Approach to the specificity and selectivity between D2 and D3 receptors by mutagenesis and binding experiments part I: Expression and characterization of D2 and D3 receptor mutants

Abstract

D3/D2 sub‐specificity is a complex problem to solve. Indeed, in the absence of easy structural biology of the G‐protein coupled receptors, and despite key progresses in this area, the systematic knowledge of the ligand/receptor relationship is difficult to obtain. Due to these structural biology limitations concerning membrane proteins, we favored the use of directed mutagenesis to document a rational towards the discovery of markedly specific D3 ligands over D2 ligands together with basic binding experiments. Using our methodology of stable expression of receptors in HEK cells, we constructed the gene encoding for 24 mutants and 4 chimeras of either D2 or D3 receptors and expressed them stably. Those cell lines, expressing a single copy of one receptor mutant each, were stably constructed, selected, amplified and the membranes from them were prepared. Binding data at those receptors were obtained using standard binding conditions for D2 and D3 dopamine receptors. We generated 26 new molecules derived from D2 or D3 ligands. Using 8 reference compounds and those 26 molecules, we characterized their binding at those mutants and chimeras, exemplifying an approach to better understand the difference at the molecular level of the D2 and D3 receptors. Although all the individual results are presented and could be used for minute analyses, the present report does not discuss the differences between D2 and D3 data. It simply shows the feasibility of the approach and its potential.

1. INTRODUCTION

Dopaminergic neurotransmission is mediated and signaled by five dopamine (DA) receptors (D1–D5). These receptors are G‐protein‐coupled receptors (GPCRs). The main DA receptors present in the brain are those that are D1‐like (D1 and D5 receptor subtypes) and those that are D2‐like (D2, D3, and D4 receptor subtypes). ¹ , ² Their expression patterns are widely distributed throughout the body.

D2 and D3 receptors, have a markedly different distributions in the brain: The D3 receptor is abundant in the nucleus accumbens, substantia nigra and the Calleja island of mesolimbic region, whereas D2 expression is more particularly found in the striatum of the midbrain. ³ , ⁴ The D2 receptor is also implicated in various diseases such as Parkinson disease, ⁵ schizophrenia ⁶ and depression. ⁷

One of the main goals of molecular pharmacology is to discover and described compounds that are as specific as possible of a given targets. Especially, as in the case of D2 and D3 when the targets are strongly linked to a given pathology. Indeed, Dopamine D3 receptor (D3) is linked to emotions, learning and the control of movements. Thus, these properties make D3 a target of choice for clinical treatments of Parkinson's disease, schizophrenia, and depression. ⁸ , ⁹ It is clear that these particular features fueled to the discovery of many D3 ligands. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ In line with the previous statement, ligands at D3 are often D2 ligands as well. In other words, specific compounds to one or the other of those receptors have been difficult to access. These difficulties can be explained by the fact that D2 and D3 subtypes have a high homology (circa 50%) in their amino acid sequences, and the homology increases to 78% in the helical transmembrane domains. ¹⁷ , ¹⁸ This feature is often shared by isoforms within a common family of GPCRs, such as in melatonin or serotonin ones. Of course, one has to remember that it is in those transmembrane domains that reside the active binding sites.

In previous works, it has been clearly stated that compounds with relative preference for D3 over D2 would display strengthened efficacy against cognitive symptoms. ¹⁹ , ²⁰

We wondered of one of the reasons for the absence of finding specific ligands to D2 over D3 or D3 over D2 was not due to the fact that the molecular tools were not perfectly adapted to such a task. Thus, after having developed an approach using meganuclease cellular surgery, ²¹ we applied this technology to directed mutagenesis. Using this approach led to Several improvements over the standard, transitory techniques used in most of the directed mutagenesis papers; (i) the cell lines were engineered in a way were the region of insertion of the transgene is inert, meaning that tall the transgenes for the different cell lines were inserted in an identical region, leading to an absence of activation of off‐targets proteins (ref), (ii) the cell lines are stable, a feature which translates into the possibility to run independent experiments in different days, without having to transfect again the transgene into a new set of cells. This has been checked and validated previously, using a series of transgenes (from GPCRs to enzymes) (ref), and provided in past experiments–on other targets–a robust series of results.

Indeed, once the gene for a mutant enters the process, we obtain a stable cell line that can be used with a superior consistency, compared to the transient approach. ²¹ , ²² Furthermore, this technique leads to the obtention of cellular clones expressing a single copy of the transfected gene, to the opposite of standard obtained clones that could expressed 4–10 copies, leading to difficulties to actually compare Bmax values between clones.

In the present paper, we approached the specificity characteristics of a large series of diverse compounds at cloned human D3 over D2 subtypes, while using a fairly large collection of precisely designed mutants of those receptors. The approach, combining the integration of the mutated gene in a seemingly neutral region of the cell genome, together with the stable nature of the expressed mutants, led to a larger capacity to screen for the binding displacement due to a large series of molecules. Overall, 24 mutants and 4 chimeras were built and expressed on which 34 molecules were tested. The rationale behind this work was to effectively understand the influence of key amino acids in the D2 and D3 isotype sequences onto the binding of those compounds. Based on these results, we could develop a molecular modelling analysis that will be presented in the part 2 report.

2. EXPERIMENTAL PROCEDURES

2.1. Reagents and ligands

[N‐Methyl‐³H]‐methylspiperone (85.5 Ci/mmol) were purchased from Perkin Elmer (Boston, USA). Methiotepine was purchased from RBI (Biotrend, Köln, Germany), raclopride and (+)‐PD 128907 hydrochloride were purchased from Tocris Bioscience (Bristol, UK), and SB‐258585 dihydrochloride, eticlopride, dopamine and other salts and compounds were purchased from Sigma (St Louis, USA). Methiotepine, raclopride, SB‐258585 dihydrochloride, eticlopride, dopamine and S compounds were diluted in 100% DMSO and (+)‐PD 128907 hydrochloride in distillated water, at a stock concentration of 10 mM and stored at −20°C.

The reference compounds are presented in Figure 1. All new pharmacological compound structures are presented in Figure 2. They belong to three main families Their synthesis will be detailed elsewhere (Rojas, Ortuno, in preparation). In Table 1, a brief description of the various compounds and the initial reports of the reference compound synthesis are given.

FIGURE 1.

Reference compound structures. Eticlopride: the D‐2 antagonist; ²³ S33138: a dopamine D3 receptor antagonist; ¹⁵ PD128907: D3‐receptor agonists; ²⁴ S33084 selective D3 receptor antagonists; ²⁵ Pardoprunox (SLV‐308) is a D2 (pKi = 8.1) and D3 receptor (pKi = 8.6) partial agonist ²⁶ , ²⁷ SB‐742457 is a potent 5‐HT ⁶ antagonist; ²⁸ methyl‐spiperone is a classical dopamine receptor ligand; ²⁹ GDN337 is another 5‐HT6 potent ligand. ³⁰

FIGURE 2.

Structures of the D2/D3 compounds

TABLE 1.

Description of the three families of compounds and a series of references used in the study

References	SRE1118 a	W1109 b	SRE1114 c
Dopamine	S54935	S53078	S54436
PD128907	S57095	S55878	S57721
GDN337	S57860	S55920	S58399
Intepirdine	S58399	S56087
Pardoprunox	S59188	S56203
S33084	S59422	S56414
S33138	S59913	S57223
Eticlopride	S60568	S58806
	S60928	S58857
	S60929	S58976
	S61350	S60123
		S62653

Note: The color‐code refers to the dots in Figures 7, 8 and 10 The structures of the reference compounds are shown in Figure 1; those of the other compounds are depicted in Figure 2.

^a The series of compounds SRE1118 is derived from this 7‐bicyclo[4.2.0]octa‐1,(6)2,4‐trienylmethanamine.

^b The W1109 series of compounds is derived from this 2,3,4,7‐tetrahydro‐1H‐pyrrolo[2,3‐h]isoquinoline.

^c The SRE 1114 series is derived from 3H‐1,3‐benzoxazol‐2‐one.

2.2. Molecular cloning of wild‐type receptors

The genes coding for the human DRD2 and the human DRD3 were synthesized by Gencust (Boynes, France) and were cloned into the pIM LP1A EF1a vector (Cellectis, Paris). All constructs were fully sequence‐verified by dideoxy chain termination methods. The sequence of the human DRD2 and the human DRD3 used in this study corresponds respectively to GenBank accession n° NM_000795 and NM_000796. All plasmids DNA used for mutagenesis and transfections was prepared using resin‐based mega‐prep purifications following the manufacturer's protocol (Macherey‐Nagel).

2.3. Generation of site‐directed mutants

Mutant receptors were created using the QuickChange XLII site‐directed mutagenesis kit (Stratagene). Oligonucleotides of approximately 30–33 bases in length were designed with the desired base substitutions (see Table S1, in supplementary materials). The oligonucleotides were annealed to the heat‐denatured double‐stranded plasmid containing the gene for the receptor and a high‐fidelity DNA polymerase was used to extend the oligonucleotides. The parental DNA was then digested by a methylation‐specific nuclease, resulting in a population of plasmids enriched with the mutated receptor gene. The plasmids were then purified from transformed bacteria One Shot® MAX Efficiency™DH10B™T1 Phage Resistant cells (Invitrogen) using the NucleoBond®PC 2000 EF (Macherey‐Nagel) and mutations were confirmed with the 3730XL DNA analyzer. All mutants were cloned into the same pIM LP1A EF1a vector.

2.4. Construction of the synthetic DNA encoding chimeric D2‐D3 receptors

Further to strict mutants, as shown in Table 2, we constructed a series of chimeras between DRD2 and DRD3, according to the sequences listed in the Table 3.

TABLE 2.

schematic presentation of the various constructions used in the present work

Mutation	Localisation	D2L	D3	Ballesteros‐Weinstein
Alanine	TM5	S193A	S192A	5.42
	TM5	S194A	S193A	5.43
	TM5	S197A	S196A	5.46
	TM6	H393A	H349A	6.55
	TM7	Y408A	Y365A	7.35
	TM7	T412A	T369A	7.39
D2‐ > D3 and D3‐ > D2	TM4	I166V	V164I	4.56
	TM7	F411T	T368F	7.38
	ECL2	E181V	V180E	45.49?
	ECL2	I183S	S182I	45.51
	ECL1	E99V	V95E	23.49
	TM1	L41Y	Y36L	1.39

TABLE 3.

Protein sequence of the chimers between D2 and D3 dopamine receptors constructed in the present work

Mut 32 chimera Nter hDRD3 _hDRD2:    MASLSQLSGHLNYTCGAENSTGASQA PHYNYYATLLTLLIAVIVFGNVLVCMAVSREKALQTTTNYLIVSLAVADLLVATLVMPWVVYLEVVGEWKFSRIHCDIFVTLDVMMCTASILNLCAISIDRYTAVAMPMLYNTRYSSKRRVTVMISIVWVLSFTISCPLLFGLNNADQNECIIANPAFVVYSSIVSFYVPFIVTLLVYIKIYIVLRRRRKRVNTKRSSRAFRAHLRAPLKGNCTHPEDMKLCTVIMKSNGSFPVNRRRVEAARRAQELEMEMLSSTSPPERTRYSPIPPSHHQLTLPDPSHHGLHSTPDSPAKPEKNGHAKDHPKIAKIFEIQTMPNGKTRTSLKTMSRRKLSQQKEKKATQMLAIVLGVFIICWLPFFITHILNIHCDCNIPPVLYSAFTWLGYVNSAVNPIIYTTFNIEFRKAFLKILHC*

Mut 33 chimera Nter hDRD2 _hDRD3:    MDPLNLSWYDDDLERQNWSRPFNGSDGKADR RPHAYYALSYCALILAIVFGNGLVCMAVLKERALQTTTNYLVVSLAVADLLVATLVMPWVVYLEVTGGVWNFSRICCDVFVTLDVMMCTASILNLCAISIDRYTAVVMPVHYQHGTGQSSCRRVALMITAVWVLAFAVSCPLLFGFNTTGDPTVCSISNPDFVIYSSVVSFYLPFGVTVLVYARIYVVLKQRRRKRILTRQNSQCNSVRPGFPQQTLSPDPAHLELKRYYSICQDTALGGPGFQERGGELKREEKTRNSLSPTIAPKLSLEVRKLSNGRLSTSLKLGPLQPRGVPLREKKATQMVAIVLGAFIVCWLPFFLTHVLNTHCQTCHVSPELYSATTWLGYVNSALNPVIYTTFNIEFRKAFLKILSC*

Mut 34 chimera TM1 hDRD2_hDRD3:    MASLSQLSGHLNYTCGAENSTGASQA PHYNYYATLLTLLIAVIVFGNVLVCMAVSR ERALQTTTNYLVVSLAVADLLVATLVMPWVVYLEVTGGVWNFSRICCDVFVTLDVMMCTASILNLCAISIDRYTAVVMPVHYQHGTGQSSCRRVALMITAVWVLAFAVSCPLLFGFNTTGDPTVCSISNPDFVIYSSVVSFYLPFGVTVLVYARIYVVLKQRRRKRILTRQNSQCNSVRPGFPQQTLSPDPAHLELKRYYSICQDTALGGPGFQERGGELKREEKTRNSLSPTIAPKLSLEVRKLSNGRLSTSLKLGPLQPRGVPLREKKATQMVAIVLGAFIVCWLPFFLTHVLNTHCQTCHVSPELYSATTWLGYVNSALNPVIYTTFNIEFRKAFLKILSC*

Mut 35 chimera TM1 hDRD3_hDRD2:    MDPLNLSWYDDDLERQNWSRPFNGSDGKADRR PHAYYALSYCALILAIVFGNGLVCMAVLK EKALQTTTNYLIVSLAVADLLVATLVMPWVVYLEVVGEWKFSRIHCDIFVTLDVMMCTASILNLCAISIDRYTAVAMPMLYNTRYSSKRRVTVMISIVWVLSFTISCPLLFGLNNADQNECIIANPAFVVYSSIVSFYVPFIVTLLVYIKIYIVLRRRRKRVNTKRSSRAFRAHLRAPLKGNCTHPEDMKLCTVIMKSNGSFPVNRRRVEAARRAQELEMEMLSSTSPPERTRYSPIPPSHHQLTLPDPSHHGLHSTPDSPAKPEKNGHAKDHPKIAKIFEIQTMPNGKTRTSLKTMSRRKLSQQKEKKATQMLAIVLGVFIICWLPFFITHILNIHCDCNIPPVLYSAFTWLGYVNSAVNPIIYTTFNIEFRKAFLKILHC*

All chimeras were obtained by complete synthesis of the corresponding genes (Gencust). A sketch of the chimera constructions is shown on Figures 3 and 4.

FIGURE 3.

Summary of the D2 (a) and D3 (b) point mutations produced in the present work. The individual mutations are colored, in order to better see them on the figures. The same color‐code was used in Figures 5, 6, 7 and 8

FIGURE 4.

Schematic presentation of D2–D3 chimeras produced in the present work

2.5. Transfection, selection, and validation of the transfected cells

All targeted integration will be performed in CBR cGPS HEK293 cell line. These cells were cultivated inCell line cGPS HEK293: DMEM Glutamax™ (Gibco) with FCS 10% (Eurobio) containing Penicillin/Streptomycin (0.5%) and Hygromycin B (100 μg/ml). Final conditions of culture will be the same for established cellular clones except that Hygromycine B is replaced by 0.4 mg/ml of G418 and 1 μg/ml of puromycin. Inbrief the protocols were as follows: One day prior to transfection, the cGPS HEK‐293 cells are seeded in 10 cm tissue culture dishes (1 × 10⁶ cells per dish) in complete medium without hygromycin. On transfection day (D), 1 μg of the LP1A Integration Matrix (pIM.LP1A.EF1al) containing your gene of interest, or 1 μg of the LP1A Integration Matrix EF1al LacZ (pIM.LP1A.EF1a.LacZ), and 1 μg of I‐CreI Meganuclease Plasmid were diluted in 300 μl of DMEM GlutaMAX™ medium only (serum‐free and antibiotic‐free). The Lipofectamine™ 2000 reagent (10 μl) was diluted in in 290 μl of DMEM GlutaMAX™ medium only (serum‐free and antibiotic‐free). Both mixes were incubated separately for 5 min at room temperature before adding the diluted DNA to the diluted Lipofectamine™ 2000 reagent. The mixture was homogenized by gentle tube inversion. After 20 min of incubation at room temperature, the mixture was dispensed over plated cells. Dishes were then incubated in a 37°C, 5% CO₂ humidified incubator. After 24 hr, the medium was replaced with 10 ml of complete medium supplemented with 0.4 mg/ml of G418. The culture medium was changed every third day thereafter. Starting after a week, and then, every other day, the medium was replaced with fresh complete medium supplemented with G418 (0.4 mg/ml) and puromycin (1 μg/ml). Two weeks after transfection, single colony clones of sufficient size were picked and seeded in a 96‐well plate containing fresh complete medium supplemented with G418 (0.4 mg/ml) and puromycin (1 μg/ml). When confluence was reached, the cells were trypsinized and transferred to, a 6‐well plate, in fresh complete medium supplemented with G418 (0.4 mg/ml) and puromycin (1 μg/ml). Four further days were needed to amplify each single cell clone in a 6‐well plate format, and about 4 extra days were needed to reach confluence after transfer to 10 cm tissue dishes. In order to control that those transfected cells included the correct targeted integration genes, control experiments were performed by Southern blot analyses. In brief, Genomic DNA (gDNA) was purified from 10⁷ cells. Five to ten 10 μg of gDNA were digested with a 10‐fold excess of restriction enzyme by overnight incubation. For a full characterization, we checked both sides of the targeted integration by choosing restriction enzymes that cleaved once in the cGPS HEK‐293 chromosomal locus and once in the integrated gene. For example, the targeted integration of the GOI is checked in the 3′ side with a NotI digestion. The probe is located within the Neo resistance gene. A targeted integration will bring the NotI site from the Integration matrix in the vicinity of the NotI site downstream of the Neo gene. Upon NotI digestion, a 1.9 kb DNA fragment is generated and identified with the Neo probe. To the contrary, when digested with NotI, the untargeted LP1A Landing Pad gives a band higher than 10 kb. The same approach can be used for the 5′ side of the targeted integration. Furthermore, 24 clones per integration matrices were screened by PCR in order to eliminate multiple targeted events. This PCR screen is based on the amplification of sequences located on the backbone of the integration matrix (Ampicillin resistance gene and pUC Origin of Replication ORI).

2.6. Amplification and cell membrane pellet preparation

One vial of each clone was used to cell membrane preparation. Briefly, cells were amplified in complete culture medium in order to reach 1 × 10⁹ cells. Cells were then rinsed twice with PBS without Ca²⁺/Mg²⁺. Cells were treated with PBS containing 5 mM of EDTA and incubated 5 to 10 min at 37°C until complete cell detachment. The cell suspensions were washed thoroughly with PBS and cell pellets were obtained by gentle centrifugation 10 min at 1500 rpm. The final pellets were suspended in 50 ml of PBS without Ca²⁺/Mg²⁺, counted and finally pelleted to obtain a dry material that was used to obtain the membrane suspension.

HEK‐cGPS cell lines stably expressing D2 and D3 wild‐type and mutated receptors, were grown at confluence and harvested in cold PBS buffer (GIBCO, Invitrogen, France) containing 5 mM EDTA and centrifuged at 1000 × g for 20 min (4°C). The resulting pellet was suspended in HEPES 20 mM‐NaCl 150 mM, pH 7.4, and homogenized using Kinematica polytron. The homogenate was then centrifuged (20,000 x g, 30 min, 4°C), and the resulting pellet was suspended in HEPES 20 mM‐NaCl 150 mM pH 7.4, cold buffer. Determination of protein content was performed according to Bradford ³¹ using the Biorad kit (Bio‐Rad SA, Ivry‐sur‐Seine, France). Aliquots of membrane preparations were stored at −80°C until use.

2.7. Radioactive binding assay

First, membrane preparations were characterized in saturation assay to determine Kd and receptor density (Bmax), in at least two independent experiments. Membranes were incubated for 45 min at 30°C in incubation buffer (TRIS Base 50 mM, NaCl 120 mM, MgCl2 5 mM, KCl 5 mM, CaCl2 2 mM, BSA 0.2%, pH 7.4) in a final volume of 250 μl containing various concentrations of radioligand, from 10 pM to 2 nM for [N‐Methyl‐³H]‐methylspiperone. ³² Non‐specific binding was defined with 10 μM raclopride. Membranes were incubated for 45 min at 30°C, time to reach the equilibrium of the mass‐action law. Reaction was stopped by rapid filtration through GF/B unifilters PEI 0.3% treated, followed by three successive washes with ice‐cold buffer (TRIS 50 mM pH 7.4). Data were analyzed by using the PRISM software (GraphPad Software Inc., San Diego, CA). For saturation assay, the density of binding sites Bmax and the dissociation constant of the radioligand (Kd) values were calculated according to the method of Scatchard. ³³ , ³⁴

Second, following the same protocol as saturation assay, pharmacological profile of each mutant was performed generally in three independent experiments (in duplicate measurements), using a set of 34 compounds, tested from 10⁻⁵ to 10⁻¹⁰ M (6 concentrations). Membranes of D2‐R and D3‐R were used at the final concentration of 51 fmol/mg and 17.5 fmol/mg of proteins respectively. [N‐Methyl‐3H]‐methylspiperone was used at the final concentration of 20 nM. Non‐specific binding was defined with 10 μM raclopride. Reaction was stopped by rapid filtration through GF/B unifilters PEI 0.3% treated, followed by three successive washes with ice‐cold Tris 50 mM buffer. Inhibition constants (Ki) were calculated according to the Cheng‐Prussof equation: Ki = IC₅₀/[1 + (L/Kd)], where IC₅₀ is the Inhibitory concentration 50% and L is the concentration of radioligand. ³⁵ Many individual data are gathered as supplementary information in Tables S2, S3, S4 and S5.

3. RESULTS AND DISCUSSION

3.1. Binding validation

Our first goal was to ensure the quality of the approach by qualifying the binding characteristics. In order to do so, we generated two experimental sets of data obtained independently, by different days, different operators and different sets of membranes derived from clones expressing a single copy of the gene, for most of the compounds tested we found a correlation (r²) between independent experiments of 0.96 for D2 mutants, 0.95 for D3 mutants, and 0.97 for the D2/D3 chimera. This correlation was calculated using the R Squared Calculator online system (www.ncalculator.com) in the simplest possible way, as it is not used to assert the significativity of differences between data, but only to point at the robustness of the data. Individual correlations are presented in Figure S3 (D2) and S4 (D3). This shows that the approach of stable gene expression is superior to transient expression to obtain robust data from large numbers of compounds that cannot be done at once in a single set of experiments, as previously discussed. ¹⁸ More consistent data, using a stably expressed transgene in a neutral region of the cell genome will favor the robustness of the results as previously listed several occasions. ³⁶ The meganuclease‐based approach was also reviewed at several occasions as a preferred approach in expressing proteins for therapeutical purposes (see Datta et al ³⁷ and Noh et al, ³⁸ among others).

3.2. Binding characterization of wild‐type receptors

Wild‐type D2 receptor affinity (pKd) for D2 for methyl‐spiperone was reported to be in the 10.2 range. ²⁹ , ³⁹ We found 10.1. For D3 receptor, we found a pKi of 9.8 as opposed to a pKi of 9.9 reported by Hoare et al. ³⁹ Overall, our results for the reference compounds are similar to those reported in the literature (see the summary on Table 4). A known source of major differences in the Bmax between various clones, is when clones expressed various copies of the gene, as it is often the case after standard transfections. ²¹ Here, the Bmax's were about 75 fmol.mg prot⁻¹ for D3 while in the 1.2 pmol.mg protein⁻¹ range for D2. One should recall that expression levels is often dictated by the sequence of the transgene. Immense differences can be found between two different proteins, even though they are not far from each other sequence‐wise. Again, as previously shown, ²¹ the use of the meganuclease technology ensures the incorporation of a single copy of the transgene.

TABLE 4.

Comparison of the pKi obtained in the present report using some reference compounds

Name	Alias	D2		D3
		Reference	Present report	Reference	Present report
Dopamine		5.8	5.2	7.0	7.1
S33138		7.1	7.3	8.1	8.9
PD128907		6.0	5.9	7.7	8.2
S33084		6.8	7.4	8.7	9.3
JLP223	Pardoprunox	8.1	7.9	8.7	8.4
GDN 337	E‐6801 analogue	<4	<4	<5	<5
Eticlopride		9.2	9.8	8.8	9.6

Note: The reference values (green cells) were extracted from the IUPHAR database for dopamine receptors D2 and D3. GDN337 is a loose analogue of E‐6801, a standard compound for 5HT‐6 receptor studies, used here as a negative control, due to the chemical proximity of this compound with some compounds of the W1109 series.

3.3. Binding characteristics of the mutants

As mentioned before, we produced 11 D2 mutants and 11 D3 mutants, together with the wild‐type D2 and D3 receptors, obtained in the same conditions than the various mutants. The mutations are summarized in Figure 3. In brief, we chose amino acids that were either known to play a key role in the binding of ligands (dopamine or spiperone) of the receptors, or amino acids that were highly different for similar positions in D2 versus D3. Furthermore, the mutations were also chosen on a series of parameters among which, the similarity and/or the differences in positions between D2 and D3 receptors. A series of curves obtained in some mutants with some molecules are presented as examples of the quality of the present data in Figures S1 and S2.

We first checked the affinity all the mutants for spiperone. Figure 5a,b clearly indicate that most of the mutations had a limited impact on the affinity of the ligand to the receptors, with the interesting exceptions of S193A D2 for which the pKd is half a log higher than for the wild‐type (10.7 vs. 10.1) and the S128I D3 for which the pKd decreases from 9.8 to 8.7. Overall, if the mutations do not change drastically the pKd for spiperone of D2 various constructs, mutations at D3 seemed to have a major decreasing impact on most of the mutants. As can be seen no major changes occurred on this parameter. The level of expression is known to have potentially an influence on the characteristics of the binding at GPCRs. Thus, we examined the Bmax values of the mutants. Figure 6a,b summarizes them. Despite a scattering of the mutant Bmax's at D2, they were all well‐expressed, similarly to the wild‐type receptor (Mut4), while almost all the D3 mutants were less expressed than the wild‐type (Mut14).

FIGURE 5.

Characterization of the [3H]methyl‐spiperone binding of all the mutants produced in the present work. The dashed red line is the standard pKd for methylspiperone for the wild‐type receptors named Mut 4 (D2) or Mut 14 (D3). The colors used are identical to the color of the amino acid mutated, as shown in Figures 3, 7 and 8. The individual data are gathered in Table S2

FIGURE 6.

Measurement of the D2 and D3 mutants B_max. The dashed red line is the standard Bmax using methyl‐spiperone as ligand at the wild‐type receptors named Mut 4 (D2) or Mut 14 (D3). The colors used are identical to the color of the amino acid mutated, as shown in Figures 3, 7 and 8. The individual data are gathered in Table S2

3.4. Molecular pharmacology of the D2L mutants

As stated at several occasions in the present report, we have at disposal a large family of stable cell lines expressing mutated D2 receptors, under controlled conditions: that is in a neutral region of the cell genome, limiting the activation of off‐target genes, with a single copy of the gene. Altogether, those features lead to a comparable expression for the various transgenes in the cellular clones, as can be appreciated from the Bmax values. One should point out that with standard transfection/expression techniques, such a constancy is almost impossible to reach, rendering obligatory the multiplications of transfection steps and of the control measurements on wildtype receptor for the sake of internal comparison. Indeed, we could characterize the various mutants with a series of compounds representing three families of compounds (see Figure 2 for structures) as well as some reference compounds (Figure 1). The details of the results are presented in Figure 7. Overall, not a single mutation has a major impact on the binding of the different compounds of the present studies. Indeed, the following mutations have limited influences on the global binding picture of those 34 compounds L41Y, E181V, I183S, Y408A, T412A, S194A, I166V and E99V have a mean pKi similar to the WT. This is an observation that can serve as a basis to understand the way the compounds of the different series adopted pauses onto the receptor, strongly suggesting that those amino acids are not key to the biding area for the methyl‐spriperone binding. To the contrary, the mutation H393A changes considerably the binding affinity of most of the compounds, indicating the key role of this histidine in the binding process. Similarly, the mutation S193A, and to some extent, the S197A one, dramatically influenced the binding, particularly the series based on the 2,3,4,7‐tetrahydro‐1H‐pyrrolo[2,3‐h]‐isoquinoline moiety (W1109), suggesting that this series of compounds interact with these two residues. (See Figure 2 and Table S3 for structures and individual data, respectively)., a similar influence was observed for the binding of 5‐0H‐DPAT. ⁴⁰

FIGURE 7.

D2 receptor mutant binding profiles. Reference compounds are in blue diamonds; SRE1118 series, in magenta; W1109, in green and SRE1114 in orange. See Figures 1 and 2 for structures. Continuous line is the slope 1; dashed line is the limit of +/− 0.5 log. All individual data are gathered in Table S3. To help the reader to focus on a given mutation, each one is presented in a different color for the amino acid and the arrow that should help to be directed to the data associated with the mutation

3.5. Molecular pharmacology of the D3 mutants

The observation on the D3 receptor mutant results is more complex (Figure 8). Considering the mutants of the D3 receptor, particularly V95E, V164I, V180E, Y365A, T369A, their pKi for most of the molecules only shifted slightly, and only a minute observation of those data could help understand the way these mutations changed the binding of all the compounds. For the remaining mutations (Y36L, S182I, S192A, S193A, S196A, H349A) the influence of the amino acid changes was important (particularly for S196A and S192A), but no clear path can be recorded globally at the examination of those data. It would be necessary to analyze in deep those results in order to extract from the data a more practical path. In particular, it would be required to analyze each family of compounds separately to be able to explain the changes in affinity induced by the mutations. This analysis will be found in the accompanying paper. (See Figure 2 and Table S4 for structures and individual data, respectively).

FIGURE 8.

D3 receptor mutant binding profiles. Reference compounds are in blue diamonds; SRE1118 series, in magenta; W1109, in green and SRE1114 in orange. See Figures 1 and 2 for structures. Continuous line is the slope 1; dashed line is the limit of +/− 0.5 log. All individual data are gathered in Table S4. To help the reader to focus on a given mutation, each one is presented in a different color for the amino acid and the arrow that should help to be directed to the data associated with the mutation

3.6. Binding and expression characteristics of the chimeras

As a last step of the present work, we “chimerized” D2 and D3 with portions of the alter ego sequences. We expressed the D3 N‐terminus with the whole sequence of D2–D2 N‐terminus omitted in mut32 (See Figure 4), or vice versa in mut33. We also built an artificial sequence in which only the TM1 of D2 was present in D3 sequence, instead of the native (D3 TM1 mut34) or vice versa (Mut35).

Expression (Bmax) and affinity for the methyl‐spiperone (Kd) are reported in Figure 9.

FIGURE 9.

Expression and binding characteristics of the D2/D3 chimeric mutants

3.7. Molecular pharmacology of the chimeras

Finally, we checked the influence of those changes onto the binding of the 31 compounds used in the present work.

First, we have to plot the data from both wild‐type receptors against each other. As can be seen from Figure 10, there is a slight scattering of the data whenever the binding results were looked at.

FIGURE 10.

Comparative pharmacology of the wild‐type D3 and wild‐type D2 dopamine receptors. Reference compounds are in blue diamonds; SRE1118 series, in magenta; W1109, in green and SRE1114 in orange. See Figures 1 and 2 for structures. Continuous line is the slope 1; dashed line is the limit of +/− 0.5 log. All the individual data are gathered in Tables S3 and S4

Thus, the notion that it exists different features driving the specificities of the WT receptors is obvious from a simple examination of the graph. Furthermore, a tendency seems to emerge whenever the families of compounds are looked at independently to each other (purple dots, green dots, and to a certain extend the orange ones, although this particular series of chemicals is underpopulated). If necessary, this also leads to the hope of finding actual structural determinants that would permit to enhance the specificity of ligands at D2 versus at D3, and vice‐versa.

It is interesting, then, to look at the influence of the various mutations onto this apparent scattering of the data. Indeed, as depicted in Figure 11, the influence of the D3 N‐terminus to the binding of the compounds compared to the D2 WT seems to be nihil. To the contrary, as suspected, the presence of this N‐terminus onto the sequence of D2 does not turn a D2 into a D3 receptor. Symmetrically, the change of the N‐terminus in the D3 sequence for the D2 N‐terminus, does influence the binding profile compared to D3 WT, and does not transform the D3 profile into a D2 one. Interestingly, a similar situation can be observed for the two last chimeras: The TM1 of D2 does not turn the profile of the chimera into a D2‐like one, and vice‐versa.

FIGURE 11.

Pharmacology of D2/D3 chimeras. Chimera schematic representation, domains of D3 receptor as labelled in red, and D2 receptor in green. Reference compounds are in blue diamonds; SRE1118 series, in magenta; W1109, in green and SRE1114 in orange. See Figures 1 and 2 for structures. Continuous line is the slope 1; dashed line is the limit of +/− 0.5 log. All individual data are gathered in Table S5

4. CONCLUSIVE REMARKS

As conclusive remarks, one should recall that an enormous amount of work has been done to develop D3 selective agonists and to identify key pharmacophoric features responsible for the selectivity at the D3 receptor over the D2 receptor. ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ Among many other examples, a large number of D3‐R agonists were derived from pramipexole, ⁶⁴ with high‐affinity D3R agonists with pKi values as high as 8.0. ⁶⁵ Other authors attempted to find selectivity toward one or the other of D2 and D3 receptors with new ligands derived from eticlopride ⁶⁶ or cariprazine. ⁶⁷ The recent availability of crystal structures for the D2 or D3 dopamine receptors ⁶⁸ , ⁶⁹ , ⁷⁰ will also help developing new approaches for specific ligands, either agonists or antagonists. The structure of D3R antagonist bound was crystallized in 2010, ⁷⁰ further nourishing the approaches of molecular modeling aiming to the same goal: finding ligands that are D3 preferred over D2. ⁶² , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ Furthermore, the selectivity issue is also to be discussed as it becomes obvious that many compounds are not binding only to dopamine receptors, but also to other GPCR family(es). ⁷⁵

Although directed mutagenesis is a key player in the understanding of the structure–activity of receptors, together with co‐crystallography of pure receptor‐ligand complexes, the number of such mutants for D2 or D3 is rather scarce in the literature (see Table 5 for a summary).

TABLE 5.

Summary of the directed mutagenesis publications on dopamine receptors D2 and D3

D2			D3			Ref.
Mutations	sta or tra c	Number of compounds in the study	Mutations	sta or tra a	Number of compounds in the study
D114N, D114G, M116L, M117C, M117G, S194A, S194A/S197A	tra	9				41
K251V, D249V, K3R‐V, K5R‐V, P264G, S259/262A, D271V	tra	1				42
F198A, F389A, F390A, F411A, L387A	tra	4				43
S193A, S194A, S197A	tra	23				44
			C114S		6	45
W90L, V91F, L94S, F110L, V111M, F164A/S167A, L170V/F172C, F189Y, V196C, T392V, Y408V, D2F411V, V91F/F110L, V91F/Y408V, F110L/Y408V, V91F/F110L/Y408V, V91F/F110L/V111M/Y408V, W90L/V91F/L94S/F110L/V111M/Y408V	sta b	6				46
			C147K, C147A, C147S, C147E, C147R	sta?	0	47
P310S, S311C, T351A	tra?	6				48
			D3.32E, F6.51W	tra	16	49
K241A	tra	0				50
			D110N, S192A, T369V	tra	6	40
H393A	tra	1				51
			D3.32E, V2.61F, L2.64F, FV3.28, 3.29LM, H6.55A, S5.42A, S5.46A, F6.52W, Y7.43A, Y7.43F	tra	4	52
			D110N, S192A, S193A, W342A, F346A, H349A, H354L, T369F S361A, F345A	tra	3	53
Y37A, L41A, V91A, L94A, E95A, G98A, K101A, F110A, D114E, F164A, C168A, L174A, N180A, E181A, I183A, I184A, A185S, N186A, S193A, S194A, S197A, F390A, H393A, H393F, Y408A, S409A, F411A, T412A, Y416A	sta	12 c				54
			S182I, E90L, Y373F, E90Q, V189I, C114L, Y365L, Y36L/S182I/V350I, Y36L, I183F, V189A, Y36F	tra	4	55
L41A, W90A, V91A, E95A, F110A, V115A, C118A, T119A, I122A, L174A, E181A, I183A, I184A, A185S, N186A, V190A, S193A, S194A, S197A, F202A, F360A, F361A, H364A, H364F, N367A, I368A, S380A, T383A	sta	8				56
W100A, W100L, F110A, F110L, V115I, C118S, C118V, I184F, I184S, S193A, S197A, H393N, Y408A, Y408F, Y408V, T412V, Y416W, V154I, S311C	tra	5				57
L41Y, E99V, I166V, E181V, I183S, S193A, S194A, S197A, H393A, Y408A, F411T, T412A	sta	34	Y36L V95E, V180E, S182I, S192A, S193A, S196A, H349A, Y365A, T368F, T369A, V164I	sta	34	Present report

^a Stable (sta) or transient (tra).

^b Not clear in the mat and met section.

^c Only on some mutants.

As stated earlier, one possible reason for this situation could be due to (i) the difficulty to obtain stable cellular clones expressing the mutated receptor that renders the structure–activity study easier and more robust–because also, one has to have similar expression levels (Bmax) of the mutants from one mutated clone to the next and pKi's to the (radio)ligand, also similar from one to the next; (ii) the fact that one or two mutant(s) afford limited information onto the impact of those mutations, considering the large number of amino acids playing a role in the receptor‐ligand relationship. This would enhance the cost of the experiments in a significant way; and (iii) the necessity to have access to rationally designed series of compounds that can document the variations due to the mutations. In other words, it is quite rare that all these obstacles can be overcome in a given laboratory. As can be seen from the Table 5 of the directed mutagenesis reports on D2 or D3, several studies comprised a large number of mutants: four on D2, ⁴⁶ , ⁵⁴ , ⁵⁶ , ⁵⁷ two on D3. ⁵² , ⁵³ Most of these studies, though, contained a relatively low number of molecules. The present report is the first one gathering 11 mutations on both receptors that were characterized with 34 molecules. The minute observations of the impact on binding affinity of the various compounds on the various mutants will be deeply discussed from a molecular modeling point of view in the part 2 of the work (Gohier, Theret, Rojas, Ferry, Boutin, in preparation).

It is obviously a combination of all the modern approaches of molecular pharmacology that will permit to understand the minute structural differences between the D2 and D3 receptors. Based on this knowledge, it will be possible that new ligands with enhanced affinity for one of the receptors, as opposed to the other, might be derived, synthesized, and characterized.

On the other hand, the use of stable mutants comprising unique mutations, and maybe with several mutants each presenting different amino acids at a single position might also help deciphering this specificity and help design a next generation of specific tools and/or drug candidates. To avoid those possible biases, one could be reminded of the formidable, nevertheless under‐regarded nuclease approach, as exemplified inhere. To complete the available pharmacological tools of the molecular biology, it seems that those approaches remain at the core of future research.

AUTHOR CONTRIBUTIONS

Celine Legros: Conceptualization (equal); data curation (equal); formal analysis (equal); validation (equal); writing – original draft (equal); writing – review and editing (equal). Anne Rojas: Methodology (equal); resources (equal); writing – review and editing (equal). Clemence Dupré: Methodology (equal); resources (equal). Chantal Brasseur: Methodology (equal); resources (equal). Isabelle Riest‐Fery: Methodology (equal); resources (equal). Olivier Muller: Formal analysis (equal); methodology (equal); resources (equal); supervision (equal); validation (equal). Jean‐Claude Ortuno: Formal analysis (equal); methodology (equal); resources (equal); validation (equal). Olivier Nosjean: Resources (equal); supervision (equal); validation (equal). Sophie‐Pénélope Guenin: Methodology (equal); resources (equal); supervision (equal); validation (equal). Gilles Ferry: Investigation (equal); methodology (equal); supervision (equal); writing – original draft (equal); writing – review and editing (equal). Jean A. Boutin: Conceptualization (equal); resources (equal); supervision (equal); validation (equal); writing – original draft (equal); writing – review and editing (equal).

CONFLICT OF INTEREST

The authors declare no conflict of interests.

Supporting information

Figure S1: Examples of the saturation curves obtained with D2 Wild‐type (A), mutant 5 (B), mutant 6 (C) and D3 wild‐type.

Figure S2: Examples of the competition curves obtained with D2 wild‐type (A, B–26 compounds), mutant 5 (C, D–31 compounds), and mutant 6 (E, F31 compounds).

Figure S3: Correlations between N = 1 and N = 2 of the individual data for the binding displacement of all the compounds at dopamine D2 receptors

Figure S4: Correlations between N = 1 and N = 2 of the individual data for the binding displacement of all the compounds at dopamine D3 receptors

Table S1: sequences of the oligonucleotides designed to introduce a point mutation in the mutant receptor.

Table S2: pKi's and Bmax's for the receptors generated in the present report.

Legros C, Rojas A, Dupré C, Brasseur C, Riest‐Fery I, Muller O, et al. Approach to the specificity and selectivity between D2 and D3 receptors by mutagenesis and binding experiments part I: Expression and characterization of D2 and D3 receptor mutants. Protein Science. 2022;31(12):e4459. 10.1002/pro.4459

DATA AVAILABILITY STATEMENT

Data available on request from the authors.