PHYSIOLOGICAL REVIEWS   Vol. 78 No. 1 January 1998, pp. 189-225
Copyright ©1998 by the American Physiological Society

Dopamine Receptors: From Structure to Function

CRISTINA MISSALE, S. RUSSEL NASH, SUSAN W. ROBINSON, MOHAMED JABER, AND MARC G. CARON

Departments of Cell Biology and Medicine, Howard Hughes Medical Institute Laboratories, Duke University Medical Center, Durham, North Carolina

I. INTRODUCTION
II. CLASSIFICATION OF DOPAMINE RECEPTORS
III. GENE STRUCTURE
IV. STRUCTURE OF DOPAMINE RECEPTORS
V. RECEPTOR VARIANTS
    A. D₂ Receptor
    B. D₃ Receptor
    C. D₄ Receptor
VI. PHARMACOLOGICAL PROPERTIES OF DOPAMINE RECEPTORS
VII. SIGNAL TRANSDUCTION OF DOPAMINE RECEPTORS
    A. Adenylyl Cyclase
    B. Calcium Channels
    C. Potassium Channels
    D. Arachidonic Acid
    E. Na⁺/H⁺ Exchange
    F. Na⁺-K⁺-ATPase
    G. Additional Signal Transduction Pathways Involved in Mitogenesis
VIII. DOPAMINE RECEPTORS IN THE BRAIN
    A. Distribution of Dopamine Receptors
    B. Function of Brain Dopamine Receptors
    C. Dopamine Receptors and Regulation of Gene Expression
    D. Development of Transgenic Animals in the Study of Dopamine Receptor Physiology
    E. Clinical and Pharmacological Implications of Multiple Dopamine Receptors
IX. DOPAMINE RECEPTORS IN THE PITUITARY
X. PERIPHERAL DOPAMINE RECEPTORS
    A. Dopamine Receptors in Blood Vessels
    B. Dopamine Receptors Controlling the Renin-Angiotensin-Aldosterone System
    C. Dopamine Receptors Controlling Catecholamine Release
    D. Dopamine Receptors in the Kidney
XI. DOPAMINERGIC SIGNAL TRANSDUCTION AND HYPERTENSION
    A. Human Hypertension
    B. Animal Models of Hypertension
XII. CONCLUDING REMARKS
REFERENCES

	ABSTRACT

Top References

Missale, Cristina, S. Russel Nash, Susan W. Robinson, Mohamed Jaber, and Marc G. Caron. Dopamine Receptors: From Structure to Function. Physiol. Rev. 78: 189-225, 1998.  The diverse physiological actions of dopamine are mediated by at least five distinct G protein-coupled receptor subtypes. Two D₁-like receptor subtypes (D₁ and D₅) couple to the G protein G_s and activate adenylyl cyclase. The other receptor subtypes belong to the D₂-like subfamily (D₂ , D₃ , and D₄) and are prototypic of G protein-coupled receptors that inhibit adenylyl cyclase and activate K⁺ channels. The genes for the D₁ and D₅ receptors are intronless, but pseudogenes of the D₅ exist. The D₂ and D₃ receptors vary in certain tissues and species as a result of alternative splicing, and the human D₄ receptor gene exhibits extensive polymorphic variation. In the central nervous system, dopamine receptors are widely expressed because they are involved in the control of locomotion, cognition, emotion, and affect as well as neuroendocrine secretion. In the periphery, dopamine receptors are present more prominently in kidney, vasculature, and pituitary, where they affect mainly sodium homeostasis, vascular tone, and hormone secretion. Numerous genetic linkage analysis studies have failed so far to reveal unequivocal evidence for the involvement of one of these receptors in the etiology of various central nervous system disorders. However, targeted deletion of several of these dopamine receptor genes in mice should provide valuable information about their physiological functions.

	I. INTRODUCTION

Top Next References

Dopamine (DA) is the predominant catecholamine neurotransmitter in the mammalian brain, where it controls a variety of functions including locomotor activity, cognition, emotion, positive reinforcement, food intake, and endocrine regulation. This catecholamine also plays multiple roles in the periphery as a modulator of cardiovascular function, catecholamine release, hormone secretion, vascular tone, renal function, and gastrointestinal motility.

	II. CLASSIFICATION OF DOPAMINE RECEPTORS

Top Previous Next References

The first evidence for the existence of DA receptors in the CNS came in 1972 from biochemical studies showing that DA was able to stimulate adenylyl cyclase (AC) (reviewed in Ref. 226). In 1978, DA receptors were first proposed, on the basis of pharmacological and biochemical evidence, to exist as two discrete populations, one positively coupled to AC and the other one independent of the adenosine 3',5'-cyclic monophosphate (cAMP)-generating system (424). It was shown, in fact, that in the pituitary DA inhibited prolactin secretion but did not stimulate cAMP formation (59; reviewed in Ref. 226) and that although the antipsychotic drug sulpiride was a DA antagonist when tested in the anterior pituitary, it was not able to block the striatal DA-sensitive AC (reviewed in Refs. 226, 424). In 1979, Kebabian and Calne (226) summarized these observations and suggested to call D₁ the receptor that stimulated AC and D₂ the one that was not coupled to this effector.

Subsequent studies confirmed this classification scheme, and D₁ and D₂ receptors were clearly differentiated pharmacologically, biochemically, physiologically, and by their anatomic distribution.

Concurrently in the late 1970s, by means of functional tests such as renal blood flow and cardiac acceleration measurements in the dog, the existence of specific peripheral receptors for DA was demonstrated. These receptors were named DA₁ and DA₂ on the basis of some pharmacological properties distinguishing them from their central counterparts (reviewed in Ref. 166). This led to a long-standing controversy concerning the identity or nonidentity of peripheral versus central receptors. However, subsequent biochemical and molecular biology studies in peripheral tissues pointed to extensive similarities between central and peripheral DA receptors so that the DA₁/DA₂ classification has been dropped (reviewed in Refs. 7, 307, 326, 336).

For a decade, the dual receptor concept served as the foundation for the study of DA receptors. However, after the introduction of gene cloning procedures, three novel DA receptors subtypes have been characterized over the past five years. These have been called D₃ (420), D₄ (450), and D₅/D_1b (431, 441).

Detailed structural, pharmacological, and biochemical studies pointed out that all DA receptor subtypes fall into one of the two initially recognized receptor categories. The D₁ and D₅/D_1b receptors share, in fact, a very high homology in their transmembrane domains. Similarly, the transmembrane sequences are highly conserved among D₂ , D₃ , and D₄ receptors (reviewed in Refs. 78, 159, 217, 401, 421). Pharmacologically, although the profiles of D₁ and D₂ receptors are substantially different, the D₅/D_1b receptor exhibits the classical ligand-binding characteristics of D₁ receptors, and the D₃ and D₄ receptors bind the hallmark D₂-selective ligands with relatively high affinity (reviewed in Refs. 78, 159, 217, 401, 421). In addition, the initial distinction between D₁ and D₂ receptors in terms of signaling events, that is, positive and negative coupling to AC, appears to apply, in broad terms, also to the novel members of the DA receptor family, the D₅/D_1b receptor being coupled to stimulation of AC (95, 169, 431, 441) and the D₃ (75, 287, 360, 377) and D₄ receptors (74, 80, 287, 290, 438) to inhibition of cAMP formation.

The D₁/D₂ classification concept developed in the late 1970s thus is still valid, and D₁ and D₅/D_1b receptors are classified as D₁-like and D₂ , D₃ , and D₄ receptor subtypes as D₂-like. The mammalian D_1b receptor, originally named on the basis of its high homology with the D₁ receptor, is now commonly referred to as the D₅ receptor.

	III. GENE STRUCTURE

Top Previous Next References

The D₂ receptor cDNA was first isolated in 1988 (47) and subsequently, in 1989, the existence of splice variants of this receptor was demonstrated (91, 162, 315). The D₃ receptor was identified by screening a rat cDNA library with the D₂ receptor sequence followed by PCR extension and genomic library screening (420). The D₄ receptor was cloned by screening a library from the human neuroblastoma cell line SK-N-MC (450).

The D₁ receptor was cloned by using either low-stringency screening of libraries or PCR based on the sequence of the D₂ receptor (95, 314, 480). The second member of the D₁-like receptor family was isolated using the sequence of the D₁ receptor and was referred to as D₅ (431), D_1b (441) and D₁ (464). It is now well established that the D₅ and D_1b are the human and rat equivalents of the same receptor.

The genomic organization of the DA receptors supports the concept that they derive from the divergence of two gene families that mainly differ in the absence or the presence of introns in their coding sequences. As summarized in Table 1, the D₁ and D₅ receptor genes do not contain introns in their coding regions (reviewed in Refs. 78, 159, 337), a characteristic shared with most G protein-coupled receptors (112). In contrast, and by analogy with the gene for rhodopsin (327), the genes encoding the D₂-like receptors are interrupted by introns (reviewed in Refs. 78, 159, 337). It appears likely that many of the genes in the G protein-coupled receptor family have arisen from a single primordial gene, suspected to be one of the opsin genes, that lost its introns by a gene-processing event (reviewed in Ref. 337). Two main evolutionary mechanisms may have created and amplified the molecular diversification within the two gene families: 1) gene duplication mechanisms that gave rise to different, but nevertheless similar, sister genes encoding receptor subtypes or pseudogenes and 2) speciation that originated species homologs and the development of genetic polymorphism that provided receptor variants found in individuals within the same species (reviewed in Ref. 452).

Analysis of the gene structure of D₂-like receptors revealed that the D₂ receptor coding region contains six introns (91, 162, 168, 315), the D₃ receptor coding region five (420), and the D₄ receptor three (450). Interestingly, the localization of introns is similar in the three receptor genes and in the opsin gene. The D₃ receptor lacks the fourth intron of the D₂ , and the D₄ receptor lacks the third and fourth introns of the D₂ . The third intron of the D₄ gene has an unusual intron/exon junction in which the conventional splice junction donor and acceptor sites are missing (450, 451).

The presence of introns within the coding region of D₂-like receptors allows the generation of receptor variants. Indeed, the D₂ receptor has two main variants, called D_2S and D_2L , which are generated by alternative splicing of a 87-bp exon between introns 4 and 5 (91, 162, 315; reviewed in Ref. 159). Splice variants of the D₃ receptor encoding nonfunctional proteins have been also identified (137, 161, 418). Analysis of the gene for the human D₄ receptor revealed the existence of polymorphic variations within the coding sequence. A 48-bp sequence in the third cytoplasmic loop exists either as a direct repeat sequence, as a fourfold repeat, or a sevenfold repeat. D₄ receptors containing up to 11 repeats have been found (451).

The D₅ receptor has two related pseudogenes on human chromosomes 1 and 2. They are 98% identical to each other and 95% identical to the human D₅ receptor and code for truncated, nonfunctional forms of the D₅ receptor (169, 464).

	IV. STRUCTURE OF DOPAMINE RECEPTORS

Top Previous Next References

Analysis of the primary structure of the cloned DA receptors revealed that they are members of the seven transmembrane (TM) domain G protein-coupled receptor family and share most of their structural characteristics (Fig. 1). Members of this family display considerable amino acid sequence conservation within TM domains (reviewed in Ref. 362).

View larger version (58K): [in this window] [in a new window]   FIG. 1.   Dopamine receptor structure. Structural features of D1-like receptors are represented. D2-like receptors are characterized by a shorter COOH-terminal tail and by a bigger 3rd intracellular loop. Residues involved in dopamine binding are highlighted in transmembrane domains. Potential phosphorylation sites are represented on 3rd intracellular loop (I3) and on COOH terminus. Potential glycosylation sites are represented on NH2 terminal. E1-E3, extracellular loops; 1-7, transmembrane domains; I2-I3, intracellular loops.

Analysis of DA receptor structure pointed to similarities and dissimilarities between D₁-like and D₂-like receptors (78, 159, 217, 337). Members of the same family share considerable homology. The D₁ and D₅ receptors share a 80% identity in their TM domains. The D₂ and D₃ receptors have a 75% identity in their TM domains, and the D₂ and D₄ receptors share a 53% identity in the TM domains.

The NH₂-terminal stretch has a similar number of amino acids in all the receptor subtypes and carries a variable number of consensus N-glycosylation sites. The D₁ and D₅ receptors possess two such sites, one in the NH₂ terminal and the other one in the second extracellular loop. The D₂ receptor has four potential glycosylation sites, the D₃ has three, and the D₄ possesses only one (reviewed in Refs. 78, 159, 217, 337).

The COOH terminal is about seven times longer for the D₁-like receptors than for the D₂-like receptors, is rich in serine and threonine residues, and contains a cysteine residue that is conserved in all G protein-coupled receptors and that has been shown to be palmitoylated in the -adrenergic receptors and in rhodopsin probably to anchor the cytoplasmic tail to the membrane (338, 347). In the D₁-like receptors, this cysteine residue is located near the beginning of the COOH terminus, whereas in the D₂-like receptors, the COOH terminus ends with this cysteine residue (Fig. 1). Likewise, as in all G protein-coupled receptors, DA receptors possess two cysteine residues in extracellular loops 2 and 3 (reviewed in Refs. 78, 159, 217, 337), which have been suggested to form an intramolecular disulfide bridge to stabilize the receptor structure (111, 142). The D₂-like receptors have a long third intracellular loop, a feature which is common to receptors interacting with G_i proteins to inhibit AC, whereas the D₁-like receptors are characterized by a short third loop as in many receptors coupled to G_s protein (reviewed in Refs. 78, 159, 337).

The D₁ and D₅ receptor third intracellular loop and the COOH terminus are similar in size but divergent in their sequence. In contrast, the small cytoplasmic loops 1 and 2 are highly conserved so that any difference in the biology of these receptors can be probably related to the third cytoplasmic loop and the COOH-terminal tail (reviewed in Refs. 78, 159, 337). The external loop between TM4 and TM5 is considerably different in the two receptor subtypes, being shorter (27 amino acids) in the D₁ receptor than in the D₅ receptor (41 amino acids). The amino acid sequence of this loop, in addition, is divergent in the D₅ and in its rat counterpart D_1b (431, 441).

Site-directed mutagenesis for catecholamine receptors (233, 426, 427) and protein modeling with the ₂-, ₂-, and D₂ receptors (189, 190, 443) suggested that the agonist binding likely occurs within the hydrophobic TM domains (Fig. 1). Highly conserved residues are present in the core of the protein and define a narrow binding pocket that most probably corresponds to the agonist binding site (190). In particular, an aspartate residue in TM3 is most probably involved in binding the amine group of the catecholamine side chain (190, 427). Two serine residues in TM5 have been shown to be hydrogen bond donors to bind the hydroxyl groups of the catechol moiety for the ₂- (426), ₂- (458), D₂ (85, 282), and D₁ (442) receptors. A phenylalanine in TM6 is highly conserved in all receptors interacting with catecholamine neurotransmitters and can make a stabilizing orthogonal interaction with the aromatic moiety of the ligand. A highly conserved aspartate residue in TM2 has been shown to play a crucial role in ₂-adrenergic (77, 190, 427), ₂-adrenergic (458), and D₁ (442) and D₂ dopaminergic (328) receptor activation and to affect agonist binding (190, 414, 442). It has been suggested that the interaction between this aspartate and the agonist is allosteric and can be modulated by Na⁺ or H⁺ (189, 196, 328). A number of cytoplasmic residues, such as the DRY sequence in the second intracellular loop or the alanine residue in the third intracellular loop of the -adrenoceptor, also play a role in receptor activation (233, 427).

	V. RECEPTOR VARIANTS

Top Previous Next References

The D₂ receptor exists as two alternatively spliced isoforms differing in the insertion of a stretch of 29 amino acids in the third intracellular loop (D_2S and D_2L) (91, 162, 315). Because this loop seems to play a central role in receptor coupling, the existence of a splicing mechanism at this level could imply functional diversity. However, in spite of the efforts of several groups, no obvious differences have emerged so far between the two D₂ receptor isoforms. Both variants share the same distribution pattern, with the shorter form less abundantly transcribed (91, 162, 315, 328). Both isoforms revealed the same pharmacological profile, even if a marginal difference in the affinity of some substituted benzamides has been reported (66, 278). When expressed in host cell lines, both isoforms inhibited AC (91, 162, 315). However, the D_2S receptor isoform displayed higher affinity than the D_2L in this effect (91, 317). Both isoforms mediate a phosphatidylinositol-linked mobilization of intracellular calcium in mouse Ltk fibroblasts. Protein kinase C (PKC), however, differentially modulates D_2S- and D_2L-activated transmembrane signaling in this system with a selective inhibitory effect on the D_2S-mediated response (265). Attempts to identify the preferred G protein -subunit for D_2S and D_2L have led to conflicting results. One group suggested, in fact, that the 29-amino acid insertion in the D_2L receptor directs its interaction with G_i-2 (175, 318), whereas another report showed that in transfected cell lines the D_2S isoform signals preferentially through G_i-2 and the D_2L through G_i-3 (405). The two receptor variants, in addition, appear to differ in their mode of regulation (240, 283, 479).

Splice variants of the D₃ receptor have also been identified. One transcript carries a 113-bp deletion in TM3 and a frame shift in the coding sequences generating a stop codon shortly after the deletion and encodes a 100-amino acid-long truncated form of the receptor (418). A second variant derives from a deletion of 54 bp between TM5 and TM6 of the D₃ receptor. Although this structure may be compatible with the occurrence of seven transmembrane domains, cell lines transfected with this cDNA failed to show any binding (161). Two alternatively spliced forms of the D₃ receptors have been identified in the mouse (137), but not in other species (161). These differ by a stretch of 21 amino acids in the third intracellular loop and are generated by a splicing mechanism that uses an internal acceptor site inside an exon, rather than a separate exon like the D₂ receptor. Both isoforms bind dopaminergic ligands with a D₃ pharmacological profile and have the same distribution pattern with the longer form predominant (137).

Analysis of the deduced amino acid sequence of the D₄ receptor reveals that it is the most distantly related of the D₂-like receptors. In human polymorphic variants, the D₄ receptor exists with different insertions in the third intracellular loop. This loop contains repeat sequences of 16 amino acids with the number of repeats differing in the different forms of the receptor. The four-repeat form (D_4.4) is the predominant in the human population (60%). The D_4.7 variant is present in 14% of the population and the D_4.2 in 10% (401, 451). Receptor forms with up to 10 repeats have also been identified but are much less frequent (401). The functional significance of these variants has not been elucidated. They display a slightly different affinity for the neuroleptic clozapine, but none of them has been related to an increased incidence of schizophrenia (401, 451).

	VI. PHARMACOLOGICAL PROPERTIES OF DOPAMINE RECEPTORS

Top Previous Next References

Although the pharmacological profiles of D₁-like and D₂-like receptors are substantially different, the main pharmacological differences described so far within each receptor subfamily are only represented by a variable shift in the affinity of certain agonists and antagonists (Table 2).

So far, it has not been possible to pharmacologically differentiate D₁ and D₅ receptors. The sensitivities of these two receptor subtypes to antagonists are similar. Nevertheless, these compounds generally show a slightly higher affinity for the D₁ than for the D₅ , with (+)-butaclamol as the most discriminating (Table 2) (401, 441). The affinity of agonists at D₁ and D₅ receptors is almost identical. The most consistent difference is represented by DA itself, which has ~10 times higher affinity for the D₅ than the D₁ (Table 2) (431, 441). Cell lines expressing the D₅/D_1b receptor show a higher basal AC activity than those expressing the D₁ (440). This property, together with the observations that DA has a higher affinity for the D₅ than for the D₁ receptor and that some antagonists display negative efficacy at the D₅ , but not at the D₁ , make the D₅ receptor similar to various mutated G protein-coupled receptors that exhibit constitutive activity (250, 386). Functionally, whether the D₅ receptor represents a naturally occurring constitutively active counterpart of the D₁ receptor remains to be clarified.

Analysis of the pharmacological profiles of D₂-like receptors shows that there are no compounds that discriminate between the short and the long variants of the D₂ receptor. A marginal difference in the affinities of sulpiride and raclopride for the two isoforms has been described (66, 278). With respect to the D₃ and D₄ receptors, it has been shown that although they bind hallmark D₂-selective ligands with high affinity, nevertheless both of these receptors have distinguishing pharmacological characteristics (Table 2).

The pharmacological profile of the D₃ receptor reveals that some agonists and antagonists can distinguish it from the D₂ . Dopamine itself has 20 times higher affinity at the D₃ than at the D₂ receptor (420), and this has been related in part to their sequence differences in the third intracellular loop (378). Among agonists, although apomorphine and bromocriptine display similar affinities for both receptors, TL-99, pergolide, quinpirole, and 7-hydroxy-dipropylaminotetralin (7-OH-DPAT) bind with higher affinity at the D₃ than at the D₂ . Quinpirole and 7-OH-DPAT are the most discriminating compounds, with affinities 100 and 10 times higher than at the D₂ , respectively (Table 2) (420). Most neuroleptics display nanomolar affinity at both receptors. However, haloperidol and spiperone show 10- to 20-fold higher affinity at the D₂ than at the D₃ , whereas ()-sulpiride, clozapine, and raclopride do not substantially discriminate between the two receptor subtypes (420). The antagonists UH-232 and AJ-76 have been shown to have three to four times higher affinity at the D₃ than at the D₂ (420). Antagonists with some selectivity for the D₃ receptor (10-30 times difference) were recently developed, such as nafadotride (reviewed in Ref. 421), S-14297 (371, 421), and U-99194A (460).

The pharmacological profile of the D₄ receptor closely resembles those of D₂ and D₃ receptors, but specific differences have emerged (450). The most important feature distinguishing the D₄ from D₂ and D₃ receptors is its higher affinity for clozapine (450). Raclopride, remoxipride, and chlorpromazine, on the other hand, exhibit 10-20 times lower affinity at the D₄ than at the D₂ and D₃ (Table 2) (401, 450). The D₄ receptors have been indirectly measured in brain tissues using [³H]nemonapride, which readily labels all three receptor subtypes, and [³H]raclopride, which labels D₂ and D₃ but to a much lesser extent D₄ receptors. The difference in binding densities of these two ligands has been proposed to reflect specific D₄ receptor binding (400).

	VII. SIGNAL TRANSDUCTION OF DOPAMINE RECEPTORS

Top Previous Next References

The coupling of DA receptors to second messenger pathways has been a subject of intense interest ever since their existence was recognized. Originally, studies of this subject were carried out in preparations from brain or in some cases using purified reconstituted receptors. However, since DA receptor cloning, their coupling properties have been studied predominantly in cell lines transfected with each receptor cDNA. This gave the advantage of working with a pure population of receptor, whereas most brain regions express multiple DA receptor subtypes. However, heterologous expression systems have the disadvantage of mostly being fibroblast in nature, whereas the DA receptors are endogenously expressed primarily in neuronal cells. This raises the possibility of the receptors being expressed in an environment that may contain different complements of G proteins, effectors, and other molecules than they are in vivo. As a result, the use of different heterologous expression systems has often led to apparently conflicting results.

As early as the 1970s, it was recognized that DA receptors could influence the activity of AC (reviewed in Ref. 226). The existence of a D₁ receptor-stimulated AC was recognized in most dopaminergic brain regions, such as striatum, nucleus accumbens, and olfactory tubercle (299). After the cloning of the D₁ receptor in 1990, it was possible to examine its signaling properties in transfected cell lines. In a variety of cell culture lines, it was shown that the D₁ receptor robustly stimulated cAMP accumulation (95, 314, 480). Upon the cloning of the second D₁-like receptor, the D₅ was also found to be coupled to stimulation of AC, as was predicted from its structural similarity to the D₁ receptor (169, 431, 441, 464). Interestingly, the D₅ receptor appears to exhibit an increased agonist-independent activity when compared with the D₁ receptor in 293 cells (440) and raises the questions of whether this is a naturally occurring constitutively active receptor and whether this feature is of relevance to its physiological role. Recent cloning of two more nonmammalian D₁-like receptor subtypes has indicated that these subtypes also stimulate cAMP accumulation in COS-7 cells (101, 430). Thus activation of AC seems to be a general property of all D₁-like receptors.

It is generally assumed that the activation of AC by D₁-like receptors is mediated by the G_s subunit of G proteins. However, it has also been shown that G_olf, which also stimulates AC, is expressed in caudate, nucleus accumbens, and olfactory tubercle and is more abundant in these tissues than G_s (188). This suggests that the D₁ receptor in particular, which is highly expressed in these brain areas, may couple to AC by previously unappreciated mechanisms. Recent studies suggested that D₁-like receptors can also couple to G proteins different from G_s. In particular, striatal D₁ receptors appear to be associated with G_i proteins when reconstituted in phospholipid vesicles (413). In transfected GH₄C₁ cells, D₁ receptors interact with an inhibitory G_i/G_o protein that has not been better identified (229). In addition, immunoprecipitation with antibodies specific for different G protein subtypes revealed that the D₁ receptor coprecipitates with both G_o (230) and G_q (459), the latter coupling D₁ receptors to phosphoinositide metabolism (459, 477).

That the D₂ receptor can inhibit AC was shown in the early 1980s in the pituitary (99, 121, 289, 345) and in the CNS (344). As expected, the cloning of the D₂ receptor confirmed these observations (reviewed in Ref. 159).

Although not immediately apparent, the D₃ receptor has been shown to be coupled to inhibition of AC. Initially, it was reported that this receptor did not inhibit AC in cell lines and did not couple to G proteins as shown by the lack of a guanine nucleotide shift of agonist binding curves (420). Similarly, subsequent observations also indicated that the D₃ receptor did not inhibit cAMP accumulation in various cell lines (144, 274, 438). However, more recently it has been shown that the D₃ receptor does weakly inhibit AC in some cell lines (75, 287, 360, 377).

On the other hand, that the D₄ receptor can inhibit cAMP accumulation was reported in retina (80) and a variety of cell culture lines (74, 287, 290, 438). Thus inhibition of AC seems to be a general property of the D₂-like receptors.

The D₁-like receptors appear to modulate intracellular calcium levels by a variety of mechanisms. One mechanism is via the stimulation of phosphatidylinositol (PI) hydrolysis by phospholipase C (PLC), resulting in the production of inositol 1,4,5-trisphosphate, which mobilizes intracellular calcium stores. There have been conflicting reports as to whether D₁-like receptors are capable of stimulating PI hydrolysis. Upon the cloning of each of the D₁-like receptors, it was reported that these receptors could not stimulate PI turnover in COS-7 cells (95, 101, 430, 441). In addition, Pedersen et al. (351) reported that neither D₁ nor D₅ receptors affected intracellular calcium concentration in Chinese hamster ovary (CHO) or baby hamster kidney cells. In contrast, it has been shown that D₁-like receptor agonists cause increases in PI metabolism in various brain regions (444, 445). However, it should be noted that greater than 100 µM agonist is required to see this effect, calling into question the physiological relevance of this response. Other results have suggested indirectly that D₁-like receptors activate PKC via a PLC-mediated mechanism. The D₁ agonists cause neurite retraction of catfish horizontal cells in culture, and this effect is mimicked by activators of PKC such as phorbol esters or diacylglycerol (379). In addition, the D₁ receptor stimulates PI hydrolysis in Ltk cells (266). In both of these cases, a significant effect was observed at 1 µM DA, suggesting that coupling to PLC may be a real mechanism of D₁-like receptor signaling, at least in some cases.

On the other hand, the D₁ receptor appears to stimulate release of intracellular calcium stores via a mechanism other than stimulation of PI turnover. D₁ receptor-induced increase in intracellular calcium levels in 293 cells (140, 263) is mimicked, in fact, by other means of increasing cAMP levels (263), and thus is probably the result of activation of protein kinase A (PKA).

Finally, the D₁ receptor appears to affect the activity of calcium channels. In both rat striatal neurons and D₁ receptor-transfected GH₄C₁ cells, D₁ agonists increase calcium currents by L-type calcium channels. In both cases, the effect is mimicked by cAMP analogs (266, 432) and blocked by PKA inhibitors (432), suggesting that it may be the result of phosphorylation of calcium channels by PKA. In addition, in rat striatal neurons, D₁ agonists reduce calcium currents carried by N- and P-type calcium channels. This activity of the D₁ receptor was also mimicked by cAMP analogs and blocked by PKA inhibitors as well as the phosphatase inhibitor okadaic acid (432). The proposed model is that D₁ receptors reduce these currents by PKA stimulation of a phosphatase which, in turn, dephosphorylates the channels leading to their inactivation. Thus the regulation of calcium by D₁-like receptors appears to be quite complex and occurs through a variety of mechanisms.

D₂-like receptors also mediate changes in intracellular calcium levels. In some transfected cell systems, the D₂ receptor produces an increase in intracellular calcium via stimulation of PI hydrolysis. This has been observed in Ltk cells (448) and in CCL1.3 cells (438). However, in many other cell lines, the D₂ receptor has been shown not to couple to this second messenger. In addition, D₂ receptors in the pituitary have been shown to inhibit PI metabolism (52, 122, 416). Neither D₃ nor D₄ receptors increase PI hydrolysis in any cell line tested. D₂ receptors have also been shown to cause release of intracellular calcium stores in NG108-15 cells, although the mechanism for this release has not been examined (64).

D₂-like receptors can also cause a decrease in intracellular calcium levels by inhibition of inward calcium currents. This is the case for the D₂ receptor in GH₄C₁ cells (396, 448), pituitary lactotrophs (268), melanotrophs (468), and differentiated NG108-15 cells (397). D₃ receptors also inhibit calcium currents in differentiated NG108-15 cells (395), whereas D₄ receptors have this effect in GH₄C₁ cells (396). Two mechanisms may be responsible for this effect: D₂-like receptor-induced activation of potassium currents leading to alterations in membrane potential (reviewed in Ref. 447) and activation of G proteins that directly inhibit some calcium channels. In both pituitary lactotrophs and GH₄C₁ cells, inactivation of G_o subunits by antisense oligonucleotides abolishes inhibition of calcium currents by D₂ receptors (15, 267). In contrast, in pituitary cells, alterations in potassium currents by the D₂ receptor appear to be mediated via G_i-3 subunits (15, 268), suggesting that the D₂ modulation of calcium currents is independent of changes in potassium conductance. Thus, similar to the D₁-like receptors, the D₂ receptor seems to alter intracellular calcium levels through multiple mechanisms, whereas to date, the D₃ and D₄ receptors have only been shown to inhibit calcium currents.

Dopamine receptors have been shown to influence the activity of potassium channels. This has not been well documented in the case of the D₁-like receptors. D₁-like agonists were shown to increase potassium efflux from chick retinal cells via a cAMP-independent mechanism (243). In contrast, a D₁-like agonist inhibited a potassium current in rat striatal neurons (232).

The role of D₂-like receptors in modulating potassium currents has been more extensively studied. In many preparations, it has been shown that D₂-like receptors increase outward potassium currents, leading to cell hyperpolarization. Such effects have been observed in rat striatal and mesencephalic neurons as well as in anterior pituitary (65, 119, 172, 232, 264, 467). This activation of potassium currents appears to be modulated by G protein mechanisms. The effect of DA on potassium currents in melanotrophs is in fact abolished by pertussis toxin (PTX) treatment (264, 467). In addition, treatment of cells with G protein antibodies or antisense oligonucleotides blocks the D₂ receptor stimulation of potassium currents. In pituitary, activation of potassium currents appears to be mediated by G_i-3 (15, 268), whereas in rat mesencephalon cultures, by G_o (264). Such discrepancies may be the result of varying G protein subunit expression between different cells, or may reflect the modulation of different potassium conductances by D₂ receptors.

The functional significance of cell hyperpolarization appears to be the inhibition of DA release by autoreceptors in the brain and of prolactin secretion in the pituitary. Blockade of potassium channels with 4-aminopyridine (4-AP) or tetraethylammonium (TEA) abolished the inhibition of evoked DA release by D₂-like agonists in striatal slices or synaptosomes (39, 63). Furthermore, in transfected MN9D cells, D₂ or D₃ receptor-mediated inhibition of DA release was also blocked by 4-AP and TEA (439).

In 1991, several groups showed that in CHO cells, the D₂ receptor potentiates the release of arachidonic acid (AA) evoked by calcium (125, 227, 356). These results were confirmed later by Freedman et al. (144) and MacKenzie et al. (274). In addition, in primary striatal neuron cultures, D₂-like agonists also cause potentiation of calcium-evoked AA release (392). The D₃ receptor does not appear to have this effect in cultured cell lines, but the D₄ receptor does potentiate AA release in CHO cells (74). This pathway is sensitive to PTX (74, 356), suggesting that G_i subunits are involved. The mechanism by which D₂-like receptors potentiate AA release is not clear. In some reports, this effect is not related to changes in cAMP levels that might be mediated by the D₂ or D₄ receptors (74, 225). Although Piomelli et al. (356) observed an enhancement of the D₂ receptor potentiation of AA release in the presence of a cAMP analog, PKC seems more likely to play a role in this signaling system. Downregulation of PKC by 24-h treatment of cells with phorbol 12-myristate 13-acetate blocks the D₂ and D₄ effect (74, 225) as does treatment with the PKC inhibitor staurosporine (125). In addition, activation of PKC increases the maximal AA release in the presence of D₂ agonists and calcium ionophore and increases the potency of agonist to elicit this response (109). This evidence suggests that potentiation of AA release is mediated by alterations in PKC activity.

There is little evidence that D₁-like receptors affect AA release. Piomelli et al. (356) reported that in CHO cells, the D₁ receptor did not affect calcium-evoked AA release. However, when D₁ and D₂ receptors were expressed simultaneously in CHO cells, a combination of D₁ and D₂ agonists caused a greater potentiation of AA release than D₂ agonists alone. In contrast, in primary cultures of striatal neurons, D₁-like agonists caused an inhibition of calcium-evoked AA release (392), an effect that was mimicked by forskolin, suggesting the involvement of PKA in this response.

Dopamine receptors also appear to affect the activity of amiloride-sensitive Na⁺/H⁺ exchangers, which are responsible for regulation of intracellular pH and cell volume. This exchanger is also the major player in sodium absorption in many epithelia (478). The activity of the Na⁺/H⁺ exchanger is regulated by multiple mechanisms, including phosphorylation-dependent and -independent events and direct regulation by the G_i-3 subunit (104, 478). In preparations of renal brush-border membrane vesicles, D₁ receptor agonists cause an inhibition of the activity of the Na⁺/H⁺ exchanger by both cAMP-dependent and cAMP-independent mechanisms (123, 125).

In contrast, the D₂ receptor activates a Na⁺/H⁺ exchanger in many cells. This has been observed in renal brush-border membrane vesicles (123) in transfected C6 glioma and Ltk cells (329) and in primary cultures of anterior pituitary cells (147). In these systems, the observed increase in extracellular acidification was not blocked by PTX, suggesting that a mechanism other than G_i was involved. However, in CHO cells, D₂ , D₃ , and D₄ receptors all increase extracellular acidification rates in a PTX-sensitive manner (74). These conflicting reports are presumably the result of the existence of multiple subtypes of amiloride-sensitive Na⁺/H⁺ exchangers as well as multiple mechanisms for their regulation.

The Na⁺-K⁺-ATPase, which pumps sodium out of cells and potassium in, is essential for maintaining the electrochemical gradient that is responsible for the excitability of nerve and muscle cells and drives the transport of fluid and solutes across epithelial membranes. It has been known that DA receptors influence the activity of this ion pump. In this manner, DA regulates fluid absorption in the kidney and neuronal excitability in the brain. Most work has suggested that DA effects on the Na⁺-K⁺-ATPase are mediated through the D₁ receptor. However, some reports also suggested that activation of both D₁ and D₂ receptors may be required. In a preparation of dissociated striatal neurons, Bertorello et al. (34) found that inhibition of the Na⁺-K⁺-ATPase required the presence of DA or a combination of D₁ and D₂ agonists. However, other studies have suggested that D₁ receptors alone are sufficient to evoke an inhibition of the Na⁺-K⁺-ATPase. In the chick retina, DA inhibits Na⁺-K⁺-ATPase activity, and this has been suggested to be mediated by D₁ receptors (243). In addition, in renal proximal tubule preparations and the Madin-Darby canine kidney cell culture model of cortical collecting tubule, D₁-like agonists cause an inhibition of the Na⁺-K⁺-ATPase, whereas D₂-like agonists have no effect (70, 407). In the kidney, the effects of DA receptor activation on Na⁺-K⁺-ATPase activity appear to be the result of phosphorylation cascades involving both PKA and PKC. However, in Ltk fibroblast cells transfected with the D₁ receptor, D₁-like agonists inhibit Na⁺-K⁺-ATPase activity in a PKA-dependent manner (195). Therefore, although in general it appears that D₁ receptors are responsible for regulation of the Na⁺-K⁺-ATPase, the mechanism may vary according to the tissue examined.

Recent evidence has suggested that in some cases D₂-like receptors are involved in mitogenesis and cell differentiation. The D₃ receptor stimulates [³H]thymidine incorporation in NG108-15 cells (355), and both D₂ and D₃ receptors have this effect in CHO cells (75, 244, 434). This effect is blocked by PTX and appears to be independent of alterations in cAMP levels. Lajiness et al. (244) found that the D₂ mitogenic effect was accompanied by an increase in tyrosine phosphorylation levels and was blocked by the tyrosine kinase inhibitor genistein, suggesting that this receptor may cause the activation of the mitogen-activated protein kinase pathway.

In contrast to the above results, the D₂ receptor has also been shown to inhibit cell growth in some cell lines. GH₄C₁ cells transfected with the D₂ receptor respond to agonists with a decrease in [³H]thymidine uptake. Florio et al. (138) found that this effect was abolished by PTX, was accompanied by an increase in phosphotyrosine phosphatase (PTP) activity, and was blocked by the PTP inhibitor vanadate. In contrast, another study found that in GH₄C₁ , the D₂ receptor-mediated inhibition of [³H]thymidine uptake was not blocked by PTX but was blocked by downregulation of PKC and by PKC inhibitors (406). Thus, although the mechanism of inhibition of mitogenesis by D₂ receptors in GH₄C₁ cells is not clear, it may result from PKC-mediated activation of a phosphatase. The effects of D₂ receptor activation on cell growth appear to highly depend on the cell type examined.

Finally, the D₂-like receptors may promote some aspects of cell differentiation. When D₂ , D₃ , or D₄ receptors were expressed in the mesencephalic cell line MN9D, agonists caused increases in neurite number and length as well as total neuritic extent (435). However, in a study using primary cultures of rat mesencephalon neurons, a D₂-like receptor agonist did not affect survival or differentiation of these cells (449). The role of D₂-like receptors in neuronal differentiation thus remains to be clarified.

In conclusion, much effort has gone into studying the signal transduction of the DA receptors during the last 20 years. Many second messengers for these receptors have been identified, including cAMP, calcium, potassium, and AA. In addition, these receptors modulate other "effectors" by more indirect means, including Na⁺/H⁺ exchangers, the Na⁺-K⁺-ATPase, and cell growth and differentiation pathways (Fig. 2). However, in many cases, there is conflicting evidence in the literature for the modulation of various messengers, or the mechanism by which an effector is modulated by the DA receptors. Many of these discrepancies probably arise from the use of different tissues or cell culture lines. It is now known that many of the components of signal transduction pathways have multiple isoforms, including receptors, G proteins, and effectors, and that these have differing patterns of expression and regulatory properties. Defining which of these specific signal transduction events is involved in the various physiological actions of DA may require the development of specific pharmacological agents or genetic animals models.

View larger version (37K): [in this window] [in a new window]   FIG. 2.   Signal transduction of dopamine receptors. AC, adenylate cyclase; PLC, phospholipase C.

	VIII. DOPAMINE RECEPTORS IN THE BRAIN

Top Previous Next References

The D₁ receptor is the most widespread DA receptor and is expressed at higher levels than any other DA receptor (95, 145, 463). D₁ mRNA has been found in the striatum, the nucleus accumbens, and the olfactory tubercle. In addition, D₁ receptors have been detected in the limbic system, hypothalamus, and thalamus. On the other hand, in other areas where the D₁ receptor protein is highly expressed such as the entopeducular nucleus and the substantia nigra pars reticulata (153, 255), no mRNA has been detected (95, 145, 463). This suggests that in these areas the D₁ receptor is mainly present in projections (95, 145, 463). As a matter of fact, D₁ receptors in the entopeduncular nucleus and in the substantia nigra pars reticulata are preferentially localized on striatal GABAergic neurons coexpressing substance P (153, 255).

The D₅ receptor is poorly expressed in rat brain when compared with the D₁ receptor. A distribution restricted to the hippocampus, the lateral mamillary nucleus, and the parafascicular nucleus of the thalamus, where the D₁ receptor is not significantly expressed, was originally reported (293, 441), with little or no message detected in the dorsal striatum, nucleus accumbens, and olfactory tubercle. Upon further examination, D₅ receptor mRNA has been found in several rostral forebrain regions including cerebral cortex, lateral thalamus, diagonal band area, striatum, and, to a lesser extent, substantia nigra, medial thalamus, and hippocampus (76, 203, 366).

The development of specific antibodies against DA receptor subtypes recently made it possible to define their cellular and subcellular localization in different regions of primate brain. Both D₁ and D₅ receptors are coexpressed in pyramidal neurons of prefrontal, premotor, cingulate and entorhinal cortex, the hippocampus, and the dentate gyrus (27, 28, 197, 417). Electron microscopy analysis demonstrated that in the prefrontal cortex and the hippocampus, D₁ and D₅ receptors have both pre- and postsynaptic localization, with the postsynaptic one more frequently observed. Ultrastructural analysis suggested that within individual pyramidal neurons, D₁ and D₅ receptors have a different localization with the D₁ concentrated in dentritic spines and the D₅ in dendritic shafts (28, 417). In the olfactor bulb D₁ receptors are restricted to the internal granular and plexiform layers and in the amygdala in the intercalated and basolateral nuclei (260). In the caudate nucleus, D₁ and D₅ receptors are mostly localized within medium-sized GABAergic neurons (28, 197, 417). D₅ but not D₁ receptors are present also in large cholinergic interneurons (28). Ultrastructural analysis suggested that D₁ receptors are present on spines postsynaptic to asymmetrical synapses, that both D₁ and D₅ receptors are at postsynaptic densities of small synapses characteristics of DA terminals, and that presynaptic D₁ and D₅ receptors are on axons forming asymmetrical synapses (28, 187, 260, 417). D₁ receptors have been localized in the entopeduncular nucleus and in the pars reticulata of the substantia nigra, where D₅ receptors are undetectable (28, 187, 260). This observation suggests that if D₁ and D₅ receptors are colocalized in medium-sized spiny neurons of caudate, only the D₁ receptor is transported to striatonigral terminals. These differences in the cellular and subcellular localization thus suggest that although D₁ and D₅ receptors exhibit similar pharmacology, they are not functionally redundant.

The D₂ receptor has been found mainly in the striatum, in the olfactory tubercle, in the core of nucleus accumbens (38; reviewed in Ref. 217), where it is expressed by GABAergic neurons coexpressing enkephalins (253, 256), and in the septal pole of the shell of the nucleus accumbens where it is expressed by neurotensin-containing neurons (105). D₂ receptor mRNA is also present in the prefrontal, cingulate, temporal, and enthorinal cortex, in the septal region, in the amygdala, and in the granule cells of the hippocampal formation (38; reviewed in Ref. 217). It is also found in the hypothalamus, in the substantia nigra pars compacta, and in the ventral tegmental area, where it is expressed by dopaminergic neurons (38, 292, 463). Immunohistochemical analysis with specific antibodies revealed that D₂ receptors are present in medium spiny neurons of the striatum where they are more concentrated in spiny dendrites and spine heads than in the somata. Colocalization with D₁ receptors is rare. D₂ immunoreactive terminals are frequently detectable, forming symmetrical, rather than asymmetrical, synapses (187, 260). The D₂ receptors are present in perikarya and dendrites within the substantia nigra pars compacta and are much more concentrated in the external segment of the globus pallidus than in other striatal projections (260). D₂ receptor immunoreactivity has been detected in the glomerular and internal plexiform layers of the olfactory nerve and in the central nucleus of the amygdala (260).

The D₃ receptor has a specific distribution to limbic areas (245, 246) such as the ventromedial shell of the nucleus accumbens (38) where it is expressed by substance P and neurotensin neurons projecting to the ventral pallidum (105, 106), the olfactory tubercle, and the islands of Calleja (38, 258). In contrast, it is poorly expressed in the dorsal striatum (38, 258, 420). The D₃ mRNA was also found in the substantia nigra pars compacta, in the ventral tegmental area, where it is expressed in a minority of dopaminergic neurons when compared with the D₂ receptors and in the cerebellum (105, 106). In the islands of Calleja, both D₃ receptor binding and mRNA are present in granule cells (106, 258), which are known to make sparse contacts with dopaminergic axons. Purkinje cells in lobules 9 and 10 of the archicerebellum express D₃ mRNA, whereas binding sites were detectable only in the molecular layer (106, 258). No dopaminergic projections are present in this area, suggesting that the D₃ receptor may respond to DA diffusing extrasynaptically (106). The D₃ receptor was also found at low expression levels in the hippocampus, in the septal area, and in various cortical layers and subregions of the medial-temporal lobe (38).

Low levels of the D₄ receptor mRNA have been found in the basal ganglia. In contrast, this receptor appears to be highly expressed in the frontal cortex, amygdala, hippocampus, hypothalamus, and mesencephalon (343, 450). Significant levels of D₄ mRNA were also found in the retina (80). Recently, a specific antibody directed against the D₄ receptor has been developed. Immunohistochemical and electron microscopy analysis revealed that in both the cerebral cortex and hippocampus, D₄ receptors are present in pyramidal and nonpyramidal neurons that have been identified as GABAergic interneurons (319). In the cerebral cortex and hippocampus, D₄ receptors thus modulate the GABAergic transmission. D₄ receptors have been also found in GABAergic neurons of both segments of globus pallidus and of the substantia nigra pars reticulata and in the reticular nucleus of the thalamus (319).

The behavioral effects of DA have been extensively reviewed (92, 217, 235, 469). Here we briefly summarize some of the functional effects of DA with particular attention to some behaviors where the role of the different DA receptor subtypes has been investigated.

The effects of DA on motor activity have been extensively investigated (reviewed in Refs. 79, 217, 456, 457). The degree of forward locomotion is primarily controlled by the ventral striatum through activation of D₁ , D₂ , and D₃ receptors. Activation of D₂ autoreceptors, which results in decreased DA release, has been shown to decrease locomotor activity (reviewed in Ref. 217), whereas activation of postsynaptic D₂ receptors slightly increases locomotion. Activation of D₁ receptors has little or no effect on locomotor activity (155; reviewed in Ref. 217). However, it is now clear that there is synergistic interaction between D₁ and D₂ receptors in determining forward locomotion so that concomitant stimulation of D₁ receptors is essential for D₂ agonists to produce maximal locomotor stimulation (41, 116; reviewed in Refs. 79, 217, 456, 457). As discussed in section VIII D, these pharmacological observations have been explicitly confirmed by targeted inactivation of the D₁ receptor gene in the mouse (471, 472).

The D₃ receptor, which has been shown to be mainly postsynaptically located in the nucleus accumbens (106), seems to play an inhibitory role on locomotion. D₃-preferring agonists inhibit, in fact, locomotor activity (93; reviewed in Ref. 421), whereas D₃-preferring antagonists evoke motor activation (reviewed in Refs. 421, 461). The opposing effects of D₂ and D₃ receptors on locomotor activity may find a neurochemical correlate in their opposite effects on neurotensin gene expression in the nucleus accumbens (105).

Mesolimbocortical DA is implicated in reward and reinforcement mechanisms as shown by the observation that administration of psychostimulants and drugs of abuse elicits an increase of DA release in the mesolimbic areas, whereas withdrawal of these drugs results in a reduction of dopaminergic transmission. A vast amount of literature has been written in this area (108, 252, 365, 404, 469). Various experimental models have been developed such as intracranial self-stimulation and drug self-administration. In the intracranial self-stimulation paradigm, rats work to obtain electrical stimulation that has rewarding properties and results in DA release in the prefrontal cortex and nucleus accumbens (reviewed in Ref. 217). Pharmacological studies clearly show that both D₁ and D₂ receptors are involved in this behavior, with agonists at both receptors stimulating and antagonists inhibiting the behavior (141, 234).

In the case of drug self-administration, it has been shown that both D₁ and D₂ receptors are involved in the reinforcing properties of different drugs of abuse, with D₂ receptors mediating the stimulant drug reinforcement and D₁ receptors playing a permissive role (25, 277, 354, 403). Stimulation of D₁ receptors by endogenous DA is thus required for the expression of D₂ receptor-mediated behaviors and gene regulation. A recent study suggested that although D₁-like and D₂-like receptor agonists are themselves reinforcing and can both substitute for cocaine in drug discrimination tests, they nevertheless may mediate qualitatively different aspects of the reinforcing stimulus produced by cocaine. In particular, activation of D₂-like receptors has been shown to mediate the incentive to seek further cocaine reinforcement in an animal model of cocaine-seeking behavior. In contrast, D₁-like receptors appear to mediate a reduction in the drive to seek further cocaine reinforcement (403). Agonists of D₁-like receptors may thus be evaluated as a possible therapy of cocaine addiction. Recently, it has been shown that D₃ receptor stimulation inhibits cocaine self-administration in the rat in a way indicating an enhancement of cocaine reinforcement (51, 349).

Although some inconsistencies are present in the literature, there is a general agreement that mesolimbocortical DA plays a role in learning and memory. In the monkey, DA neurons in the A10 area have been reported to be involved with transient changes of impulsive activity in basic attention and motivational processes underlying learning and cognitive behavior (394). Pharmacological studies have shown that both D₁ and D₂ receptors mediate the effects of DA on learning and memory. Activation of both D₁ and D₂ receptors in the hippocampus improves acquisition and retention of different working memory tasks in the rat (261, 348, 465, 466). In the monkey, activation of both D₁ and D₂ receptors in the prefrontal cortex has been reported to improve performance in a working memory task (12, 390, 391). Because of the lack of true agonists and antagonists discriminating among D₁-like and D₂-like receptors, the role of DA receptor subtypes in learning and memory has not been investigated. However, it is worth noting that although the D₁ receptor is poorly expressed in the hippocampal formation, the D₅ receptor is highly expressed in this area so that the D₅ , more than the D₁ receptor, is likely to mediate the effects of D₁ agonists on learning and memory. Similarly, D₃ and D₄ receptors are expressed in the hippocampus, and D₃ receptors are present in the septal area, suggesting a possible contribution of these receptor subtypes to the behavioral effects of D₂ agonists. In contrast, because of their distribution at the cortical level, a central role of D₁ and D₂ receptors can be proposed in the prefrontal cortex-mediated behaviors.

The role of D₃ and D₄ receptors in the physiology of dopaminergic system is still mostly unknown. They are specifically expressed in limbic and cortical regions involved in the control of cognition and emotion and, to a lesser extent in the dorsal striatum, and this makes them attractive and promising targets for new generations of antipsychotic drugs with low incidence of extrapyramidal side effects.