Molecular Biology of the Dopamine Receptor Subtypes

Olivier Civelli

Until recently, our understanding of the dopaminergic system has been based on the interactions of one neurotransmitter, dopamine, with two receptors, the D1 and D2 receptors. In the last few years, the application of molecular biological techniques has led to the identification of three new dopamine receptors (for review see ref. 11); see also Dopamine Receptor Expression in the Central Nervous System, Dopamine Receptors: Clinical Correlates, and Signal Transduction Pathways for Catecholamine Receptors). The discovery of these "unexpected" dopamine receptor subtypes has had a revolutionary impact on the study of the dopaminergic system and their implications in human disorders.

Dopamine is present in most parts of the central nervous system (CNS) but in particular in the nigrostriatal pathway comprising the neurons of the substantia nigra (A9) and projecting to neurons of the neostriatum and the mesocorticolimbic pathway composed of neurons of the ventral tegmental area (A10) connecting with those of the limbic cortex and other limbic structures (5).

The involvement of the dopaminergic nigrostriatal pathway in extrapyramidal dysfunctions was shown by the discovery that degeneration of this pathway occurs in the brains of patients afflicted with Parkinson's disease (18, 49). The depletion of dopamine resulting from the degeneration of the nigrostriatal pathway led to the development of dopamine-replacement therapies which are successful in alleviating Parkinson's disease (4, 29). The hypothesis that dopamine is involved in the pathogenesis of psychosis, in particular schizophrenia, rests on the finding that most antipsychotic drugs are dopamine receptor antagonists and that agents which cause excessive release of dopamine mimic schizophrenia-like states (7, 8, 12, 50). The mesocorticolimbic pathway has been implicated as the principal dopaminergic pathway involved in the etiology of psychoses. These data explain the dilemma associated with dopamine-related drug therapies: The blockade of the dopaminergic system, desired for reducing psychoses, induces extrapyramidal dysfunctions and vice versa.

In 1979, Kebabian and Calne found that dopamine exerts its effects by binding to two receptors, known as the D1 and D2 receptors (30). These receptors could be differentiated pharmacologically, biologically, physiologically, and by their anatomical distribution (for review see ref. 13). Pharmacologically, the hallmark of the D1 receptor is to bind the benzazepine antagonist SCH 23390, while that of the D2 receptor is to recognize with high affinity the butyrophenones: spiperone and haloperidol. These two receptors exert their biological actions by coupling to and activating different G protein complexes. The D1 receptor interacts with the Gs complex to activate adenylyl cyclase, whereas the D2 interacts with Gi to inhibit cAMP production. The anatomical distributions of these two receptors overlap in the CNS, yet their quantitative ratios differ significantly in particular anatomical areas. With respect to mental disorders, it is noteworthy that both D1 and D2 receptors are present in the nigrostriatal and mesocorticolimbic pathways.

For 10 years, this two-subtype classification has accounted for most of the activities attributed to the dopaminergic system. The existence of other dopamine receptors has been proposed but had been refuted when the "new" receptors were recognized to represent different affinity states of the canonical D1 and D2 receptor (2, 31). However, this classification was dramatically changed with the application of recombinant DNA technology to the molecular characterization of the dopamine receptors.

The cloning of the D2 dopamine receptor resulted from the recognition that, on the basis of its inhibitory activity on adenylyl cyclase, it would belong to the supergene family of the G-protein-coupled receptors (17, 27; see also Signal Transduction Pathways for Catecholamine Receptors). Consequently, the use of a cloning strategy based on the sequence homology known to exist among G-protein-coupled receptors could lead to the molecular characterization of the D2 receptor. The D2 dopamine receptor was cloned using the hamster b2-adrenergic receptor coding sequence as hybridization probe under conditions which would detect sequentially related DNA fragments (6). Via genomic and cDNA screenings, a rat brain cDNA was identified and shown to encode a protein featuring the characteristics expected for a G-protein-coupled receptor. The receptor encoded by this cDNA had the pharmacological profile and biological activity of the dopamine D2 receptor found in the brain and pituitary, demonstrating that this cloned receptor is the same D2 receptor as the one described in 1979 (1, 6, 45).

The success of the homology approach in the cloning of the D2 receptor opened the door for the cloning of other dopamine receptors. Successful cloning of the D1 receptor was reported by several groups (15, 41, 56, 63). The sequences derived from these clones share the characteristics expected of G-protein-coupled receptors in general and of the catecholamine receptors in particular (63). These putative receptors were expressed by DNA transfection and were shown to bind D1 receptor ligands and to stimulate adenylyl cyclase activity, the two hallmarks of the D1 receptor. Molecular characterization of the D1 receptor had been achieved.

The generality of the homology approach allowed for the search of other unexpected dopamine receptors. Using a D2-receptor-specific DNA fragment as probe under low-stringency hybridization conditions, Sokoloff et al. (54) identified another dopamine receptor, the D3 receptor. When expressed in eukaryotic cells, this receptor was shown to bind D2 but not D1 ligands. Its structure and binding characteristics thus permitted its classification as a new dopamine receptor called the D3 receptor. Noteworthy is its ability to affect second messenger systems, which has thus far not been demonstrated.

Furthermore, by analyzing the mRNAs of human neuroepithelioma SK-N-MC cells with D2 receptor cDNA probes under conditions of low stringency, another D2-related mRNA was detected (58). The corresponding cDNA and gene analyses led to the characterization of the D4 receptor. The D4 receptor, when expressed in COS-7 cells, binds D2 antagonists with a pharmacological profile that is distinct but reminiscent of that of the D2 receptor. The D4 receptor was shown to couple to G proteins, although its potential at inducing second messenger systems is still being determined.

Finally, the D1 receptor clone was used as a hybridization probe to identify D1-related genes. A human D5 and a rat D1b receptor have been characterized (26, 55, 57)). They display the same pharmacological profile, reminiscent of that of the D1 receptor, and are able to stimulate adenylyl cyclase activity. On the basis of their sequences, the D5 and D1b receptors are human and rat equivalents of the same receptor, respectively.

Thus the application of homology screening techniques not only led to the deciphering of the molecular structures of the D1 and D2 receptors, but also led to the characterization of three new dopamine receptors: D3, D4, and D5. These discoveries have, of course, medical implications. For example, most of what is known about dopamine agonists' and antagonists' actions has to be reevaluated in view of the existence of the different dopamine receptors. Our renewed knowledge of the dopaminergic system begins with the study of the dopaminergic receptor family.

In their putative transmembrane domains, the D1 and D5 receptors are 79% identical but are only 40–45% identical to the D2, D3, and D4 receptors. Conversely, the D2, D3, and D4 receptors are between 75% and 51% identical to each other, the first indication that the five receptors can be divided into the D1-like and D2-like receptor subfamilies. The topologies of the five dopamine receptors are predicted to be the same as all the other G-protein-coupled receptors. They should contain seven putative membrane-spanning helices which would form a narrow dihedral hydrophobic cleft surrounded by three extracellular and three intracellular loops. The receptor polypeptides are probably further anchored to the membranes through palmitoylation of a conserved Cys residue found in their C-tails (347 in D1, the C-terminus in D2-like receptors) (46). The dopamine receptors are probably glycosylated in their N-terminal domains; in addition, the D1-like subtypes have potential glycosylation sites in their first extracytoplasmic loop.

The genomic organization of the dopamine receptors also supports the notion that they derive from the divergence of two gene subfamilies, the D1-like and D2-like receptor genes. The D1 and D5 receptor genes do not contain introns in their protein coding regions, whereas the D2, D3, and D4 genes do. Furthermore, most of the introns in the D2-like receptor genes are located in similar positions (25, 54, 56, 58, 63).

The cloned dopamine receptors, when expressed by transfection, exhibit binding profiles which can also differentiate them into the D1-like and D2-like subfamilies. The D1-like receptors bind with high-affinity D1 and not D2 antagonists. A prototypic ligand for the D1-like receptors is the benzazepine SCH23390 (Kis < 1 nM); on the other hand, they bind the butyrophenone spiperone with low affinity (Kis in the micromolar range). In contrast, the D2-like receptors efficiently bind spiperone (Kis < 1 nM) and not SCH23390 (Kis for D2 in the micromolar range); they also recognize most of the neuroleptics. Because there are 21 amino acid residues which differentiate D1-like from D2-like receptors in the transmembrane domains, these might participate in the selective recognition process. While there presently exists no ligand to differentiate the D1 from the D5 receptor, several D2 antagonists can distinguish the different D2-like receptors. The compound 7-OH-DPAT is selective for the D3 receptor (33), whereas clozapine has the highest affinity for the D4 receptor. It is noteworthy that dopamine binds to the D3, D4, and D5 receptors with nanomolar or submicromolar affinity constants, while its corresponding constants for the D1 and D2 receptors are in the micromolar ranges.

The predominant biological activities associated with D1 and D2 receptor stimulation are the activation and inhibition of adenylyl cyclase activity, respectively. Stimulation of the D1 and D5 receptors in transfected cells has been shown to result in activation of adenylyl cyclase, indicating similar pathways of second messenger induction for the D1-like receptors. On the other hand, the D3 and D4 receptors have, thus far, not been shown to induce second messenger systems, thus preventing their subfamily classification based on biological activity. However, because receptors' interactions with G proteins involve the cytoplasmic loops (16, 32) and because D2-like receptors have a large third cytoplasmic loop and a short C-terminal tail representative of the catecholamine receptors coupled to Gi proteins, the D2-like relative homology suggests that they might couple to the same set of G proteins.

Thus, on the basis of their primary sequences, of their genomic organization, and of their pharmacological and, at least partly, biological activities, the different dopamine receptors can be classified into the D1-like and D2-like subtypes. This, and the fact that the D3, D4, and D5 receptors are present in significantly lower amounts than are the D1 and D2 receptors, suggest that the existence of the former ones could not be found by pharmacological analyses.

As mentioned above, no selective ligand has been described which is able to differentiate the D1 from the D5 receptor. On the other hand, the pharmacological profiles of the D3 or D4 receptors show distinct striking differences when compared to that of the D2 receptor.

Most neuroleptics were developed as D2 receptor antagonists and thus are expected to bind to the D2 receptor with higher affinity than to the D3 and D4 receptors. This is true for the majority of the neuroleptics, which implies that those neuroleptics are acting predominantly at D2 receptors in the human brain. However, a few neuroleptics have been found to show selectivity for the D3 or D4 receptors; through these, some aspects of the functions of the D3 and D4 receptors may be revealed.

Two antagonists, UH232 and AJ76, bind to the D3 receptor with a higher affinity than they do to the D2 receptor (54). These compounds are classified as selective for presynaptic receptors or for autoreceptors. In addition, it was found that dopamine binds the D3 receptor with a 20-fold higher affinity than the D2 receptor, a characteristic expected for autoreceptors. Furthermore, the presence of D3 receptor mRNA in the substantia nigra, a center of dopamine production, supports the hypothesis that the D3 receptor may be a presynaptic receptor. Noteworthy is that the D2 receptor mRNA is the predominant dopamine receptor mRNA in the substantia nigra (38) and that, as for the D3 receptor, 6-OHDA lesions show its presence in the dopamine-secreting neurons (9, 21, 34, 35, 54). Therefore both the D2 and the D3 receptors are autoreceptors. Interestingly, the recent involvement of the D3 receptor in modulating cocaine self-administration has also been associated with its autoreceptor properties.

Clozapine, an "atypical" neuroleptic (i.e., a neuroleptic whose actions are not accompanied by adverse motor control side effects), shows a higher selectivity for the D4 receptor than for any other D2-like receptors. In schizophrenia therapy, clozapine is administered at a concentration 10-fold lower than its affinity constant for the D2 receptor, indicating that clozapine may not be primarily acting at the D2 receptor. The D4 receptor binds clozapine with a 10-fold higher affinity than does the D2 receptor (58). Therefore the D4 receptor may be the specific target of clozapine. A corollary of this is that antagonism of dopamine binding to the D4 receptor could be an important step in prevention of psychoses, a hypothesis reinforced by the low abundance of D4 mRNA in the striatum (58). Thus the lack of extrapyramidal side effects observed with clozapine treatment may be a reflection of D4 receptor localization in the CNS. These observations point to the D4 receptor as an important molecule in the etiology of psychoses (see also New Developments in Dopamine and Schizophrenia and Atypical Antipsychotic Drugs).

Because there are no current antibodies against all the different dopamine receptors, our knowledge of their tissue distribution comes primarily from in situ hybridization experiments. In the CNS, the five dopamine receptors exhibit overlapping but also distinct localizations. In the periphery, the different receptors are mostly expressed in a tissue-specific fashion.

The tissue distribution of the D1 and D2 mRNAs in the CNS supports their participation in the different aspects of dopaminergic neurotransmission which have been described on the basis of ligand binding and receptor autoradiography experiments. The D1 and D2 receptor mRNAs are present in all dopaminoceptive regions of the rat brain (20, 34, 36, 38, 39, 44, 62). High levels of D1 and D2 mRNAs are present in the caudate-putamen, nucleus accumbens, and olfactory tubercule, and lower levels are present in the septum, hypothalamus, and cortex. Regions where D2 but no D1 mRNAs were detected are the substantia nigra and ventral tegmental area, where the D2 mRNA is expressed at a high level, and the hippocampus. Conversely, the amygdala contains D1 mRNA but little, if any, D2 mRNA.

The D3, D4, and D5 receptor mRNAs are mostly present in tissues where the D1 and/or the D2 mRNAs are also expressed. However, their relative abundances are one to two orders of magnitude lower than that of the D1 or D2 mRNAs (54, 58). It has been shown that, relative to the D1 or D2 receptors, the D3 and D4 receptors are more selectively associated with the "limbic" brain, a region which receives its dopamine input from the ventral tegmental area and is known to be associated with cognitive, emotional, and endocrine functions. The location of the D5 receptor mRNA, on the other hand, is highly specific. The D5 mRNA is found only in the hippocampus, the hypothalamus, and the parafascicular nucleus of the thalamus and thus might be involved in affective, neuroendocrine, or pain-related aspects of dopaminergic functions (37). Finally, using in situ hybridization experiments, it has also been possible to demonstrate that D1 and D2 mRNA are colocalized in 26–40% of all caudate-putamen cells and in about 50% of all dopamine receptor mRNA-positive cells (38).

Dopamine receptor reactivities have also been described in several peripheral organs. mRNA detection by Northern blot analyses have shown that neither D1 nor D3 receptor mRNA are detectable outside the CNS (54, 63). On the other hand, the D2 receptor mRNA is expressed at high levels in the pituitary (6) and in the adrenal gland and also in the retina. Of particular interest are the kidney and the heart in which both D1- and D2-like activities have been described (2, 19). The D5 receptor mRNA is expressed, albeit at low levels, in the kidney (J. H. Meador-Woodruff and D. K. Grandy, unpublished data, 1992). Whether it is the expected D1-like receptor has yet to be demonstrated. None of the cloned D2-like receptor mRNAs is present in the kidney. On the other hand, the D4 mRNA is expressed in the heart (47) and might account for the expected D2-like reactivity reported for this tissue. None of the D1-like receptor mRNAs exists in significant amount in the heart. These data open the possibility that the D4 and D5 receptors carry the dopamine receptor reactivities detected in the kidney and the heart.

In conclusion, one can foresee that an advantage for the organism of having heterogeneous population of receptors is that it permits tissue-specific expression. mRNA detection experiments show that the different dopamine receptors exhibit specificity in their tissue distribution in the periphery, while in the CNS they often share tissue locations and, possibly, individual neurons as in the case of the D1 and D2 receptors. Although selectivity in cellular distributions has also been found in the CNS, it does not seem to be the rule for the different receptor subtypes. Another factor to consider in our understanding of the importance of the receptor diversity is the comparison of the relative abundance of the subtypes. Variable levels of distinct receptors, added to the fact that interactions between different dopamine receptor subtypes exist (3, 51, 60, 61), generate a high degree of diversity in responses that reflect the broad spectrum of the physiological activities known to be regulated by dopamine.

Although the human genome contains five dopamine receptor genes, the number of dopamine receptor mRNA species that it encodes is higher. This results from the fact that polymorphism and alternative splicing events play a role in dopamine receptor gene expression and leads to the existence of more than five different receptor binding sites.

First was the discovery that there exist two forms of D2 dopamine receptors (10, 14, 23, 25, 40, 42, 48, 52). These two forms differ in 29 amino acid residues located in the putative third cytoplasmic loop of the receptor. They are generated by an alternative splicing event which occurs during the maturation of the D2 receptor pre-mRNA (14, 25, 48). The two D2 receptor forms are neither species- nor tissue-specific; they coexist in all tissues analyzed but at a highly variable ratio. Because of its location in the third cytoplasmic loop, the 29-residue addition was expected to affect G protein coupling and consequently second messenger systems. It has been shown that both forms can inhibit cAMP accumulation (14) and that their efficiencies are somewhat variable (28,43). Alternative splicing events have also been shown to occur during the maturation of the D3 dopamine receptor pre-mRNA (22 , 53).

The existence of different variants of the human D4 receptor has also been demonstrated, although their generation is not by alternative splicing. These variants differ in the number of 48 base-pair repeats contained in their putative third cytoplasmic loop (59) and they have been detected in the genomes of different individuals, showing that a genetic polymorphism is responsible for the generation of the D4 receptor variants. These repeats are not present in the rat gene, making the polymorphism specific to humans. When expressed by DNA transfection, the variants containing 2, 4, and 7 repeats bind clozapine with equal affinities in the presence of sodium chloride. In the absence of sodium ions, however, the variants containing 2 and 4 repeats had a six- to eightfold lower dissociation constant for clozapine, while the affinity of the variant containing seven repeats was practically unaffected (59). Although it is not understood what effects the sodium ions have on receptors, these data indicate that the variants can behave differently with respect to the mechanism of ligand recognition.

Finally, the D5 receptor gene is peculiar among the G-protein-coupled receptors because it is associated with two pseudogenes in the human genome (26). The three D5-related genes are found on different chromosomes (24). Only one gene (DRD5, chromosome 4 q15.1-q15.3) codes for the active receptor; the two others contain an 8-base-pair insertion which leads to a frame shift and are genuine pseudogenes. Interestingly, these pseudogenes appear to be specific to humans, suggesting that the evolution of the D5 pseudogenes is a very recent event which may be restricted to primates.

The discovery of the "unexpected" dopamine receptors has and will continue to impact our understanding of the dopaminergic system. Of immediate interest is whether agonists or antagonists to the new dopamine receptors can be of therapeutic value like the D2 receptor antagonists are. The D3 and D4 receptors have two similar particularities. They bind most of the neuroleptics with less affinity than the D2 receptor, which indicates that, as commonly carried out, neuroleptic treatments may have not affected their activities. Furthermore, the D3 and D4 receptors are found predominantly in the limbic system and are relatively absent in the nigrostriatal system, and thus are associated preferentially with the etiology of psychoses instead of locomotion dysfunctions. The D4 receptor carries the further characteristic of binding clozapine with an affinity corresponding to its therapeutic concentration. Although clozapine can also bind to other receptors, its affinity to the D4 receptor might begin to explain its activity in the dopaminergic system. Consequently, the atypical effect of clozapine may be derived from the relative absence of D4 receptor in the basal ganglia. Whether this hypothesis proves valid will require the synthesis of specific antagonists. Finally, the D5 receptor may also be of therapeutic interest. It is present in very low amounts and is restricted only to a few tissues in the CNS. Its interest may stem from its presence in the kidney, whose function is improved by dopamine in cases of shock and low cardiac input. Thus a D5 agonist with low affinity for other catecholamine receptors could be valuable.

In conclusion, the discovery of the new dopamine receptors is too new to be conclusively evaluated with regard to potential therapies. Yet the few data that have already been obtained show promising characteristics and will hopefully lead to the development of tools which, in turn, will help further our understanding of the dopaminergic system and of its physiological implications.

published 2000