densities in dendritic spines (Fig. 2A).

Figure 1

Immunolocalization of calcyon and D1 DA receptors. Selected areas in the macaque brain labeled with subtype-specific antibodies to the D1 receptor (A,C, and E) and calcyon (B, D, and F). (A and B) Neurons in the CA3 region of the hippocampus; (C and D) caudate nucleus; (E and F) layer III neurons in area 46. Contrast and brightness of the images have been adjusted. Arrows in (C) point to labeled cell bodies.

Figure 2

Ultrastructural analysis of calcyon protein and colocalization of D1 DA receptors and calcyon in primate brain. (A) Immunogold localization of calcyon in two cortical dendritic spines (s), which receive asymmetric input (arrowheads) from unlabeled axons (a). Note extra- and perisynaptic membrane localization of immunogold particles (arrows). Immunopositive astrocytic process protruding between two spines is labeled with asterisks. Bars = 0.2 μm. (B) Calcyon protein localization in the small dendrites (d) of the cerebral cortex. Immunogold particles are associated with the plasma membranes (arrows) and cytoplasm (small arrow). Confocal (C) and epifluorescent (D) detection of D1 receptor and calcyon antibody double labeling of prefrontal cortex (C) and caudate nucleus (D).

Brain sections were double-labeled with calcyon and D1 receptor antibodies (11, 13), followed by a cocktail of species-specific Cy3-conjugated and fluorescein isothiocyanate (FITC)–conjugated immunoglobulin G (IgG). Overlays of the FITC and Cy3 fluorescent staining of prefrontal cortex indicated coexpression of D1 receptors and calcyon in the same population of pyramidal neurons (Fig. 2C). In the caudate, calcyon localized to a subpopulation of D1 receptor expressing medium spiny neurons (Fig. 2D), in which the D1 antibody labeled the neuropil so densely that cell body labeling was obscured (Fig. 1C).

Deletion of the COOH-terminal 11 and NH2-terminal 55 amino acids had no apparent effect on the ability of the D1 receptor bait to interact with calcyon (14), indicating that residues 421 to 435 of the D1 receptor comprise a core domain sufficient for interaction. We tested the ability of S-calcyon, a bacterial fusion protein containing the COOH-terminus of calcyon (residues 93 to 217), to associate with a GSTD1 (15), a glutathione S-transferase fusion protein containing the D1 receptor bait sequence or with full-length D1 receptors expressed in HEK293 cells (16). GSTD1 was bound by S-calcyon but not by the negative control.

Inclusion of a peptide containing D1 receptor residues 421 to 435 (pep421–435) in the pull-down reaction, but not an unrelated 17-residue peptide (pep2) (17), prevented detection of GSTD1. Full-length D1 receptor polypeptide migrates with a molecular mass of 48 to 50 kD on immunoblots of D1 HEK293 cell lysates. Bands of similar size were pulled down from D1 HEK293 cell lysates by S-calcyon, but not by S-β-gal. The rat D1 monoclonal antibody (13) coimmunoprecipitated calcyon from pCI-calcyon–transfected cells (18) (Fig. 4C).

We tested whether calcyon plays a role in D1 receptor signaling in D1 HEK293 cells after activation of endogenous G-protein–coupled receptors (GPCRs). The D1 agonist SKF81297 (10 μM), when applied after stimulating endogenous P2Y purinergic receptors (19) with adenosine triphosphate (ATP) (50 μM), triggered an immediate increase in Ca2+ i (Fig. 3A) (20) in cells transfected with either pEnhanced green fluorescent protein (EGFP)–calcyon or pCIcalcyon plasmids (18). SKF81297 produced a gradual but small increase in Ca2+ i in untransfected and vector-transfected cells (Fig. 3B). Without priming, the D1 agonist-stimulated responses were undetectable (21). Similar responses were observed in transfected cells bathed in Ca2+-free medium containing EGTA, indicating release of Ca2+ from intracellular stores. As D1 receptors and the Gq/ll G protein physically associate (22), priming may increase the affinity of the D1 receptor for Gq/11. In calcyon-transfected cells, D1 receptor–stimulated Ca2+ i release was comparable in magnitude to that elicited by the conventional Gq/ll- coupled P2Y receptor in the same cells (92.6% ± 8.3% of the ATP response; eight independent transfection experiments). However, D1-stimulated cAMP levels were unaltered by calcyon expression (Fig. 3C) (23). These findings could be explained by the ability of calcyon to further strengthen coupling of D1 receptors to Gq/11 or to regulate enzymes or substrates that influence Gq/11 signaling (24).

Figure 3

Calcyon potentiates D1 receptor-stimulated of Ca2+ i release in HEK293 cells. (A, B, D to F, andH) Ligand-induced Ca2+ i release in fura-2–loaded D1 HEK293 or HEK293 cells expressing EGFP-calcyon, EGFP, or both EGFP-calcyon and EYFP-D1421–435. Substances were applied for the times indicated by horizontal bars. Ca2+ i signals are reported as the mean of six to eight transfected cells. Similar results were obtained from four to eight additional independent transfection experiments. (C) cAMP accumulation in transfected D1 HEK293 cells treated with 50 μM ATP followed by 10 μM SKF81297 10 min later (t = 10 min). Reactions were stopped at t = 25 min. Results are reported as the average of three independent transfection experiments assayed in triplicate. (G) Grouped data from experiments similar to that shown in (F). The magnitude of the SKF81297 stimulated response in D1 HEK293 cells transfected with pEGFP-calcyon (0.02 pmol) and either pEYFP or pEYFP-D1421–435 in a 1:1 or 1:3 molar ratio expressed relative to the size of the response elicited by ATP in the same cells. Bars show mean ± SE of (+EYFP) 1:1, n = 6; 1:3,n = 4 or (+EYFP-D1421–435) 1:1,n = 3; 1:3, n = 3 independent transfection experiments. *P < 0.01; independent t test.

Functional interaction of Calcyon and D1 receptors was required for maximal SKF81297-stimulated Ca2+ i release as the response was blocked by the D1 receptor antagonist SCH23390 (50 μM) (Fig. 3D) and undetectable in HEK293 cells expressing calcyon only (Fig. 3E). D1 HEK293 cells were cotransfected with expression plasmids encoding EGFP-calcyon, and either enhanced yellow fluorescent protein (EYFP) or EYFP fused to the core calcyon-interaction domain (EYFP-D1421–435) (25). The SKF81297-evoked response of cells cotransfected with 1:1 and 3:1 molar ratios of pEYFP-D1421–435 to pEGFP-calcyon was reduced by about 60% and 75%, respectively (P < 0.01, independent t test) compared with cells cotransfected with similar ratios of pEYFP and pEGFP-calcyon (Fig. 3, F and G). Presumably, expression of D1421–435 decreased the SKF81297-stimulated response in a DNA concentration-dependent manner by competitively inhibiting formation of the calcyon-D1 receptor protein complex.

Priming or cross talk with Gq/ll-coupled receptors apparently is necessary to activate the calcyon-D1 receptor complex as D1 agonists triggered a large increase in Ca2+ iin calcyon expressing cells if applied after stimulating endogenous M1 muscarinic receptors (19) (Fig. 3H), but not after stimulation of endogenous β-adrenergic receptors (19). Pretreatment with two different protein kinase C (PKC) inhibitors significantly attenuated D1-stimulated Ca2+ irelease in calcyon-transfected cells (Fig. 4, A and B). Whereas the calcyon-D1 receptor complex forms in the absence of agonist (+HBS), interaction between calcyon and the receptor appears to be strengthened in the presence of agonists and reduced, but not abolished, by treatment with the PKC inhibitor bisindolylmaleimide (Fig. 4C) (18). The cytoplasmic domain of calcyon, which contains consensus PKC phosphorylation sites at Ser154 and Ser196, can be phosphorylated by purified PKC isoforms (Fig. 4D) (26). Preliminary data indicate that this region of calcyon can bind the acidic phospholipid phosphoinositol-4,5-bisphosphate (PIP2) (27).

Figure 4

PKC dependence of calcyon function. (A) Ligand-induced Ca2+ i release in fura-2–loaded D1 HEK293 cells expressing EGFP-calcyon. Substances were applied for the times indicated by horizontal bar. Ca2+ i signals are reported as the mean of six to eight cells. Cells were treated with the PKC inhibitor (2 μM bisindolylmaleimide) (+BisI) or 10 μM myristoylated PKC (19–27) (+myrPKC) for 10 or 60 min, respectively, before application of ATP. (B) Grouped data from five to eight experiments similar to that shown in (A). The magnitude of the SKF81297 response is expressed relative to that elicited by ATP in the same cells. The response elicited by SKF81297 in the absence of PKC inhibitors is shown for comparison. Bars show mean ± SE. *P < 0.01, independent t test; untreated, n = 8; +BisI, n = 5; +myrPKC, n = 5. (C) Calcyon antibody reaction with immunoblots containing solubilized proteins (20 μg) from pCI or pCI-calcyon–transfected D1 HEK293 cells (lysate) and proteins immunoprecipitated by the rat D1 receptor monoclonal antibody (+D1 mab) (17). Transfected cells were untreated (+HBS), treated with bisindolylmaleimide (2 μM) (+BisI), or treated with 50 μM ATP followed by 10 μM SKF81297 5 min later (t = 5 min) (+ATP/SKF) and before solubilization at t = 10 min. Calcyon was detected with rabbit calcyon antibodies and HRP-conjugated antirabbit (Fc fragment specific). Position of rat monoclonal antibody heavy chain is indicated by IgG, and positions of the molecular mass markers in kilodaltons are indicated on the left. (D) Autoradiographic detection of S-calcyon (0, 2, and 4 μg) after incubation with purified PKC and [γ-32P]ATP. Proteins were separated on an SDS–12% polyacrylamide gel.

Calcyon appears to represent a prototype for a GPCR cross talk–specific accessory protein. Calcyon also may interact with other DA receptor subtypes and/or other GPCRs. Indeed, preliminary functional data suggest that calcyon can interact with the D5 DA receptor, which contains a region similar in sequence to the D1 receptor's core calcyon binding domain. The D1 receptor is of particular interest because it is the most prominent DA receptor in the cerebral cortex (28), and it modulates excitatory transmission in prefrontal neurons during working memory performance (1). The mechanism by which calcyon alters D1 receptor signaling after Gq/11-coupled receptor priming in HEK293 cells provides a molecular framework for understanding D1 receptor–mediated neuromodulation. Other neurotransmitters may prime the D1 receptor–stimulated Ca2+ i release, as D1 receptors and calcyon localize to pyramidal cell dendritic spines, which are the site of excitatory amino acid input. Several electrophysiological as well as molecular models of synaptic plasticity require both D1 receptor activation andN-methyl-d-aspartate receptor-mediated glutamate transmission (29). Furthermore, M1 muscarinic receptors also localize to spines of pyramidal neurons (30). The cross talk between muscarinic/dopaminergic signaling shown here thus may be relevant in vivo (31).

  • * To whom correspondence should be addressed. E-mail: cbergson{at}mail.mcg.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article