Unveiling the Differences in Signaling and Regulatory Mechanisms between Dopamine D2 and D3 Receptors and Their Impact on Behavioral Sensitization

Dopamine receptors are classified into five subtypes, with D2R and D3R playing a crucial role in regulating mood, motivation, reward, and movement. Whereas D2R are distributed widely across the brain, including regions responsible for motor functions, D3R are primarily found in specific areas related to cognitive and emotional functions, such as the nucleus accumbens, limbic system, and prefrontal cortex. Despite their high sequence homology and similar signaling pathways, D2R and D3R have distinct regulatory properties involving desensitization, endocytosis, posttranslational modification, and interactions with other cellular components. In vivo, D3R is closely associated with behavioral sensitization, which leads to increased dopaminergic responses. Behavioral sensitization is believed to result from D3R desensitization, which removes the inhibitory effect of D3R on related behaviors. Whereas D2R maintains continuous signal transduction through agonist-induced receptor phosphorylation, arrestin recruitment, and endocytosis, which recycle and resensitize desensitized receptors, D3R rarely undergoes agonist-induced endocytosis and instead is desensitized after repeated agonist exposure. In addition, D3R undergoes more extensive posttranslational modifications, such as glycosylation and palmitoylation, which are needed for its desensitization. Overall, a series of biochemical settings more closely related to D3R could be linked to D3R-mediated behavioral sensitization.


Introduction
Dopamine (DA), an ethylamine with an attached catechol group (a phenyl group with two hydroxyl groups attached to meta and para positions) is released from nerve endings upon axonal stimulation. Released DA acts on postsynaptic and presynaptic receptors at the synapse and is mostly taken up back into nerve endings by the dopamine transporter protein, which belongs to solute carrier transporter family.
DA modulates the neuronal activities in the brain regions that are innervated by three major dopaminergic pathways: nigrostriatal, mesocorticolimbic, and tuberoinfundibular. Drugs acting on D 2 R and D 3 R have traditionally been utilized for the treatment of various disorders including Parkinson's disease [1], schizophrenia [2], and hyperprolactinemia [3].
The dopamine receptors are pharmacologically classified into D 1 -like and D 2 -like families [4][5][6]. D 1 -like receptors comprise D 1 and D 5 receptors (D 1 R, D 5 R) [7][8][9], whereas D 2 -like receptors consist of D 2 , D 3 , and D 4 receptors (D 2 R, D 3 R, D 4 R) [10][11][12]. D 1 R and D 5 R mediate the stimulation of adenylyl cyclase by coupling to G s , whereas D 2 R, D 3 R, and D 4 R exert inhibitory effects on this enzyme by coupling to G i/o . D 2 -like receptors have larger third cytoplasmic loops and smaller carboxyl tails compared with D 1 -like receptors.
Whereas D 2 R and D 3 R exhibit a high degree of similarity in sequence homology and signaling properties, they differ in some other respects. For example, the transcript levels of D 3 R are much lower than those of D 2 R, differing by several orders of magnitude in their expression levels [25]. However, D 3 R shows about 150 times higher affinity for DA [26]. In addition, the distribution of D 3 R in the brain is more restricted than that of D 2 R, that is, D 3 R shows preferential expression in brain regions responsible for emotional and cognitive functions, such as the nucleus accumbens and the islands of Calleja and has been suggested as a therapeutic target for the affective and mood-associated diseases [27][28][29]. Moreover, significant variations exist between D 2 R and D 3 R with respect to their endocytosis, desensitization, and posttranslational modifications [30][31][32].
The main objective of this review is to comprehend the functional implications of the distinct regulatory processes of D 2 R and D 3 R by highlighting their differences. Functional in vivo roles and ligands of D 2 R and D 3 R are described in other excellent reviews [33][34][35][36][37]. This review focuses on the regulatory functions of D 2 R and D 3 R in biochemical pathways.

Desensitization
Desensitization of G protein-coupled receptors (GPCRs) represents a gradual attenuation of receptor responsiveness by continuous or repeated exposure to agonists [38][39][40]. GPCR desensitization is generally classified into two categories in which either GRKs/ arrestins or second messenger-regulated kinases are involved. The former desensitizes only agonist-activated receptors, whereas the latter can desensitize receptors both in an activation-dependent and activation-independent manner [40]. Typical second messengerregulated kinases are protein kinase A (PKA) and protein kinase C (PKC) [41][42][43].
The desensitization of GPCRs has been most extensively characterized by the conventional or steric-hindrance-based uncoupling model. When these receptors are activated by agonists, they undergo conformational changes that involve the outward movement of transmembrane helix VI. This results in the formation of a cavity that allows for the binding of G proteins [44][45][46]. The (E/D)x(I/L)xxxGL motif, which is commonly present in the C-termini of Gα and arrestins, is believed to participate in their interaction with the active GPCRs at the same binding crevice [47,48]. The binding of arrestins to GPCRs is likely to occur in competition with G proteins. By binding to the same binding crevice, arrestins can obstruct further coupling of G proteins via steric hindrance, which ultimately leads to a decrease in receptor signaling [49].
Even though D 2 R and D 3 R have similar structural features and signaling pathways, they display distinct properties of desensitization and intracellular trafficking. In response to agonistic stimulation, D 2 R is phosphorylated in a manner dependent on GRK2/3 and in part on GRK5 and mediates arrestin translocation to the plasma membrane accompanied by receptor endocytosis [30]. In contrast, D 3 R is rarely phosphorylated and undergoes only a minute amount of endocytosis.
According to the steric-hindrance-based uncoupling model, D 2 R but not D 3 R would show agonist-induced desensitization. However, subsequent studies have reported the desensitization of D 3 R but not D 2 R [50][51][52], suggesting that the steric-hindrance-based uncoupling model may not be applicable to the desensitization of D 2 R and D 3 R or desensitized D 2 R might be rapidly resensitized [51,53].
Recent studies conducted on D 2 -like receptors (D 2 R, D 3 R, D 4 R) and the β 2 adrenoceptor (β 2 AR) indicate that receptors that undergo desensitization have a predisposition to form a stable complex with Gβγ and enable a basal interaction between Gβγ and arrestins [54]. This arrangement leads to the recruitment of Mdm2 to the cytoplasm by these receptors in their basal state, ultimately resulting in the constitutive ubiquitination of arrestins in the cytoplasm [55].    Circles filled with yellow indicate the amino acids responsible for agonist-induced desensitization; the two S residues highlighted in red indicate PKC-mediated regulatory processes.

Endocytosis
The process of endocytosis involves the uptake of extracellular materials by cells through the inward budding of vesicles that originate from the plasma membrane [60]. Endocytosis of the receptor occurs without failure when agonist stimulation induces receptor phosphorylation and arrestin translocation. Translocated arrestins mediate receptor endocytosis [61,62] and additional components of endocytic machinery, such as dy-

Endocytosis
The process of endocytosis involves the uptake of extracellular materials by cells through the inward budding of vesicles that originate from the plasma membrane [60]. Endocytosis of the receptor occurs without failure when agonist stimulation induces receptor phosphorylation and arrestin translocation. Translocated arrestins mediate receptor endocytosis [61,62] and additional components of endocytic machinery, such as dynamin and β2-adaptin, are also involved in this process [63,64]. The interaction of arrestins with receptors in endocytic vesicles is dependent on the affinity between the two, with varying levels of association observed [65]. Class A receptors, such as the β 2 AR and D 1 R, exhibit lower affinity towards arrestins compared with class B receptors. As a result, these receptors tend to dissociate from arrestins when they are incorporated into endocytic vesicles. In contrast, class B receptors, such as the angiotensin II type 1A receptor and vasopressin V2 receptor, maintain their interaction with arrestins during endocytosis. This ability of arrestins to remain associated with class B receptors is facilitated by specific clusters of serine and threonine residues located in the carboxyl-terminal tails of the receptors [66].
Similar to β 2 AR, D 2 R undergoes agonist-induced endocytosis. GRK2/3 and GRK5 in part mediate the phosphorylation of D 2 R. In contrast, GRK-mediated phosphorylation rarely occurs in D 3 R, which undergoes only a minute amount of endocytosis. It has been reported that S/T residues located in the second (T134, T144, S147, S148) and the third (T225) intracellular loops of D 2 R are involved in agonist-induced D 2 R endocytosis (Cho et al., 2010) ( Figure 1). Another study has shown that some other S/T residues located within the third intracellular loop of D 2 R (S256, S257, T258, S259, T264, S282, S288, S292) are also phosphorylated in a GRK-mediated manner and that they are required for the agonist-induced (GRK-mediated) recycling of endocytosed D 2 R [53] (Table 1, Figure 1). It is interesting that different sets of S/T residues mediate distinct intracellular trafficking processes. The resensitization of D 2 R occurs when the receptor dissociates from arrestin via agonist-induced endocytosis [51].
In addition to GRK/arrestin-mediated regulation, D 2 R and D 3 R are also regulated by PKC. Robust phosphorylation of D 2 R is induced by PMA (phorbol 12-myristate 13acetate) but the resulting endocytosis is weaker than the agonist-induced endocytosis of D 2 R [43,51]. The phosphorylation, endocytosis, and desensitization of D 3 R induced by PMA are more robust than those of D 2 R [43,51,56]. It has been reported that treatment with PMA results in the ubiquitination of D 3 R and its subsequent degradation through the lysosomal pathway. The degradation of D 3 R by PKC is dependent on clathrin-mediated endocytosis. However, this process is distinct from the desensitization of D 3 R and is not affected by its inhibition [59]. The D 2 R residues T225, S228, S229, T322, T324, and S325 are responsible for PMA-induced receptor phosphorylation and endocytosis [56,67] (Table 1, Figure 1). PKC-mediated phosphorylation and endocytosis of D 3 R are mediated by the S229 and S257 residues [43] (Table 1, Figure 2). Even though D 3 R does not undergo typical endocytosis, which involves the movement of receptors from the plasma membrane to the cytosol, it undergoes another type of intracellular trafficking called pharmacological sequestration [52,68]. Upon desensitization of D 3 R induced by an adequate amount of agonist, it translocates to the hydrophobic region of the membrane where the hydrophilic agonist has reduced binding efficiency. The time course of D 3 R desensitization and pharmacological sequestration are similar, and both depend on the presence of Gβγ and arrestin, indicating that the phenomenon of pharmacological sequestration may account for D 3 R desensitization.
The movement of GFP-tagged receptors from the plasma membrane to the cytosol upon agonist treatment is a convenient way to confirm the endocytosis of receptor proteins. Alternatively, biochemical assays such as fluorescence-activated single cell sorting (FACS) or enzyme-linked immunosorbent assay (ELISA) can be employed using constructs with HA or FLAG epitopes attached to the N-terminus of the receptor. Hydrophilic radioligands can also be used to assess receptor endocytosis as they cannot bind receptors in the cytoplasm. However, caution should be exercised when dealing with receptors that have high affinity for the agonist, such as DA and D 3 R. In such cases, it is difficult to completely wash away the agonist from cells expressing D 3 R unless the cells are washed under harsh conditions, such as with a low pH buffer. If cells are treated with a radioligand under conditions where a pre-applied agonist is not completely washed, it may give a false impression that receptor endocytosis is occurring (approximately 30%) because the radioligand cannot properly bind to the receptor that is already occupied by the tightly bound agonist.

Roles of Small G Proteins in the Signaling and Endocytosis of D 2 R and D 3 R
There are two families of GTP binding proteins: heterotrimeric large G proteins composed of three subunits (α, β, and γ) and small G proteins with a single subunit. Small G proteins mediate more versatile functions than trimeric G proteins due to their larger number in the cell [69,70]. Small GTPases can be categorized into five main families, namely Ras, Rho, Rab, ADP-ribosylation factor (Arf), and Ran, based on both sequence and functional criteria [70].
Small GTPases, including Rab and ARF proteins, as well as the large GTPase dynamin, play regulatory roles in vesicular transport [71]. Several Rab proteins, including Rab5 and Rab23, are involved in the regulation of membrane trafficking and recycling of cell surface proteins from the Golgi to the plasma membrane [72,73].
Whereas RalA and ARF6 small G proteins are known to regulate D 2 R and D 3 R [67,74,75], the involvement of Rab5 and Rab23 in the endocytosis of D 2 R remains controversial [67,76].

RalA
RalA belongs to the Ras family of small G proteins. The GTP-bound form (G23V, active) of RalA is known to inhibit the endocytosis of D 2 R in a manner that is independent of the previously reported downstream effectors of RalA, such as Ral-binding protein 1 and phospholipase D [74]. The endocytosis of GPCRs, including D 2 R, can be inhibited by active RalA through its high affinity for GRK2, which sequesters GRK2 away from the GPCRs. RalA is converted from its active to inactive state by the translocation of Ral GDP dissociation stimulator-like (RGL), a guanine nucleotide exchange factor, from the plasma membrane to the cytosol in a complex with Gβγ. Thus, the agonist-induced Gβγ-mediated conversion of RalA from its GTP-bound to GDP-bound form is thought to trigger and facilitate the endocytosis of GPCRs induced by agonist stimulation [74].
Filamin A (FLNA), an actin-binding protein, functions as a signaling and intracellular trafficking scaffold for various GPCRs, including D 2 R and D 3 R. FLNA plays an essential role in maintaining the proper expression of D 2 R on the plasma membrane [77] and is also necessary for signaling through both D 2 R and D 3 R [78][79][80]. FLNA is known to constitutively inhibit RalA activity, thus ensuring proper receptor trafficking and signaling. In the absence of FLNA, RalA is activated and sequesters GRK2 from the receptor, resulting in desensitization of D 2 R due to its inhibition of recycling [51,53]. On the other hand, active RalA inhibits the signaling of D 3 R, which does not undergo agonist-induced endocytosis, in an arrestin-dependent manner. Thus, it is suggested that active RalA modulates receptor signaling differently depending on whether GRK2 or arrestin is involved in the functional regulation of the respective receptor.

ARF6
The ARF (ADP-ribosylation factor) family of small GTPases comprises six members, among which ARF6 is the most extensively studied with regards to its role in intracellular trafficking of membrane proteins. ARF6 is known to accumulate in clathrin-coated pits (CCPs) in a GTP-dependent manner and facilitate the rapid recycling of plasma membrane receptors [81,82]. Furthermore, research suggests that the activation of ARF6 via arrestinmediated mechanisms is involved in the endocytosis of β 2 AR [83].
According to a report, the endocytosis of D 2 R is not influenced by ARF6, as neither constitutively active nor a dominant-negative mutant of ARF6 has any effect [67]. However, the recycling of endocytosed D 2 R is inhibited by a constitutively GTP-bound mutant of ARF6 (Q67L), indicating that the conversion of ARF6 from the GTP-bound to GDP-bound form is crucial for the proper recycling of endocytosed vesicles.

Roles of Regulators of G Protein Signaling (RGS) in the Signaling of D 2 R and D 3 R
In its inactive state, the Gα subunit is bound to GDP, whereas activation of the GPCR by an agonist triggers the exchange of GDP for GTP on the Gα subunit, resulting in its dissociation from the Gβγ heterodimer [84]. Both the Gα and Gβγ subunits are capable of regulating downstream effector proteins. The Gα subunit possesses intrinsic GTPase activity, which converts the GTP-bound Gα to its inactive form, Gα-GDP, allowing it to reassociate with the Gβγ heterodimers and effectively terminate signaling via both Gα-GTP and Gβγ.
The rapid physiological timing of GPCR signal transduction observed in vivo contrasts with the slow rates of GTP hydrolysis exhibited by purified Gα subunits in vitro. This discrepancy has been explained by the existence of the GTPase-accelerating protein (GAP) family [85,86]. Unlike the five main families of GAPs that regulate the Ras superfamily, regulators of G protein signaling (RGSs) specifically target heterotrimeric G proteins [87]. RGS proteins bind to and regulate the Gαi and Gαq subfamilies of proteins by increasing their GTPase activity. However, the Gαs subfamily already exhibits a sufficiently high intrinsic GTPase activity and is not sensitive to RGS protein-mediated GAP activity.
RGS2, RGS4, and RGS9-2 are RGS subtypes that have been associated with the dopaminergic nervous system. In a previous study utilizing in situ hybridization, it was suggested that D 1 R and D 2 R co-localize with RGS2 and RGS4, respectively [88]. It has also been shown that D 1 R and D 2 R regulate the expression of RGS2 and RGS4 in opposite directions. For instance, quinpirole, a D 2 -like receptor agonist, induces the downregulation of RGS2 and upregulation of RGS4 [89]. Interestingly, male rats require RGS4, but not RGS2, for the rewarding effects of cocaine [90]. Further studies have demonstrated that the N-terminal domain of RGS4 interacts with D 2 R and D 3 R and inhibits receptor signaling via the RGS domain [57].
RGS9-2 exhibits high levels of expression in the striatum and nucleus accumbens, regions that also express D 2 -like receptors. Studies suggest that RGS9-2 plays a functional role in regulating D 2 -like receptors. Specifically, introducing RGS9-2 through viral expression in brain regions such as the nucleus accumbens or introducing RGS9-2 proteins into striatal cholinergic interneurons reduces the behavioral or electrophysiological response to D 2 -like receptor stimulation [91][92][93]. In contrast, when RGS9-2 is knocked out, there is an augmentation in the behavioral response to D 2 -like receptor activation [91,93].
RGS9-2 is a member of the R7 RGS subfamily and possesses three domains: RGS, disheveled-EGL10-pleckstrin (DEP) homology, and G-gamma-like (GGL). The RGS domain binds to the Gα subunit and facilitates GAP activity. The DEP domain anchors the protein to the membrane and allows for interaction with the C-terminal tail of GPCRs. Finally, the GGL domain ensures protein stability by dimerizing with Gβ5 [94][95][96][97].
In HEK-293 cells or C6 glioma cells, the expression of RGS9-2 specifically hinders D 3 R signaling while leaving D 2 R/D 4 R signaling unaffected. This can be attributed to the varying affinities of the receptors for arrestin3, which facilitates the creation of a complex consisting of RGS9's DEP domain, Gβ5, R7-binding protein (R7BP), and D 3 R [98].

Biased Signaling
GPCRs carry out their functions through various signaling pathways, which can be categorized into G protein-dependent and G protein-independent pathways (Wisler et al., 2018). The G protein-independent pathways mainly involve arrestins, which were previously believed to be involved in agonist-triggered receptor desensitization and endocytosis [61,99,100]. Unlike conventional balanced agonists that stimulate both G protein and arrestin pathways, newly synthesized ligands have been reported to selectively act on one of the two pathways [101][102][103]. It is speculated that biased agonists stabilize GPCRs in a specific conformation [104,105].
There has been great interest in the development of biased ligands because they are proposed to be superior to traditional balanced agonists in terms of efficacy and adverse effects in the treatment of certain disorders [106,107]. Usually, biased ligands have been designed based on data obtained through structure-activity analyses of currently available ligands [35,108]. Biased signaling can occur not only in response to tailored ligands but also through genetically modified GPCRs, which can provide new insights for designing novel biased ligands. Several biased D 2 Rs have been reported, and some of them are discussed below (Figure 1, Table 2). A135R/M140D G protein < Arrestin [111] Structure-activity analyses of D 2 R agonists revealed that a hydrophobic pocket at the interface of the second extracellular loop and fifth transmembrane segment of D 2 R is involved in biased signaling [108]. One of the point mutations examined in D 2 R, specifically the mutation at F189, which is equivalent to F 5.38 according to Ballesteros-Weinstein GPCR numbering system [112], showed a preference for G protein coupling. When tested for G protein signaling, the DA potency for F189A-D 2 R decreased about 500 times compared with that for WT-D 2 R; however, the interaction with arrestin2 was almost completely abolished, resulting in impaired arrestin recruitment while maintaining G protein signaling to some extent.
The Asp-Arg-Tyr (DRY) motif is located at the beginning of the second intracellular loop. The Arg residue within this motif plays vital roles in G protein coupling, receptor phosphorylation, and arrestin recruitment by forming intramolecular interactions [113,114]. To create a biased D 2 R, a point mutation was introduced at R132 in the DRY motif (R 3.50 ), so that both G protein coupling and arrestin translocation were suppressed. An additional mutation was then introduced at L 3.41 (L123W), which determines the thermostability [115], to increase the potency for DA [109]. The resulting D 2 R mutant (R132L/L123W) produced displayed biased signaling through G proteins. This is because the L123W mutation, in addition to R132L, significantly restored the signaling ability of R132L-D 2 R through G proteins while having a lesser impact on its ability to recruit arrestins.
Research also revealed that modifying four amino acid residues within the N-terminal region of the third intracellular loop ( 212 IYIV 215 ) resulted in a signaling-biased receptor. The receptor maintained its ability to bind ligands and couple with and activate G proteins but demonstrated impaired ability to mediate arrestin3 translocation to the plasma membrane [110].
Using the evolutionary trace (ET) method [116], another study aimed to develop a biased D 2 R. Peterson et al. successfully identified D 2 R mutants that exhibited selective signaling through either G protein or arrestins [111,117]. The team discovered two D 2 R mutants with dual amino acid substitutions, L125N/Y133L and A135R/M140D, which showed a preference for G protein-mediated and arrestin-dependent pathways, respectively. These modified D 2 R mutants were then used for in vivo studies to investigate the roles of the arrestin pathway in amphetamine-induced locomotion potentiation [117].
Identifying the sites that determine biased signaling in D 3 R might not be straightforward, as agonist treatment rarely induces arrestin recruitment [30]. To measure arrestindependent signaling, the carboxyl terminus of the vasopressin type 2 receptor is typically attached to the receptor for stable arrestin translocation [118]. In the case of D 2 R, this approach leads to a complementary interaction with the DRY motif, whereas in the case of D 3 R it causes a shift in the arrestin translocation pattern from type A to type B [113,114]. In type A GPCRs, arrestin is released from the receptor as endocytosis progresses, whereas in type B GPCRs arrestin remains bound to the receptor and accompanies it in the endocytic vesicle [65].
A previous study has demonstrated that a G protein-biased ligand interacts with Asp110 on transmembrane 3 and His349 on transmembrane 6, but it is unclear whether these amino acids are the decisive factors for the biased signaling of D 3 R [119]. Although the involvement of the arrestin-dependent pathway has been suggested in D 3 R signaling, identifying the amino acid residues responsible for biased signaling and creating new biased agonists will require both conceptual and methodological advancements [120].
There are several potential reasons for the discrepancy observed in the effects of arrestin depletion on ERK activation. One possible explanation is that the G protein and arrestin pathways are complementary for ERK activation, meaning that the remaining pathway can compensate for the absence of the other. Another possibility is the dual role of arrestins in mediating both their own pathway, such as ERK activation, and desensitizing the G protein pathway. Therefore, removing endogenous arrestin may block arrestindependent ERK activation but enhance G protein-dependent ERK activation. Another consideration is the technical limitations of siRNA-or shRNA-mediated knockdown, which may not completely remove the target protein. If a small amount of arrestin is sufficient to support the arrestin signaling pathway, ERK activation may not be affected even with significant arrestin depletion.
The identification of the factors that determine ligand binding and understanding the 3D structure of D 2 R and D 3 R can greatly aid in developing selective or biased ligands for these receptors. Specifically, the structural analyses of D 2 R or D 3 R in complex with G protein or arrestin can have significant implications for developing biased ligands. Recent research has revealed the crystal structures of D 2 R bound to risperidone and D 3 R bound to eticlopride [124,125], as well as the 3D structures of D 2 R or D 3 R in complex with Gi through cryo-electron microscopy analysis [126,127]. Additional studies linking computational techniques with experimental results have identified factors involved in ligand binding, further advancing structure-based drug design [128,129]. Together, these findings enhance our understanding of the structural and functional properties of D 2 R and D 3 R and underscore the potential of structure-based drug design to develop novel therapeutics targeting these receptors.

Dimerization of Dopamine D 2 and D 3 Receptors
GPCRs have the ability to form dimers or oligomers with similar or different types of GPCRs. Coactivation of D 1 R and D 2 R, which couple to Gs and Gi/o, respectively, has been demonstrated to activate Gq protein, leading to intracellular calcium release through the activation of phospholipase C and IP3 [130][131][132].
The D 1 R and D 2 R interact through adjacent glutamic acids in the carboxyl tail of D 1 R (404E/405E) and two adjacent arginine residues in the third intracellular loop of D 2 R (245R/246R) [133]. Previously, the third intracellular loop of D 2 R, specifically the region 217 RRRRKR 222 , was identified as a potential site for heteromer interaction, forming heteromers with D 1 R [134], 5-HT 2A R [135], or adenosine A2A receptors [136].
Studies have demonstrated both direct and functional interactions between D 1 R and D 3 R. Depending on the cell type or signaling, the two receptor subtypes may affect neurons in either a synergistic or opposing manner [137][138][139]. Furthermore, D 3 R stimulation has been shown to enhance D 1 R agonist affinity, which potentiates D 1 R-mediated behavioral effects [140,141].
Dimerization between these two receptors has also been observed to alter endocytic properties. Heterodimerization with D 3 R, for instance, abolishes agonist-induced endocytosis of D 1 R but allows the endocytosis of the D 1 R/D 3 R complex in response to simultaneous agonistic stimulation of both receptors [141]. It is noteworthy that G proteinindependent signaling can occur with stimulation of D 1 R/D 3 R heteromers in the nucleus accumbens [142].
There have been several studies indicating the functional interactions between D 3 R and nAChR in vivo. For instance, D 3 R and α4β2 nAChR form heteromers in dopaminergic neurons that are crucial for the neurotrophic effects of nicotine [143]. Furthermore, the development of nicotine sensitization is accompanied by an increase in the expression of D 3 R [144,145], and it has been suggested that D 3 R ligands can be used to treat tobacco dependence [146]. A recent in vitro study showed that nicotinic stimulation of α4β2 nAChR leads to Src activation in an arrestin2-and 14-3-3η-dependent manner. The activated Src phosphorylates the tyrosine residue(s) on Syk molecules that then interact with phospholipase Cγ1 to trigger the translocation of PKCβII to the cell membrane by elevating cellular diacylglycerol levels [147].

Posttranslational Modifications
Posttranslational modifications (PTMs) are chemical modifications of amino acid side chains that occur after protein synthesis is complete. These modifications usually take place in the endoplasmic reticulum (ER) and the Golgi apparatus [148]. In the case of GPCRs, PTMs can occur in all regions of the receptor except for the transmembrane domains. These modifications are important for regulating receptor folding, maturation, trafficking, and signaling, which increases functional diversity and fine-tunes signaling pathways [149,150]. The most extensively characterized PTMs include glycosylation, phosphorylation, ubiquitination, and palmitoylation [150,151]. The sites and functional roles of PTMs of D 2 R and D 3 R are summarized in Table 1 and Figure 3.

Glycosylation
Glycosylation takes place prevalently at the N-termini or extracellular loops of GPCRs. N-glycosylation, the major form of glycosylation in GPCRs, links a sugar molecule to the nitrogen of the Asn (N) residue in the consensus motif N-X-S/T (X = P) [152,153]. Detection of N-glycosylation in target proteins has traditionally involved enzymatic cleavage with peptide N-glycosidase F (PNGase F) or endoglycosidase H (Endo H) along with sitedirected mutagenesis in the consensus sequence.
The process of glycosylation commences with dolichol, which is a polymerized isoprene molecule. In the ER membrane, dolichol, with its hydroxyl group facing the cytoplasmic side, is converted to dolichol phosphate [154]. Subsequently, N-acetylglucosamine (NAG)-UDP and β-D-mannopyranose are added in succession to dolichol phosphate. Dolichol then flips so that the entire glycosyl moiety faces the ER lumen and additional mannose (M) moieties along with glucose (G) molecules are added to form a large complex of dolichol diphosphate-NAG-NAG-M/Gs, which is now ready for the glycosylation of the target protein. Glycosylation takes place when the NAG-NAG-M/G moiety, detached from dolichol diphosphate, interacts with the nitrogen atom on the side chain of the asparagine residue in the target protein.
Glycosylation is involved in the modulation of various aspects of receptor function, such as maturation, trafficking, ligand binding, and cell signaling, as reported in several studies [155][156][157]. The D 2 R possesses three N-linked glycosylation sites at its N-terminus (N5, N17, N23), whereas the D 3 R has four N-linked glycosylation sites. Two of these sites are located at the N-terminus (N12, N19), whereas the other two are present in the second and third extracellular loops (N97 and N173, respectively).
Surface expression of both D 2 R and D 3 R is commonly affected by glycosylation at the N-terminus, with D 2 R undergoing caveolae endocytosis and D 3 R undergoing clathrin-mediated endocytosis. In the case of D 3 R, glycosylation at the N-terminus is necessary for basal signaling and desensitization, whereas glycosylation at N97 and N173 in the second and third extracellular domains, respectively, is involved in receptor endocytosis [57,58,158].
Posttranslational modifications (PTMs) are chemical modifications of amino acid side chains that occur after protein synthesis is complete. These modifications usually take place in the endoplasmic reticulum (ER) and the Golgi apparatus [148]. In the case of GPCRs, PTMs can occur in all regions of the receptor except for the transmembrane domains. These modifications are important for regulating receptor folding, maturation, trafficking, and signaling, which increases functional diversity and fine-tunes signaling pathways [149,150]. The most extensively characterized PTMs include glycosylation, phosphorylation, ubiquitination, and palmitoylation [150,151]. The sites and functional roles of PTMs of D2R and D3R are summarized in Table 1 and Figure 3.

Phosphorylation
Following agonist binding, GPCRs undergo conformational changes and interact with a specific heterotrimeric G protein, which enhances its GTPase activity, resulting in the separation of the G protein into α and βγ subunits that subsequently modulate downstream signaling pathways [84]. Concurrently, regulatory processes are initiated following receptor activation, with receptor phosphorylation playing a crucial role [159,160].
The intracellular regions of GPCRs contain numerous phosphorylation sites that are predominantly targeted by two classes of serine/threonine kinases: GPCR kinases (GRKs) and second-messenger-dependent protein kinases, such as protein kinase A (PKA) and protein kinase C (PKC) [40,42,43,161]. Techniques such as mass spectrometry, phosphorspecific immunoblotting, and site-directed mutagenesis of the consensus sites in the target proteins have been utilized to identify GPCR phosphorylation sites [162,163].
Protein kinases, including PKA and PKC, recognize specific substrate sequences independent of receptor activation [49,164]. Furthermore, the activation of these kinases is reliant on second messengers that can spread throughout the cell [42,43,165]. As a result, it is likely that both occupied and unoccupied neighboring receptors are phosphorylated by these kinases.
GRKs have a preference for phosphorylating GPCRs when the receptors are in an active state, which means that they are occupied by agonists [166]. This selectivity is due to the dynamic interactions between GRKs and GPCRs, as both proteins undergo conformational changes upon agonist stimulation of GPCRs [167,168]. The interaction and activation of GRKs are mainly mediated by the overall topologic structure of the activated receptor rather than the amino acid sequence surrounding the phosphorylation site [167,[169][170][171]. Activated receptors also enhance the enzymatic activity of GRKs, making them both a substrate and activator of GRKs [167,170,172].
The intracellular regions of D 2 R and D 3 R contain a number of serine and threonine residues. The short isoform of D 2 R (D 2S R) has three threonine residues located in the first intracellular loop, two serine and two threonine residues in the second loop, and a total of thirteen serine and ten threonine residues in the third loop. On the other hand, D 3 R includes three threonine residues in the first intracellular loop, two serine and two threonine residues in the second loop, ten serine and seven threonine residues in the third loop, and an additional serine residue in the carboxyl terminal region.
Phosphorylation of D 2 R is induced by agonists at various serine and threonine residues, including S256, S257, T258, S259, T264, S282, S288, and S292 [53]. PKC-mediated phosphorylation of D 2S R, which is inferred from D 2L R results, mainly occurs at residues 228, 229, and 325, with additional contributions of threonine residues at 322 and 324 [56]. D 3 R is phosphorylated at S229 and S259 upon PMA stimulation, and these phosphorylations play a significant role in regulating desensitization and intracellular trafficking of the receptor [43].

Ubiquitination
Ubiquitin is a polypeptide consisting of 76 amino acids that is found in eukaryotes. The process of ubiquitylation involves the formation of an isopeptide bond that connects an internal lysine residue of the target protein with the C-terminal glycine (glycine 76) of ubiquitin (Hershko, 2005). Ubiquitination requires three sets of enzymes: Ub-activating (E1), Ub-conjugating (E2), and Ub-ligating enzymes (E3) [173]. E1 and E2 sequentially form thioester bonds with ubiquitin, which involves condensation and conjugation between the thiol groups (usually of cysteine) of E1 and E2 and the carboxyl group of Ub (glycine 76). E2 then associates with an E3, through which ubiquitin is transferred from E2 to E3. Finally, E3 catalyzes the conjugation between the amino group of a lysine residue on the target substrate and the carboxyl group of Ub via an isopeptide bond. E3 plays a crucial role in ubiquitination by conferring substrate specificity [174].
There are three different ways in which substrate proteins can be attached to Ub, namely monoubiquitination, multi-monoubiquitination, and polyubiquitination. To form polyubiquitin chains, eight residues in ubiquitin can be employed, including K6, K11, K27, K29, K33, K48, K63, and M1 (linear). Depending on the type of linkage, the modified proteins are directed towards diverse cellular outcomes [175,176].
Although PMA treatment has been found to induce ubiquitination and lysosomal degradation of D 3 R [59], there is no evidence of agonist-induced ubiquitination of D 2 R and D 3 R.

Palmitoylation
The initial step of palmitoylation involves creating a thioester linkage between the carboxyl group of palmitic acid (hexadecanoic acid, C16) and the thiol group of coenzyme A (CoA). S-palmitoylation, which is more frequent than N-palmitoylation, involves the thiol group of a serine residue in the target protein attacking the carbonyl group of the thioester linkage formed in the first step [180]. This results in the formation of a fresh thioester linkage between the target protein and palmitic acid. Meanwhile, CoA is released and can be reused. The precise consensus sequences surrounding the cysteinyl residues that undergo palmitoylation have yet to be determined [181].
When methionine, the first amino acid of a polypeptide, is removed, N-palmitoylation occurs when the first amino acid is either cysteine, glycine, or lysine [182]. Similar to S-palmitoylation, the free amino group of these amino acids in the target protein reacts with the carbonyl group of the thioester linkage established in the first step. As a result, a new amide linkage is formed between the target protein and palmitic acid.
S-palmitoylation can occur spontaneously or can be catalyzed by multi-span transmembrane integral proteins known as protein acyl transferases (PATs). PATs possess zinc-finger and aspartate-histidine-histidine-cysteine (zDHHC) domains [183]. The intracellular DHHC motif plays a crucial role in the S-acylating activity of PATs, allowing protein palmitoylation to occur at the cytoplasmic face of membranes in the secretory pathway, including the ER and Golgi apparatus as well as the plasma membrane.
Palmitoylation is a reversible process, and depalmitoylation is mediated by protein thioesterases or depalmitoylases. Currently, three classes of depalmitoylases have been identified; acyl-protein thioesterases (APTs), α/β hydrolase domain-containing 17 proteins (ABHD17s), and palmitoyl-protein thioesterases (PPTs) [184]. APTs located in the cytosol are responsible for depalmitoylating targets of Gα and Ras [185,186]. In contrast, PPTs found in the lysosome play a role in regulating target protein degradation [187].
During the biosynthesis of GPCRs, palmitoylation often occurs at their C-termini and sometimes at intracellular loops. This modification is usually a basal process but can also be induced by agonist stimulation in some cases [188]. Palmitoylation is essential for various functions of GPCRs, including trafficking, cell surface localization, dimerization, and signaling [188,189]. D 2 R and D 3 R share a conserved cysteine residue at their carboxyl termini and have a highly similar sequence for the last 10 amino acid residues except for the second-to-last residue (histidine in D 2 R and serine in D 3 R). Despite this similarity, D 3 R is constitutively palmitoylated at C400, whereas D 2S R is not palmitoylated in HEK-293 cells [31]. Palmitoylation is a critical modification for proper localization of D 3 R to the cell surface, as well as for PKC-mediated endocytosis, agonist affinity, and desensitization of the receptor [31].
Studies have shown that the palmitoylation of D 2 R may vary depending on the expression system. D 2S R was not palmitoylated in HEK-293 cells [31], whereas other studies have reported that both D 2S R and D 2L R are palmitoylated when expressed in the baculovirus/Sf9 cell system [190,191]. Moreover, conflicting results have been reported on the palmitoylation site of D 2 R, with one study suggesting C443 as the palmitoylation site [192] and another study reporting no palmitoylation at this site [31]. These discrepancies may have resulted from differences in experimental strategies and conditions, such as differences in the concentration and duration of 2-bromopalmitate treatment and the use of deletion or point mutation strategies to disrupt palmitoylation.
Taking into account the variations in experimental conditions, it is probable that only D 3 R (not D 2 R) undergoes palmitoylation at the C-terminus. Furthermore, although palmitoylation may occur in D 2 R, it appears that D 3 R is more heavily palmitoylated. Additionally, palmitoylation seems to have a more crucial impact on the regulation of D 3 R's functions than those of D 2 R.
Although the consensus sequence for palmitoylation is not yet fully established, a comparison between D 2 R and D 3 R yielded some insight. Specifically, the presence of a Ser residue preceding the Cys residue appears to be crucial for palmitoylation of D 3 R. For instance, when the C-terminal "HC" sequence of D 2 R, which is typically not palmitoylated, is modified to "SC" as in D 3 R, D 2 R becomes palmitoylated, and the reverse is also true [31].

Conclusions
Establishing a correlation between the in vivo roles of certain GPCRs and their biochemical characteristics in vitro is challenging. Both D 2 R and D 3 R are involved in regulating locomotion, reward, and addiction, but D 3 Rs appear to have a more specific role in regulating emotional and cognitive processes that involve behavioral sensitization [37]. Evidence suggests that behavioral sensitization is closely related to the plasticity of limbic neurochemical systems [193,194], which are also implicated in a range of psychiatric and substance use disorders.
The roles of D 3 R in the development of behavioral sensitization have been proposed based on various observations [195]. For instance, D 3 R antagonists can block behavioral sensitization, and both transcriptional and translational activities involved in D 3 R expression are reduced following behavioral sensitization [144,196].
Previous reports have demonstrated that D 3 R has inhibitory effects on the synthesis and release of DA in the striatum [197][198][199]. D 3 R also appears to mediate the inhibition of neuronal firing [200,201], and selective D 3 R agonists can lower extracellular levels of DA, leading to the inhibition of locomotion [202,203].
Although the signaling pathways of D 2 R and D 3 R largely overlap, the intensity of D 2 R signaling is stronger than that of D 3 R, except in the inhibition of dopamine release [22,204,205]. Therefore, the development of the desensitization of D 3 R after repeated neuronal stimulation may result in the loss of a D 3 R-mediated 'brake' on dopamine release, leading to increases in locomotion and behavioral sensitization [195,206].
Posttranslational modifications play critical roles in maintaining the desensitizing properties of D 2 R and D 3 R. Among the regulatory processes, such as agonist-induced receptor phosphorylation, arrestin translocation, and receptor endocytosis, only those related to intracellular trafficking occur more strongly with D 2 R than with D 3 R. Considering that receptor endocytosis is necessary for the resensitization of desensitized receptors, D 3 R is likely more susceptible to desensitization than D 2 R. In contrast, glycosylation and palmitoylation of D 3 R are necessary for maintaining its capacity to undergo desensitization.
Overall, D 3 R is predominantly expressed in limbic brain regions that are associated with various psychiatric and substance use disorders, where the plasticity of limbic neurochemical systems plays a significant role. In vitro studies have demonstrated that D 3 R undergoes dynamic desensitization that necessitates specific PTMs. As D 3 R appears to function as a negative regulator for various behavioral aspects that are linked to the dopaminergic nervous system, it is justifiable to propose a theoretical model of behavioral sensitization that views it as the desensitization of inhibitory systems, where D 3 R has a vital function.